UNCLASSIFIED

AD NUMBER
AD001531
LIMITATION CHANGES
TO:
Approved for public release; distribution is
unlimited.

FROM:
Distribution authorized to U.S. Gov't. agencies
and their contractors;
Administrative/Operational Use; 30 SEP 1951.
Other requests shall be referred to Ballistic
Research Laboratory, Aberdeen Proving Ground,
MD 21005.

AUTHORITY
DCSL D/A notice dtd 18 Jun 1980

THIS PAGE IS UNCLASSIFIED

 ·- ...
~ ..

C~09

iur q
of Sj mp os------ I·
-1
Charges
Held at the

Laboratories
Aberdeen Proving Ground
(.
__d _ _ ---------·Maryland \ _
-- ~,-""""" ""'GRADING
4
~-.,._.,.;n-';~~ !\_rrt-~C P0~" ..."
-::~cmcat LII&DY .
('Bldg. IUllb!.iUP'EP.
.. :.tDI.-LB · 55 lCJ\.tiON

-'
~- PIIEW --GfOIDm.L llll~
--IJIG
---- -·--
November 13 -16 , 1951
, ~. • r "· : .,, '• ...
e i··~ <• f ,~,- , -,., ,, • •

r :c:·.· L• ' : c . ''•""·' "··· li?-$,

J, •. _,
:c.'> ··.r,,. •·~,r,,..~•<: :•1 '''·
1'""'•'--,r 1~ (<'1 ur.•Jiil

AB ER DE Er DR OV IN t: t:D nu un " MA RY LA ND

~~· . 1''
~-... ~·

~--
E.::~,; :~~u~~~k~~1v 2s1~a --............."~
. ~ --- - .. . ._
 ..

.
. '

.. ---
--- -

--------------- -· --- ---- --- --

-- ----·-· --

'.

ONClASS1flf~
 UNCLASSIFIE
., -
L <. ,,·., , ,_-·, t _ :)
This symposium was the outcome of a meeting at the Naval Ordnance Laboratory in
July 1951 between representatives of The Naval Ordnance Laboratory, Carnegie Institute
of Technology, New York University, and The Ballistic Research Laboratories (Army
Ordnance),

Mr, Louis Zernow, Ballistic Research Laboratories, was appointed chairman of a

cammittee consisting of Mr. Anson w. Solem, Naval Ordnance Laboratory, and Hr. Robert
.Eichelberger, Carnegie Institute of Technology, The technical program was arranged by
this committee. -

The local arrangements committee at the Ballistic Research Laboratories consisted

of Mr. Louis Zern:ow, Mr. John L, Squier, !11". Joseph !1, Regan, Mr. Irving Liebeman,
Lt. John c. Rhodes, Lt. John c. \Mtmore, and :Mr. Robert 1¥. Foster, with the assistance
or Mr. W. D. Dickinson and his staff.

PrDeoject No. TBJ-OlJ)j for Research and

velopment Div1s1on,
· · Ordnance Corps

r
r iii

\
 UNCLASSIFIED
\
Ut~Gl~;)~antll ..

"

J
{"' ,.
-
~ :!'·~..:· • ~
'I
- l I ---
 TABLE OF CONTENTS
,,
FOml'ORD • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • iii

\'IELCOME by Maj. Gen. E. E. MacMorland, Ccmunanding General,

Aberdeen Proving Ground, Mar,yland •••••••••••••••••••••••••••••••••••••••• 1
DITRODUCTORY RmARKS by Col, Alden P, Taber, Director,
Ballistic Research Laboratories, Aberdeen Proving Ground, Maryland •••• , •• 3

Sessions l and 2, FmiDAMEtiTALS

REMARKS ON SOME FUNDAMENTAL FEATURES OF DETONATION

By s. J, Jacobs, u.s. Naval Ordnance Laborato:cy,
White Oak, Silver Spring, Maryland • • • • • • • • • • . • • • • • • • • • • • • • • • • • • • • • • • • • • • • 5
THEORY OF LINED HOLUM CHARGES
By Dnerson M. Pugh, Department of Physics,
Carnegie Institute of Technology, Pittsburgh, Pennsylvania •••••• ••••• •••• ll

MULTIPLE FRAGMENT IMPACT EFFECTS IN SHAPED CPJ.RGE PEMETRATION

By John s. Ilinehart, Michelson Laborato:cy,
u.s. Naval Ordnance Test Station, Inyokern,
China Lake, California •··•~•••••••••••••••••••••••••••••••••••••••••••••• 33
INITIAL STUDY OF THE EFFECTS OF ANNEALING ON THE PENETRATION
PERFORMANCE OF COPPER SHAPED CHARGE LnlERS
By R. L. Phebus and w. o. Rassenfoss,
Ballistic Research Laboratories, Aberdeen Proviq; Ground, Mazyland ...... , 43
. GENERALIZATIONS CONCERNING THE HOTIO!l OF A THIN SHAPED
CHARGE LINER v/ITH AN ARBITRARY INITIAL CONTOUR
By G. E, Hudson and C, Gardner, New York University,
Nell York, Nw York ••••••••••••••••••••••••••••••••••••••••••••••••••••••• 61
SOURCES OF DISPERSION IN SHAPED CHARGE PERFORMANCE
By R. v, Heine-Geldern, Department of Physics,
Carnegie Institute of Technology, Pittsburgh, Pennsylvania .... ........... 75
PERFORMANCE OF PERIPHERALLY INITIATED SHAPED CHARGES
By A. D. Solem and W, T. August, u. S, N!Wal Ordnance
Laborator,r, White Oak, Silver Spring, MBr,yland ••••••••••••••••••••••••••• 83

THE PERFORMANCE OF PRECISION-MADE CONICAL LINERS

.
;
By J. Dewey, H. I. Breidenbach, Jr., J, P~zarella, and
J, Longobardi, Ballistic Research Laboratories,
Aberdeen Proving Ground, Mar,yland •••••••••••••••••••••••••••••••••••••••• 97

I
 TABLE OF CONTENTS - Continued

Sessions 3 and 4. INSTRUl1ENTATION

LCM VOLTAGE FLASH RADIOORAPHY ...

By J. J. Paszek, B. c. T~rlor, and J. L. Squier,
Ballistic Research Laboratories, Aberdeen Proving Ground, Mary·land ••••••• 107

FLASH RADIOORAPHIC STUDY OF JETS FRON UUROTATJID 105 1-lM

SHAPED CHARGES
By L. Zernow, s. Krorunan, J. Paczek, and B. T~rlor,
Ballistic Research Laboratories, Aberdeen Proving Ground, Maryland ••••• •• 119

FLASH RADIOORAPHIC STUDY OF JETS FROM ROTATED 105 MM

SHAPED CHARGES
By L. Zernow, s. Krorunan, F. Rayfield, J. Paszek, and
B. Taylor, Ballistic Research Laboratories,
Aberdeen Proving Grrn.md, l~ary-land ••••••.•.•..••••.• • • • • • • • .. • • • • • •. • • •• • • 133

EXPENDABLE FLASH X-RAY TUBE

By G. Hauver and G. Bryan, Ballistic Research Laboratories,
Aberdeen Proving Ground, }1aryland ••••• , ••..•...•••••..•.••••. ~ .•...•••• !I'. 151

THE KERR CELL CAMERA AND ITS APPLICATIONS

By E. C. ~lutschler, Department of Peysics,
Carnegie Institute of Technology, Pittsburgh, Pennsylvania ••••••••••••••• 157
THE PIN TECHNIQUE FOR VELOCITY }JEASUIID1ENTS
By H. Dean Mallory, U. s·. Naval Ordnance Laboratory,
White Oak, Silver Spring, Maryland ••.•.•..•••••••••.•.•.•.•••••••••••••••• 175

HIGH SPEED PHOTOGRAPHY WITH AN IMAGE CONVERTER TUBE

By R. D. Drosd, T. P. Liddiard, and B. N. Singleton, Jr.,
· U. s. Naval Ordnance Laboratory, White Oak,
Silver Spring, :P.Iaeyland ••••••••• , •••••••••• , • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 181

HIGH SPF.ED HIGH RESOLUTION STREAK PHOTOORAPHY

By c. T. Linder, Carnegie Institute of Technology,
Pittsburgh, Pennsylvania •••••••..••.•.•••••.••••..•••• • •.•...•.•.••.•...• 189

EXPERIMENTS ON DETONATION PHENOMENA

By G. R. Walker, Canadian· Armament Research and
Development Establishment, Valcartier, Quebec •••••••••••••••••••••••••••• 195
.:......

<:ONy:IDENTIAL •'
/
/
;
v
 '
.,1£ .LC (L£$ 4
-.
UNCLASSiflt~
TABLE OF CONTENTS - Continued

Sessions 5 and 6. WEAPONS

CmlPARATIVE EFFECTIVENESS OF ARMOR-DEFEATING Al~!UNITION

By A. Hurlich, Watertown Arsenal Laboratory, (:- .~iV'
Watertown Arsenal, Watertown, Massachusetts •••••••••••••••••••••••••••••• 19JI
REviEW OF THE PRESENT POSITION OF HOLLOW CHARGE AND SQUASH
HEAD RESEARCH AND DEVELOPMENT JN UNITED KINGDOH
By w. E. Soper, Armament Research Establishment, v'
Fol"t Hal.stead., Kmt •• , ••••••• , ••••••••••••••••. , . "' .•• , •..•••...•••••••••• 217

HELLER "WARHEAD DEVELOPMENT

By R. w. Foster, Canadian Annarnent Research and ·
Development Establishment, Valcartier, Quebec •••••••••••••••••••••••••••• 225
CURRENT PROGRAH ON HEAT AND REP ARTILLERY PROJECTILES
By Lt. Col. R. E. Rayle, Office, Chief of Ordnance, (;~-;;-):;./
Washington, D. c. •....................................... · • · · · · · · · · · · · · ~
DEVELOPMENT OF SHOULDER FIRED SHAPED CHARGE ROCKET HEADS
By H. s. Weintraub, s. Fleischnick, and I. B. Gluclanan,
Picatinqy Arsenal, Dover, New Jerse.y ••••••••••••••••••••••••••••••••••••• 247 ~
THE PRESENT PERFORMANCE AND PROBLEMS OF THE 105 Ml-1 BAT RJio'LE
By H. P. Manning,· c. w. Musser, and H. W. Euker, (../
Frankford Arsenal, Philadelphia, Pennsylvania ••••.•••••••••••••••••.••••• 269
DESIGN VARIABLES AFFECTING THE PERFORMANCl!: OF
THE BAT HEAT ROUND·
By Earl w. Ford, Defense Research Division, v'
Firestone Tire and Rubber Co., Akron, Ohio ••.•.•..••••••.•.••••••••••••••• 279
APPLICATION OF THEXlRY TO DESIGN
By Nonnan Rostoker, Carnegie Institute of Technology,
Pittsburgh, Pmnsylvania ~-······••••••••••••••••••••••••••••••••••••••••• 291
FOREIGN DEVELOPMENTS IN SHAPED CHARGES
By Harry Bechtol, Development and Proof Services,
Aberdeen Proving Ground, Mary-land •.•••••••••••.•••••.••••••.•.•••.•.••••• 299 ~.
RNR REPORT ON 120 MM SPIN STABILIZED PROJECTILE \~ITH
NON-ROTATING SHAPED CHAHGE
By Stanley Dubroff, ArtilJ,ery Ammunition Department,
Frankford ~senal, "Philadelphia, Pennsylvania •••••••••••••••••••••••••••• 305 ~

vii
~~cu~~~~nt~ .
" '• t·J .· ~.::z;0
,-._. :.............-~
,., -
JtL'
 UHGl~St~Hl~!
'

TABLE OF CONTENTS - Continued

Session 7. ROTATION

STUDY OF THE EFFECTS OF ROTATION UPON THE PENETRATION OF

JETS FROM 105 MM SHAPED CHARGES
By L. Zernow, J. Regan, J. Simon, and I. Lieberman,
Ballistic Research Laboratories, Aberdeen Proving Ground, Maryland ••• •••• 319

SPIN COMPENSATION
By E. L. Litchi"ield, Carnegie Institute of Technology,
Pittsburgh, Peilnsylvaxda • , • , • , •. , ....•••..•.. , •••••.•.•••....•.•••..•..•• 331

lliNIMIZING THE EFFECT OF ROTATION UPON THE PERFORMANCE

OF LINED CAVITY CHARGES
By Hugh Winn, Defense Research Division,
Firestone Tire and Rubber Co., Akron,_ Ohio ••••• ••• •••••••.•••••••••.....• .339

A ZERO ORDER THEORY OF THE INITIAL MOTION OF FLUTED

HO~l CHA!lGE LINERS
By L. H. Thomas, Watson Scientific Computing Laboratory,
New York, New York ••••••••••••••••••••••••••••••••••••••••••••••••••••••• .353

Session 8. EFFECTIVENESS OF, AND DEFENSE AGAINST

SHAPED CHARGES

SHAPED CHARGE OOIAGE BEYOND ARMOR

By D. R. Kennedy, Uo' s. NltVal Ordnance Test Station, . · .
Iny'okem,, China Lake, California • , •••••••••••.•••.••••••••••....•••....•• 359

v THE OO'.AGE EFFECTIVENESS OF SHAPED CHARGES AGAINST TANKS

By F. I. Hill; Ballistic Research Laboratories, r · \ 1
Aberdeen Proving Grotmd, Maryland •••••••••••••••••••••.•••••••••••••••••• ~· 363 )

SHAPED CHARGE :DA.V.AGE TO AmCRAFT STRUCTURES

By G. C. Throner, Uo' S. Naval. Ordnance Test Station,
Iny-okern, China Lake 1 California •••••••••••••••••••••••••••..• , ••.••••••• 369
v DEFENSES AGAINST HOLL~l CHARGES
By R. J. Eichelberger, Carnegie Institue of Technology,
Pittsburgh, Pennsylvarrla • • • • • • • • • • • • • • • • • . • • . . . . . • • • • . . • • . . • . • . • . . . . . • • • • 313

BmLIOGRAPHY AS DIS'!RmUTED BY EMERSON M. PUGH .AT THE SIMPOSIUM: BRIEF

HISTORY AND BmLIOGRAPHY OF CLASSIFIED REPORTS LEADING TO THE PRESENT THEORY
OF MET~IC JET CHARGES, ••••••••••••••••••..••••••••••••.•••.•••••••••••••••••• 385

viii
 WELCOME BY

MA.JOR GENERAL E. E. MwJORLAND

' )
Commanding General
Aberdeen Proving Ground

I would like to extend a cordial welcDII!e to all of you assembled here

for this important conference. The problems you are going to di11cuss are
both pres11ing and fundamental. I expect that you will all be in a better
position to improve our shaped charge weapons as a result of this discussion.
I would therefore like to express If13' hope that you meeting will be both

fruitful and enjoyable.

\ll19~ :~ l •• ~ ~ ~ { ~ ~ ~
&I kid&;
. '

'
 UNliLA~~Itltii
221!.

INTRODUCTORY REMARKS

COLONEL ALDEN Po TABER

Director
Ballistic Research Laboratories

The Ballistic Research Laboratories have a great interest in the shaped charge
program. Although the shaped charge type of munition is o~ one of several different
means to defeat he119Y armor or fortifications, it has certain distinct advantages over
other types of weapons. Genera.ll;r, the shaped charge r'ound is more efficient than
others or the same calibreJ and often the lwnching weapon mq be liglrt.er. In the
hands or an en811\Y, the- shaped charge munition presents UB with a very difficult defen-
sive problBD. In fact,. a defense against this type or attack may never be solved to
the satisfaction or our field forces.
Our efforts in the shaped charge program in the BRL are directed toward three
principal objectives, as followsa
1. Increasing the efficiency or penetration or a round.
2. Compensating tor the spin or a round fired .from a conventional cannon.

)o Providing a satisfactory defensive material or design against enei1\V" shaped

charge munitions.
I should like to comment bri~ on these objectives.
An inarease of 100% in the efficiency of penetration of shaped charge Dl1lllitions
would resul.t in lighter and more potent weapons in the hands or our soldiers. The
present bazooka could be smaller and more effective. Our light tanks with their
relativel;r small. calibre guns would be capable of defeating anything encountered on
the battlefield.
Compensating tor the spin or a shaped charge round tired tram a couventional can-
non voul.d eliminate special weapons and inarease greatl;r the effectiveness or this type
or artillery ammunition.

As mentioned before, the defense against shaped charge type munitions is a diffi-
cult problBD. A suitable material or materials tor this purpose may never be roum..
The problem may resolve itself into one or design o~.

3
UNGLASSifU~U
 In order to accomplish the objectives, that I hBVe mentioned, o_ne requires a
better understanding of what actually happens in jet ronnation, the behmor of the
jet in its trBVel in the air and in the target material, and the beh!Nior of the tar-
get material. A .great deal of basic and "a;>plied research is necessary to deterridne
and evaluate the many variables involved. For example: Why ie a copper cone better
\han a steel cone? What are the effects of diff"erent changes in the geometry of a
cone? How important are h!irdness heat treatment, micro-structure or atomic structure
of cones? Why doee glass offer more resistance to penetration than other target ,,
~~aterials of comparable density? There are tens or perhaps hundreds of other similar
questions that need and are undergoing investigation. The anSli'ers to these questions
JIIUSt come from research. A great deal of research on this subject is in progress.
As an end result, it iS hoped that a suitable theory or rational explanation or ob-
served facts will be attained.
Once a suitable theory is obtained and the fundamentals on the mechanism of j.et
formation and its behavior through different media are Wlderstood, a great step will
have been taken toward solution of present objectives.

.,.

•: d£b2EZI ·
 c.------··
REMARKS ON SOME FUNDAMENTAL FEATURES OF DETONATION
...
S, J, Jacobs
U, S, Naval Ordnance Laboratory, White Oak, Silver Spring, Maryland
ABSTRACT

Computations £or the shaped charge e££ect require that one

know the state o£ detonation product gases and the isentropic re-
lation £or expansion to a reasonable degree o£ precision, These
functions have been or can be derived £rom basic data £or a number
o£ compoWlds and mixtures using more or less approximate equations
o£ state, Experimental checks indicate that the important para-
meters £or calculations o£ boundary motions are capable o£ being
predicted with reasonable precision, say 5 or 10%, Though this
does not mean that all is well in our state o£ knowledge in this
£ield, the new worker should not be required to share the pessimism
which often is voiced when the subject is discussed from the view-
point o£ equation o£ state, The discussion will present same usef'ul
approximations to aid in canputations,

Mr. Chairman, Ladies and Gentlemen: It is, indeed, a privilege to be given the
o £irst opportunity to speak at this symposium on shaped charges, The reason £or opening
the discussion with some comments on detonation is, I believe, due to the realization
that progress in extending our knowledge toward improving shaped charge weapons re-
quires, among other things, that we understand the detonation process and use this
knowledge as a tool in improving our understanding o£ the shaped charge e££ect. I
shall endeavor, in the limited time I have, to mention the results o£ important ele-
ments o£ progress which have occurred in the past decade.

Detonation may be defined as the decomposition of an energy-rich explosive com-

pound; that decomposition being effected by a shock wave in which initiation occurs such
as to complete the energy release in a short time. The unit of time £or such a process
is the microsecond. The shock velocity for solid explosives, as most of you know,
ranges from about 4000 to 9000 meters per second depending on the specific energy and
the loading density. The prediction of detonation velocity from thermodynamic and hydro-
dynamic conservation laws using themo chemical data has been worked out to give satis-
factory answers for most of the explosives of interest to this group. · The calculations
are made through the use at steady-state assumptions, The results predict the pressure,
sound speed, mass velocity and temperature in addition to the propagation rate of the
wave. The fundamental way in which calculations are made and the numerous assumptions
required to obtain the answers has led many to distrust the results obtained, It is
usually considered that relative values are rather good for pressure, etc. but that
absolute values may be in large error. A paper by H. Jones presented at the 1948 Sym-
posium on Combustion, etc. iS of great help in this regard in that it places a fi~re
o£ precision on the absolute magnitudes of pressure and particle velocity. I shall
return to this paper in a few minutes.

27 &dlhJ£
 The present view of the shape of the reaction region leading to the postulate of
an initiation •spike", was presented b:1 von Neumann in this countcy and by Zeldovich
in Russia. A first statement of a non-steady state detonation in which a rarefaction
wave follows the reaction shock was presented by Sir Geoffrey Taylor. This is a step
in the direction toward obtaining the flow pattern for the detonation products for
general boundary conditions and general explosive confi~ration. Some work has been
published- on gas rarefaction in two dimension steady flow b:1 Doring and Burkhardt.
This trend of studying flow behind detonations should be followed more diligently for
it is here where the external effect of an explosion is coupled to the nature of the
explosion products and the energy release. The data that is needed from the detonation
is the pressure and the adiabatic p-v relation for the gas expansion.

The basic piston in a cylinder of gas model most commonly used to develop one di-
mensional shock theory is a most illustrative way to show the equations of steady deto-
nation and the effect of rarefaction. (Figure 1) If the •gas• is considered to be a
solid explosive which has been started to detonate at t a 0; x • 0 with both the
piston and the end of the reaction, (2) st·arting to move from x • o, and if (1) refers
to the state ahead of the propagating wave, (2) to the state immediately behind the
end of reaction (dotted line) and (3) to the state at the piston we can easily write
the conservation laws for mass, momentum and energy when u • ~ for then the entire
3
gas between (3) and (2) will be of uniform state and moving with velocity u • Re-
2
ferring to p as density, p as pressure, D as shock velocity, us as "gas" velocity,
c as local sound epeed and E as internal e~rgy (including chemical); the following
equations can be deduced without difficulty:

D !>l • (D - u)p 2

It can be ehown that u can only equal u when:

3 2

UJ + CJ • ~ + c2 ~ Do

Equality to D implies the Chapman-Jouquet condition; inequality implies an over

boostered detonation. An interesting problem was brought up at NOL the other day
which, I think, has :IJIIportant bearing on the proof of the Chapman-Jouquet condition.
Tbis is to consider what would happen if the piston, initially moving at a uniform
velocity u • D - c , were suddenly stopped at a t:IJIIe T,. o. According to the Riemann
3 3
theocy a rarefaction wave would move from the piston (now at rest) with its head
moving at a velocity u + c a D. But the detonation is also moving at velocity, D,
2 2
so that a region of constant velocity, ~· having a constant length (D - u ) T will
2
be following the detonation. The detonation is not distrubed. If now the time of
stopping the piston is reduced to an infinites:IJIIally small value greater than zero,
we would end up with a detonation propagating without being affected by a simple wave
which is attached to it. This is the problem solved first by G. I. Taylor and leade
to a picture of a propagating wave as shown in Figure 2.
6
 0
X

u3
:
I PI, P1, E I
I
I
I u1 = a, c1
:
\
" p3. P3 • E3

u3 • ()3
P2' P2• E2

u2 • 0 2

Figure 1-Piston in a cylinder model for presenting conservation laws.

X•Ot+a
o II
x = ot u" u SPIKE

u = 0
'.
Figure 2-Detonation in closed tube with back boundary at rest.

ll

ppr ..
4
The existence in gaseous detonation of the rest point shown in Figure 2 has been
pointed out by Patterson in a photograph taken by Bone and Fraser. Time does not
pennit ·my saying anYthing about .the effect of expansion to the sides of a finite
charge such as a cylinder. I do believe that the problem is one which can be tackled
by the use of characteristics in steady'-two dimensional nov. The one solution which
exists to my knowledge applies to an ideal gas case. We need to generalize it to the
case for the imperfect gases from solid explosives.

Returning now to calculations of pressures, particle velocity, sound speed and

density at the Chapman-Jouquet plane for dense explosives, there are two usef'ul approrl-
JIIations for quick answers. One is the ass11111ption that particle velocity is 1/4 of
detonation velocity. For cast or high density pressed explosives, this is probably
better than!:. 20,:. The other is the use of H. Jones equation which may be writtenr

B
u
• (2 + a) (1 + :,!
u
~)
upl
where a .is a thermodynamic quantity sliown to be positive and of magnitude of about
0.2!). ~ is the slope of the detonation velocity vs. density curve at density, p1 •
pl
Thus detonation data can give an estimate for~ from which other detonation data can be
derived with an estimated accuracy of S-lO:C based on an assumed error of SO% in a.
One other result, not too familiar in this country is the use of a polytropic equation
of state for the adiabatic expansion after detonation. This is of the form:

p • A(s)p.t' (s referring to entropy constant), and is a usef'ul &p):lroxl-

mation for computing other functions suoh as the Riemium function J £d p with
c2 • (*)II • It A p K -1

The usefulness of this equation is partly due to the relation between IC and ¥which
can easily be shown to be:

Thus, much of the data needed for determining the effect of an explosive in detonation
can be obtained from the measured detonation velocity. A table is included to show a
comparison of parameters so calculated with (a) calculations made by Snay using the
Kistiakowsky-Wilson equation of state and least squared co-volume data, and (b) results
calculated from 111easurements on velocities set up in Aluminum targets when they are im-
pacted normally by a plane detonation.

In conclusion, I would like to emphasize that there seems to be no reason to be-

lieve that the calculations one makes on detonation parameters lead to any less accu•
rate data than similar calculations for low pressure phenomena. The conservation of
momentum and mass plays such an ilnportant role in these shock effects that the equation
of state is relegated to a secondary role. If velocities, pressures and densities are
the require.d ftmctions from detonation data, thes.e can be obtained with accuracy
satisfactory for our present needs. If equilibrium or temperature are wanted, then the
picture is different. This for us is fortunately .not the immediate need.
8
 £!£ LEI lid dE 3i ti ~

TABLE I

Explosive and
Method of calculation P]. D u p2
E E "
c~ B i6oL!!Q ~ 1.68 7790 3085
u • D/4 assumption 255 1950 3 2.24
Jones equation 272 2085 2.76 2.295
snay•s calculation 243 1780 3.38 2.18
Al Target e:xp. 275 2ll5 2.69 2.30 ;,

~ 1.58 6880 3225

u • D/4 assumption 187 1720 3 2.10
Jones equation 191 176o 2.91 2.12
Snay•s calculations 166 1530 3.49 2.04
Al Target e:xp. 184 1690 3.07 2.10

Note: pressure in ld.lobars (1 k.bar ~ 103 atm.) velocities in meters/second.

REFERENCES

1. J. von Neuman, "On the Theory of Stationary Detonation Waves, 11 OSRD 1140 (1942),
(or see Reference ll, Page 227).

2. Y. B. Zeldovich, "On the Theory of the Propagation of Detonation in Gaseous

S;ystems 11 , Translated in NACA, Technical Memorandum 1261 (1950). (Originally
published in Russian in 1940.)

3. H. Jones, "Third Symposium on CCI!Ibustion, Flame and Explosion Phenomena", page 590 -
594, Williams and Wilkins Canpany, Baltimore, Maryland (1949).

4. Brinkley, s. and Wilson, E. B., "Revised Method of Predicting the Detonation Velocity
in Solids", OSRD 905 1 (1942), also OSRD 1707 (1943).
5. stewart, Patterson, "The Hydrodynamic Theory of Detonation",. Part II, Research, b
51 (1948) Page 221-234.

6. G. B. Kistiakowsk;y and E. B. Wilson, "Final Report on the Hydrodynamic Theory of

Detonation and Shock Waves" 1 OSRD l.l.4 (1941).

7. E. A. Christian and H. G. Snaya, "AnaJ.Jrsis of Experimental Data on Detonation Ve-

locities", NavOrd Report No. 1508 (1951).

8. H. Snay and I. Stegun, "Auxilliary Functions for Thermodynamic Calculations Based

on the K-W Equation of Staten, NavOrd Report No. 1732, (1951).

9
 , . . . ._ _ _ _ _ •o

9. Sir Geoffrey Taylor, 11 The Dynamics of the Combustion Products Behind Plare and
Spherical Detonation Fronts in E:xplosives 11 , Proc. Roy. Soc. A. 200, 1061, page 235-
247 (1950).
lD. s. Patterson, "One-dimensional Flow Behind a Steady Plane Detonation," Research ,1,
page 99, (1950).

11. R. H. Cole, "Underwater E:xplosions", Princeton University Press; Princeton, New

Jersey ( 1948) •

12. R. Courant and K. o. Friedricks, "Supersonic Flaw and Shock Waves", Interscience
Publishers, New York, (1948).

13. w. Doring and G. Burkhardt, "Contributions to the Theory of Detonation", (Trans-

lated by Brown University). Technical Report No. F-TS-1227-IA (GDAM A9 - T - 46)
Air Material Command, Wright-Patterson AFB, Dayton, Ohio (1949).

14. R. w. Goranson, H. D. Mallory and s. J. Jacobs, "Detonation Pressures in Solid

Explosives by the Measurement of Shock and Surface Velocities Induced in Metal Tar-
gets", NavOrd 1883, (in preparation).

15. A series of papers on a Symposium on Detonation held in Great Britain appears in

Proc. Royal Society, A, 204, 1-30, (November 1950).

Note: Abstracts on Detonation generally appear in Physics Abstracts, Sec. A under the
class numbers 541.11 or 541.12. (Physical Chemistry, Chemical Reactions.)

lD

I.
 -- - - ·-
- I - - 0 o• - 0

THEORY OF LINED HOIJ.al CHARGES

,.,
Emerson M. Pugh

Department of Physics, Carnegie Institute of Technology, Pittsburgh, Pennsylvania

ABSTRACT

An extension of the theory of cone collapse and jet formation

published in the Journal of Applied Physical is presented. By
assuming a time gradient in the velocity or collrsse or the cone
walls, the long rear end of the jet is elll:!lainedl without recourse
to a slug extrusion theory. Experimentsll:l verifying this theory
are discussed. The theory of penetrationl is reviewed to see how
well it explains the experimental results obtained with liners of
different materials and with targets of different materials. A
release wave hypothesis~ based upon characteristic surfaces of
Courant and Friedriche2-', is briefly mentioned. Though manifestly
inadequate, it provides a very rough correlation between the per-
formances of explosives of quite different shape.

In discussing the present state of the theory of lined hollow charges with you,
.. I will presume that you have all read the article published in the Journal of
Applied Physical. While this was as good a summary as could be declassified for
publication at the time, nmch was necessarily omitted. Individuals cleared for
access to the original sources should most certainly consult _these sources. They
contain nmch valuable material that could not be repeated in any swmnary or reasonable
length. A bibliography of reports important to the development of the theory, to-
gether with some brief historical notes is being prepared to aid investigators in
this field. A first draft is being distributed here in the hope that you will call
our attention to reports that have been missed* and help us to make it a more complete
and usei'ul document.

The theory of lined cavity charges will be presented in what may appear to be a
backwards order; namely, target penetration first, jet formation second and the
interaction of the explosive w1 th the liner third. This order seems most natural
to me, since it is the order in which I learned the subject.

1 Journ. Appl. Phys. Vol. 19, 6 pp. 563-582, June 1948.

* It has been noted, since this list was made, that nmch has been missed by
lumping the duPont reports into one item. Several of these are of out-
standing importance and will be listed separately. One glaring flaw is the
omission, in this summary, or the development or the theory, or important
papers by M. A. Cook.
r

11

ss;r 7 iT
"
 ~----·-·
PENETRATION

Early researches by MacDougall2 and his collaborators established that conical

steel liners produced long, small diameter jets of steel fragments having velocities
up to 10,000 meters per second at the front and little more than 1,000 meters per
second at the rear. Perforation of a target plate by one of these jets used up the
front of the jet but the rear of the jet passed through the hole unaffected by the
plate. The volumes of the holes made in targets were shown to be roughly proportional
to the kinetic energy of the jet used up and dependent upon the strength properties of
the target material. Qualitative reasoning suggested that the improved penetration
with standoff was due to the jet lengthening as it traveled. By considering the
momentum transfer from this stream of high speed particles to the target (whose strength
was assumed to be negligibl~ for these high velocity particles), they derived the
relation, pjVj(Vj - U) • pU /2, between jet velocity Vj and penetration velocity u,
where pj and p were the jet and target densities respectively. Their drum camera
measurements established that the ratio Vj/U was dependent upon the density in rough
agreement with predictions of this equation. These researches paved the way for the
more complete theory3 of penetration given by the author in May, 1944.

Reconsideration of the momentum transfer relations convinced the autho~ that the
relationship between the velocities should be of the form, 1/2 pjAj(Vj - U) • pAU2/2,
where4 A/Aj • 1 for perfectly made charges and greater than one for imperfect charges.
This relationship leads :Immediately to V/U .. 1 + const.JP, which was beautifully
verified by the experimental data of G. H. Messerly and D. P. MacDougall, and to
/P.A .
P • I :J:l which immediately suggested why penetration P increased with standoff.
/Pi
The length I of the jet increased because of its large gradient in velocity, and
therefore, its-penetration increased with increasing standoff. Because the jet
density pj was inversely proportional to I, penetration for this particle jet could
increase onl]' as, /I and therefore as IS • Carefully determined P vs. S curves
showed that with reasonable constants rs could not possibly increase fast enough
to account for the rapid increase in P at low standoffs. It was, therefore, postulated
that pj remained constant during the early stages of jet travel ~i.e. within the cone
and just beyond its base) and that the jet later broke up into particles. Thus
penetration (for good charges in which A/Aj • 1) started proportional to I or S and
after break-up became proportional to /j or/S • This made it possible to fit the
P vs. 8 curves very well and a whole series of quantitative correlations3 followed.

2 G. B. Kistiakowsky, MacDougall arid Messerly, 11 The Mechanism of Action of Cavitjr

Charges,• O.S.R.D. 1338, April 1943.

3 E. M. Pugh, 11 A Theory of Target Penetration by Jets,• O.S.R.D. 3752, May 1944.

4 The relationship between jet area Aj and effective area of impact on the target
vas suggested by the fact the unpublished radiographs of J. C• Clark and L. B.
Seely generally shafled misaligned jete. See BRL Bpt. 368.
12

......._ _ _ _II.
 <Jittml!ktlllll I I a ,.
Hill, Mott and PackS independently derived the relations ~j (Vj • u)
2
•

e../- + a and P • f />.p /P when a ..-.< >.pl //2.

The strength of the target material
1e represented by a. They used conservation of momentWII and energy relations to
obtain the velocity equation with particle jete and Bernoulli's equation to obtain
this equation for fluid jets, They were able to combine the two into one equation
where A • 1 for fluid jets and A • 2 for particle jets, They assW!Ied that low
melting point metals formed fluid jets while high melting point metals formed particle
jets, While this point of view vas not too successful in explaining the different
penetrating powers of different metals as liners, it did point out an error in the
author's original formulation of the velocity equation for the contiunous jets. It
also introduced a term to take account of target strength. A term a proportional to
ultimate target strength vas added, Since the H-M-P formulation of the Vl!llocity
equation vas the most rigorous, it vas adopted generally. While their basic velocity
equations strictly applied to steady-state conditions only, experimental evidence
indicated that the predictions obtained from their application to the non-stea4Y con-
ditions in these jets agreed with the e:xpeiiments,

In the author's treatment continuous jets6, whether fluid or solid, are asBW!Ied
to follow the Bernoulli equation ('A • 1), Particle jets require 'A ~ 2, since they
cannot support internal pressures, The correction required by the H-M-P fomulation
vas included into the original theory3 by introducing what might be called a pseudo
density J • >.pj Aj/A. For perfectly symmetrical charges J • >.pj' The velocity equa-
tion is then J (Vj - u) 2 • pU 2 + 2a and the penetration due to each jet element .
(when a is small) is given by dP • d{ JJ/p , As in the original theory, the penetration
of each jet element is proportional to dl when J is constant as in ductile drawing
('A and Pj both const.) and proportional to ldf after the jet is tully broken up and
p j becomes proportional to 1/d{, The only effect of these changes in the original
penetration theory3 vas to make invalid the original estimates of where the jets broke
up, All of the correlations but one of the original theory were still valid,
Immediately after break-up A starts to increase and pj starts to decrease, Thus
J • >.p j remains nearly constant for a time after the break-up, This means that the
break-up must take place somewhere along the linear7 portion of the P vs. S curve,

5 R, Hill, N. F. Mott and D. C, Pack, IIPenetration by Munroe Jets,• Armament

Research Dept,, AC Rpt. No, 5756, Feb. 1944.

6 For simplicity, solid and fluid metals are assumed to be incompressible,

7 Integrated values of P • J d{ ./J7P

are not quite linear, since the elements
break up at different points in space and also may have different alignments.

127
instead of at the point where the curve becomes parabolic, as was originally
believed, While the original correlations were obtained by the use of average
6
values instead of by actual integration, they were later confirmed when Fireman
integrated the equation P • fd//JlP and reached essentially these same conclusions,
The Qorrections indicated above were introduced into the penetration theor,y at a
Symposium>' on Shaped Charges in May 1945. At the same t:lme it was postulated that the
remarkably large penetrations produced at long standoff by charges having copper or
aluminum liners, were due to the high ductility of these materials under these con-
ditions, Figure 1 is taken directly from an old duPont report10, The P vs, S curves
for conical liners of copper, aluminum and steel are typical of the IIIB.ey" curves that
have been obtained rlth these metals. At low standoffs the penetrations made by
cones of copper and steel are pretty much the same but penetrations made by cones of
aluminum are smaller because of its lower density. Copper and aluminum penetrations
rise continuously with standoff whereas steel penetrations fall off at rather low
standoffs, The explanation proposed9 was that these copper and aluminum jets remained
continuous longer than the more brittle steel jets. This prediction made in 1945 has
been verified rece!ltly in the pictures taken· by R. Heine-Geldern and E. Mutschler.
Figure 2 shows photographs of jets from two M9Al size charges, identical except that
the one on the left had a steel liner and the one on the right had a copper liner,
Taken at the instant when each had traveled about 12 in. beyond the base of their
liner, the steel jet is broken up while the copper jet is still continuous. Figure 3
shows a copper jet from the same ld.nd of charge after it had traveled 30 in. beyond
the base tl:its liner, At this stage the copper jet has finally broken up into
particles • Kerr cell photographs have also been obtained of aluminum and brass
liners, The aluminum jets remain continuous to long standoffs whereas the brass jets
breakup at low standoff, as should be e:xpected fran their penetration vs. standoff
characteristics, The straightness of the 30 in. long copper jet is worthy of note,
The copper jet shows little evidence of radial spreading at 20 diameters standoff,
Steel jets spread more quiek:cy than copper, but with well made charges there is little
evidence of spreading of steel jets within 10 diam, standoff, Thus ~ reduction in
the average penetration at standoffs less than 10 diam, IIDlSt be attributed to lack of
symmetr,y in the charge assembly rather than to unavoidable radial spreading of the jet.
Figure 4 shows a penetration vs, standoff curve for apparently identical charges lined
with steel cones made to the ear}Jr manufacturing tolerances, The vide spread in the
penetrations due to different asymmetries is clearly seen, When charges are as well
made as they can be today 1 average penetrations at these low standoffs do not fall. off
as shown in this figure. For simplicity the remainder of IllY remarks will be l:IJIIited
to perfectly made charges where these asymmetries would not exist,

6 E, Fireman and Emerson M, Pugh, "Fundamentals of Penetration by Jets,n OSRD

4629h, March 1945, These calculations were necessarily based upon certain
assumptions. It is possible that these assumptions could be altered to allow
steel jets to break up before reaching the liner's base; i.e. at negative s.

9 Emerson M. Pugh, "Theories of Penetration," pp 79-101, in OSRD 5754, May- 9, 1945.

lD C, O, Davis and W, R, Burke, "Cavity Effect with Cones of Various Pure Metals,n
E, I. duPont de Nemours report for March 1943. · ·

ll After this paper vas presented orally, Louis Zemow presented some beaut:l1'ul flash
radiographs that showed more clearly than these Kerr Cell pictures that steel liners
break up at low standoff whereas copper liners remain continuous 11111ch longer and only
break up at long standoff,
 r__-:_.,..,.-.::-___"T_,_..._. ,...,!r-'.-r_...,....,...,._,.....,,...~
l!r-"1"""'1,.....,i-·'l"'l. -r~""'""'~'1-or-;_ """1:-""';. -..."',_""'~~"~.,.,..,.~, ;
.u ,:., i ·: I ,..L, i. ·.1 ;!. ) .1. . r-"1""!
... ,
.. .... : .... .......... ,- ... ..
:
9
.. ..

:
7

....

.. .. ... ..; ....... :. . :

. . . ...
............. ...... ;...
-~
·'
- ...
:, :r : i
.
.. ·+ ...
~ ;

: j
;
: : .. i •
: :
... ;
•
~~---~'-·---i--·-i··-~~-----~----t-~--·~---~----.1''··--·~---.---·~---·-·
I ... ... . ' ' • ...... ~.... ,.. f "
... .. .. .. :

0~----~--~~--~-1~r-----±-I--·--~---~~--~:~_L ...··_JI
f NDO~F
... ....
6 7
I - ltVCH£$ ,. ... e; J·~ ll

'"'" /

Figure 1-Penetration vs. standoff curves for liners of various pure metals
from duPont report for March 1943.

15
fBbi!L&
 Figure 2--Kerr cell photographs of 12 in. long jets from a steel cone (left)
and a copper cone (right). The steel jet is broken up
whereas the copper is still continuous.

Figure 3-Kerr cell photograph of a 30 in. long copper jet. At this long standoff
the copper jet is finally broken into particles.

16

seumnetm' 'j
 ., y_, j.

AVERAGE, 14AXIJ4Uh1 AND MINIWLI PENETRATIONs

VEBSUS STANDOI'F
C.I.T. CHARGJ!: COrlE DIAJI. 1.63 IN.
50/50 PENTOLIT!S WT. CHARGE 115 GRAMS
on
r-.
!")
0
...... ..,. I:'
•;) on
0>
..... ..,. 0
.....
Cll
C\llf)
0
.....
0
......
O(ll
E-t
p •0
0:12
:<;til
0
<.1)

...~
L) 7
;z;
....
(j
.-I
1.<1
~ ~
A A
A
e.
~
E-t ~ ~
(/} 5
.....
..... A·
-1
....:a
4.
z:
.... 0
:z;
3
....0 G ~
•
~E-4 2
~ 0

1%1
;z;
re l
STAh'DOPP - INCHES
11 12 13 8

Figure 4-Penetration vs. standoff· curve for CIT standard charges made with early
manufaeturi~ techniques. The average penetration is small at long standoff because
of lac of symmetry in the charges. Occasional charges performed well
at long standoffs.
 TIME IS MEASURED FROM WHEN THE
DETONATION WAVE WAS AT THE CONE

B A

6
2 0 J.l sec. -----

Pa
c B

3 3 . 5 J.l s e c.
- - - -------
' ....
00

•
--- c

.105.6J.i sec.
--
.'

2 3 4 5 6 7 8 9 10 II
INCHES~

SEVERAL STAGES OF PENETRATION OF A MILD STEEL TARGET BY A STANDARD

CHARGE CONTAINING A M 9AI STEEL CONE
Figure 5-Diagram illustrating penetration of mild steel~"' bf jeta from CIT
standard charges. The contributions to the jet, of three )eDgtll 11011&) elements
of the liner, are shown at three instants of time during eir travel to the target.
The lengthening of these elements with travel is due to their veloeity gradient& ,

,, •'
·I
 . ., ,, L

'

- - - - ~- - - Vj _-,:f tor ho110plate

----~tor r:aild .. ceal

- lf ~ for lead

f.W. JS'l'

Figure 6-Diagram showing how the effective jet length depends upon the target.
The low velocity rear elements cannot penetrate strong targets.
-
 _i_
I
~:w.t~
I
T
,;n.
l~n. !L,. r
i//1'~ /II/// r,J'/, -. I
I~ 1/1.

1---lf---11--1 "'in.
81n. ?in.
r 1
v ~,,
s "'-
lJ

l
l,.,. 1
6in.
7in. T I
\. ~
I
.,,
\ I
1\)
0
- l \ 7
Bin.

t.
41'1.

l
\I
W//ffff?l.im
In.
J.
.J
I
1//h.

1~.'
..1.
\I
l
-'·n. f'l'///~& lin.
...1.
~
v
T
t. .,t:#d plqt.,
F;q. 7. Hole prort!e.s 117 com6/naft"on lead- plafe, orn?or-plt:lfe
for9ef.s. The dtJfance rrorn fhe fe~p ol"fhe pl/e fo fhe fop ol"fiM
armor plofe i.J de.st9nofed A in the dt.1CtJ.s,.,on.

Figure 7-Eichelberger experiment. Illustrates the stopping of slow

rear elements of jet by hard homoplate.

.,
 '- - .. - ' .. ;~ - -
' - <
,,_ '

If perfectly made chargee were shot through vacuum, the penetrations into steel
should atop increasing with standoff when the particles of the jet are spread so far
apart longitudinally that the penetration due to one particle is completed before the
next particle arrives. This means that with perfectly made charges, the penetration
(produced by any given element of the jet) should; first rise linearly with s, second
rise linearly with.J"S and later become constant. Penetrations at small standoffs
are in excellent agreement with this theory. Here air resistance is of little im-
portance. At long standoffs the situation is quite different. Air resistance must
cause radial spreading and may COJ11Pletely stop most of the finer particles. For long
standoff penetrations the effects of asymmetries become exaggerated.
In Figure 5 three zonal elements A, B and C of a cone liner are shown. The jet
elements from these three cone elements are shown at three different times in their
travel to a target. At 20 ~ sec. jet element C is not yet formed and the front of jet
element A has just reached the target. The shaded portion shows the part of the hole
in the target that is made by jet element A. At 33.5 ~ sec. jet element C is completely
formed, the tip of B has just reached the target and A is used up. The shaded portion
shows the part of the hole made by jet element B. At 105.6 ~ sec. the tip or C has
just reached the target and A and B are used up. Notice that both A and B lengthen
while they travel but that the increase in length is much the greatest for C. While
this Figure 5 has been drawn from calculations made with the theory, the results have
been verified with jet velocity and penetration va. time measurements.
Several thousand shots through some SO to 100 different materials have shown
that, with the exception of one class of materialsl2, all target materials are
penetrated by the high velocity front end of these jete in accordance with the two
!
relations dP • d/. ./JlP and J(Vj - u)2 • p u2. The first of these relations will be
taken up by R. J. Eichelberger in hie paper on protection under •Residual Penetration
Theory.• For the slower moving elelnente at the rear of the jets the situation is
different. As one passes from the fast moving front elements toward the slower moving
rear elements to the region w,here Vj approaches the value ffa1J , the si111ple relations
break down. When Vj • J 2a/J , penetration stops and the remainder of the jet has no
effect upon the target. The effective length or the jet then depends upon the strength
of the target material. This accounts for. the total penetration into armor plate being
just a little less than in mild steel.
Tbis fact was firet stated qualitatively by MacDougall. The effective shorten-
ing of the jet length is illustrated in Figure 6. A jet (at a given instant of time)
is drawn along the axis of abscissa and the velocity of the various elements is
plotted as ordinates. Dotted horizontal lines illustrate the locations or vj •)2a/J
for targets or homoplate, mild steel and lead. The vertical dotted lines show the
point in the jet where penetration in the particular target will be stopped. The
effective length or the jet is that portion or the whole jet lying between the front
tip (at the right) and the vertical dotted line corresponding to the particular target
material.

12 Glasses and rocks containing high percentages of quartzite are the only material&
that have been found to date that provide significantly different resistance to
the frqnt or these jets than is predicted by these relations. These materials
provide much better than the theoretical protectioq when used in large blocks.
'·
In finely divided form they are ineffective for in this form they too follow the
•density law.•
21

iii If itudi it P
 'hifli!QGIJI 4 $62£6 '·'"'•

The truth qf these statements is strikingly illustrated by an experiment devised

by Eichelberger3 and illustrated in Figure 7, One in, thick homoplate was moved
down through stacks of 1 in. lead plates from top to bottom. At the top of the stack
the effect of armor plate was little different than the effect of an equivalent thick-
ness of lead, As the armor plate was moved down through the lead stack its effect
became more· and more noticeable. When it reached the point where~ntercepted on1y
those parts of the jet where Vj was no longer large compared to/2a/J for armor plate,
the nemainder of the jet was stopped in this plate, Here the total penetration was
less than in a stack of armor plate. As the armor was moved below this position the
slow moving jet particles penetrated the soft lead and still reached the armor plate,
The fact that the average penetration in the last stack was greater than in the one
before, even though the jet did not reach the armor in either case, may have been due
to one of two causes, First the averages are unreliable because only two shots per
stack were made, and second the armor may support the lead and contribute some
strength properties to it,
JET FORMATION

Theories ot target penetration developed rather logically from the researches

of the Explosives Research Laboratory group starting in late 1941 under MacDougall
to the formulation in 1945 of the modern version by the author, ·
On the other hand, attempts to formulate theories for the formation of these
jets met with little success until the spring of 1943, During this period J, L. Tuctl4
in England and independently J, c. Clark and L, B. Seelyl5 at the Aberdeen Proving
Ground obtained nash radiographs of collapsing conical liners. With the infonnation
obtained from these radiographs, Sir Geoffrey Taylor16 and independently Garrett
Birkhoftl7 obtained their well known hydrodynamic theory of jet formation. This theory
was very successful in explaining the appearances of the radiographs in the early stages
of jet formation, It was based upon the assumption that steady-state conditions
•
existed in a moving frame of reference. The radiographs seemed to justify these
assumptions, since the collapse angles appeared to be nearly constant and the slug
and jet appeared to be the same length •throughout the collapse process,• However,
the jet appeared •to be issuing from the slug long after the collapse process was
completed," and the jet was found to have a gradient in the velocity of its elements
that caused it to stretch out to great lengths, No satisfactory theory was available
to explain these facts, nor to explain the fact that more of the liner mass near the
base of the cone went into forming the jet than was predicted by the steady-state

13 R. J, Eichelberger, •Fundamental Principles of Jet Penetration" Carnegie Tech

Project AH-1, Monthly Rpt., Sept. 15, 1944 or "Protection Against Shaped Charges,•
Final apt, O,S,R.D. 6364, Nov. 10, 1945,
J, L, Tuck, •studies of Shaped Charges by Flash Radiography I. Preliminary,•
AC 3654, March 15, 1943,
15 Sir Geoffrey Taylor, "A Formulation or Mr. Tuck's Conception of Munroe Jets,•
AC 3724, May 27, 1943.
16 L. B. Seely and J, c. Clark, "High Speed Radiographic Studies of Controlled
Fra@ftentation, I. The Collapse of Steel Cavity Liners,• BRL Rpt, ,366,
June 16, 1943.
17 Garrett Birkhoff' "Mathematical Jet Theory of Lined Hollow Charges• BRL Rpt.
370, June 16, 19~3.
22

£tiib&i!i2L
theory. The possibility that these discrepancies could be explained, if the process
of collapse of the liner was not truly steady-state, was considered by many. It vas
thought, however, that this possibility was ruled out by the radiographic evidence
quoted above.
Several facts were overlooked in the above interpretation of the radiographic
negatives. While the collapse angle appears roughly constant during the time when the
detonation wave is sweeping the liner, it is far from constant during the period in the
collapse process that follows. The zonal elements of the cone near the base are pro-
jected inward at a much lower velocity than are those near the apex, the complete
collapse of the liner requires many times as long as the time required for the
detonation wave to sweep the liner. The observation that "the jet appears to be
issuing from the slug long after the collapse process is completed" means only that the
last jet and slug element that are formed travel forward at the same speed. The
remainder of the jet stretches out rapidly because of the large velocity gradient.
The radiographs show that the collapsing liner is never perfectly conical and
deviations from the conical shape increase toward the end of the process.
By introducing into the Taylor-Birkhoff theory of jet formation the assumption
that the collapse velocity d§creases from apex to base of the cone, it has been
possible to derive a theoryl that accounts for all of these discrepancies. This
theory accounts for the mass distribution in the whole jet, for its velocity gradient,
for the appearance of the radiographs throughout the process of ·jet formation. In
the first attempt at formulating this theory the additional assumption was made that
the detonation velocity also decreased as it swept across the cone from apex to base.
Subsequent measurements have shown, however, that this detonation velocity changes less
than a fraction of a percent. Calculations then showed that the theory agreed better
with experiment when the detonation velocity was assumed constant and the collapse
velocity was assumed variable than when both were assumed variable.
A number of the relations in the steady-state theory of jet formation can be
carried over directly to the non-steady theory. As in the steady-state theory it is
assumed that the liner acts like a perfect fluid. It is further considered that this
perfect fluid assumption means that adjacent elements move independently so that the
laws of conservation of mass, momentum and energy may by applied to each element
separately. The difference between the steady-state and the non-steady theories is
illustrated in Figure 8. Consider a zonal element of liner originally at P, In
either theory the detonati-on wave has moved from P to Q in unit tiMe, hence the
distance PQ .. U • UD cos a, where UD is the detonation velocity. In the smne unit
time the liner element has moved with a velocity V0 from P to J. In the steady-state
V0 is the same for each element and P1 would have moved to ~· Thus in the steady-
state the contour of the collapsing liner is a true co~e indicated by the dotted line
~Q. In the non-steady case V0 is lese for the element P1 than for P and hence P
1

18 E)nerson M. Pugh, ll. J, Eichelberger and Norman Rostoker, "Theory of the

Formation of Jets by Charges with Lined Conical Cavities;" CIT-ORO 21, June 30,
1949 and CIT-ORD 31, Feb, 28, 1951.

23
 • ·rrr ;; 3 iliilllli a

DETONATION
WAVE I
I
I
I
Ql

COLLAPSING
LINER

CONE AXIS

Figure 8--Liner collaPse and jet fonnation. Dotted line shows the contour of the
collapsing liner on the steady-state theory and the solid line
shows this contour on the non-steady theory. .

J} sin J = V./zv •

. !91 COS z IU. = dmYrJm

..:!l r= uesc ~ cos{a<- ~ +J')
:tf (J = sinp*-x sinrx{l-tunllfan
/Qfl .
£ %
Y./. r.J., 4 .J.
CDSf3 +X Sin« t1Jn n +ton~ /V.
jl"= 2$ +0( 2/I=(J*+o( 11 = lfi,J
~.,(a:)

Figure 9-lmportant equations of the

non-steady theory of jet fonnation.
 $14Z&l!t •a

has moved oncy to D. The collapsing liner is not conical but curved as shown by
the line ~. The important equatio:os of the non-steady theory are shown in
Figure 9 (rmmbered as in the papersltl on this theory). All of the variable
properties of the elements of the liner, slug and jet, are expressed in tel'llls of
x, the coordinate determining the location of .the element in the original liner.
The first two equations, (1) and (10) are identical with those of the steady-state
theory. The angle between the axis and the collapsing liner element is ~+ in the
steady-state theory and ~ in the non-steady theory. In accordance with a theorem
given by Sir Geoffrey Taylor, the collapse velocity V makes an angle 8 with the
0
nomal to the original liner, which is equal to half the angle FQJ of the isoceles
triangle l'QJJ hence Eq. (1) follows .f'rolll simple trigonometry. In this non-steady
theory the division of the liner mass between jet and slug expressed by Eq. (10),
and the jet velocity expressed by Eq. (7), depend upon ~ instead of upon~+ as in
the steady-state theory. The angle ~ is given by Eq. (18), where 2A • ~+ + a and
v~ • dV/dx·
UDder certain conditions the predictions of the non-steady theory may differ
radically from those of the steady-state theory. This is due to the fact that the
mass and the veloc.ity of the jet elements d~nd so criticall,y upon the collapse
angles and to the fact that these collapse angles ~ and ~+ may differ so greatly.
For e~le, either in the CIT standard charge or in the charge used by Seely and
Clarkl9 in their flash radiographs, the elements of the liner near the base may have
relatively small values of V and relatively large negative values of V~. ·
0

Consider the case where V << 2U and V <<

0 0
xV',
0
which are true for the basal
elements of the liner in most charge designs. Here ~ approaches 90° + a, while ~+
approaches a. Putting the values into Eq. (7) and into the equation forVS (Eq. (B),
footnote 18) gives Vj • VS and the elements of jet and slug travel along the axis to-
~ gather at the same velocity.

At this stage nearly two-thirds of the mass of this rather massive liner element
may be going. into the jet. Since the front of the jet travels at many times this
velocity, the radiographs create the illllsion that the jet is being extruded from
the slug. This phenomenon, which is clearly observed in fiash radiographs, gave rise
to the oft quoted statementl "an 'after jet• continues to be emitted (from the slllg)

19 Lo B. Seely and J. c. Clsrk, "High Speed R!¥iiogrsphic Studies of Controlled

Frag~~JSntation. I. Collapse of Steel Cavity Charge liners," BRL Rpt. 368,
June 16, 1943.
long after the walls have completely collapsed.• 20

The four equations in. Fig. 9 have been used to obtain an experimental
verification of the non-steady theory. They contain just three unknowns ( 11, V and
8 ; all functions of x) that cannot be determined by experiMent. Eqs. (1) anS- (7)
can be used to eliminate V0 and 8 f~m (18), leaving two independent Eqs. {10) and
{18) in the single unknown, ~. Figure 10 shows the result of solving these two
equations for ~· The circled points were obtained by solving Eq. (18) graphically
and the solid line is obtained from Eq. {10) using slug recovery data. The values
in Figure 10 were obtained with a CIT standard charge of pentoUte, lined with an
M9Al steel cone. Similar results have been obtained with other chPrges. The close
agreement between the two sets of values for ~ provides strong support for"'"tfie non-
steady th;OZ of c'Oii'e CO!l'ap'Se iiid jet fomat!on. DUring the earlfstages 'Orc"'nil
coilapse oesn~change very rap!aiy. This helped to justify the original steadY'"
state theory. The angle ~. however, increases very rapidly for elements near the base
of the cone. This is clearly shown in the series of flash radiographs published by
Clarkzl. The second radiograph (the first of the collapsing liner) shows a ~ near 40°
~~~~~!h~h~~~di~a~;::~ht~h~st! ~~ !hpng~:terT~~~~~retation of the last is

20 On the basis of the radiographic evidence that appeared to show jet issuing from
the slug, at least two slug extrusion theories have been presented.

Birkhoff visualized the slug as having a nearly fluid core, which was squirted
out by the application of a residual gas pressure converging on the slug. A
mathematical analysis showed that the core of the slug would be much hotter than
the outer layers.

M. A. Cook1 •Investigation of Cavity Effect" E.· I. duPont de Nemours Co., Contract

W-670-0RD-4331, Apr. lS, 1944, devised an ingenious theory in which the core of '
the slug contained a quasi-fluid of fine metal particles. The pressure required
for the extrusion came from residual inward kinetic energy of the slug walls.
The gradient in the jet was produced by friction of the outer layers of the
extruding jet along the inner walls of the slug.

Recent slug recovery experiments, together with other evidence, appear to rule
out both of these theories. It would be difficult for either of these theories
to explain the slug recovery results obtained with charges having large belts
of explosive around the base of the cone, which are discussed in the closing
paragraphs of this paper.
I

21 J. C. Clark "Flash Radiography Applied to Ordnance Problems •" Journal of

Applied Physics, Vol. 20, No. 4, April 1949, Figure 11.

26

lbNE ibkrt ¥IM!

 1!40 -Voluea colculoled from Slug Recovery
0 Voluea measured from Contau,.

120

"
-...•... iOO

-
~
~
... 80
c"
•
I.
.!! 60
8

40 • 0

20

0 Q20 0.40 0.60 0.80 1.00 1.20 lAO 1.60

Distance tram tap at Cane. I in.)
Figure 10---Independent determinations of the collapse angle /3. The agreement
between the solid line and the circled points provides a verification
of the non-steady extension of the theory.

Figure 11-Flash radiographs by J. C. Clark showing three stages of liner collapse and
jet formation. These collapse angles agree qualitath-ely with the values
determined in Figure 10.

27

a a&
 Values of U ror Eq. (1) were obtained with a rotating mirror camera. Values
of Vj for Eq. ( 7} were also obtained with the rotating mirror camera. However, to
obtain these velocities at various points along the jet, it was necessary to pass
these jets through target plates to knock orr the desired portion or the forward
jet elements. The mass of the remaining elemepta were determined by collecting and
weighing the portion or the jet that passed through. Thus a curve or Vj vs. mj was
determined. To identify these velocities with the original cone element from which
they came, it was necessary to use slug recovery data. Slug sections have less mass
than the cone section from which they came. The difference is the mass forced into
the jet, mj • m - •s• Thus mj was determined as a f'unction of x and, since Vj vs. mj
was known, Vj versus x was determined. The slug recovery data also determined dms/dm
for use in Eq. (10) • For slug recovery eJCperiments22 , the charge liners were made in
two parts, divided along a plane perpendicular to the axis. When these charges were
shot into water or other recovery mediUIII, two slugs were recovered. The weight or
these compared to the li'eight or the original cone sections provided a point on the ms
vs. m curve shown in Figure 12. This curve then was ditferentiated to obtain dm/dm•

v0 vs. x can be calculated from the preceding information. The values for the
CIT standard charge are shown in Figure 13. It is desirable to be able to calculate
V (x} directly from the properties or the eJCplosive and the goemetry of the charge as-
0
sembly. A semi-empirical release wave procedure is now being developed for this
purpose.

The release wave theory is being developed by Eichelberger, Linder and Dressen.
It employs ideas gleaned from the much more exact treatment of Courant and Friedrichs
with a treatment or a sp~cial case given by Gurney and Sterne. The ailll is to develop
a simplified procedure for analyzing the effects produced by different eJCplosive
geometries. Several effects must be neglected. An empirical approach is used to
determine which effects may be safely ignored. Briefly the theory visuallz~s a high
pressure zone in the gases behind a detonation wave that is cut orr by release waves
initiated at any free surface t.ouched by the detonation wave.

In its present unfinished state this procedure has successfUlly predicted liner
velocities for several charge geometries. Charges having conical cavities lined with
metal present a difficult problem.. Nevertheless, predictions or V0 (x) have been made
for two radically different charge geometries (with lined conical cavities} that
agree qualitatively with the results obtained from slug recovery and jet velocity
measurements.

These two chllrges are illustrated in Figure 14. Charge S is the standard CIT
charge while charge c has belts of eXplosive around the liner of uniform thickness
from apex to base •

22 This type or eJCperimental analysis was originally devised by D. P. MacDougall

and M. A. Paul. The technique has been moditied by R. J. Eichelberger to
produce precision results.
23 R. Courant and K. o. Friedrichs, "Supersonic Flow and Shock Waves,•
Interscience Publishers, 1948.

28
 ,,, ('
...

""'•
El
~ 14
..,
..,.,"
..,
.0:

....0 12
.,.., M9Al Steel Cone
:a
"'fa e Refined Teohnictue
10 0
bO
:1
.....
t.
u.l '\/ Old Methods
., t.
..,
.0:
8
....0 0

~
.,.
N
'Cl
::1 6
.,
..,
~
.....
....."
~ 4
a
2

X (in.)
0 0.9 1.0 1.1 1.2 1.3 1.4
0
0 2 16

CuJJUlative Mass m or Ori,e:inal Cone ( !;Ill•)

Figure lli-Division of steel cone between jet and slug. Results obtained by
recovering slugs from sectioned cones; using CIT standard charges.
 5 c
0.7
VELOCITY 0.3
0.6

.,~ 0.5 d"?J /-~s

....:t. dm dm 0.2
....0 :&
C) 0.4
c
,..
; !::
0
9
0.3 COMPENSATED
CHAR7 -
"'>
0.2 Pin. vc.c
STANDARD CHARGE s 7.9 1/.8 11
0.1
c 4.5 /'1.0
0~~~--~--~~~~--~--~--~~

..
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Cone llxi.r
DISTANCE FROM TOP OF GONE (IN.)
0 .3 9 1.3• 1.1 ;
Figure 13--Jet velocities and liner collapse Figure 14-Percent of cone liner going into
velocities for the S and C type charges jet fonnation vs. the distance of the original
shown in Figure 14. zonal element from the apex; for the S and
C type charges pictured•

•
 s mmaca . amm
The values of Vj(x) and V0 (x} for these two charges are shown in Figure 13.
These are in rough qualitative agreement with the predictions of the incomplete re-
lease wave theory. The average penetrations with the s- charge were 1.9 in. and with
the c- charge were only 4.S in. The large value of V~ in the s- charge resulted in a
long jet, which therefore penetrated deeply. The large belt of explosive around the
base in the c- charge made the value of V0 large and V~ small at the base. The
resulting short jet did not penetrate deeply, though it did make a bigger hole (19.0
cc against 11.8 for the s- charge} because of its greater energy. The slug recovery
results shown in Figure 14 are most instructive. Towards the base of the cone, the
percentage of the mass of the liner going into the jet, increases rapidly for the s-
charge (up to 31%) but very ~lowly for the a-charge (to only 17%). From Eq. (18) p
was large for the basal elements in the s- charge but small for those in the c- charge.
~~=~!:0~ (10) dm8 /dm became small for the s- but almost constant for the c-

On the other hand, it appears to be practically impossible to justify this result

on the basis of any slug extrusion theory. Before slug recovery results were available
on charges having large belts of explosive around the base of the cone, it was
generally believed that the reduced penetrating power was caused by the exploded
gases breaking out around the liner base to disrupt the rear of the jet. This analyses
appeared to be supported by radiographic evidence, which showed that the connection
between the jet and slug, commonly observed with other charges, was missing.

The obvious explanation with the non-steady theory is that V0 is not much smaller
than xV~ and, therefore, Vj f V8 but Vj > V8 • Thus the rear of the jet separates
from the slug and the radiographs do not show jet in contact with the slug.
CONCJlJSION
The present non-steady hydrodynamic theories of jet formation and target
penetration appear to be capable of interpreting the experimental results obtained
with conical metallic liners, provided their thicknesses are not too far from the
optimum and their apex angles lie between 300 and 90°. Undoubtedly these theories
would also apply to cones with angles outside these limits, provided the liners arP
sufficientL~thin.
..c:ls•••••••IIIII!J~IIIfGIIII'alllllliiiMis::

dOi!I£2£
 )tEZ#tziZ. sszuo 21!2. a

MULTIPLE-FRAGMENT-IMPACT EFFECTS IN SHAPED CHARGE PENETRATION

John s. Rinehart
Michelson Laboratory

J u. s. Naval Ordnance Teat Station, In;yokern, China Lake, California

ABSTRACT

It is aaSUIIIed that a shaped charge jet conta:ills a relatively small

group of discrete ~erveloci ty solid fra~ents. The probable cumulative
effects of the successive impacts of the several fra~ents are predicted
from known facts concerning the impacts of very high velocity- fra~ents.
The predictions are caapared with experimental observations on the inter-
actions between targets and shaped charge jets. The conclusion reached
is that the cumulative effects of multiple impacts appear, in some cases,
to p~ an important role in the mechanics of shaped charge penetration.
It is urged that further theoretical and experimental studies be pursued
in order to establish the relative place that these multiple-fra~ent­
impact effects must occupy- in a canplete treatment of shaped charge pene-
tration.

INTRODUCTION

It has been well established that a shaped charge jet uaual4' contains as an im-
portant component, a relatively small group of discrete ~ervelocity solid fra~nts.·
The total n11111ber of fragments that will be found within a particular jet will depend
upon the size of the charge,. cone angle, liner thickness, liner material, and many other
factors. The masses of the fragments and their respective velocities will be governed
b;r the same vQ'iables. Observations on jet-target interactions suggest that the cUIIIU-
lative effects of the multiple impacts of the fragments of the group could and probab:cy-
do p~ an important role in the mechanics of shaped charge penetration. The primary
purpose of this paper is to present a type of approach that appears to offer IIIUch
promise toward establishing the relative place that multiple-fragment-impact effects
DIUSt occupy in a complete treatment of shaped charge penetration. Known. facta con-
cerning the effects produced on a target b.r the impact of a single ~ervelocity rrag-
ment are used to predict the CUIIIUlative effect of several successive impacts. It is
assumed that the masses of the fragments with which we are concerned lie between a few
hundredths of a gram and a few grams and that their velocities lie in the range fran.
10 1 000 to 30,000 feet per second.

EFFECT OF SINGLE FRAGMENT IMPACT

When a fra~ent strikes a target, the most characteristic thing that it does is to
,_ produce a hole or crater. The shape and size of the crater will depend upon the shape,
mass, and velocity of the fl'lll!1Uent and upon the ph;ysical properties of the target. The
general character of the variation of shape of crater with. impacting velocity- is illus-
trated in Figare 1. At low velocity- (below about 4,000 feet per second for a steel tar-
pt) the crater is aimpl,y a straight-aided hole whose cross-section is similar to that

33

I
 7 PR 7 5 ., I 5 • 4

LOW VELOCITY HIGH VELOCITY HYPER VELOCITY

<4000 ft/sec 4000- 10,000 ft/sec >10, 000 ft/sec
Figure !-Variation of shape of crater with impacting velocity.

SURFACE
OF
FAILURE

Figure 2-Distribution of shearing stress prqdueed by application

of load over small area.

34
 of the impacting fragment. At higher velocities cavitation sets in and the profile of
the hole is roughly conical or bell shaped and its cross-section is more or less circu-
lar and considerably greater than that of the impacting fragment. At ve~ high veloci-
.•, ties (greater than about 10,000 feet per second) the crater will have a cup-shaped
appearance. We will be concerned here only with the latter type of crater.
The mechanisms involved in crater formation at the very high velocities are still
somewhat obscure. A most profitable way· of looking at the problem is to suppose that
the size and shape of a crater will depend primarily upon the stress-distribution ex-
isting in the target during and immediately following deceleration of the fragment.
Stresses that are undoubtedly much higher than the plastic limit for steel must exist
during this time in the region close to the area of impact. The probable shape of the
crater is arrived at by assuming (a) that the fragment is stopped in a negligibly short
distance, (b) that the force of the impact distributes itself within the target in
accordance with the same geometry as the stresses produced by a static load, and (c)
that the target material will fail within a region in which the shearing stress exceeds
a certain critical value.
When a semi-infinite solid body is subjected to a static load over a small area,
the distribution of shearing stress is approximately that shown in Figure 2. Each of
the circles in the figure corresponds to the intercept of a surface of constant shearing
stress with the plane of the figure. As one moves away from the area of application of
force the magnitude of the shearing stress decreases and eventually becomes less than
that required to cause the material to fail. There will exist a surface of failure
that corresponds to a critical limiting shearing stress. Such an assumed surface has
been drawn as a heavier line in the figure. We might expect that the crater would have
approximately the shape of this limiting surface of failure with the exception that ma-
terial near the free surface of the block would be pushed upward and outward since it
is under little restraint.
"' The profiles of craters produced by hypervelocity pellets a.re in substantial
agreement with the above predictions. Tracings of the profiles of two craters produced
in a lead target are shown in Figure 3. The crater on the lett was made by a 8,000
feet per second, 2 gram, steel pellet and that on the right by a 15,000 feet per second,
2/3 gram, aluminWI1 pellet. The profile of the crater made by the aluminum pellet has
been superimposed onto a stress distribution of the· type discussed above in Figure 4.
The latter drawing lends considerable support to the concept of a surface of failure
that corresponds to a critical limiting shearing stress. More precisely, we should
probably use here the von Mises criterion of failure. To the accuracy needed the
critical shearing stress criterion seems to be adequate. Although craters in lead
targets have been used here for illustrative purposes, it is well known that the crater
made by a shaped charge fragment striking a steel target has essentially the same shape.
It has been recognized for many years that the volume of the crater produced by
an impacting missile or fragment is directly proportional to the kinetic energy of the
missile or fragment. The depth of penetrati~n of a fragment of specified mass and
velocity can be computed if this observation is used and, if in addition, the shape of
the crater is specified.
A first approximation to the shape of the crater made by a hypervelocity fragment
is a sphere that is tangent to the point of impact. For this shape of crater, the
depth, p, to which an impacting fragment will cause the material to fail is given by

JS
 j• I in. ~

TARGET: LEAD

STEEL PELLET ALUMINUM PELLET

DIAMETER: ...v0.5 in. DIAMETER: ..... 0.5 in.
MASS: 2. 0 gm. MASS: 0.7 gm.
VELOCITY: 8,000 ftlsec VELOCITY: 15,000 ft/sec
VOLUME OF CRATER: 14cc. VOLUME OF CRATER: 17 cc.
K. E. OF AL./ K.E. OF STEEL = I. 2

Figure 3-Tracings of profiles of craters made- in lead by hypervelor:ity pellets

-PELLET

l''igure 4-Superposition of crater profile

and probable shearing stress distribution.
Crater made in lead target by 15,000 ft/sec,
0.7 gm, aluminum pellet.

36
 (l)
where V is the volume a£ the crater. Now the volume in terms of the kinetic energy ot
the fragment will be given by
(2)
.'
where c is a constant that depends upon the target material and 111 and v are the mass
1
and velocity, respectively, a£ the fragment. Substitution of the value of the volume
V given by Eq. · (2} into Eq. (1) gives

(.3)

where
.3~ 1/.3
cl!: • <-;-> (4)
It is ertdent trom Eq. (.3) that the penetration ot a very high velocity tra~nt is
proportional to the one-third power of its mass and the two-thirds power of its veloci-
ty. At 1011 :Impacting velocities penetration is proportional to the square of the ve-
locity.
The constant ~ ot Eq. (2) is the volume ot target ~~~&terial displaced per unit ot
ld.netic energy a£ the impacting fragment. Its value and, hence, the value a£ c2 ot
Eq. (.3), depends upon the target material. Experimentally determined values ot c;_ and
computed values of relative penetration are listed tor several target materials in Table
I. steel has been taken as unity. It is significant that the penetration variss as the
cube root of ~. Although the volume of the crater formed by a particular eypervelocity
fragment will be nearly ten times as great in lead as in steel, the depth to which it
will penetrate will only be 2 •.3 t:tmes as great.

CUMOLATIVE EFFECT OF MULTIPLE IMPACTS

The etrect of nmltiple impacts on a target can be readily arrived at. Consider
f~st the situation shown in Figure 5. SeVeral tra~nts of respective masses, ~·
DJ:!• ••• and velocitiss vl' v 2 , • • • separated in space, are all traveling in the
same strai(!Jlt line toward a target. When the fir!lt strikes the target, it produces a
crater of depth, pl' equal to ~~l/.3v /.3. The n&xt tra~ent then strikes and pro-
2
1
duces another crater. The result· at this inetant is· shown in Figure 5b. This process
cont:l,.nues until all ot the fragments have struck (Figure 5c). The total .depth of
penetration, P, will be given by ·

.37
 4t&iil&IW If W?!h?

,.•
t

''
'"i

(A) (B) (c)

Figure 5---Diagrams that show effect of multiple impacts on target (a} just before
impact, (b) after two particles have struck, and (c) total effect of all six particles.

1 em -1 1-

Figure 6--Trace of outline of hole produced

by steel lined Mk. 1 shaped charge.

38
 ··rasmsTn £6 !Iiibl!ii •

(5)

The contour of the hole will be the envelope of the several craters and should
have a general sausage~like shape.

It is interesting to consider at this point the relative penetrations produced by

a single fragment of mass 1 m1 and velocity1 v 1 and by the same mass brokell up into n
equal parts-, all of which have the same velocity but strike one after the other, The
penetration• p, of the single fragment will be

p • ~ml/3v2/3 (6)

The penetration. P1 of the n fragrnentswill be

2/3 m l/3 (7)
p. c2v n(n)
The relative penetrations will be given by

(8)
It can be shown mathematically that division into n equal parts yields the maximum
relative penetration,

EXPERIMENTAL OBSERVATION

In this section the above concepts will be related to available experimental data
on shaped charge penetration, The following two experimental facts are well established.
· First, the depth of penetration produced by most lined shaped charges increases as the
distance between the charge and the target is increased frQII zero to some optimum dis-
tance that is of the order of several charge diameters, At greater than optimum stand-
off. depths of penetration are likely to be erratic, Charges which are presumably well
aligned will sometimes give depths equal to those obtained at optimum standoff, Second 1
the volume of the hole remains substantially constan~ as standoff is increased to the
optimum for depth of penetration, Beyond the optimum, the average hole volume decreases
with increasing standoff,

A third observation is that the contours of the holes made by shaped charge jets
SJ!hibit a characteristic irregularity. This irregularity is characterized by a quite
noticeable pattern of swellings and constrictions, The contour of a typical crater made
by a steel lined Mk:, 1 cone in a steel target has been traced in Figure 6, The same
type of irregularity is also observable in most of the photographs of cavities that have
been published,

Qualitatively• the concept that the cumulative effect of the impacts of a group of
discrete fragments is an important aspect in the penetration of a shaped charge jet 1
fits the experimental observations well.

39
 .f sn a •: I r

4)

(A)

(B)

•
•
•
••
• •

(C)

Figure 7-Diagrams that show cha.nge in penetration with standoff

(a) elose standoff, (b) optimum standoff, and (c) long standoff.

40
 First, over short distances, the fragments will lose little velocity so that the
total volume of the cavity, since it depends upon the total energy of all of the frag-
ments, ought to remain practical~ constant. At longer distances, it ought to de-
crease somewhat because of retardation of the fra~ents.

Second, an optimum standoff is to be expected. This optimum will occur at a dis-

tance from the charge such that the fra~ents are separated just enough to produce
essentially individual craters, all of which lie very nearly along the same line. Be-
yond this optimum standoff, the fragments may be poorly collimated so that their indi-
vidual penetrations will not be additive. As noted above, hawever, very good pene-
trations are sometimes observed at long standoff. This probably occurs in those rare
cases in which the jet remains well collimated. The three situations, close standoff,
opt:imwn standoff, and long standoff, are illustrated schematically in Figure 7.

The fairly regular pattern of swellings and constrictions described above, and
evident in Figure 6, m~zy be explained by assuming that each bulge corresponds to the
impact of an individual fragment. .I£ this were strictly true, the bulging would be
very regular; however, it is not. The non-uniformity of the bulging undoubtedly arises
from the simultaneous, or near simultaneous of two or more fra~ents; the skewness, fran
fragments striking slightly off axis.

The number of bulges ought to correspond roughly to the number of fra~ents. Photo-
graphs of Mk 1 shaped charge jets taken with a streak camera usual~ exhibit some 15 to
25 traces, each one of which corresponds to a discrete fra~ent of appreciable size.
Experiments have been carried out in which the shaped charge was moving laterally at the
,, time of firing. Under such conditions the jet fragments produced sane 15 or 20 indi-
vidual small craters in a steel plate placed a short distance from the charge. The hole
in Figure 6 ought, therefore, to contain some 15 to 25 bulges. Even though the hole is
irregular and an accurate count is not possible, the llWllber of observable bulges appears
to be in agreement with the number expected.

The diameter of the crater that will be made by a hypervelocity i'ragment can be
computed directly from Eq. {3) since for the shape of crater as~ed ~e diameter and
depth are numerically equal. Taking a value for '1. of 0.29 x 10- in /ft-lb, we find
that a 20,000 feet per second, 0.1 gram, steel fra~ent, the very approximate velocity
and mass, respectively, of Mk 1 shaped charge i'ragments would produce a crater about·
1 em in diameter. The diameter of the crater profile, drawn in Figure 6, ranges i'rom a
minimum of 0.4 em to a max:lmwn of 1. 7 em. It is not unreasonable to expect that it
could have been excavated by successive impacts of several fra~ents of the above mass
and velocity.

SUltURY

The several foregoing rather general observations suggest strongly that the cumu-
lative effects of multiple impacts of hypervelocity fragments m~zy play an important role
in the mechanics of shaped charge penetration. The data and observations are still too
meager to establish unambiguously the relative place that these multiple-i'ra~ent-impact
effects must occupy in a c<m~plete treatment of shaped charge penetration. It is felt
that the general method of approach is most promising and should be pursued vigorously.
 TABLE I

Experimentally Determined Values of Material

- Displaced Per Unit of Energy.

VolUIIIS
displaced
per unit
of energy Relative
cl penetration
(Steel taken
Material (in3/rt-lb x 104) as reference)

Steel
Lead
Copper
24ST Al.

GENERAL BIBLIOGRAPHY

1. G. B:l.rkhorr, D. P. MacDougall, E. M. Pugh and G. I. T~lor, "Explosives with Lined

Cavities", J. Appl. Phys., 19 563 (1948), :-'

2• J. S, Rinehart, "Same Observations on High Speed Illlpact", Pop. Astron, LVIII 458
(1950). "

3. J. S, Rinehart and \-l, C. White, "Shapes of Craters Formed in Plaster of Paris Tar-
gets by Ultra-Speed Pellets", American J, of Pbys, (In Press),

4• Felix Helle, "Traite de ba11stique experimentale", Dwnaine, Paris (1840).

5. w. M. Evans and A. R. Ubbelohde 4 "Formation of !1unroe Jets and Their Action on

Massive Targets", Research 331 \1950).

6. W,---M. Evans and A. R. Ubbelohde, "Some Kinematic Properties of Munroe Jets", Research
376 (1950).

7. R, Hill, "An Analysis of R.R.L. Exper:iJnents on Cavitation in Mild Steel", Armament

Research Department Report 11/48, Theoretical Research Report l/48, March 1948
(Confidential).

42
 •

INITIAL STUDY OF THE EFFECTS OF ANNEALING ON THE PENETRATION PERFORMANCE OF COPPER

_, SHAPED CHARGE LINERS
R. L. Phebus
w. o. ~ssenfoss

Ordnance Engineering Laborator;r, .Ballistic Research Laboratories,

Aberdeen Proving Ground, Hl.ryl.and

ABSTRACT

Previous studies of various metals as shaped charge liners

are summarized. The need. !or additional stuey o! metallurgl.cal
aspects o! liner performance is indicated.

Copper liners or various hardness vera fired to study the

ei'!ect o! the variable upon stand-o.fi'-penetraUon curves. These
tests indicate that annealing within this recovery region or grain
growth region will improve the penetration performance o! copper
liners. No improvement will be observed when copper liners are
annealed. within the recrystallization region.

A proposed iuvestigation is outlined to study the e!!ect o!

second phases and their distribution upon liner performance.

The SiiiiPle Jvclrod;,ynaml.c theor;y o! jet !orDBtion from metal cavity liners as pro-
posed b,)r Birkho!£1 Pugh, M!lcDougal and Taylor (l) has considered only density as a
liner pal'llllleter. Deviations in the per!orDBnce o! metal cavity liners !rom pre-
dictions o! this theor;y have indicated that other parameters DBY be involved.
The search !or these other parameters has in a large measure 'been unsuccessf'ul..
other properties which have been considered are; boiling point, melting point, (2)
ductility, tensile strength, and hardness. Experiments involving studies o! these
properties have not been tc;>o successful in explaining the deviations !rem predicted
performance. ·

Among the more comprehensive investigations ot the performance of shaped charges

are those conducted b,)r the DuPont Company (3)(4) using 1.63" diameter liners. These
investigations report stand-o!! penetration curves !or liners o! various cone angles
and wall thicknesses. The DuPont data will be summarized below.

Steel

Figure ~ shows the ei'!ect o! various wall thicknesses on the stand-of! penetration
curve !ar 80 drawn steel liners. These curves indicate that maximum penetration
ocsurs at approximately 3 to 4 diameters stand-ott and that the optimum thickness !or
80 drawn steel liners is apprax::l.mately 0.037" - 0.048".

In Figure 2 the effect o! various cone angles upon penetration is studied. These
data show that maximum penetration occurs at 2 l/2 cone diameters. Examination o!

43
 1 F ?PIE' Mil} Iki , ...
T

10" STEEL LUERS

l.ea• DIA.
• 0.027 c 0.080
• 0.04& 0 0.088
A 0.080
oL------7------~2 --_l--~.------~.r------;._jL-__06~-----i,------1.

Figure 1-80° Steel liners, 1.63" dia.

•
0

i
I.. •
Q,

STEEL LINERS
L&3 DIA. ,...,..---
-------------e..__
0
____ ":"~
2
• eo• 0.043

• 100" 0.0150

4 120" 0.068

0 140" 0.061

•a ••• 20"
0.037
0.024
2
• 4
STANDOff
•
t INCHES)
• T 8
•
J

Figure 2--Steelliners, 1.63" dia.

44
 zemsaa a s aM z a

these data shows that the optimwn cone angle !or drawn steel liners is IJSI.
The curves shown in Figures 1 and 2 present the overall picture or the etfecta or
cone angle, and wall thiclmess upon penetration or steel liners. However, upon closer
examination these curves do not present the complete picture when observed from the
standpoint of a metallurgical selection of a liner naterial. The first question which
arises is 1'What steel was used"? The second question to be raised is "What is the
condition or the steel?"

In answer to these questiollB, DuPont (4) in 1944 conducted evaluation tests or a

number of deep-cirl!wing steels used !or the manufacture of metal cavity liners. Metal
cavity liners, 16° cone angle, lo63" diameter were nade from various steels and tested
in !our condi tiona&. as received, annealed lOOOOF, annealed 1250<?, and annealed 16500F.
Liners DBde from Armco Irons appeared to be superior in all conditions tested. And
this iuvestigation recommended that future steel cones be equivalent or Armco Iron.
These data IBl'e presented in Table I.

What is the expl.anlltion !or the apparent superior per!ornance o£ Armco Irons?·
Armco Irons are essentially single phase alloys (ferrite with iron oxide inclusions);
that is, the carbon is in solution. The other DBterials tested are two phase alloys
(f'errite and cementite); carbon is present in sufficient quantity to .form the component
Fe c Which precipitates as a second phase. If the collapse of the liner, and jet
3
!ormation are related to the .flow characteristics or the IIBterial, then the presence of
a second phase nay become extremely important.

- Aluml.num
DuPont has iuvestigated spun aluminum .for use as 1.63° diameter liners. Figure 3
shows the e.f!ect o.f wall thicknesses upon the stand-o.f.f penetration curve .for 450 spun
aluml.nwn liners. These data show penetration to increase With increasing stand-ott,
With an apparent DBXimwn at 5 cone diameters. Also it appears that the optimum wall
thiclaless .for spun aluminum liners is approxiDBtely 0.035°.

The e.ffect of cone angle upon tha stand-of£ penetration curve is shown in Figure 4.
Examination or these data indicates that maximum penetration occurs at approxii!Btely
5 cone d:l&meters and that the opti11111111 cone angle tor spun aluml.nwn cones is approxi-
mtely )OO - ~.

Since alumi.IDlDI is a work hardening naterial, it would be very interesting to study

the e.t.rect o£ working upon the stand-ott penetration curve for variO\IB cone angles and
wall thicknesses. Ma.ey aluml.nwn allo;ys are o.f the precipitation type and a stud;y of the
e.f.fect o.f age hardening upon penetration could also be very interesting.

Copper
These curves show that copper is ~rior to steel and aluml.nwn and apparently bas
its DBXilllwn ef.fect at approxiDBtely 4-l/2 cone diameters. It is interesting to note
that While the penetration or alumi.DUDI continues to increase with an increase in stand-
o.r.r, Figure 3 shows that even at 6 cone diameters, the DBximom penetration is less than
that observed for steel.
 1.63" STAND-OFF 45• .025" WALL THICKNESS TABLE I

ANNEALED ANNEALED ANNEALED MATERIAL

AS REC'D.
1000" F 1250° F 1650° F
I
AV. DEPTH AV. DEPTH AV. DEPTH AV. DEPTH
LOT NO. c MN p s Sl CR Nl
(INCHES) (INCHES) (INGHESl (INCHES)

lA 5.44 5.20 5.79 5.47 .022 .055 .003 .029 .010 .013 .029 TO CAN IRON

18 5.60 4.82 5.42 5.49 .025 .055 .006 .OZT .006 .010 .033 TO CAN IRON

2A 5.76 5.49 5.97 5.24 .081 .36 .008 .033 .011 .005 .025 RIMMED STEEL

28 5.48 5.80 5.47 5.19 .075 .36 .008 .028 .002 .025 .024 RIMMED STEEL

4 5.38 5.57 5.45 5.81 .13 .33 .008 .034 .010 .017 .020 KILLED STEEL

5 5.85 5.94 6.10 5.20 .17 .33 .0 II .036 .012 .050 .024 KILLED STEEL
ARMCO
6 5.93 !1.85 5.72 5.63 .037 .022 .004 .026 .Oil - .025 ENt.rlij-ING
7 5.90 5.79 6.00 5.65 .030 .048 .004 .035 .010 .003 .010 ARMCO IRON
ARMCO INGOT
8 6.10 6.00 6.44 5.59 .038 .028 .006 .023 .022 .003 .026 IRON
9A 6.24 5.74 5.73 5.48 .052 .46 .008 .018 - .06 - UNKOWN

98 6.31 DE CAR 8 URIZED LINERS .085 .46 .007 .018 - .04 -

 C&as t s&!diiL§ !iiibl2!5 -

Q
~
I
I
I
I
I
I

I
I "'
I
I
I
\i CD

II
Ul
...z
It:

-'

o!
:a
CD .,'
"l
....
...
iii
l: 5
0 :§
z
on
IL
~
IL
0 ]
0
z <
•
~
. Ul
l"'
~

...-~
...

.... on N 0

( S3HONI) NOil'd!ll3N3d

u7
 TJWF'T --ucit§ liitOi

,x;<- I /
~-

~
I
"iS
ia
~:;;ssa
0 0 ...
c:iciooo
I
~

/f\
D ~
.._,
12
;!;

o
I

..
ii~~~
s;;~.,c
1 1 1 I I
1 I I

~
..d!
I

;;;
"'~
•
~
,;
:a
.....

1t. \ .. :.,;sf
~
-
..10 JI.. <IQ)CO •
~

0
~ "
lf
\ I \ \ \
.,~
~
'\~~
\ \ N

0: " "'"
~~'
~
\ '
w'c, '~
~ ~1
~
• a .. N ~
(S'3HONI) N01l'IH13N3d

•
l ~

..
:a
.2s:a
1
. . ..
1 I
:,
I •
.,•
"!
.....
" I!
I- :!
"'g
• . ·§
~

0
I.
• a
" 1
[
i>::
N

(c::IHONI)
.. NOilWI:IllN:Icl
.. ..
41l

3Ji !SELl I
 TABLE n
CONE CONE WALL STAND·
TYPE PENE.
DIA. ANGLE THICK. OFF

A 2.0" 45° 0.048" . 2.0" 3.90" TYPE A

2.10" DENTAL AMALGAM

9 2.0" 45° 0.048 • 2.0" 3.9" TYPE B

3.8" 50% CD., 50%HG.

c 2.0" 45° 0.048" 2.0" 5.5 " TYPE c

5.1" 45•/o CD.
47'11. HG.
A 2.5" 45° 0.048" 2.0" 4.2" 2.5'11. PB.
4.7" 5.6°/o SN.

s 2.5" 45° 0.048" 2.0" 5.1"

11
4.9
.••
c 2.5" 45° 0.048" 2.0" 4.3"
11
4.6

TABLE m
STEEL PURE COPPER COPPER + 0.1"/o AG.

PENETRATION PENETRATION PENETRATION

RD.* RD.* RD.*
MM MM MM

I 149 6 207 II 206

2 160 7 200 12 214
3 179 8 217 13 211
4 165 9 205 14 200
5 161 10 192 15 198

AV. 163 203 207

L9
 , cam 62£22§ I

Copper is now generally accepted as the most suitable material for the manufacture
of metal cavity liners. Figure 5 shows the stand-off penetration curves for various
pure metals. The liners used to obtain these curves were 45 11 cones having a base
diameter of 1.6311 and a wall thickness of 0.037 11 •

Copper, like aluminum, is a work hardening material and studies of the effects o:r
working upon penetration may prove enlightening.

Other Material

In addition to copper, aluminum, and steel, the pure metals zinc and lead have been
used as metal cavity liners. Stand-off penetration curves for these metals are shown
in Figure 5. The curves for these materials show them to be inferior to copper and
steel.

Various allc.ys have been used as liner material. Results of firing tests for these
materials are recorded in Tables IT and III and also in Figure 6. The penetration data
for the dental amalgam and cadmium liners (5) when compared with data for the smaller
diameter liners in Figure 5 show these materials to be inferior to copper and steel.

Figure 6 compares the stand-off penetration curves for copper, silver, red brass,
and steel (5). Since these curves are drawn from only two points, the information pre-
sented is only qualitatively useful. It may be seen in this figure that at all stand-
offs, red brass appears superior to steel, and copper and silver appear better than red
brass.

Table III compares the penetration data for steel, copper, and copper - 0.1% silver
(6). These data indicate that the copper-silver alloy is as good ILB copper. This is
the first indication that alloy liners perform as well as pure metal liners. Examining
the phase diagram for this system it may be shown that the solubility of silver in
copper is approximately 1% at 300°C; hence, this alloy may be considered a single phase
alloy. Previous tests with alloys have been conducted with compositions within the
multiphase regions. studies of alloy systems h;iving a large solid solubility range at
room temperatures should be extremely interesting. A study of such a system (copper-
al1DIIinum) :i.s now being undertaken.

Hardness

Hardness is a variable which has been studied to see i f it has an effect upon the
stand-off penetration curve. .In Figure 7 a portion of the stand-off penetration curves
for drawn steel liners of various hardness are shown. These data show that the softer
material gives deeper penetrations. Annealing of copper liners has been tried. Data
have been obtained with these liners and have shown no significant improvement in pene-
tration; however, the spread in penetration obtained appeared to be reduced when
annealed liners were used.

Recently in the course of other experiments a large difference was observed between
two lots of coppEir liners supplied by one manufacturer. These were identical in all
respects with the exception of the geometry of the fiange at the base. This difference

50
 9 c

•
8 6 r

• " --- -- --· -

..-- ---- ---
--- X

___
5

.----------
6
---" ,._
;;;
"'"' 4 o-_
l!
fj,_ ----
z
5
"':z:
ILl ---~
.
2
1-
0:
1-
0
z
o------
3
~
.."'"'
z 4 Z'
0
....
<(
a:
1-
3 ILl
z 2 ~
ILl
CL
45° STEEL LINERS
1.63" DIA.
2

45" LINERS 1.63" DIA. I ~ • AS REC'D. 65 - 72 B

X CASE HARD. 29- 33 c
• S ILV[R 0.039
t; 38- 43 c
X RED BRASS 0.034 .
0 47- IIIC
A STEEL 0.050 0 57- 59C
0 COPPER 0.040 .I

6
0 2 3 4
2 .3 4

STAND- OfF (INCHES} STAND- OFF (INCHES)

Figure 6---45" Liners, 1.63" dia. Figure 7---45 • Steel liners, 1.68" dia.
 TPWW? Edt; una: 7

Figure 8--Sketch showing cross section of two lots of copper liners-

sharp cone shown on left-round cone on right.

RIM
4
• 8 10 12 14 16
APEX'

Figure 9-Variation in hardness of two Jots of eopper liners. Measurements

made along slant height of liner from rim to apex.

52
a ETP·
in geometey is shown in Figure 8. Tests were conducted in lihich the geometry at the
base was changed so that both lots were identical. These tests indicated that the
rounding of the base could not completely account for the poorer perfomance.

Chemical analysis and metallographic examination revealed no outstanding differ-

ences. Standard Rockwell "F'' hardness measurements revealed a five point difference
in hardness between the two lots as is shown in Figure 9. While it was felt that this
difference in hardness was not necessarily responsible for the large difference in
performance observed, it was deemed advisable to initiate a study of the effect of
hardness upon performance. of copper liners.
The hardness program was conducted with 105mm 45° copper liners with spit-back
tubes. In these tests, liners supplied by a contractor were heat treated to various
hardness and stand-off penetration curves were determined. Figure 10 canpares the
softening curves for various reduction in area of cold rolled copper. Commercial
cold drawn copper bar stock sufficiently large to machine 105mm liners has been sub-
jected to approximately 10% reduction. These curves indicate that three phencanena
are occurring during the annealing of copper. (1) recovery (or stress relief), (2)
recrystallization and growth, and (3) grain growth. During stage 1 the working
stresses are relieved with .no apparent change in properties as may be shown in Figure
ll. In stage 2 recrystallization occurs. In this process, the more heavily deformed
grains break up into smaller new grains in a lower energy state having a somewhat pre-
ferred orientation. These newly formed grains grow at the e:xpense of the less heavily
deformed grains. During this stage, there exists a highly unstable energy state with
maey different orientations. This may be shown by the hardness measurements shown in
Figure 12. In stage 3, migration of the grain boundaries occur as the material seeks
to attain a m:l.n:imum of orientations. Given the proper time and temperature conditions,
it is possible to obtain a single crystal. with a single preferred orientation. This
uniformity of the energy condition and somewhat preferred orientation of .the grains mq
be seen in the hardness curves of Figure 13.

By annealing at 300'T, 6oo~ and 900~, liners in each of the three conditions
described above were obtained.

The hardness of the liners was measured on a standard Rockwell tester using a
special fixture as shown in Figure 14. Normal variation in hardness was observed in
the liners heat treated at 3QOOp' and 900CIF. Erratic hardness was obtained with the
liners annealed at 6oo0po. The e:xplanation of this phencanenon is that during the cold
drawi.ilg operation, uniform working of the bar was not accomplished. Therefore, from
the curves in Figure 10 we see that in the more severely worked portion of the bar,
complete recrystallization and partial grain growth hne occurred, while in the less
severely worked portions of the bar recovery and partial. recrystallization have occurred.
Stand-off penetration curves were obtained ·by firing five liners frCBII each heat
treatment at each of three different stand-offs. These re81ll.ts were compared with data
obtained with nnannealed liners at the same stand-off.

53
 ) &diu nr tWIG Iili&liiili&i

200 400 600 800 •c

......, REouc'T oN 'ey ROLL NG;
r93-to
,~
(/)
Q. 140
rso•""
0
0
0 ,,. v----zs•.4

(/)
. 120
(/)
11.1
z ~1o•""
0
a:: 100
C( 3 .,.
J: ~
.J
.J 80 \
11.1
z
a:: \ '
'
CD
60
0 400 800 1200 1600
ANNEALING TEMPERATURE, °F.

THE DECREASE IN HA.RONESS PRODUC.ED BY THE

ANNEALING OF COPPER SHEET THAT PREVIOUSLY
RECEIVED DIFFERENT REDUCTIONS BY ROLLING.
(KOESTER)

Figure 10--The decrease in hardneBB produced by the annealing of copper sheet

that previously received different reductions by rolling.

54

$2 !LIZ&!ia:M
 _,

''

80

,.

.. •o ii
•
"
• .. ~
• "

,....
~
i0
•
~
. ""
I• . ~
•
..
V'l
V'l ..
I
;;

•'
a'
••• "
0

••
""
••
~
..
"
""
!
. 20
i!

II 20

........ . ,
;.
0
0 01' LINERS
• 00
tv.RDNE"SS
'AMNI!ALED 600° f' 00
AS Rf:C(IVEO UIEIIS

....
MI. .AUO 100°' LUilRS
~
•
".... DlnAliCE ( INCKESI
.... ....L-----------~----------~--.--:-,-----.....
DISTANCZ (IICMESI
.... •
DI&TMICE (UICHESI

Figure 11-Hardness of as reeeived liners Figure 12--Hardnesa of liners Figure 13-Hardness of liners
and annealed 300' F liners. annealed 600' F. annealed 900' F.
&sass ' 5 I
:' 7

56
 88£ FFNTTOJ Ss ·· , t"e

The resul.ts of these tests are shown in Figure 15. Here we 11ee that when copper
is annealed either within the recovery- region or the gr'ain growth region, the stand-off
penetration curves are essentially identical. A statistical comparison of these data
with as received data reveals a significant effect of annealing. These differences in
penetration are significant at a 0.05 level, this is to say that there is only a 5%
chance that the penetration data for the annealed liners came from the same population
as the received liners.

However, we have not considered the curve for the 6oo'7 treatment. These linere
are e:xpected to be in some intennediate state o:f strain. In this region, there exists
areas of high strain which attempt to reach a lower level of strain. These nuclie cause
the original gr'ain to break up and smaller gr'ains are formed which gr'Oil at the expense
of the larger gr'ains until the material has completely recrystallized after which nonl&l
gr'ain gr'Oilth begins. The liners treated at 6oollF were in this state of unbalanced
stress, and the poorer results are for the present attributed to this condition.
We cautiously conclude that i f work hardened capper liners are annealed within the
recover;y region or gr'ain gr'Oilth region (which for practical purposes are strain-free
regions) an improvement· in penetration will be obtained. If the liners are annealed
within the recrystallization region, no improvement in penetration appears to be ob-
served. These results need further verification before they can be considered con-
clusive.

Multi-Phase AlloY

A systematic metallurgical investigation to study the effects of second phases

and their distribution upon liner performance is being conducted. The capper-aluminum
system was selected for this study. In Figure 16 is shown the copper rich side of the
cappar-aluminum phase diagr'sm. From this diagram we see that with a :few compositions
it is possible to study several single phase regions and several combinations of two
phase regions. In these studies sufficient liners of each phase combination will be
obtained so that a stand-off penetration curve may be plotted.

Conclusions

It seems likely at this stage that aQy metalurgical investigation of material for
use as metal cavity liners should consider the following as variables; (1) Cry-stal
structure, inasmuch as various structures lend themselves more readily to slip or fi011.
(2) The degree of cold work, since this has an effect upon the yield strength of the
metal. This is particular~ true for the work hardening metals such as capper and
aluminum. Grain size might also be considered as the now characteristics of some
metals at various temperatures are influenced by gr'ain size. The presence of two or
more phases should be considered as it is known that the presence of a second phase
often hinders the flow properties of metals. 'lhe usual mechanical praperties such as
tensUe strength, hardness, and ductility are inter-related and may be varied within
limits; however, these properties themselves are dependent upon the fundamental prap-
erties; crystal structure, and the number of phases present, and they also depend on
the degree of working.

57
'

-·· ·~- an a
 ,5777 17 I i&ilQ iili& if& 4

...
!
L
•g N

.."".
~
c , r N
N
z I
z I
c

I
"c
Oi

0 0
N
\
I
!!!

l\
z
z I
!!

\
c
I !::

"' ;;;
..,'l!
~
!!
......
...
....
0
I!
:!: ..
c'
z .;fZ.
:;;

!!

00
~
•
.."
~

:1z
li =
!!

..
.
..

(S:JHONI) NOU'WHUN3d

58
 Wib!LL&t

L /3+L

r
13
+
r
a
+
a 13

565°
15.6
9.4 11.8
a+y2
500
ou 5 7.6 10 12.5 15 17.6 20
WEIGHT .,. ALUMINUM
Figure 16

59
 SEEMIE§ Iht&nu · -
J.E
'

REFERENCES
'
1. Bil'khoff, OJ McDougall, D. P.; Pllgh, E. M. and Ta,ylor, G. I. - "Explosives with
Liner Cavities" 1 J. App. Phy. !21 1946, pp 563-562.

2. Lennard-Jones, J.E.; and Devonshire, A. F. - "Same Notes on the Theory of Hollow

Charges", A. c. 2934.

3· DuPont - •Investigation of Cavity Effect, Final Report•, 18 Sept. 1943.

4. . DuPont - "Evaluation of Deep Drawing Steels for Manufacture of Purity Charge Cones• 1
6 Dec. 1944.

5. National Defense Research Council - "Interim Report on Shaped Charges" 1 1945 NDRC/
Dil'. 8 SC-l-SG-21.

6. Nash, T. and Evans, w. M. - "A C0111parison Betwe.en Steel, Pure Copper, and Copper/
Silver Alloy Conical Liners f'or Hollow Charges• 1 ARD/Expl. 406/46.

60
 12211 ii&MMil§ iliib!i ·c

GENERALIZATIONS CONCERNING THE MOTION OF A THIN SHAPED CHARGE LINER WITH AN

ARBrmARY INITIAL CONTOUR
+
G. E. Hudson
C. Gardner

New York University, New York, N. Y.

+Portions of this report were presented at a Sym-
posium on Shaped Charges at Aberdeen by G. E. Hudson; the
report was written by Hudson whereas much of the mathe-
matical background for the theoretical work herein dis-
cussed grew aut of conversations between Hudson and other
member or the Institute for Mathematics and Mechanics,
but particularly between Hudson and Gardner.

ABSTRACT

Upon assuming a thin liner whose particles do not exert 8IIT

forces on one another, and upon treating it as an incaapressible
fluid during its motion, general equations or motion and continuity
are derived. It is shown that these are equivalent to a Schrodinger
· t:lme-dependent type of equation with space and time variables inter-
changed. The impingement and subsequent extrusion of this liner at
an arbitrarily moving and oriented surface element are investigated.
This fOl'IIIUlation suggests several new problems, experimental,
theoretical, and mathematical, as well as the importance of i'urther
investigation of saae already considered. In the latter class are
the problems of the interaction of simple waves of finite amplitude,
and the effects of compressibility in the liner. Friedrichs, Keller
and A. Lax, and then Touart have considered the possibility of rare-
faction and compression shocks in the liner, and have thus justified
the asSIIIIIPtion by G. I. TSTlor that the liner acts as a continuously
turning stream of non-interacting incompressible fiuid particles.
Touart and Friedrichs have also shown that the observed mass distri-
bution in the jet can be accounted for by the continuing pressure
exerted by the burnt e.:xplosive on the liner as it collapses. In
addition, Touart has shown that the probable effect of the rare-
faction wave in the bu.rnt explosive as it e.:xpands into the atmos-
phere is the observed decrease in velocity fraa head to tail of the
jet. Thull the external shaping of the e.:xplosive is important in
stand-off effects.

i
 p fWPF75
1
I 7 a; ii2I
•
I, INTRODUCTION

The theories of the liner collapse and jet extrusion processes which have been
presented in the past (l) (2) (3) 1 all have the COMDIOn feature of considerin~ the pro-
cesses as steady state ones when viewed from the proper frames of reference, However,
it is known that the actual processes are not steady ones -- and that the effect of the
unsteaciiness is paramount in determining such important phenanena as the mass distri-
bution and velocity distribution in the jet. In this paper an attempt is made to re-
formulate the problem to include non-steady motions in a logical manner.

For orientation purposes 1 let us review briei'ly SOiue of the points of view in the
previous attacks on the problem,

The theory of Birldlo!! et al, avoids the di!ticult problem of determining the
pressure on the collapsing liner by means of the assumptions of an impulsive velocity
!or each liner particle 1 produced as the detonation front eweeps by, and thereafter a
uniform rectilinear motion independent of the other particles, until extrusion takes
place, In this way, questions of material strength are also avoided almost canpletelJr
and the mathematical theory can be given a very s:lmple !om indeed. These consider-
ations are considerably generalised by the work o! Friedrichs, Keller, A. Lax, and
Touart. By asSl1111ing a steady liner motion in the "detonation frame" 1 the expansion of
the gases fran the burnt explosive against the collapsing liner is treated as a can-
bination shock and simple wave problem (in two d:lmensions) so that est:lmates of the con-
tinuing pressures on the liner can be made, It is clear tha.t the assumption o! a steady
motion precludes the exact treatment of the case with cylindrical symmetry, and avoids
the problems connected with the interaction of the rarefaction wave 1 due to the ex-
pansion of the explosive into the external atmosphere, with the expansion (and can-
pression shock) wave spreading fran the liner,* In this theory, the effect of wave
motion in the liner material itself is considered, and is shown to lead to a justilica-
tion of G. I. Taylor's assumption that the liner particles do not interact with ens
another.

In· the present paper we are naturally intluenced strongly by the above investi-
gations 1 both by their achievements as well as their shortcanings. To aid in ac-
complishing our purposes here; namel;y, to include non-steady motions, and to make a
beginning in the st~ o! the stability o! shaped charge phenOIIlena, we shall intro-
duce certain compromises and limitations: the liner will be treated as an intini-
tes:lm~ thin, continuous collection of particles, as in reference (2), whose strength
properties, considered to a certain extent in reference (l), can be neglected. '!'bus
the importance of thickness and strength of materials both in experiment and in making
a proper theoretical formulation o! the problem will be emphasized in a negative way.
The problem at determining the pressure exerted by the burnt gases (and the atmosphere)
is not considered here directl;r, as it is in reference (l), but the presence or such

* C. Touart and Friedrichs have generalized the Friedrichs theor:y slightl;y by" intro-
ducing 8 quasi-steady state processes" but the error introduced by" such a trea"bllent
has not been estimated. Pugh and Rostoker have also made similar generalizations.

62

EFT.
 continuing pressures is taken into account, The f'o:rmulation presented herein is limited
to a consideration of' two-dimensional motions (generalized wedge-shaped charges) onlJr;
it is possible however to extend it, rather easily, to include cylindrically symmetric
'- motion.

We have been motivated in this formulation by the desire f'or mathematical

simplicity and by the possibility of deriving descriptions of liner motions more nearly
agreeing with observations, Yet simplicity, generality, or agreement with observations
are not the only motivating influences, It is felt that by pursuing the· problem in the
direction we have taken, important engineering que:.tions may ultimately be answered,
In this category is the question, "Given a jet having certain characteristics (e,g,, a
certain velocity distribution, or density distribution), how may one obtain an initial
liner and charge configuration that would produce this jet?" Another important question
is, 11 If' the initial configuration is altered slightly, what will the effect be on the
jet?"; in this paper we shall be much concerned with an initial exploration of' this
phase of' the problem,

II, DERIVATION OF THE EQUATIONS OF MOTION OF A LINER SECTION

A. Coordinates and Initial Conditions

Consider a shaped-char~ and liner configuration which is perfectly uniform in one

direction ( z) • In a plane ( z • 0) which is perpendicular to the direction of' uni-
formity introduce rectangular Cartesian axes (x,y). Consider a section of the liner of
unit width included between the planes z = 0 and z " 1, and the motion of the plane
trace of' this section in the plane z • o. · This trace may be considered to be made up
of' a continuous one-dimensional array of' mass particles having a linear density dis-
tribution y(x,y) (mass of' section per unit length along the trace),

Let "a" be the Lagrange coordinate of the particles of the trace; it is defined
as the mass of the section measured f'rom one end of' its trace (a • 0) up to the
particle in question, Then

a•J'yds (1)
where ds • J 2 2
dx + dy is the element of arc length, and the integral 'is taken f'ran one
end of the trace of the liner section to the particle in question, We shall always
assume that the position of' each point mass particle is completely determined by a
specification of' the Lagrange coordinate "a" and the t:lme 11 tl'; this is what we mean by
stating that the liner is infinitesimall,y thin. If' we denote the position vector to a
particle of' the liner by the cauplex number

V.. x+iy, (2)

then evidently a choice of' the complex-valued function 1j1 (a,t) completely specifies a
motion of the liner trace.

6j
 --------------------------------------

rm SCl& it§ Iiii

It is of :interest to note, as :l.n Figure 1, that the argwnent, 9 sq, of "4f a*

specifies the angle the trace at each point makes with the x-direction, while its
modulus

is the specific length of the l:l.ner section (length per unit mass),

Let us suppose that a pressure difference, acte across each element of the section
at time t; this pressure difference, multiplied by the unit section width :l.n the z-
direction will be denoted by PJ \p\ is the magnitude of the force per unit length
(l:l.near pressure) acting on each arc element ds of the trace. Clearly the force acting
on such a trace element is perpendicular to it; if p is > 0 when the force is in the
direction specified by a counter-clockwise rotation of 90° (i.e., multiplication by i)
fran the direction of :increasing values of a, then the force on an element mar be
denoted by the complex Jllllllber

ie
i9 p ds.

This DlllSt equal the mass, da, of the element times its acceleration, according to
Newton, Dividing this equality by da we see that
(3)

is the partial differential equation of motion of l:l.near trace elements.

In this equation, p would ord:l.narily be determined :l.n terms of the pressure ex-
erted by the detonation products next to the element :1.n question at tim.e t; that is, we
should :l.n general regard p as a real-valued .function of 'tlr, its derivatives, and t, It
should be pointed out, in passing, that (3) has the form of Schrodinger 1s time dependent
equation, with, however, a zero potential energy, the time and space variables inter-
changed, and a variable ratio of Planck's constant to the mass of the fundamental parti-
cle, As we shall see immediately, the type of :l.ni tial-value problem which 1i8 consider
in this report also differs fundamentally fran that considered in the quantWII-mechanical
-_case.
At time t • 0, the liner has some shape an4 sane velocity distribution specified by
l(a,O) s X(a) + iY(a)
(4)
1Vt(a,O) • U(a) + iV(a)
Until the pressure function p is specified it is clear that the problem is not yet
completely formulated, However there are certain cases of interest, some of which have
been considered :l.n the past where we mar specif'y P• In this wq we see that the
ncaucl:W' problem (3) and <4S leads to very :interesting results, which mar be indicative
of certain important practical conclusions.

* Subscripts indicate partial differentiation except where otherwise specified.

64
 e m ma:aw aces ts

iy

iei 9 pds

~~----------------------------------x
0

Figure !-Representation of the motion of· a liner by the motion of

its trace in a complex plane.

65
d&LllE&
 III, SOLUTIONS

It is possible to •cook up" a wide variety of 11 solutions 11 of the system of

equations (3), (4), corresponding to various choices of the pressure function p, and
the initial data. In this way one may hope. to describe motions which are close to
physically observable ones. We· shall consider only those functions ~ to be solutions
which are continuous everywhere, which have derivatives t t• 1jf tt' and fa' and satisfy
equations (3) and (4) except on a derrumerable set of points { ai'ti} •

A. Elementary Solutions

Case 1. p • 0 (rectilinear particle motion)

In this case the 1solution is obviously:

V!"(a,t) • ]t(a,T) + f#"t(a,T)(t- T) (5}

where the data t (a, T), If t (a, T.) are considered to be supplied at time T.

Case 2. IL: constant (central elliptic motion about the origin)

Let the initial data be

2n~
vr<a,o). Re , (0 !i a :S m)

2n~
ft(a,o) • iPRe

where m is the mass of the section; hence the liner particles form a closed circle of
radius R and are moving initially with uniform angular speed m about the origin, tan-
gentially to the circle. The solution satisfying these conditions is,
a
)t(a,t) • R (cos t
2
J ::!'
+ :I.Cil sin t Jiir,2
Jfi> e n~
(6)

which represents a circle of particles whose radius~c · ates fran R to RCil 2~ and J
back with a period one half as long as the period, ~. which it takes a particle to
P 2 n
make a eanplete elliptical circuit about the origin. If m '" 2"\ii• the solution is
i(cot + 2n!)
1f• Re m (6a)

which represents a uniformly rotating ring of particles having a constant angular

velocity Cll and radius R. If p is negative 2n is changed to -2n, then the order in
which the particles are arranged in the circle is reversed, the pressure p which
formerly was exerted inward is n011 outward according to our convention, and the motion
is no longer periodicJ this is borne out by noting that the trigoncmetric functions in
(6) beccme 'hyperbolic.

66
 t aa II 5 :r· 7 rne

Case 3. p • constant (analytic initial data)

The above solutions are particular cases of the somewhat more general one in
which tbe initial data have derivatives of all orders, which are Bilfficiently bounded.
Then an expansion in a power series in t yields the SYJ!Ibolic form
sinh t JipDa
,-ca,t) • (cosh t / ipD8 ) jl"(a,o) + ( IIPDa ) ---t(a,o) (7)
d
involving the lzyperbolic operators containing the differential operator, Da a da'• in
their arguments.

It wi11 becane evident later that, unless the initial data belong to some such
special class of f'unctions as the analytic one considered above, there actually wi11
exist no solution for constant p values. Physically, this means that a liner with the
characteristics we have assumed would almost always i'ly to pieces, except in certain
cases which are actually in practice unrealizable. This fact emphasizes the statement that
the present theory is by" no means a canplete one - it must be supplemented by further
investigations or the dependence or p on the time and liner position and shape, and of
the strength characteristics of the liner. Nevertheless, as we shall see, the fol'IIIU-
1ation m~ be used heuristically to generate solutions which describe motions approxi-
mating in some degree those actually observed.

Case 4. p • i'(a)P (analytic initial data)

In case the pressure depends analytically on the mass parameter "a11 only, and
the initial data are analytic, the problem may be reduced to the previous case. For,
set

b •f ~Ca)' so that a • a(b) (8a)

Then we DIUSt solve

(8b)

subject to t;.he given initial data

1f(a(b),o), (8c)

B. One Dimensional Motion

If the velocity l'"t and the tangent ¥"a

are proportional, with a real proportion-
ality factor >.(a,t), then all particle trajectories .coincide with a single liner trace;
the motion is not ne-cessarily stationary, although this case is included. In particular
we shall see that the elementary solution given by Birkhoff, et al, appears as a special
case of this kind of motion.

Roughly speaking, the- specialization

(9)

67
 1! 0

:implies that one may have given only a limited portion of the data arbitrarily, such as
the initial distribution of the liner trace and the initial pressure distribution --
then all other quantities will be detemined. Let us see how this canes ab011t.

If one introduces (9) into equation (3), it is easy to deduce that, because of the
reality of p and l.,

f(a) l.tt • 0
(10)
~b) (pl.2\ • (pl.)a

provided the trivial case in which the initial velocity distribution vanishes is ex-
cluded.

Equation (lOa) shows that l. must be a linear function of llt11 with coefficients de-
pendent on 11 a 11 • If one puts

ht .. pl.

2
ha • pl.
then (lOb) is satisfied and h must satis.fy

(lOc)

an equation whose characteristics (along which h is constant) are the same as those of
(9). If g(a,t) " const. is the equation of any characteristic, we see that we may take

g(a,t) .. R(a) + S(a)t (ll)

while { l.(a t) R' (a) + S' (a)t ga
' " S(a) • ~
where R(a) and S(a) are arbitrary !unctions of "a" (to be detemined perhaps fran
initial conditions). Then since h is an arbitrary !unction of g, insofar as equation
(lDc) is concerned, we see that the pressure p must have the form

~
2

p(a,t) • h 1 (g)
ga
or: 2
p(a,t) ., R'(a~ ~a~•(a)t h'(R(a) + S(a)t) (12)

where h 1 denotes the derivative of h with respect to·its argument "g". It is easy to
show that initially we must have

ll't (a,o) • S(a)eih(R)

(13)
{ V'(a,O)• ~a R•(a)eih(R)da + X(O) + iY(O)

68

53!!2 db hAL
whereas at aey later time the solution tis simply written in either of the equivalent
forms as

)t(a, t) " (14)

.
(R 1 (a) + S 1 (a)t)eih(g)da +I:t
s'(o)eih(R(O) + S(O)t)dt + X(O) + iY(O)

[ a R'(a)elh(R)da + [ t S(a)eih(g)dt + X(O) + iY(O) '..

It is quite clear now, that an arbitrary choice of the functions R, s, h (and the .
constant I(O) + iY(O)) cCIIIpletely characterizes 0118 diJJiensionalJDOtion {llllli the pressure
acting I). On the other hand, one .lllight choose a pressure function in the form of (12)
which would similarly serve to determine the appi'Qpriate initial. conditions,· and· the
motion, up to a multiplicative constant and an additive constant. Or we might have
given 'lj'(a,o), in which case R'(a) (and hence R(a) or a(R)) and then h(R(a)) as a
f'lmotion of nan (and hence h as a f'lmction of R) will be detem:lned (except for a con-
stant added to R). '!ben i f either p(a,O) or the initial speed distribution S(a) are
given, all other quantities may be frund. Examples of 11 solutions"* of these kinds are
given in the next section. ·
c. Canposite Solutions

We are now in a position to write down several solutions which may correspond to
sene extend to the actual motion of a collapsing shaped charge liner. 'lbeir chief
interest perhaps is that they may serve to collate certain e:xperimEmtal data. One of
these turns out to be the well-known Birkhoff approximation in which the particles of
the liner in succession receive impuleive accelerations by a very large pressure pulse
(detonation wave) acting for a very short time. .
Case 1. Birkhoff motion
Let us assurue that in a frame of reference whose origin coincides with the
detonation front at the point 'llhere it meets the liner 1 the collapse process appears
as a steady one, with the liner material of linear density y, fiowing into the origin
u .
at a speed S " cos a• where U is the detonation front. velocity, a is the angle the
liner makes (at the origin) with the positive x-axl.s (half-wedge angle).

. The liner particles change direction of motion as they are accelerated by the
pressure pulse 1 which we shall treat as proportional to a Dirac 6-f'unction whose peak
is situated at the origin.

If at time t • o1 particle a • o vas at the origin, then at time

a (lSa)
t -~

* "Solutions" appears in quotes here since we are considering p as a function of "a"

and "t" instead of the JDOre natural arguments x and Y•
69
particle 11 a 11 is at the origin and is being given an impulse, -I, per unit length by the
linear pressure

p " -I6(St - .!) (J5b)

y

Guided by the fact that the function h is to be dimensionless, we choose

g '" St - ~' R(a) • - ~

(15c)

\
h1 (g) " ~ 6(St - .!)
yS y

in agreement with (11) and the equation preceding (12).

Since
a
-I, t 0! ys

a
o, t < :ys
we have
I
yS + c, g 0!:0
h(g) .. (15d)

c ' g < 0

The additive constants lllllBt be chosen so that initially the liner makes an angle a with
the x-axis at the origin, and the :particles are moving toward the origin. That is,
C • a+ n. Then, according to (13},
ft{a,o) • -Seia
a ia (15e)
f ( a,o ) "-
y e

At aey time t 1!:. o, the description of the motion is given by

(.! - St)eia t ~~
y '
t(a, t) •
a i(a + I )
(-y - St)e ys , t 3: ~
Hence we see that the effect of the :lmpu].sive pressure is to change the direction of

10
 motion o£ the particles by the angle ~· while the magnitude of the velocity is un-
changed• a result which is to be expected.

Now i f we transfonn to the laboratory frame of reference. by adding the velocity

ia
Se to all the particles • then

• ~
-e
ia
• t <
a
ys
y
t= i[a+L)
(15f)

Steia + (~ - St)e y5 1 t ~~
We shall not discuss the Birkhoff theory further, but shall pass on to a generalization.
which "takes into account the aspect that the pressure exerted by the explosive actually
must act £or some time on each particle.

Case 2. (Finite Pressure Pulse)

Let the frame of reference as before be the detonation front frame, and assume
a steady motion, Define the pressure pulse. applied over an interval of the liner. by
t <a
0 ' ~
p • -P • ~ ~t s ~ +T (16a)

o, ~+'l'<t
.
where 't is the t:l:me of application of the constant pressure P on each particle, Now we
may take

g .. St - ~ (16b)
as before while:

0 I g < 0
p
hI (g) '" o!>gsSr (16c)
~·
yS
0
• sr "' g

An application o£ (lh) again, with the same initial conditions as in the previous case.
yields• after again transforming to the laboratory frame of reference:

71
 ·-
f"(a,t) • (l6d)

a ia a
-y e · ,tc:~ f

v;} i(n + a) { ip (t - a ) }
Steia + T e ~ e y.J ~ -1 , ~ ~ t ~ ~ + T

SteiB + .If-2 e i(i + a) { e i~-1} + (~ - S( t- T)) e

i(a + ~)
1 ~ + T "" t

Thill is clearly interpretable (in the detonation fr~) as .indicating that the liner
particl?ars accelerated at constant angular speed, :;s• around a segment of a circle of
radius , and then move off rectilinearly with the same speed S that they had in
. PT
approaching the origin, but in a direction making an angle 'Y!r with the original. wedge
direction. Such a solution is clearly derivable using the most elementary of physical.
concepts -- it is produced here as an illustration of the mathematical. procedure to be
followed in more complicated cases. Incidentally, i f P - ... , T - o , and PT-I,
this solution reverts to the simpler Birkhoff motion.

Thus by choosing various fonns of the function h'(g), the effect of the con-
tinuing pressures acting on the liner can be estimated. On the other hand, by varying
the i'anctions R and s, it might be hoped that effects due to the variable release of
pressure at the outer explosive edge could be estimated. This however turns out to be
fruitless, insofar as one-dimensional motion is concerned, since the initial contour
and velocity distribution depend critically on the fonn of R and s. Indeed as we· also
have intimated previously it is possible for at least two reasons that fmother investi-
gation of our fonnulation, .in the direction we have pursued in this section, maar really •
lead to illusory conclusions. One is that the dependence of p on ¥"iteelf ~and ite
derivatives) has been entirely neglected. Second, the effects of liner strength
properties have also been ignored, a consideration which in part leads to the important
phenomena pointed out in the following paragraphs.

IV. STABTI.ITY OF SOLUTIONS

It was soon evident after formulating the Cauchy problem (3) and (4), that its
solutions might exhibit a remarkable "unstable" behavior, particularly when the
pressure .function p is regarded as a function of "a" and "t". This was pointed out by
C. Gardner, who melttioned that a similar situation arises in the case of Laplace's
equation or the equation of heat conduction, when the problem of finding a solution
is posed in a certain way. (4). Although we have by no means tully investigated this
behavior, it may be illustrated by the following example.

Consider the special solution of (J) 1 with p • ~

const
.:n
•
2
i(t.rt+2n!) A . -2ni n .!
'jt(a,t) • Re .m + .:3 cosh nwt. e m (17a)
n

72
where n iS a very large integer, and A is a constant. Then initia:J.4r
a 2a
2n~ -2nin-
, (a,o) • Re + ~e 111

n
(l'Tb)

which represents a closed curve approldmating as closel:y as we like, for large enough n-
values, the initial circular distribution of the solution (6a). In fact we have merel:y
superimpoeed on. the circular distribution a periodic crinkle -- and since y a also
approldmates as closely as we like the smoothl:y turning tangent vector of the (!ircuiar
distribution, wile the rotational velocity y t has exactl:y the same values, the initial
data for the cases agree as closel:y as we like. Nevertheless, for t "o, the solntions
(6a) and (17a) diverge :f'l'om one snother by as large sn amount as we desire, i f n is
taken large enough. ThiS peculiar situation is characteristic of problems for which, in
mathematical teminology, the solutions do not depend contirmously on the initial data.
According to Hadamard the Cauchy problem (3) and (4), for p • constant at least, is
"not properl:y set", t i,e,, continuous initial data csn not be prescribed arbitraril:y
and be e:xpected to lead to a solution satis.fying (3),)

However, it is hard to escape the feeling, based on pqysical intuition, that the
initial asswnptions made (of a very thin, weak liner material, acted on by a pressure
differential 'Which could conceivably remain constant) are not so radical an approldmation
to the physical situation that no physically significant information could be gleaned
from the theory. When then is the interpretation of the instability exhibited above?
According to the writer's view it means two things: (a) in order to obtain a properly
set mathematical problem, effects of thickness and liner strength must be inclnded, and
(b) the motion of weak liners is an inherently unstable phenomenon 4 in the sense that .
unless extreme care is taken in the preparation of shaped charges ~machining, detonation)
the results obtainable with them will at best exhibit a wide scatter around a mean valne,
As partial evidence for these conclusions we may point to the tendency toward large
statistical fluctuations in the e:xper1mental results, it not too great care is taken in
the e:xper1mental conditions, and to the fact that lead, a weak material, or steel, a
.brittle material, do not appear to yield as consistent results as copper, a ductile
material stronger than lead.

V. CO~ICLUSIONS

In considering the general motion of a two-d:lmensional shaped-charge liner 1 we

have omitted man;r important phases and effects. 'We have not considered the jet ex-
trusion process or the mechanism of fomation of a velocit,y distribution in the jets,
although it is to a certain extent quite simple to treat these questions from the point
ot view of the present fonuulation. We have not taken into account the fact that real
shaped-charge liners are surfaces in three-dimensions (cones 1 hemispheres 1 etc.) and
that irregularities in the third dimension certainly must influence the motion marlcedl:y.
As mentioned before the effects of material strength and liner thickness have not been
included, It they are, as indeed they must be some time for a complete theory, tems

13
which represent the effect of constraints, contributions due to non-uniform thickness
and pressure distributions, and stretching and bending stresses in the liner, must be
added in. Finally the motion of the explosion gases, already investigated to some ex-
tent by Friedrichs, must be taken into account in order to obtain an accurate estimate ,
of the pressure difference p, as well as to complete the fozmulation of the problem.

We have, however, concentrated on certain aspects of the problem with certain aims
in mind. Several heuristically usable descriptions of the motions have been obtained,
and their relations to previous investigations pointed out. The important question of
the stability of the motion has been posed, and in some respects, from the present point
of view, considered -- as a result the importance of treating more exhaustively the
problems mentioned in the previous paragraph is accented. Finally the important engi-
neering question, 11 Given a certain jet shape and velocity distribution, what sort of
shaped-charge configuration will produce it?" can now be considered more fully. For
i f we pose part of this question 11What shaped-charge and liner configuration will pro-
duce a certain liner collapse motion?" then within the limitations of this formulation,
we may giire a partial answer. If' the given collapse motion is
y ... 1'(a, t) ... x + i:y
then x and y must satisfy

xa xtt + Ya Ytt • 0'

and the pressure which will produce this motion is

(1.) Friedrichs, K 0.; Keller, J.; Lax, A. "Remarks About the Effect of the Deto-
nation WKVe on the Liner of a Shaped Charge", Preliminary Report, I.M.M.,
May, 1950.

(2.) Birkhoff, G.; MacDougall, D. P.; Pugh, E. M.; and Taylor, G. I. 11Explosives

with Lined Cavities", Jour. App. Fhysics, 19, No. 6, June, 1948.

(3.) Touart, c. 11 Theoretical

Interior Ballistics of Lined Wedge-Shaped Charges''•
UnpUblished Report, I.M.M., July, 1951
(4.) Hadamard, J. 11 Lectures on Cauchy's Problem in Linear Partial Differential
Equations," Yale Univ. Press, 1923.

74
 SOURCES OF DISPERSION IN SHAP,ED CHARGE PERFORMANCE

R. v. Heine - Oeldern
Department of Physics, Carnegie Institute of Technology, Pittsburgh, Pennsylvania

ABsTRACT

Dispersion in the performance o$ hollow charges can be

, assigned to three distinct causes:

(a) Lack of homogeneity of the explosive

(b) Geometric liner imperfectiona

(c) Improper alignment between charge and liner

Of these three causes, only the last tw have been evaluated

quantitatively. The importance of perfect axial S)'lllllllltry in zones
perpendicular to the axis cian hardly be stressed enough. Very
little is knO'IIIl· about the effect of (a). Further attempts at re-
ducing charge variability should be directed at improving the
homogeneity of the explosive.

Anyone familiar with shaped charges, has been confronted with the fact that two
apparently identical charges can produce widely different penetrationa. A typical
experiment ie illustrated in Fig. 1, llhere the penetration produced by a single lot
of 224 charges are plotted in the form of a histogram. It will be noted that the
penetration values range from 5.2 to 9.5 inches, with an average penetration of
8.16 inches. The outstanding feature of this distribution is its marked ske1111ess,
llhich makes data of thie type very difficult to treat by established statistical
methods. If the long tail of this distribution, which contains about one-third of
the individuals, were amputated, then the remaining two-thirds of the individuals
would be distributed approximately in accordance with the normal distribution law
and the data would thus become amenable to routine statistical treatment. The first
objective of aliT work on charge variability should therefore be the removal of the
long tail of low penetrations. Only when this has been done will it be profitable
to make an attempt at reducing the spread of the remaining near-normal distribution.
In this paper, an attempt is made to show that those factors which affect the
symmetry of cone collapse are the ones which determine the degree of skewaess of the
frequency distribution. In addition, it will be sho'llll that pronounced asymmetry of
collapse gives risa to a bi-modal frequency distribution of which the skewed dis-
tribution shown in Fig. 1 is a limiting case with a relatively small number of charges
falling within the lower distribution.

For proper formation of a jet from a conical liner, it is essential that all
liner elements around a circular zone of the cone meet axially with the same momentum.
Fig. 2 shows a number of charge imperfections which tend to destroy the required

75
15 15
Ftnetlaliona ttl 224 C11orQe1 wHh M9AI 1:0pper
conn, standoff 4 ".

f
...
r
p,. 8.16 in.

r~· 11%
.. 0.89\n.

IIi
"
10 ~ 10

..."
!!

t!"
...!!!
;:

• -'
0: •

0~
5 -L-L~LL---L~~.~~Li_LULl_Li~7~--Li-WUL~LL_y~~Li-WUL~Li-Y~L----WUL~,0 8 9
PENETRATION ( TO NEAREST Ql IN. J

Figure !-Penetration of 224 charges with M9Al eopper cones, standoff 4".

J.

-ALIGNMENT-

Figure 2-Schematie of serious eharge imperfections.·

76
S)'11111111try of collapse. For the purpose of this discussion, these defects are divided
into those originating from (a) faulty detonation of the explosive charge, (b)
geometric liner imperfections, and (c) faulty alignment between explosive and liner.

The first type of defect, most readily illustrated by off-axial initiation, is

not usually enco\Ultered in a well-designed charge. Some time ago, tbil! group
obtained SO experimental charges from a commercial source. These consisted of
bakelite-confined charges with conical liners; at the top of the charge, a hole in
the bakelite allowed the insertion of a piece of primacord with its axis at right
angles to the axis of the charge. The primacord was intended to take the place of a
conventional blasting cap for initiating the charge. With such an arrangement, a
tilt of the detonation 'IIELve amounting to several degrees can be expected. The
penetration histogram obtained from SO of these charges is shown in ~. where
the bi-DDdality of the distribution can be noted. The Kerr cell pictureilin Fig. h
show the effect of off-axial detonation on the jet itself. Both of these jets
resulted from charges· detonated off-axially, the radial displacement of the detonator
amounting to 1/8 in. for the jet on the left and 1/4 in. for the one on the right.

The second type of charge imperfection, namely geometric imperfection of the

conical liner, is encountered very frequently. It is very wll illustrated by the
following e:mmple1 when this group obtained its first shipment of cones from a new
source of supply, it 'IIELS found that most of the cones displayed a skewness such that
the axis near the tip of the cone was inclined at about twenty minutes of arc with
respect to the axis near the cone base. The performance of 100 charges made with
these skewed cones is illustrated in Fig. S. The bi-modal character of this dis-
tribution is again quite evident. The next shipment of cones from the same manu-
facturer did not displa;y the skewness shown by the first one, and charges made
from them and subsequent shipments displayed the long "tail" of low penetrations
discussed earlier, but not the obvious bi-modality of the first group.

The third cause of imperfect charge performance, namely lack of alignment

between charge and liner, used to be the most frequently encountered and the most
serious of the three imperfections. It has been largely eliminated by casting the
charges in a mold, designed by this group, which is shown scheaatically in ~.
This casting procedure aligns the cone with the charge more accurately than~
possible by the old procedure of casting the charge in commercial steel tubing, where
alignment is determined by the cone flange. The development of this modified casting
procedure constitutes the biggest single step in reducing charge variability. This
is demonstrated by Fig. 7, which shows the penetration histogram of 104 charges made
by the conventional procedure with that of So charges aade by means of the C.I.T.
mold. It. will be noted that the standard deviation of the first group (0.87 in.)
is nearly three times the one of the second group (0.31 in.).

Recent evidence indicates that C.I.T. molds must be aade to very rigid specifi-
cations for optimum charge performance. In co~ing two groups of 6o charges each
aade in two different molds, one of the groups performed significantly better than
the other. This difference in pert:ormance vas eventually traced to the inclinations
of the alignment pin in the C.I.T. molds used; in the case of the mold which yielded
the better charges, this inclination amounted to 3 minutes of arc, while the other
pin ms inclined at 9 minutes of arc.

77
 urns 22Z §@E&Pil§ lillS

~. U2 lo.
,.. . 1.1' i.-.
riP • 40 ~ -
sf-- .--- .--- -I

.-"'.
~

r ..- r-
.. 10 f-- - I0
i
j .--
..
>
t--
~
...l

a::"' 5f-- -
r-r-- ..___

t--

0 I 0
1.5 3.0
PENETRATION CIN.l

Figure 3-Performan ce of 50 experimental

charges, standoff 2.5 cone diameters.

Figure 4-Effect of off-axial detonation on jet.

78

• U£0
 100 Penetrations obtained at 4 in.
standoff with charoes cost on skewed
copper cones .
P• 6.35 in.
~· 2.28 ita.
t:I"'""IP= 36%

PENETRATION (IN.)

Figure 5--100 penetrations obtained at 4" standoff with charges caat

on skewed eopper eones.

Figure 6-C.I. T. Mold.

79
 & ¥121Dil41I£& I 'smsti

104 Penetrations bJ charoes cast In

commercial steel tubing. 4 in standoff.
15• 5.48 in . '
..-•0.87 in
r/P•I6 'II.

0 3~~L----L~L-~~4L-~~~~~~~5~~~~-L~~~~6~~~-LJ---~

PENETRATION I IN.) -

25
t.... 50 Penetrations bJ charges cast In
C.I.T. Molds. 4 in. standoff •
P• 5. 79 ln.
...z
0 ,.. •0.31 in.
riP•S.4'11.
20 ...a:
-
D..

>-
z0
...::>
15 ...
0

If
...>
10 ...~
a::

0 3~--------------~4--------~L---~5~~~~-LJ-~~~6~~-L~----~

PENETRATION (IN.)-
Figure 7-(a) 104. penetrations by charges caat in commercial.ateel tubing, 4" atandolr.
(b) 50 penetrations by charges caat in C.I.T. Molds, 4" atandolr.

80
 This last example illustrates the extreme accuracy in alignment llhich is uecessar;y
for good charge performance. The requirements for liner symmetr;y are barcflT less
exacting, while the inclination of the detonation 11ave can amount to several degrees
before charge performance begins to deteriorate.

To sWIIIllll.rize, it can on4" be repeated that all those factors llhich affect
symmetr;y of the collapse process should be rigicflT controlled, while othe~ factors,
not directly concerned with charge Symtlllltry, can be allowd considerable tolerances
without impairment of charge performance.

The above considerations have some bearing on all experiments designed to compare
the performance of liners made from different materials. It is knolill, for exa~~~ple,
that steel liners give more consistent performance than copper liners. When such
liners are subjected to measurements of their contours and wall thicknesses, it is
generally found that the steel liners come closer to the requirements set forth in
this paper than the copper liners. In the literature, statements are frequently
found to the effect that metal A is a better liner material than metal B. In par-
ticular, lead has frequently been quoted as being a poor liner material. It is the
opinion of this group that such statements express nothing more than the machinability
of dif'!erent materials and that a fair comparison between two metals ·can be made only
i f both of them can be machined or worked to the same symmetr;y tolerances.

Bl

e
5 t£#41 ··-
 SCC&!L§ ,, I
CEPR'U T

..OEDWMME U> ·,
 PERFORMANCE OF PERIPHERALLY INITIATED SHAPED CHARGES

A. D. Solem
w. '1'. Angust
U. s. Naval Ordnance Laboratory, White Oak, Silver Spring, Maryland

.ABSTRAC'l'

The performance of experimental cone-lined shaped charges is

being investigated under the condition of peripheral initiation of
the charges. Peripheral initiation is simultaneous initiation of
the entire top periphery of the charge in contrast to point or
plane wave initiation. It is obtained by use of a cup-shaped
inert-filled initiator placed over the top of the charge such that
initiation is transmitted to the periphery of the charge but is
delayed in passage through the inert material in contact with the
top surface of ths charge. The behavior of penetrations from
peripherally initiated charges into mild steel targets for yariation
of charge height, stand-off, cone material, cone wall thickness,
and cone apex angle are being studied and compared with the
behavior of penetrations from point initiated charges for like
conditions. The results are described and reasons for the ob-
served behavior discussed. Experimental evidence showing why
peripherally initiated charges produce greater penetrations
are presented.

INTROIXJCTION
One of the means available for improving shaped charge penetrations and per-
formance is that of shaping the detonation wave in its passage through the explosive.
The interaction of detonation wave and liner will thereby be ·altered, and condi-
tions for increasing shaped charge performance will be enhanced. Various wave
shapings are possible; the most obvious one is that of initiating the charge simnl-
taneously around ita periphery at the top. The detonation wave then sweeps into the
charge like an expanding torus to strike the liner su.rface more nearly normal than is
obtained with ordinary point initiation. This method of wave shaping is not new,
variations of it were tested during World War II and. were tried in some weapons
(reference a* gives a list of references to this early work), but as far as is known
no systematic stu~ of the system had ever been carried out. In the autumn of 1950
the shaped charge group at the Naval Ordnance Laboratory undertook as one of. ita
programs a general investigation of the behavior of peripherally initiated shaped
charges. The purpose·s of the tests were to establish conditions optimizing shaped
charge performance under this system of initiation, to observe limitations of the
system, and to determine if peripheral initiation were feasible in practical
applications. This present paper reports same of the early results obtained in the
tests.

* Reference (a) - Nav Ord Report 1722 "Peripherally Initiated Shaped Charges•
dated 1 November 1950

83

cE2itMB a
 &. '. . ,· ~ . "'·~·· . ' " ·w1
EXPERIMENTAL CHARGES

The system used to optain peripheral initiation is given in Figure 1, which shows
a sketch or a typical assembled experimental charge. The charge proper consists of a
cylindrical charge cast over a conical liner. In the present tests the charge di81118-
ter was 1.63 inches,· and the liner dimensions ~ere those of the standard M~Al cone
(1.63 inch base diameter, 44 degree apex angle, 0.037 inch wali thickness) • The
charge height and the liner material were the factors varied in the tests. The ex-
plosive used tor these charges, except for one series of tests caqparing different
explosives, was S0/50 Pentolite.
The peripheral booster consists of a cup-shaped, 50/50 Pentolite charge which
tits over the charge proper in contact with the charge proper only at its peripher;r.
The top surface of the charge proper. :i.s separated from the booster charge by inert
filled cavity of the booster. The inert material is lead oxide, mono (other inert
materials have been tried and the results have not been as good as with the lead
oxide), and, except tor one series ot tests where the thickness was varied, the thick-
ness or the lead oxide is 0.5 inches. The assembled charge is initiated on the charge
axis by an electric detonator and is boostered ey a Tetryl pellet.
Conditione tor symmetry are much more critical tor these charges than tor ordi-
nary point initiated charges. Unless complete simultaneous peripheral initiation or
the charge proper is obtained inferior performance results. The initiator must be
situated.on the axis of the charge so that the detonation wave passing the barrier ia
essentially plane and nonnal to the charge axis; the peripheral booster must make good
uniform contact with the charge proper so there is no delay or detonation at any point
in passage into the charge proper. In test these conditions are not always met even
with care in assembly. The spread of results is somewhat greater than obtained with
experimental point initiated charges. The spread diminishes as the charge height in-
creases.
EXPERIMENTAL RESULTS

Behavior of Various Liner Materials. Figure 2 compares the penetrations obtained

With polrit, plane wave, aJid peripheral initiation of M9Al steel cone lined charges
with variation of charge height. Note the increase of penetration as the charge
height is increased when point or plane wave initiation is used as is expected. Note
also the peaking or penetration with. peripheral initiation at the 2.35 inch charge
height. This high penetration is some 30 per cent greater than obtained with a
standard point initiated charge 4.00 inches in height. This increased penetration
turnished incentive to further investigations of the effect.
Figures 3 and 4 show the penetrations obtained with other liner materials as the
charge height is varied. Liner materials considered consisted or cast iron, glass,
copper, brass, aluminum, and magnesium. Note that while there is reasonably reguliU'
variation or the penetration pattern there is considerable scatter or penetrations tor

* The work with glass cones was with 1.90 inch diameter charges; the cones had a
1.90 inch base diameter, and a 60. degree apex angle.

84
 i Initiation. Point

~
0" /38

------------
-+-
Barrier
Thickness
-----
--
- --- --
- -----
-- --
-----
--~-
--·-
- --./

0".11

~
T
C'\
'CIO
.!f c•
•
1%: r-1

'' fo
~
:0
!
0
I ....

~
Cone !0
0

1" .63 -----.1,.11'11

Figure 1-Experimental peripheral initiation charge.

BS
 ···.:...-Peripheral Initiation ...___
-ID
Gl
..r:1
7
• ~
, ;· ·-·--.:r
............. .---
-' • ....... •
,

~
.. i ..... ... ~-
...............................
Plane Wave Initiation
~

86 .... ~---···
H
.......
1:1
0
'"
~ 5
t "'
Cone
_Height _ .
-· -·Standott
4.00 inches
.__ Point Initiation ~
~
"'
Gl
I
I ... ~· ---~
R
Gl
1)..
4 .
--a. L
\
2.0 3.0 4.0
Charge Height (Inches)

Figure 2-Elfeet .of charge height variation for steel liners.

7 I "' ...' ' 11

,_..
Cast Iron
(M9Al Dimensions)
Cone-1 .... ,'
I
Height I " ' ' ,, Standoff'

Jr
_..............
...4.00 inches
~-
0
,~---
l-·--·/
Glass
~
Cone
Height I j ' - (1.9 in. base, 600)
'-I
" • Stando.ff
- _ ...
: 0
8.00 inches

I
2.0 3.0 4.0
Charge Height (Inches)

Figure 8--E1fect of ~harge height variation for cast iron and glass
liners under peTipheral initiation.
 /'14 ---~- ...... ~ .......
.,. ~

" ~,
... _ • .}• "'~opper / ... ...
I" ~
.....
~I
I
" (M9Al Dimensions) "· ~
Standoff
8 4.00 inches
""' •,....... •-. .....
,• ~""o
...
Brass
I I
••••
........, (M9Al Dilllensions)
Standoff

6 .....·• ·"
~
••
•
•
,' ~
/.,/ --
=--·- -·
____
.... ~,..
4.00 inches

/ I
-.,
II)
.c:t
' •
'~
/6
·Aluminum
(M9Al Dimensions)

-
H
g
-· /..: Standoff
8.00 inches

.. /
......... '• ~··
., .......... .
....
'
;'
• Magnesium
(M9Al Dimensions)

2 ! Standoff
i 4.00 inches
I
I
1
' - Cone
', Height

0 J 1 ' 2.0I 3.0 4.0

Charge Height (Inches)

Figure 4--Effec:t of eharge height variation for copper, brass, aluminum,

and magnesium liners under peripheral initiation.

87
 s rnsymu a 5 5
,.

Table 1

PERFORMANCE OF VARIOUS LINEIIS

44° CONES lo63n BASE DIAMETER
50/50 PENTOLITE

Point Initiatio n Periphera l

Charge Charge Percentag e
Liner Height Pentratio n Height Penetrati on Increase
Copper 4.00 8.46 3.00 9.51 12
Steel 4.00 5.60 2.35 7.~ .30
Cast Iron 4.00 7.49 2.35 7.91 6
Brass 4.00 7.08 2.50 7.28 .3
Magnesium 4.00 1.70 2.25 .3.58 106
Aluminum 4.00 5.27 3.00 6.50 23
Glass* 4.00 4.27 3.50 5.74 34

* Glass1 fJJ° Cone, 1.9• Base Diameter

88
 7775725 '3
5
I ,

Table 2

PERFORMANCE OF VARIOUS EXPLOSIVES

M9Al STEEL CONES

Penetration
(Inches)
Point Peripheral Percentage
Ellploaive Initiation Initiation Increase
50/50 Pentolite 5.56 7.41 33
Comp B 6.17 7.20 17
PTI-2 6.57 7.16 9
Comp A-3 6.22 6.82 10

BTNEU/WAX 5.64 6.79 20

B9
 327 3 I 1

8 • ~Peripheral Initt.tion
- ·----r·------1-------.·.-------·t
·•J6
!
.,.!'_ • - -1-r--- --7----- ~r- -----.. ·
- \Point lDitlation
I
PEBTRA'l'IOR
1
.

I
Studof'f'
Charge Heights: 4.00 inches
Peri~heral -· 2.3~ inches
-
PoiD ·- 4.00 inC ea
1
0 .50 .75 1.00
.25
Confinement (Inches)

,. -.- ---...------- - ______ ...._

30 1
, ,/" \Point Ini4 iation
~/
• , ..... ~----- - --.-. --- ------
--
u
u
20
.,t, ,. •'
.I ""
I~
\Peripheral Initiation

" CAVI'l'r VOLUD

0 .25 1.00

Figure 6-Eft'eet of confinement variation for M9Al steel ·cone liners

under point and peripheral initiation.

90

ssnwz •··
 MitE ·tr'"rs66 f'i tit.
the different charge heights for the cast ~ron, glass, and copper cones, Note also
that the penetrations are not typified by a simple behavior pattern, although some
order may be made from the chaos, The results are not complete but are given to shaw
the type of behavior observed,
Some comments on the results are pertinent:
a. The penetration pattern for steel liners is fairly stable, The results
have. been checked, and the peaking at the 2, 3S inch charge height has
been verified with many shots.
b, Only a few shots were fired (on the order of five) to determine each
of the points shown for the other materials. Thus the stable position
for each point may vary considerably from that shown.
c. The double-humped behavior of the copper liners is believed to be real.
This behavior has been verified qualitatively by work performed at the
Carnegie Institute of Technology, although their results do not agree
quantitatively with those shown here,
d, The variation in charge heights tested is somewhat limited and further
testing will extend the charge heights in both directions,
Table 1 summari~es the increases in penetration obtained to date in going to peripheral
initiation for the various liners tested. Some of these, of course, have no practical
applications. The results are encouraging for steel, aluminum, and glassJ they are
·• disappointing for copper, brass and cast·iron.

Behavior of Various ~losives. Another series of tests that has been considered
has been a comparison of e behaVior of various explosives on M9Al steel cones under
point and peripheral initiation, The results of these acts are shown in ·Table 2. All
point initiated charges were 4,00 inches high; all peripheral initiated charges were
2.3S inches high. Note that as the penetration under point initiation goes up the per-
centage increase when peripheral initiation is used goes down. TWo comments on these
results are offered: 1. The charge height under peripheral initiation was optimum
for SO/SO Pentolite; it might not have been optimum for the other explosives. 2. It
appears that a saturation effect is being approached, so far as peripheral initiation
is concerned, as the explosive power goes up. The choice of SO/SO Pentolit.e and steel
cones for the first peripheral initiation tests was surely a very fortunate one.
Behavior Under Confinement. Figure S shows the results of a series of tests to
observe the eftect or coritinement on point and peripherally initiated shaped charges.
standard SO/SO Pentolite charges lined with M9Al cones were used. The charge height
for the point initiated charges was 4.00 inches; the charge height for the peripherally
initiated charges was 2.JS inches. Insofar as penetrations are concerned it is con-
cluded that confinement has little or no effect on the results for the case con-
sidered. Cavity volume does increase with degree of confinement up to a nominal con-
finement of 0.2S inch thick steel cylinder. The volume increase is on the order of
300 per cent for both point and peripherally initiated charges, It should also be
) noted that cavity volumes for peripherally initiated charges are always 20 to 2S per
cent less than for point initiated charges. If cavity volume is acy measure of total
energy in the jet less energy is associated with the peripherally initiated charges,
but is so distributed to increase penetrations.

91
 c: I f etst

8
I
I.,._---"I - -·--- I
7
•0
-- - -- - -• -
• ~- ....; ~

I
/,
!, I
--..
/ Barrier Thickness NormallT Used
-
I
... \Penetration for
• Point Initiated Charges
(4.00 inch charge height)
I
I

Standoff Barrier PbO

-~
1
4.00 inches
I
0
.2 .4 .6 .s 1.0
Thickness at Barrier (Inches)
Figure 6-Effeet of varying barrier thickness for M9Al steel cone
liners under peripheral initiation.

92

..III.--·IW'...' •..
:'l.~[!I!SIJ.
 I
•
·.;
.......
J• Peripheral Initiation
~ "' ~~ 2.35 inch charge height
',, I "" STEEL COllES
'2.....>
j
L. I /

.......
.,
fl
..-1 I
·,. . .~. . . . 'f.......
K ............... Standoff
....... 41---
Poil)t Ipitiationl
(CIT} .... ...
~·.___--..... .
4 00 inches
• I
8 1---...j _ "
: 5.00 inch charge height _
. '- .,._ --1-f--_-___
~ 2

2 4 0
Steel Target Penetrated (Inches)

' ',!~!/I
COPPER CONI!8
........ ~~
/ r- ........, -...._..,. 111 Standort
Point Initiation f-- _ •---...!..._ ~.00 inches

-- -.
1
(CIT - I·-
t-- 5.00 inch charge height ' • .... __ j---... "- _ ~
.... __, -"
0 E
2 4
Steel Target Penetrated (Inchea)

Figo.re 'l-Initial and emergent j_et velociti~ from~~ ~d copper

M9Al cone liners under pomt and per1pheral1mtiation.

93
 ,c ................, ,. ...._LXI4 L. ... SJ

Behavior Under Variation of Barrier Thickness. Figure 6 shows the results of a

aeries of tests to determine the amount of barrier required to produce peripheral
initiation. Standard 50/50 Pentolite charges 2.35 inches in height cast over M9Al
steel cones were used in the test. The barrier material was lead oxide, mono. Note
the uniformity of penetrations as the barrier thickness is increased beyond 0.2 inches.
The standard barrier thickness used in all other tests was 0~5 inch. It is therefore
concluded that true peripheral initiation was obtained. The exact nature of initi-
ation for the 0.2 inch barrier thickness is not known, but it is speculated that it
approaches true peripheral initiation. For the thinner barriers initiation is surely
tranSIIIitted through the inert material.
JET VELOCITY STUDIES
A final series of tests to be reported here is some preliminary work on jet
velocities which it is hoped will explain the behavior of peripherally initiated shaped
charges. Figure 1 shows the initial and emergent jet velocities after penetration o£
a given target thickness .for both point and peripherally initiated charges. Results
are shown for both steel and copper cones. The velocities for the point initiated
charges were taken from Carnegie Institute of Technology data (reference b)*. Charges
5.00 inches in height were used in each case. The initial and terminal velocities
were checked at the Naval Ordnance Laboratory for both the copper and steel cones.
Agreement within 2 per cent of the Carnegie data was obtained, and it was concluded
that intermediate velocities would also be duplicated. The velocities for the
peripherally initiated charges obtained. at the Naval Ordnance Laboratory came from
charges 2.35 inches high for the steel cones and 3.00 inches high for the copper cones.
These charge heights were selected as the ones giving the greatest increase of pene-
tration when using peripheral initiation of the charges.
Note in each case that the emergent jet velocities were always higher for
peripherally initiated charges than for the point initiated charges. The increase is
much greater for the steel cones (shorter charge height) than for the copper cones
(longer charge height). From these results one can say that peripheral initiation
does increase jet velocities and hence jet velocity gradients; however, the exact
bearing this .has on penetration will depend upon the velocity and mass distribution
spectra of the jets which have not yet been obtained. Efforts are being directed
toward obtaining these distributions but they will not be inunediately forthcoming
because of complications of the derivations. Initial jet velocities for other charge
heights have been obtained to observe how they vary with charge height. The results
are tabulated below:

INITIAL JET VELOCITIES

(meters/second)
Charge Height Steel cones Copper cones
(inches)
3.00 8800
2.35 9330 9410
2.15 11440
2.00 12660

* Reference (b) - Sixth Bimonthly Report (CIT-ORD-9) Annual Summary, "Fundamentals

of Shaped Charges• dated 30 June 1947 Carnegie Institute of Technology.
94
The initial jet velocities at least increase as the charge height is decreased.
The reasons for shaped charge behavior under peripheral initiation (increased
penetrations) cannot be definitely explained by these tests, but one might speculate
on the mechanism which increases penetration. For steel liners where the penetrations
peak for a single charge height one set of factors (jet velocity, velocity gradient,
jet mass, and energy) is optimized at the one charge height. For copper liners on the
other hand where penetrations peak for two charge heights two sets of factors must be
optimized. For the short charge height such things as jet velocity, mass and.energy
are optimized; for the longer charge height factors such as jet velocity gradient and
jet ductility are optimized. With the short charge the velocity gradient is too
strong and the jet is pulled into particles so that fUll benefit of the optimum jet
velocity, mass and energy is not realized. These factors are not optimum for the
longer charge height, but this is over balanced by conditions of optimum velocity
gradient. These comments are of course only speculations; their possible validity
must depend upon fUrther experimental results.

95
 c. 1 .
THE PERFORMANCE OF PRECISION-MADE CONICAL LINERS

J. Dewey'

H. I. Breidenbach, Jr.

J. Panzarella
J. Longobardi

Terminal Ballistic Laboratory, Bal.listic Research Laboratories,

Aberdeen Proving Ground, Maryland

ABSTRACT

SDBl.l drawn, cast, and machined cones in bare Pantolite charges

have been investigated for a determination of fabrication control
necessary for good reproducibility of performance. Explosive cast-
ing vas carefully- controlled. Deliberately introduced flaws in
cones produced effects which are usUally those predictable frQm
consideration of the time at llbich each portion of the cone reaches·
the axis. Flash radiography of the jets supplemented observation
of target blocks. Precision cast cones gave about the same
reproducibility a,s drawn cones. Three fabricators made electro-
formed and DBchined copper cones of wideq differing precision and
performance. Warping and 11111tallurgical non-uniformity wre sources
of large spread in the performance of the products of tw fabri-
cators. The third manufacturer produced 3/4" copper cones to 0.1
mil to:).erances frcim. llhich straight jets of uniform penetration were
obtained. A 3% cone gave 6.8 cone diameters penetration.

The Terminal Ballistic Labol'atory has been investigating the effects of variations
in liners on jet performance. In QJ'der to radiograph cone collapee, as well &I! the jet,
cones of 3/4" and 1 8 base diameters have been used. The uniformity of the explosive
lBS investigated by- the technique developed in the plastics industry, low voltage
radiography of the bare charge. Except very near the edge of a cylinder, where the
density gradient is high, this method reveals 2% variations in mass per unit area. In
preliminar,y work, just completed, drawn steel cones and copper electroformed cones of
slightly better construction than drawn ones were used. As these have random vari-
ations in thickness of tbf! order of o.OOl", large de-liberate variations and the parti-
cular defectS characteristic of the method of manufacture wre the only ones studied.
Figures 1-6 show some results, starting with a jet from a charge having a cavity- near
the base, for comparison. All the effects shown can be explained O;y assuming that

97

££_ .. . <II
r-------------------------~,,.,.----·~

thin portions or cones collapse rapidly, not meeting the other portions so as to !om~
a jet. warped cones, such as are sometimes obtained by drawing and when electro!ol"lll8d
cones are removed !rom a mandrel without annealing, give very poor penetrations.

Very recently' cones have been obtained which meet specitications to within a rev
ten-thousandths or an inch. These are or sp!t copper, or uni!orn~ high density and large
grain size. They give straight jets which reach a considerably greater length before
breakup than jets !rom drawn cones, as shown in Figure 7. Figure 8 shows the penetration
ot jets !rom these liners, with some Bruceton and Carnegie Institute o! Technology
results. The 3% cones show considerable spread at optimum stand-ort. This stand-ort
is twice as great as !or drawn cones. The results suggest that the rapid decrease in
penetrati.on With stand-orr usually' observed results !rom wandering, as well as !rom
break-up o! ths jet. The di!!erence in the shape or ths holes in the targets contirms
this. Figure 9 shows radiographs .or blocks penetrated by jets !rom 'Various types o!
cones. The precisely' made liners· !orn1 a slug which always enters the hole symmetrically',
tilling it. With the 3% liners it was necessar,y to saw ths block to see the hole. When
tapered cones or this type 11ere !ired, the conical slug !oriiiBd hung together better,
giving a more open hole.

Another dit!erence between liners is seen when stacked or spaced plates are the
targets. Penetration is not reduced unless the spacing is large, and no copper is seen
around the exit holes. (Fi.gure 10)

98 •

a
 Cavity 1n Cost H.E.

t
01spersed Jet Slug

Dispersion and Curvature of Some Jet

2 Microseconds Later

NOTE: Radiographs I 8 2 were mode from x-ray

sources 90• aport.
Figure 1-46" copper liner having .015" radial and .0007" vertical variation
in wall thickness. Air bubble in cast H.E. visible.

99

IE U·
 DOUBLE FLASH RADIOGRAPHIC
STUDY OF ELECTROFORMED COPPER CONES
3/4 • base dia. , 46°, 0.02 5" wall

1-'
8 Static VAIUATIOMS IM COME WALL
THICKNESS
4
VUTIOAL ( 10" 1N.l ltAOIAL
Round
No. 12:00 a:oo •:oo t:oo . 25 .110 .715

194 4 t I I I

Penetration : 3-5/8" Jet Curvature : 7. 5°

Collapse Anoles : Upper 33.5°
Lower 36.25°

F igure 2-Double flash radiographic study of electroformed copper cones,

%" base dia., 46• , 0.025" wa ll.
 ZGitl IDEit ana

THE EFFECT OF DEFORMED CONES UPON PENETRATION.

43~. 037': t'' 1.0., J}t~ HIGH STEEL CONES.
PICTURED BELOW ARE THE BASE DEFORMATION OF
THE CONES AND THE EFFECT UPON THEIR RESPECTIVE
TARGETS
SHADE DENOTES FILED SECtiONS ON INSIDE' OF CONES. AVG .. Oro•
FILED OUT.

~!

Figure 3-The effect of deformed cones upon penetra t ion.

THE EFFECT OF OEFORME 0 C ONES UP ON PE NETRATI ON.

43~ .037," 1•• J. D., J'l,,,n HIG H ST EEL CON F.: S.
P~TURED BELOW ARE THE BASE DEFO R MA T ION Qt;

THE CONES AND THE EFFECT UPON T HE IR RES PECT iv E

TARGET 5

Figure 4-The effect of deformed cones upon penet rat ion.

101

04 s t!IAE a
 THE EFFECT OF DEFORMED CONES UPON PENETRATION .
43~ . 037': I'' 1.0., 1>1~ HIGH STEEL CONES .
PICTURED BELOW ARE THE BASE DEFORMATION OF
THE CONES AND THE EFFECT UPON THEIR RESPECTIVE
TARGETS
A LINE OF WELD, APPROX. ~ THICK. ON EACH CONE
F'ftOM APEX TO BASE EDGE.

0 0

Figure 5-The effect of deformed cones upon penet ration.

EFFECT OF THIN WALL SEGMENTS ON STEEL CONES

.037" WALL THICKNESS, I" I. D.
TARGET, 2" X 2 " X 6", 1020 STEEL

NORMAL THIN
WAL L - -SEGMENT
NORMAL THIN .037" .020"
WAL L - -SEGMENT
.037" .030"

0 ''

PENETRATION, 1.5 CONE DIAMETERS PENETRATION, 3/ 8 CONE DIAMETER

F igure 6-Effect of t hin wall segments on steel cones.
102
 SLUG

..

JET JET

DEEP DRAWN COPPER

.060" Wall, 43•, I" 1. D.
Shot in vacuum , 67 micron•, 29 u• after
start of collopu of cone.
ELECTROFORMED COPPER CAST COPPER
PENETRATION, 2.75 CONE DIAMETERS
.050" Wall, 46 •, . 75" I. D. . 050" Wall , 46•, I" I. D.
34 u1 After start of collapse 3 1 u• After 1tort of collopw
of con~. of cone .
JET LENGTH, 7-3!4" JET LENGTH, 6 - 1!2"

Figure 7-Comparison of jets from different types of cones.

lOJ

em a
104
 \

COMPARISON OF JET PENETRATION

FROM VARIOUS TYPES OF CONES.
TARGET : 2" X 2" X 6", 1020 STEEL
ALL PENETRAT IONS IN CONE DIAMETERS

......
0
\1\

025" WALL, D-42 .050" WALL, D-12 TAPERED WALL, D-67 .025" WALL, S-73 .025" WALL, C-61 .050 " WALL, P-59 .037" WALL
NETRATION, 6. 40 PENETRATION, 5. 35 .025" APEX TO .050 BASE PENETRATION, 5.50 PENETRATION, 5.80 PENETRATION, 4 . 75 PENETRATION, 3.06

PENETRATION, 5.87

ELECTROFORMEO COPPER CONES CAST COPPER CONE DEEP DRAWN STEEL CONE
. 750" BASE I.D. I" O.D. I" 1.0.

Figure 9-Comparison of jet penetration from various types of cones.

 SOFT LEAD CONE
JET SPLASH

SPLASH FROM JET FROM

3.5" ROCKET HEAD
STEEL CONE DEEP DRAWN COPPER CONE ELECTROFORMED COPPER CONE
SPLASH NO SPLASH

Figure 10-Exit holes of jets from various material cones through steel target plates.
 L(Jol VOLTAGE FLASH RADIOGRAPHY

J. J. Paszek
B. d. Taylor
J. L. Squier
. Ordnance Engineering Laboratory# Ballistic Research Laboratories

Flash radiography is generally accepted as being the best means or obtaining

direct pictorial presentation of the d,ynamic perfOl'J!IBllce of shaped charges. Optical
methods# although useful ·in other studies of detonations, are not altogether suit-
able, due to both the lUillinescence of the object and the .t'act that the vi&w is o.rten
obscured by material g;l.ven orr by the jet. FlAsh radiographs give a direct picture
or the material in i'ront or the film w~le optical pictures o.rt.en require interpre-
tation. Up to now most if not all the radiographic work on shaped charges has been
done at voltages of 100 KV# or more, We believe there are advantages in using so.rt.er
x-rays. A system for taking sub-microsecond duration flash radiographs or shaped
charge phenomena using low-voltage (.34 KV) X-rays will 1!e described, The system con-
sists or a simplified low-voltage X-ray pulse generator circuit and a method of pro-
. tecting the x-ray tube and film.

Particular advantages accrue from the use of low voltage. or· prime importance
is the ability to get greater detail f'rom objects small in size or or low density.
The break-up of a jet, for instance, taken With high voltage x-rays may be lacki.Dg
in detail due to the transparency or the smaller particles of the jet to the x-rays.
LikeWise, materials or low atomic number such as aluminum which would be almost
transparent to high voltage x.-rays ·show up quite well when the voltage is reduced,
When working with low voltages, the use.f'ul lite or the X-ray tube is not directly
determined by arc-over along the inner wall of the tube. The inner wall of a flash
X-ray tube becomes coated With a metallic coating as. it is nashed and, when this
coating becomes heavy en0ugh, arc-over along the wall may occur, In addition, other
practical advantages result £rom the use or low voltages. The e'iuipment is simplified
considerably, Components are less expensive and II!Uch more readily obtainable, Corona
which plagues all high voltage equipment is all but eliminated# DBidng possible the
use of the equipment in the field even under high buml.dity conditions, The equipment
becomes !lQmpact and light in weight. A photograph or the original equipment is shOIIJl
in Figure l. By using smaller components (standard catalog items) the equipment can
be DBde small enough to fit into a S cubic root carrying case permitting i'ul.l portabilit;y.
Flash radiography is thus taken out or the category or special laboratory equipment
and becomes a general purpose tool.

It DBY be o.r interest to relate briefiy how the pre&ent circuit evolved, at the
same time giving acknowledgement to those who were or assistance to the authors. As
part or the Ordnance Engineering Laboratory's program for shaped charge ~estigation
nash l'll.diograpey was required, An order had been placed tor a nash X•ray f.12t or
the type developed by H. I. Breidenbach and described in published reports. ( • J This
c1rcuit has been in use since 1947 by the Terminal Ballistics Laboratory,

l.G7
1?77 PEES §p Pitr t'm'Sti?T

Fig ure 1

108
 *
While we were waiting for the deliver,y of 'this equipment, R. o. Fleming of Old
and Barnes, Inc. visite<i this laborator,y and delivered a paper entitled, "A Simple
Flash X-R~ Circuit"• OJ In. addition, Fleming gave us personal instruction as to how
to reproduce his c1rcu1t. An attempt to duplicate Fleming's circuit using available
' canponents was unsuccessi'ul.. It should be emphasized that the use of substitute can-
ponente prevents us from making a proper appraisal of his circuit. The low voltage
x-r~ circuit lihich is the subject of this paper came into being as the result of our
l)eing required to work with canponents at hand.
'
Brief reference has been made to three X-ray circuits. The Breidenbach circuit
is not -tuPicalq a low voltage circuit and thus will be described only brisfq. The
basic circuit is shown in Figure 2. It uses a Westinghouse WL-389 three t~lAQtrode
cold cathode field emission X-r~ tube operated at approximate:cy 100 KV. (4,5 J ntis
tube has a tr~ger electrode, mounted ver,y close to the cold cathode, which is used
to initiate c.athode-to-anode breakdown. In the circuit of Figure 2 the trigger
electrode voltage is derived .f'ran the anode voltage, so that the anode voltage must
be at least as high as that requ:l,red to break down the trigger gap. As we shall see
thie voltage 1a relatively high.

The Fleming circuit operates at a cathode to anode voi tage of 80 KV. ihe circuit
1a unique in the manner in which it makes possible the use of a two electrode tubs.
Reference to figure 3 will reveal the basic circuit. The tube is the ver,y small 45 1
·dental X-ray tube made by Amperex. While this tube· has a filament, it is used here
as a cold cai.llode tube. A capacitor is connected between anode and ground, lihil.e the
secondary of a pulse transformer 1a connected between ground and cathode. Effectiveq,
the capacitor and the pulse transformer are in series. Fleming gives B. Cassen the.
,;. credit for suggesting this connection of the transformer. '!he capacitor iS charged
to 40 KV positive by means of a small r.f. type power supp:cy. A simple pulsing
circuit (not shown) supplies a negative 10 KV pulse to the primary of the 4:1 step-
up ratio transformer. The 40 KV across the capacitor is alone insufficient to break
down the tube. The additional negative 40 KV pulse raises the total cathode to anode
voltage to 80 KV causing the tube to nash. Triggering is thus accomplished by direct
cathode to anode pulsing without the use of a trigger electrode.
Our reproduction of the Fleming circuit used substitute components. The tube was
a Machl.ett LCP-L, which is a small size dental X-r~ tube. The power supply provided
34 KV instead of 40 KV. The pulse transformer had a J.Ci:l step-up ratio and conse-
quentq a higher secondary impedance. Using these canponents we were unable to pro-
duce x-rays.

In addition to the power suppq and the pulse transformer mentioned above we had,
at ths time, several 1 mfd. capacitors rated at 2!5 KV .and one WL-389 X-ray tube. As
previouaq stated, the WL-389 is a three electrode cold cathode field emission x-ray
tube having, in addition to the cathode and the anode,. a trigger electrode. Availa-
bllity of the WL-389 suggested a return to a circuit utilizing a three electrode tube.
Direct cathode to anode pulsing is not practical inasmuch as the breakdown voltage of
the WL-389 may be as high as 100 KV or more. Properq used, the trigger electrode,
which is mounted ver,y close to the cathode, can cause the tube to flash at lower

* Now of Microt:lme Laboratories, Loa Angeles, California.

109
 t
'M

I ~011~5~1d
I
Ill

.2 ~ld II

THYRATR~N r
.
TO
-SOKV

• !SOKV
r- (OC22~
.
PUL.llf:
'-----J
-8

.2~1d: ~ TRANSI'ORIII2R

IM Ill

Figure 2

... 40 KV >40
PULSE
KV

Figure 8

nm
- 60tCV PULSE

I. MFD
+34KV

Figure 4

I.MFD
Ill 511

1.0 llfD TRIGIIER

GAP

Figure 5

110
 vol.tages. Circuits that use some f'raction or the anode voltage as the trigger voltage
require the anode voltage to be at least as high as that needed to break down the
trigger gap. Being limited by the 34 KV power supply, our problem was to !lash the
• tube With only this voltage on the anode. This was accomplished by connecting the
pulse trans!ormer directly to the· trigger electrode so as to apply a negative trigger
pulse between the trigger electrode and the cathode. This connection ads it possible
to trigger the tube independently or the anode voltage, pemitting operation or the
*
tube at an anode voltage as low as 17 KV. A 50 KV trigger pulse proved necessary
to nash the tube reliably.
The simpl.itied schematic shown in Fig. 4 wil.l help make clear what has just been
s_tated. The circuit consists or a 1 mi'd. capacitor with the positive teminal
connected to the anode or the X-ray tube, and the negative terminal connected to the
cathode. The secondary or the trigger transrormer is shown connected between the
trigger electrode and the cathode.. A small pulsing circuit (not shown) applies a
pulse across the primary winding or the trans!ol'lller. The resultant secondary pulse
breaks down the tri.gger gap, causing the tube to nash.

The operation or the complete circuit is best understood by re!erring to the

schematic diagram shown in Fig. 5. A 1 mi'd. capacitor is again shown connected be-
tween the cathode and the anode or the "WL-389 X-ray t.ube. The capacitor is charged,
through the 5 megohm isolating resistor, by the 34 KV r.r. type power supply. This
voltage is illsut!icient to break down the tube and will, there!ore, resin across the
tube until some additional energy is supplied to break it down. This is accomplished
by applying a negative 5 KV pulse to the primary or the l.Oal step-up ratio trans-
l:orrner. The resultant 50 KV pulse which appears at the trigger electrode causes an
initial breakdown between the trigger and the cathode, which in turn causes a break-
down between cathode and anode. This allows the 1 mi'd. capacitor to discharge
through the tube, producing a burst or X-rays. The WL-389 X-ray tube is thus made to
nash with an anode voltage lower than that required to break down the trigger gap.

The negative 5 XV pulse is provided by the trigger circuit which is included in

the schematic in Figure 5. A 1 mi'd. capacitor is shown connected in series with the
priiiiLry or the pulse transrormer. One side or this capacitor is grounded, under
static conditions, through the transrormer primary winding whil.e the other side is
charged to a positive 5 KV l.evel by the power supply. A 5 megohm resistor serV"es to
isolate the power supply .from the rest or the circuit during pulsing. The plate or
the 4C35 hydrogen tQyr&tron tube is connected directly to the positively charged side
or the 1 mi'd. capacitor while its cathode is connected to the grounded terminal or
the pulse transtormer primary winding. No bias is needed on the gr:l:d or the thyra-
tron tube to hold orr 5 KV. A 1 mi'd. capacitor is connected in the grid circuit or
the t!J1ratron tube. One side or the capacitor is grounded, under static conditions,
tb:rough the 56o ohm grid leak resistor; while the other side is charged to a negative
6oO volt l.evel by the 6oo vol.t powe'r supply. The trigger gap serV"es to initiate the
trigger circuit when its terminals are shorted.

Shorting the trigger gap raises the negatively charged aide or the 0.1 mi'd.
capacitor to ground potential, causing the oposite side or the capacitor to rise

* It has recently come to our attention(6) that the Germans had used a circuit
exact:cy- like this at high vol.tages (100 KV) without apparently realizing the
advantages or the l.ower vol.tages.

lll
suddenly in a positive direction producing a pulse at the grid or the thyratron tube.
The no:nnalJ¥ non-conducting thyratron tube now conducts, shorting the positive]T
charged side of the 1 mfd. capacitor to gt-ound. This produces a negative going
pulse on the opposite side of the capacitor which develops across the transfomer
primary winding. This voltage is applied thru the transfo:nner to the trigger elec-
trode which causes the tube to nash.
It might be appropriate to say a few words about the components used in the
circuit described. Except for the X-ray tube, all the components are standard
catalog itEIIIS and can be purchased for a total cost of less than $l,ooo.oo. The
power supplies are small units of the r.f. type. Less than one milliampere output
is required of these supplies. The capacitors used are fllter type capacitors. The
1 mfd. capacitor is rated at 25 KV and was obtained from surplus at a cost of ~75.00.
It h~[~s been used at voltages up to 50 KV. The pulse transfo:nner is a Westinghouse
4P-30. Although the secondary is rated at .36 KV, no trouble was &Jqlerienced as long
as the input voltage did not exceed 5 KV. Two transfo:nners falled at higher voltages.
Once the circuit was completed the immediate questions that arose were those
concerning flash duration, resolution, and X-ray output. At present these questions
are best answered by the results obtained from the application of the circuit to jet
radiography.

The first tests were made with small 3/811 diameter copper lined shaped charges
or spitters which proved ideal for testing the low-voltage flash X-ray unit. Three
radiographs of the jets from these spitters are shown in figures 6, 7, B. DuPont jet
tappers, 1-3/4" diameter shaped charges with conical copper liners, were also used to
test the circuit. Th.e radiographs in figures 91 10 1 11 show the jets fran these
charges. The spitter radiographs are the more interesting at this point in that they
provide a good deal of info:nnation on the perfo:nnance of .the circuit. Note that the
jet breaks up into small particles some of whieh are little larger than the head of
a pin. This fact permits direct est:lmation of particle resolution and :f'lash duration.
The. fine resolution and excellence of detall shown in these pictures illustrates and
lends support to our contention that low-voltage X-rays are particular]T suited to
jet studies. The tip velocity of the spitter jet is known to be approx:lmate]T 1/4"
per microsecond. Sane or the particles in the leading portion of the jet are less
than 1/16" in length, as measured along the direction of jet travel. Now the length
or the smallest particle in the leading portion of the jet is indicative of the
effective duration which can be inferred fran an est:lmate of the IIIIIXimUI'II amount of
blurring of the jet particle. The amount of blurring is est:lmated b,y assuming that
the length of a given small tip particle represents the distance through which a veey-
sma.ll particle has moved during the duration of the flash.

The spitter pictures indicate a vecy short effective nash duration, whose
upper limit is conservatively est:lmated at 1/4 microsecond. This should surprise
the reader as 111\lch as it did the authors, i f one remembers that the capacitor across
the X-ray tube has a value of 1 mi'd. and is not constructed so as to have low in-
ductance. For this reason it seems better to speak or effective flash duration
rather than actual flash duration. It is likely that the operation of the Wlr-389 at

112
 .
•
'•
•
•

'
\

•
. '
•
• ••
'

~
•
•
, •
) '
I
•

•
I

••
•

,
•

Figure 6
Figure 8

Figure 7

0 I 2 3 4
I II II I II II LI II !IIIII I I
INCHES

113

A 2122 22£26 I£8

 Q§f klbi iU&SE : •

' •I
J

'I '
'I
,' •,,

•'.

'
J
..
w
I
•
...•
._
'
~
..... I
c.,.,
•
' '
•'
'
''
•
'
J,
••

Figure 9 Figure 10 Figure 11

0 I 2 3 4
Ill 1 d 1 111l11 ul1 11 1 I I I
INCHES

114
 a low voltage causes the effective output to be or shorter duration than the total
fllisho Far instance, the thick glass wall of the tube does not transmit the very
soft X-rays. Likewise 1 the spectral response or the film and the screens may be such
as to record or respond to only a portion of the nash, thus serving to shorten the
-- effective exposure duration. The possibility that a law voltage_ arc ldtll:ln the blbe
terminates X-ray output soon after i t begins haa not been overlooked·. (4,5 J It is the
intention of the authors to examine the problem of e:xposure duration in greater detail.
Present:cy, however, it is interesting to note that even the simple and relatively
crude .flash X-ray equipment that has been described is capable o:f producing such a shcrt
effective :flash duration.

The film used in all our work is Kodak Type F Industrial X-Rizy Film. DuPont
Fluarazure (zinc suli'ide) intensii'ying screens were chosen because oi' their sensi-
tivity to soft X-rays.
Additional radiographs revealed further ini'onnation about circuit perfonnance.
It was found that approximately every -tPitd):tlash was a double flash close:cy spaced
in time (Figure 8). It has been reported(2 that reducing the resonant :frequency or
the trigger circuit has helped to eliminate multiple :flashes. We do not yet know
whether the double :flashing illdicates a retention or control by the trigger circuit,
or whether the multiple :flash is due to sane other mechanism. We hope to control the
multiple nash. However, we have :found that such double exposures .tumish a unique
and accurate method for determining the velocity gradient along the jet since they
provide two superimposed images or the jet slight:cy displaced in time.

O:t more concern is the observed fact that, at 34 KV, every :fourth :tlash is or
too low an intensity to be usefUl. This is most probah:cy due to the fact that the
WL-389 X-ray tube was originally designed to be used at approx:IJnate:cy 300 KV.
Several avenues are being explored in hope of obtaining a more suitable tube. A
modification of the WL-389 may prove to be the solutionJ however, work is being done
at this laboratory by Go Hauver and G. Bryan on a small, low cost X-ray tube. This
work is reported in the paper following this one in the Transactions or the November
1951 Shaped Charge Symposium, entitled, "Expendable Flash X-Ray Tube. 11 Present
efforts promise solution of the problems mentioned. Moreover in practice the ad-
vantages of the low voltage system :tar outweigJl its present iimitations.

The. low voltage :flash radiography circuit was original:cy developed to investi-
gate the jets from 105mm. steel cased rounds. Accomplishment of this, however,
b:inges upon devising a suita.ble method far protecting the tube and film fran frag-
ments and blast. The 105mm. steel cased Z'OUild presents a very difficult shielding
problem as it is necessary to place the film very close to the jet in order to get
good definition. This stems fran the fact that the tube is not a point source or
X-rays, so that unless the penumbra e:f:fect is minimized a shart image will not be
formed. In order to reduce the penumbra the filJD to jet distance must be made as
small as possible. ·Making the -tube to jet distance large will also improve defini-
tion, but the X-ray intensity will fall off as the inverse square ot the distance.
In practice, the film to jet distance is of the order of 6 inches and the tube to
jet distance is o:f the order of· 74 inches.

We have said that low voltage X-rays are particularly suited to jet inv'esti-
gation because they are stopped by low density materiaJ.s. But this means that very
 X-RAY TUBE

PROTECTION.

Figure 12

116
 M!S I Ilk I Et5 sea i§ c:w;:~

5 84 53 4

(/)
1&.1
X
Q
z

117
 little blast and fragment shielding material mlo/ be placed in front of the film and
in front of the tube. The problem is solved by obtaining this protection outside of
the tube to film path,

Figure 12 shows how this is done. The X-ray tube is mounted behind a steel
plate through which a small window has been cut, A second plate is mounted, hori-
zontally, against the first plate and above the window, The round is placed above
the horizontal plate and fired downward (through a small hole) past the film, which is
mounted below the plate. By this device the tube and the film are shielded fran most
of the blast and all of the fragnents from the round, The tube and the film now need
to be protected from o~ the shock produced by the high velocity jet. The pro-
tection shown in front of the tube and in front of the film is plywood, 1/2 inch thick.
This protection has proved adequate.

When the low voltage nash system was first devised, and for sane time thereafter,
our goal was td obtain radiographs of the 105mm. round, Figures 13 and 14 are radio-
graphs of 105mm. jets, Many excellent radiographs have been obtained of a type which
have not been obtained before. The system has proven its value and many new applica-
tions have suggested themselves, Refinements and modificatiops ~ich are in progress
are expected to remove most of the present limitations and extend the usefulness of
the system.

REFERENCES

1, Ballistic Research Laboratories Report No, 678, "Ultra-Fast X-Ray for Multiple
Radiography," by H, I. lh'eidenbach, lmch 1949.

2. Review of Sci'elltific Instruments, Volume 20, No. 12, December 1949, 11 A Fractional
Microsecond X-Ray Pulae Generator for Studying High Explosive Phenomena," by
H, I. Breidenbach.

Ordnance Research Division, Old and Barnes Inc,, Technical Memorandum Report No,
16, "A Simple Flash X-Ray Circuit," by Robert 0, Fleming, Jr.

4. Journal of Applied Physics, Volume 12, No. 2 (1941), "Field Emission X-Ray Tube",
by C, M, Slack and L. F, Ehrke,

5. Proc, I,R.E., Volume 35, No, 6 (1947), 11 0ne Hillionth Second Radiography," by
. G. M. Slack and D, c. Dickson.

6. 11Report
on German Scientific Establishments," by Lsslie E. Simon, Col., 'Ordnance
Department, September 1945, Office, Chief of Ordnance, Washington, D. c., Page
1114.

118

~--~----- --·-~--- ---- -· -··--

 FLASH RADIOGRAPHIC STUDY OF JETS FRG! UNROTATED l05MM SHAPED CHARGES*

L. Zernow
s. Kronman
J. Paszek
B. Tayl.or

Ordnance Engineering Laboratory, Bal.l.istic Research Laboratories,

Aberdeen Proving Ground, Maryland

ABSTRACT

Jets from conical copper and iron lined shaped charges of

1D5mm caliber are compared at various jet lengths. The early
breakup of iron into a particle jet is clearly shown. The- tatty-
like streching of the copper jet due to the velocity gradient is
also quite evident.
Jets from copper trumpet and hemispherical liners which per-
fomed poorly in penetration tests are ehown to be badly formed
and associated with an unfavorable velocity distribution along the
jet.

HISTORICAL

Bver since the first pioneering work of J. c. Clark, and independently of the
British alii the Germans, flash radiography or shaped charges has been confined to.
·small charges. The largest charges reported in the literature have been of the order
of 1.5 inches in dill!lleter with most of the work ~ing done in the 1" diameter range.
This ia attributabl.e to two thillgs. First, the IIISjor interest in the early da;ys ~
in the collapse of the liner. Secondly, the bl.ast shielding prebl.em appeared to be-
ccme rapidly more troubl.esome with the increased mass of high explosive, and it vas
therefOre easier to work with small charges.
JIOliVATION Ftll PRESENT ~RK

Examination of the best fl.ash radiographs availabl.e indicated that although _the
collapse process seemed to be as cl.early delineated as the B!llall cone size would allow1
there vas a serions gap in the avail.able radiographic detail of the jet i tselt. This
vas partly due to tbe fact that the jet from a small charge vas difficult to radiograph
clearly, especially in the early da;ys vhen the fl.ash durations were of the order of one
microsecond. Enin with never techniques cl.aiming shorter flash durations, radiographs-
of 'the jet lett much to be desired in the wa:y or detail because 1110st of the radiographs
were taken under conditions necessary for the collapse to be visible, and these were
definitely not the best tor showing up jet detail.

* Also published as Ballistic Research Laboratories Report No. 802.

ll9
 IDS£ 2
I

tier

Figure 1-Cut-away view of the 105 mm projectile used in the experiments.

120

&223 I RiG I as
 •.

BARRICADE.
WALL
ARMOUR PLATE

TUBE
FILM SHIELD
HOLDER

X-RAY
TUBE

Figure 2---Sket!:h of field set-up for flaah radiography of shaped charge jets.
 ...,~ .... - - - - -

In the work of this group with large caliber shaped charges (90mm and l0$mm cali-
ber), it had been found in the penetration studies that large improvements in both
uniformity and performance level could be achieved with larger caliber l.i.nere. We
therefore expected that if nash radiography could be applied to the stud;y of. these
large liners, one could achieve fr9m this alone an improvement in the visible details
of the jet. Furthermore i f these studies were separated from collapse studies, con-
ditions 1110re favorable to jet examination could be achieved.

The low voltage flash radio~aphic technique developed b;y Paszek and Taylorl was
well suited for field adaptation to this problem because of its comparati~ insensi-
tivit;y to conditions of high humidit;y and because of its relativel;y small bulk and high
mobilit;y. We were therefore in a good position to undertake this work with large
charges.

METHODS

The shielding of the X-Rq s;ystem and the film from the blast and fragments vas
accomplished b;y detouatinS the lD511D11 cased projectile (shown in Fig. l) above a 4•
armor plate and allowing the jet to travel through a hole in the plate past the film.
(Fig. 2 shows the set-up.) The :X-R~ tube am film were therefore subjected onl;y to
the blast effects acc0111panying the jet, to the indirect shock transmitted via the
supports and to the blast that could get thru the small jet aperture or around the top
~late. It is of course essential that tra~Poent ricochets be prevented b;y proper
orientation of the walls of the barricade i f the firing is done in an inclosed barri-
cade.

It was found possible to protect both the film and th!il :X-Rq tube aperture b;y
means of approximatel;y l/2" of pl;ywood. The film holder was designed to allow some
fiexing of t.pe shielding material without allowing the shielding to touch the film.
The holder is shown in Fig's. 3 and 4.

The triggering of the X-Ray tube in these initial experiments vas accomplished b;y
means of a eilaple donble aluminum foil swi~ which was closed b;y the passage of the
jet itself thru the two .oo5 inch Al foils.

It was also possible to measure penetrations into l[ltacks of plate for the same
jets being radiogtoaphed. The standof.(s of course would be determined b;y the jet length
desired in the radiograph.
EXPERIMENTAL PROGRAM

Machined copper liners of 45° angle with and without &pitback tubes and appro:d-
matel;y O.lD• in thickness were compared with dra'IID steel liners of the t,pe used in

1. Paszek, J., Ta;ylor, B., and Squier, J., •Low Volta~ Flash Radiogtoaph;y,n Trana-
actions of the Shaped Charge Symposium (NO'Iember 1.951).

2. Improved s;ystems with electronicall;y generated U. delays han been developed

since these e:xper:lments.

122
 ,..;
~
~
0
..a::
.§
~
....0
~
~
·s:
~
~

]
s:lo
><
~

~
• • • • ~

':;,"'
t>O
~

• •
• • • •

123
•

standard lOSmm H.E.A.T. ammunition • .In addition, the jets from trumpet and hemispheri-
cally shaped liners were radiographed. All charges were loaded with Camp B (60-40) and
were :lllitiated with tetryl Boosters. The charges were cased in steel bodies having a
wall thickness·or about 3/8".
EXPERIMENTAL RESULTS

s__ All of the radiographs included in this report have been reduced to approldmate:cy-
'IPJ of tul.l size.
a. Copper Jets from Mac:hi.lled 45° Cones.

Fig. 5 shows a series of 3. copper jets taken at 3 different conditions of

jet length. Despite the fact that these are jets from 3 different lillers, the s:lmilari-
ty of the f'ront end of the jets is quite apparent, and indicates the reproducibility
that is possible with large charges.

There are two other items of interest regarding this set of pictures.
First, they clear4" indicate that the copper jet is stretching like ta.f'.f'y .under the in-
fluence of the velocity gradient which exists along its length, and which makes .each
element of the jet move faster than the element behind it, thus lengthening the jet in
flight. This is borne out by the changing diameter of the jet, decreasing as the jet
graws longer. This contradicts the idea that the copper jet mlllf be composed or discrete
particles so close together that the radiograph doesn 1t show their separation. Such a
system WQU].d separate into individuall¥ visible particles under the action of the ex-
isting veloeity gradient and would not be expected to show a continuous jet whose dia-
meter decreases with increasing jet length. The latter implies the action of cohesive
forces.
Finally the last picture in Fig. 5 shows evidence of the beginning of
breakup of even the copper jet at long standoff.

It is well knawn3 of course that shaped charges show a standoff effect,

in which the penetration increases to a milx:lm1llll as the charge is spaced farther from
the target. Further increase of the standoff is accompanied by a decrease ii1 the pene-
tration. This has been attri:buted to two things. First, a leveling off of the pene-
tration because of the decrease in average density of the jet resulting from a breakuP
into particles. In addition, the non-collinearity of the velocity vector of each ele-
ment of the jet which anses as a result of deviations f'rom perfect symmetry in the
1:1Jler and the detonation '"'"', results in a radial dispersion of the jet which .begins
to make itself seriousq felt at the longer standoffs and hence contributes to a re-
duction of the penetration at long standoffs. It is i.nteresting that the radiograph at
this late stage shows o~Lcy' incipient jet breakup and no strong indication of radial
dispersion effects.. It is also interesting that this picture corresponds rather close:cy-
to-the condition of the jet near opt:imum standoff which is about 6 cone diameters for
these copper lillers. ·

3. Birkhorr, G., MaoDoa.gaJ., D. P. Pugh, E. M., and Ta;ylor, Geoffrey, "Explosivas with
Lined Cavities,• Journal of Applied Pb;yaics, Vol. 19, No. 6, pp 563-583, June 1948.

~4
 '~,•

CoPPER CoPPER

::0
~
(I)
....,
::0
......
Q
P1
~ tl
trl
(I)
1-'
I
;j
......
I\)
V\
(I)
~
n
() !;
...., ;:;·
tr:l '<
tl ......

, ....,
:::3
0
....
3
Ill

,
<"T
c;·
:::3

Figure 5-Jets from 105 mm copper liners with spitback tubes.

Il£222! LA am Tpfmm g'isr'

s
.......e
j

126
 A second item of interest is the 11 pre-jet11 whic~ has been found to pre-
cede the main jet from liners which contain a ~it-back tube.
' origin in the cylindrical spitback
The pre-jet is believed to have its
tube. This view is supported by the evi:ience that sinrple conical liners without spit-
back tubes do not show this characteristic pre-jet. A typical jet from a sinrple conical
liner is shown in Fig. 6. In Fig. 5 it is the pre-jet which has triggered the f'l.ash
radiograph in all .3 cases.
The presence of the pre-jet in the jet fran spit-back liners suggest that
it may be one of the contributing :ractors5 to the greater penetration of this t;ppe of
r
liner.
b. Steel Jets fran Drawn 42° Cones.
The e:xper:l.ments with drawn steel liners. resulted in the radiographs
shown in Fig. 7. In this case it is quite evident that the steel jet has started to
braak into frapents veey early in its life. The characteristic increasing separation
of the individual fragJ~~ents as the jet lengthens is clearly visible. Also evident is
the relatively constant particle diameter independent of the length of the jet. The
steel liner vas a simple liner without a spitback tube so that, in the absence of the
pre-jet the trigger vas actuated by the main jet itself. This results in a small part
Of the main jet passing thru the trigger foil prior to the X-Ra,y nash, bec81lse of the
few microseconds delay between the triggering, and the nash. The steel and copper
jets are compared in Fig. a.
'There is an interesting speculation possible at this point, which ~s
on &Gille work done recently at the Johns Hopkins University by Kramer and Maddin.
These investigators have found that there is an observable delay in the initiation of
sUp in bod;r centered cubic systems of single ceystals, (to which Fe belongs) whereas
face centered cubic system (to which copper belongs) showed Do measurable dela,y. This
means that a stress applied at a rate greater than a certain critical rate will result
in brittle fracture for the B.c.c. System (Fe) while the F.c.c. System (cU5 will yi,el.d
plastically at this s8111e strain rate. While admittedly the strilin rates in these eJqlEiri-
ments were not ccmtparable to thoile which result front a detonation ;vet they strongly
suggest what ma,y be the first real clue to the basic physical reasons for the difference
between iron and copper.
4. The spit-back tube vas orlg:lnall;y bufl.t into liners to permit the initiation of the
base booster by means of a so-called •spit-back" fuze. Such a fuze contains a
small shaped charge initiated on target impact and tir1ng backward down the "apit-
back tube• to initiate the booster. Such tu~es are much faster than inertia t;ppe
fuzes, but are being superceded by still faster electrical systems. The spit-back /
tube is being retained in manor cases however because it has been shown that ite v
presence ~e_pauetration of the shaped char_~. especially at lower stand-
offs. ·· . ----.
5. Wave shaping resulting from the effects of the cooling of the explosive (Ccnp B) by
the spitback tube during casting is considered an additional possible contributing
factor.
6. Kr8111er, I. R. and Maddin, R.~ IIJlela,y Time for the Initiation of Slip in Metal Single
CrystaJ.s,• Trans. AIMB (l.952J P• l97J Journal of Metals (February 1952).

127

2226 •
 STEEL 5T££L 5 TE:EL

li

'
a- f,.
~

J •

·; '' ~
t%J
rn
1-i
~
~

~
• Q
t%J
t%J
?
''
rn
1-i 1-'
~
1\) rn
~ CD (1)
(')
(')
1-i e;
t%J q"
tj
• t f .._.
N -..,
~
::3

t
0
~
3
....
Ill
s·

•I ::3

..• ~

-- •
f ..
•
•••

Figure 7-Jets from drawn 105 mm steel liners without spitback tubes.
TEEL CoPPE:R 5TE:£L CoPPE:.R CoPPeR

:
..

•
1

f
j •

I
t,
, .
t ...
...

•I
..

Figure 8-Flash radiographs of the jets from 105 mm shaped charges comparing copper
and steel at several standoffs. The early break-up of the steel jet and
the continuous nature of the copper jet are clearly visible.
.... ··.-· •'

c. Jets from Copper Trumpets and Hemispheres.

Fig 1s. 9 and 10 show the results that were obtained with trumpet and
hemisphere shapes. Both of these had given approxilllatezy 1/2 of the penetration ob-
tainable with cones in tests against green IU'lllOr plate. The radiographs clearzy show
the unfavorable mass distribution in these jets and qualitativezy indicate in addition
an unfavorable velocity distribution along the jet, as judged by the relativezy lllllall
amount of jet elongation. Since qualitativezy speaking a jet should be long and thin
rather than short and fat i f one wants to obtain the deepest holes 1 it is readizy seen
why these liners perform poorzy in penetration experiments.
CONCLUSION

!he flash radiographs of jets from large charges add substantial weight to the
crurrentzy- held beliefs that the superior penetration performance of conical copper
liners is associated with their ability to form long thin continuous jets, under the
action of the velocity gradient along the jet.

The taffy-like stretching of the copper jet is clearly indicated.

The break-up of the jet i'rom a drawn steel liner also strongzy corroborates the
previouszy held belisfs that this was the reason for the poorer penetration performance
of iron as canpared with copper, and is in agreement with the results obtained inde-
pendentzy by Pugh using Kerr Cell photography.

These radiggraphs show in addition that the break up occurs at a very early stage
in the process.

The poor penetration performanc~a of a set of copper trumpets and hemispheres can
be made readizy understandable by examination of the radiographs of their jets.

~A -~·,. "'"'tion .,.,.., r e - tho d.,irobW.tq ot • 1ong thlD jot :u tho

amnillnition avematches the target, and one is interested in the maxilmnn damage
behind the plate. In this case it is beginn:ing to be widezy recognized that a
jet which is poor on the penetration criterion ~ be greatzy superior on the basis
of the damage criterion.
B. Recent unpublished work in the radiography of collapsing 105IIDII liners shows even
earlier break-up than is indicated by the radiographs on this paper.

1,30
 RE ST Ri cT ED -S ecu rity
Inform at ion

: ..:!:
·..·.
.:·.··

-------------~~~·~·-~

..;
Q)

~
b
s
s
• "' 10
0,.....

s0
........
..."'
Q)

1
~
Q)
....
~
~

_________._.,_ • • ...
131

RE ST diC TE D
 7s 0
750
.. •

132

'• .
 FLASH RADIOGRAmiC S'lUDY OF JETS FROM ROTATED 105MM SHAPED CHARGES

L. Zemow
s. Kronman
F. Rayf:ield
'I
I J. Paszek
B. Taylor
Ordnance Engineering Laboratory, Ballistic Research Laboratories,
Aberdeen Proving Ground, Maryland
ABSTRACT

The development of new techniques for rotation of large pro-

jectiles and for l!Jif voltage flash radiography has made it possible
to obtain successf'ul flash radiograplzy- of jets from l05mm cased
shaped charge projectiles which are rotating. The first radio-
graphs or this type are discussed.

The effects of rotation are seen in considerable detail. TWo

distinct effects are observable in terms of which the deleterious
effects of rotation m~ be more readily anacyzed. These are (l)
axial breakup, and (2) radial dispersion. A tentative hypothesis
which distinguishes these two effects from each other is proposed.
The possibility of using the shaped charge as an experimental
device for studying.otherwise inaccessible properties of liner
materials under very high strain rates and unusual stress combina-
tions is pointed out.

The advantages of a cylindrical liner are discussed, and, in

particular, attention is called to its expected resistance to spin
deterioration. There is also noted the additional advantage that
the penetration from such a liner can be made to depend upon the
liner length rather than the projectile caliber.

HISTORICAL BACKGROUND

The first pUblished information on possible deleterious effects of rotation on

the performance of sha:geo charges seems to have come simultaneously from the British
and the Germans. Tuckll) in 1943 qualitatively explained the effects of rotation by a

(l) AC 3596 - 11 A Note on the Theory of Munroe Effect•• - James L. Tuck.

133
 QSS!JL. a s rdnr-Trfzmst' ;.
widening of the jet which could be predicted from the resultant of the collapse
velocity and the tangential velocity du~ to· rotation. His work may have been motivated
by the prelimi:t}B.fY experimental results l2} which were formally reported some months
later, Trinksl31 in April 1943 estimated "considerable reductions 11 in the penetration
due to the expansion of the jet diameter resulting from centrifugal force effects, He
also remarked that the final result would probably be a radial spray of "droplets"
rather than a simple expansion of jet diameter, At essentially the same time in May
1943 the British formally ~ported their first experimental evidence of rotation
effects in a 311 field gun,l2 J They verified that the observed deterioration ·in per- r
formance was attributable to rotation, by spinning the projectiles with an automobile '
s.tarter motor and observing similar deterioration, This appears to be the first re-
ported instance of a proje.ctile rotating device for shaped charge studies, It is also
noteworthy because of the large caliber (311) :Involved,

Following the British experiments, th~ ~tt!Q.y was taken up in this country by the
Explosive Research Laboratory of the OSRD,l4H5)l6J The British results were confirmed
am extended, Most of the work reported by ERL was done ~Tith smaller caliber M9Al cones
of 1 5/811 diameter, Considerable work with 57mm and some experiments with 75mm were re-
ported, A single case was reported of a 105mm projectile which was rotated during deto-
nation, Several years after the end of World \far II, additional contributions were made
by Carnegie Institute of Technology in connection with spin compensation studies, The
experimental work by C,I.T. also was mainly carried out with small calibers of the order
of 57mm,

All of the. work so far described :Involved studies of penetration into targets.

In a BRL Report published in 1948, Clark & Fleming,(7) described earlier work that
had been done in flash radiography of jets from cones and hemispheres of 1 3/8 11 base
diameter. The work described was carried out prior to May 1945 and was probably the
first flash radiographic work on rotated shaped charges, The radiographs in the report
of Clark and Flemiilg indicate that the jet was dispersed and broken into particles by
the rotation, but their pioneering radiographs were not clear enough to give a quantita-
tive idea of the dispersion process.

No other .flash radiographic studies of rotating shaped charges have been found,
although the Gennans reported(8) nmltiple Kerr cell studies which indicated the
llspreadingll of the jet due to rotation.

(2) AC 3987 - "Hollow Charge Rotated Projectiles11 - (M,D,I,)

(3) Dr, Walter Trinka - "Mathematical Study of Lined Hollow Chargesll - Report 43/6
Translated as OTIB No. 1484 - 30 April 1943.
(4) OSRD 1680 - ttTarget Penetration by the Jet from a Rotating Cone Charge 11 - August 5
1943.
(5) OSRD 3874 - IITarget Penetration by the Jet from a Rotating Cone Charge 11 - July 10,
1944.
(6) OSRD 5598 - IITarget Penetration by the Jet from a Rotating Cone Charge" -
November 5• 1945.
(7) J, C, Clark and R, o. Fleming, Jr. - IIEffect of Rotation Upon the Explosive Col-
lapse of Thin Metal Surfaces of Revolution" - BRL Report No, 671, July 1948.
(8) Leslie E, Simon, Col,, Ord Dept. - "Report on Genuan Scientific Establishments" -
Chief of Ordnance, Washington, D. c., Sept. 1945, Page 97.

JJ4

ttmwmwar· .
 A theoretical treatmept of the et'f\lcts of rotation, based on conventional hydro-
dynamic theory, was given bY Birkhorrl9J vho predicted a hollow jet under rotational
forces. He also suggested that the jet ~~~~zy break into droplets.
The uncertainty regarding the details or· the "spreading" process and the need for
additional experimental data which could be used as a basis for further development of
the theory motivated the undertaking of the present wrk. It is believed that only by
means of an understanding of the detailed mechanism of spin deterioration can the most
• efficient attack be planned on the problem of spin canpensation.

BASIS OF JRESENT V.URK

The development of the wire-driven projectile rotator(lD) so simplified the prob-

lem of rotating large caliber shaped charge projectiles that lD5mm projectUes could be
easily rotated •t frequencies in excess of 350 revolutions per second. The only ex-
pended element in this rotation system is the steel drive wire and contactor assembly.
Figs. 1 1 2 and 3 show the pertinent details of· the wire driven rotation system, 'Which
is .located on Spesutia Island and has been nicknamed "Big Mo11 because of its massive
construction. The rotating projectile is so stable in this system that during the
SUJillller t:iJIIe, flies repeadeuly try to alight on the spinning projectile, only to be
violently thrown away.

The second essential development was that of the s:lmgle)low voltage system of
nash radiography described by Paszek, Taylor and Squier.1ll

"-· These two recent developments have been cambined and have made possible this
study of jets frCIII rotated 105mm shaped charges. These liners are more than twice as
large as any previously reported in radiographic studies. The advantages of the large
caliber are two-fold in these experiments:

1. There is a gain in detail and reproducibility resulting from the larger

dimensions of the jet and the cilinensional uniformity easily obtained with large cones;,

2. The greaten sensitivity of large calibers to rotational ·deterioration

makes it possible to cover almost the entire range of interestillg phenomena in the
frequency range frCIII 0 to 350 rps. Smaller calibers wo:uld require correspondiil~UY
higher rotational frequencies, so that if, for example, 1 3/4" diameter liners. were
used, rotational frequencies up to 100 rps would be required to bring out corresponding
effects.

(9) Garrett Birkhoff - "Hollow Charge Anti-Tank Projectiles• - BRL Report No. 623,
lD February 1947.
(10) s. Kroman, and L. Zernow - "A Wire Driven Projectile Rotating Device for Hollow
Charge Investigations" - BRL Report No. 798, March 1952.
(11) J. J. Paszek, B. c. Taylor and J. L. Squier - 11 Low Voltage Flash Radiogra:pey'1 -
Transactions of SymposiUm on Shaped Charges, held November 1951 at BRL.

1.35

••
 t'

Figure 1- 0verall view of wire driven rotator.

136
Figure 2-View of motor and monitor Figure 3-Closeup view of 105 mm projectile
of spin frequency. on wire, and electrical firing system.
 f!tJNt lbiN IIXE±seea: ll§ Lib! iil&Hbft

Figure 4-Cut-away view of the 105 mm projectile used in the experiments.

138
 EXPERIMENTAL PROGRAM

The work described here was done d•.tring the summer and fall of 1951 and represents
the initial effort to study the jet from rotated liners at various rotational fre- ·
quencies. Flash radiographs were taken of 105111111 caliber 4.5° apex angle machined copper
liners rotating at o, .30, 45 and 6o rps. In addition some steel liners were fired for
comparison.

Fig. 4 shows the details of the projectile used in these experiments;

The details of the flash radiographic; S(lt-up are essentially the same as
those reported ln the study of unrotated jets.ll2 l The projectile is hung from the
motor driven wire in these experiments, as shown in Fig. 3.

ANALYSIS OF .EXPERIMENTAL RESULTS

Fig. 5a compares capper jets from an unrotated projectile and a projectile rotated
at 60 rps. FOE' the purpose of reproduction the radiographs have been reduced to ab011t
'1$1% of natural size. The tip of the rotated jet is seen to be essential]Jr unaffected b;y
the rotation. At point A we see the radial break up of the jet into what looks like
two separate particle jets (bifurcation). This effect is seen to become more: pronounced
from point A upward until final]Jr at point .B we see what appears to be still further
break-up (po~roation) of the already bifurcated jet, with the radial spread of the
fra~ents still increasing. At point C the fra!P'Ient spread has reached a maximum. The
significance of the bifurcation process will be discussed in a following paper.

The radial disturbance of the jet seen in the radiograph is at least partial]Jr
understandable in terms of the simplest notions of jet formation. The t:ip of the jet
has its origin at the apex of the inner liner surface. The pm'tions of the jet which
follow the tip cane from portions of the liner successive]Jr further fran the apex, and
hence fran successively greater radial dietances from the axis of rotation, as . shown in
Fig. !)b, Hence, it is to be expected that the front tip of the jet would be una.ffecteC
b;y rotation, and that the rearward portions of the jet 110uld bs progressively more
serious]Jr disturbed in a radial direction. This t;ype of disturbance extends to point C
(Fig. 5a). The region from C to D will be discussed later. .

The ear]Jr break-up of the rotated jet into particles in the axial direction is a
mora subtle phenomenon. The apparent trend toward uniformity of the size of these
particles suggests that they' may have their origin in multip]Jr reflected shocks either
in the liner* ar the jet. The jet from an unrotated liner of the same type (Fig. 5a)
shows incipient jet breakup in the form of a ps~o-periodic necking down of the jet.
'
The next question that arises concerns the reasons for the marked effects of
rotation upon the axial breakup process. We believe that the explanation involves
the introduction, as a consequence of rotation, of a rotational velocit,y gradient. in
addition to the ordinary axial linear velocity gradient. This rotational velocit,y

(12) L. Zernow, s. Kroman, J. Paszek, and B. Taylor - "Flash Radiographic Study of

Jets from Unrotated 105DDII Shaped Charges" - Transactioll8 of Symposi11111 on Shaped
Charges, held N<>V8111bar 1951 at BRL.
Dr. A, w. Hull, 1111ong others. has suggested that such vibrations in the liner may
be "'l'blipoll81b1e.
139
 UNROTATED ROTATED AT
60 r.p. s.

---0

~-
'1\
c
.,,
,~~

.,..t!E-
ft.';t (")
0
~
~.~ .....
0
·~.. ~ t%.1
z~
(')
0 ~t.~
\ -..~,
' .....
z>rJ .•J
f
,,
.~.-:..··'
.....
0
t%.1
~ CJ)
(!>
z 0
• •••
0

~
.....
~
·-... .
• 5.
~

'<
.....
. ' f ::l
.....
0
(II
01:
~ ·~ "1
3
Ill
j ' I ~

s·
"' ,r ::l

"
f'· t ~
.- .:l
'·.,.,1
•" 8

t I

,
FLASH RADIOGRAPHS
OF JETS FROM 105mm
SHAPED CHARGES
A
• fig. 5A

Figure 5a- Flash radiographs of jets from l 05 mm shaped charges.

 INITIAL POSITION
OF LINER

Figure 6a

LOCATION ON JET OF AN INTERIOR CONE ELEMENT

(WITHOUT ROTATION)

® RELATIVE POSITION OF
JET ELEMENTS OF AN
UNROTATED JET AFTER
()j\/"S: / s: ...· •
LOCATION ON JET OF AN INTERIOR CONE ELEMENT
)
.. DIRECTION
OF JET
DETONATION
( WI'TH ROTATION 8 WITHOUT AXIAL VELOCITY GRADIENT)

FLIGHT
Figure 6b

os:: ,,. "s

LOCATION
< ........

ON JET OF AN
"SO . ."'
INTERIOR CONE ELEMENT
(WITH ROTATION a AXIAL VELOCITY GRADIENT)

Figure 6c

Figure 5b-Sketch showing how successive jet elements come from FOR THESE SKETCHES THE JET IS ASSU~ED TO BE A
greater radial distances. CONTINUOUS ELASTIC SYSTEN
gradient could arise naturally from the variations in the angular velocity or a jet
element at the extrusion point due to the variations in the radial· distance of its
parent ring in the liner. Tlms, in the light of this concept, the jet, i f it remained
continuous, would not only be stretching due to the ordinary axial liner velocity
gradient, but would also be undergoing a torsional deformation. The forward tip
would have an angular velocity about the jet axis initially equal to C&l01 the pro-
jectile 1 s angular velocity, because it originates at the apex of the liner. Successive
rearward portions of the jet could have increasingly larger angular velocities about the
jet axis. The jet would thus be stretching and winding itself up simultaneously as
illustrated schematically in Figs. 6a and 6b. It is interesting that as shown in Fig.
6c the axial linear velocity gradient would tend to decrease the variation in pitch of
spiral representing the angular displacement of an interior cone element making it more
neady a uniform spiral.

The "winding up" process would superimpose an additional torsional stress on a

continuous jet. This torsional stress would increase with time i f the continuous jet
were considered to be an elastic system. Now, if this time dependent stress system is
superimposed on the one resulting from the axial linear velocity gradient, in which the
stresses also would increase with tL~e, it would be possible to account for the
accelerated axial breakup of the rotated jet on the basis of the known reduction in
torsional ductility of a wire subjected simultaneously to tension and torsion. A
uniform rotational velocity gradient along the jet could explain the unifom logitudinal
breaku:p.

It might also be pertinent to consider the effects on the angular velocity gradient
in the vicinity of the extrusion point which would result both from the initial con-
traction and from the subsequent expansion of the jet following its attainment of mini-
mum diameter at what might .be called the "vena contracta" by analogy. This process of
contraction and subsequent expansion (which is 'stron~ly suggested by both the radio-
graphs and the target data to be reported separately) occurs i:>_ver a fairly small axial
distance and therefore a large gr.adient of angular velocity will be established. in that
region. It is interesting to note that the jet appears to be continuous in the vicinity
of the "vena contracta 11 and shows breakup clearly only after the expansion. Also, the
expansion is nruch more marked in the rotated jet than, in the unrotated one. Since it is
likely that the jet is more nearly a plastic system than an elastic system, these con-
siderations would therefore need to be modified from this point of. view.

Let us now return to the region C-D of Fig. 5a. Here the jet appears to be be-
having anomalously since the rearmost portions of the jet seem to be less affected,
although they should be more dispersed. Various explanations may be given, including
the possibility that because of their later extrusion, these parts have not had time to
spread to the full extent. However, the results from a single radiograph are not
sufficiently conclusive. A multiple nash system now under development should furnish
the required additional information.

MEASUREMENT OF VELOCITY DISTRIBUTION ALONG THE JET

Fig. 7 compares the nash radiographs of copper lined charges rotated at 3', 45
and 60 rps. The 30 rps radiograph is an especiall;; useful one of the double nash
type, lihich makes it possible to measure easil:y" the velocity gradient along the jet.
 .. ~ .
.~ ...... .
. ... ,
.,: • I
'-

.. ..
"' 30 RPS 45 RPS 60 RFS
••

~~
\ ...
JETS OF 105 ... CHARGES
. . . ,.
..( 45° COPPER CONES WITH
SPITBACK TUBES AT VARIOUS
~\
;) ...
-wr RATES OF ROTATION . .
··~·.
·~~
~
~-
""
0.
~#4
\ ~
• J.
., I

..- j.
-~
·•••••
~.~··'
Cl
0 ~
0

~\.
0 '

, •• t :-~
... '
.....,
',,,
~-'
~~
s::-
\.IJ
.. ,,
..,,,. ,..
;.a ir .,,
~·~

,,.
l'·' ~...

~·
• •
-~~ ~

·''\•·..
-.-
~
,\ .
,, o.-)--

r
I

~
,t: '·It
~
01
't. ttN
•

F igur e 7-Jets of 105 nun charges. 45° copper cones with·spitback

t ubes at various rates of rotation.
?? ££2 §@@dill§

Figure 8
The results of such measurements are plotted in Fig. 6. It is evident that the velocity-
does not vary linearly- along the jet. The deviations fran linearity are believed to be
real. However, these deviations are fairly- small. In addition the average of the two
rounds closely- approaches linearity-. It can therefore be assumed that pending a more
detailed analysis and extension of the experimental data thiiJ ~sult is not inconsistent
with that fran measurements made by- the Carnegie Tech •. groupll3J using the rotating
mirror technique in conjunction with slug recovery experiments. The simplicity- of these
measurements fran a double · flash radiograph constitutes a distinct advantage. However,
this is partly- offset by" the following disadvantages: (1) the double nash method is
not as readily- applied to a continuous jet and (2J the double nash, at present, is a
random occurrence.

EFFECTS OF VARYING ROTATIO!W. FR~UENCY

The first effect of rotation appears to be the axial breakup process, which appears
to be. thep~ominent process at 30 rps, Evidence for il}.cipient bifurcation can be
seen in this radiograph although there is other evidenceUhJ that it may- already be
well under way at the rearmost part or the jet.
At 16 rps the bifurcation is clearly- aPJ)arent. In addition, the axial breakup
appears to be much more severe than at 30 rps. However, a single radiograph could
conceivably- fall to show the bifurcation if it occurred in a plane perpendicular to
the film. At 6o rps, the breakup of the jet appears to be not appreciably- greater than
at 4S rps. · In fact in this particular radiograph the dispersion of the particles seems
to be smaller. This is also attributable to radomness in the orientation of the plane
of bifurcation relative to the J.:ilm planeo AlthoUgh radiographS· of jets of 30 rps
have.npt shdwn a discernible bifurcation, all radiographs or jets at 4S or 6o rps show
the bifurcation clearly-.
If the 30 rps radiograph is considered to be one showing incipient bifurcation,
the breakup or the jet under rotation may be interpreted as resembling a catastrophic
process, in which up to a critical rotational frequency there is no breakup. Past
that critical frequency the onset of axial breakup occurs. F1na1ly at a higher fre-
quency bifurcation of a ·singificant portion of the jet length occurs. This hypothesis
is qualitatively- presented in Fig. 9. Two critical frequencies are postulated; cO , for
longitudinal breakup and co,, for significant bifurcation. There is no apparent pf!isical
reason that both axial breiiJrup and significant bifurcation should occur at the same
critical frequency. However, that a critical frequency (I) may- exist below which no
marked effects are a:PParent can be shown from phy-sical col1siderations. It is to be
expected that such considerations would involve the elastic and plastic properties of
the liner. Because or the potential importance or these phenomena, they are being in-
vestigated in greater detail. ·

COMPARISON OF COPPER AND STEEL LINERS

Fig. lD canpares copper and steel liners rotated at 45 rps. The unifomly poor
behavior of the steel liner is in agroeement with the radiographs or jets from unrotated

(13) C.I.T. - ORD-31 - p. 33.

(1.4) L. Zernaw, J. Regan, J, Simon, and I. Liebennan - "Study- of the Effects p.f
Rotation upon the Penetration or Jets from IDSDDII Shaped Charges 11 - Transactions
or Symposium on Shaped Charges, held November 1951 at BRL.

-·
 CURVE DEPICTING
PROPOSED HYPOTHESIS OF CRITICAL
BREAKUP FREQUENCIES

c..l.=CRITICAL FREQUENCY FOR

INCIDENCE OF LONGITUDINAL
BREAKUP

WI =::RITICAL FREQUENCY FOR

INCIDENCE OF BIFURCATION
OVER MAJOR PART OF
z JET LENGTH
0
l-
et
a::
1-
w
z.
w
Q.

ROTATIONAL FREQUENCY

Figure 9-Curve depicting proposed hypothesis of critical break-up frequencies,

146'
 • <

•.
charges as reported in a previous paper, (l2) The unrotated steel liner breaks up into
particles at a very earzy stage. Rotation seems to hONe resulted in a further break up
of the particles which now are smaller on the average than those in the unrotated steel
jet,

The massive jet tip from the steel liner is attributable to the peculiar con-
struction of the particular steel liners 1 which have a blunt hemispherical apex of about
1/411 radius.
'lbe earzy breakup of the steel liner apparentzy results in a rather unifom radial
spread of particles, 'Which suggests that the rotated steel jet may already be broken up
at the 11 vena contracta11 • The material from the blunt apex, however, has obscured the
shape of the jet tip sufficiently to make the anazysis of radial dispersion somewhat
dubious, These experiments will be repeated with better steel liners before further
analysis is undertaken.

THE SHAPED CHARGE AS AN EXPERD!ENTAL HlGH STRAIN RATE DEVICE

The fact that the shaped charge can be used as an experimental means of appzying
_very high strain rates and unusual. stress combinations to a liner seems not to have
been generally recognized, A good start has been made in the quantitative study of the
interactions of s:\.J.!~Ble (!labs of explosive and slabs of metal., for example, as in the
work of Rinehart, ~1~,16} and it has been noted that the processes observed are to same
extent characteristic of the high strain rates as well as the high pressure amplitudes.
The relativezy simple geometry of these experiments is' a desirable feature, both from
the experimental. and the analytical points of view. However, it appears from. these
initial studies of rotational effects that by the use of the shaped charge additional
infol'lllation may be obtained which may not otherwise be accessible; for eX!IliiPle, rotation
apparently pemits the superposition of a torsional. and tensional stress system. The
liner need not be a cone, but for simplicity ma;y- be a cylinder.

THE CYLINDRICAL LINER

Experiments* have been perfomed with cylinders 'fmich indicate that a jet can be
obtained, Also, there is flash radiographic evidence of jets from cylindrical. liners;
for example, such as reported in a previous paper. (J2 ) The cylinder, in addition to its
simplicity has several other potential. advantages: (1) The cylindrical liner should be
relatively insensitive to rotational effects since all of the matex:ial entering the jet
can be made to came from near the axis of rotation. This is of very great, importance
frau the viewpoint of application. (2) The penetration from a cylindrical. liner could
'
(15) J. S, Rinehart- J, Applied.Physics, 22, 555 (1951)
(16) J, s. Rinehart - "Scabbing of Metal. Under Explosive Attack - Multiple Scabbing" -
NOTS 'I'M No. 348.
* This general problem of cylindrical. liners has been discussed with Prof. Richard
Courant or N.Y.U. who has independent!¥ arrived at essentially the same con-
clusions, which he discussed briefzy at the ~osium in an unrecorded invited
paper.

147
 a r MIL §n stu h'sr 1 1

7.2'•
.. , I ll •
.. 1
•

.. \ .. - ~·. • '·~~..~...... Jilllllt ilfll

.ra.....J,.._,.., ,. .
'~·• ,..~. ·~• · , ~, ·: .'·~· .~• .., ··- ~-~
\..
·..'II~ ~: ·1·\ ..~ .~.
..
1

..J
UJ (f)
UJ (/)
(f)
Q..
~ 0:: ..J
(/)
I- w w 0::
w z w l()
J t- v
LL
0
..J (f)

0::
..
0 <tj
w
z w 0::
(I)
0 ~ w ..J
(f) t- Q.. ~
-0:: 0 Q.. u
0:: 0
~
0:: Q.. 0 ~
w ~ ~
~
Q.. 0 LL
Q..
0 0:: 0 l()
0
u LL 0
0

.s 8 4 .S84

Figure 10--Comparison of jets from rotated liners of copper and steel

105 mm caliber, 45 RPS~

148

?FEE EYnW:
 be made a function of the liner length rather than of the diameter of the projectile.
This characteristic therefore makes it possible in principle to design a long projectile
of small diameter which can pierce a target much thicker than that which can be defeated
by a cone of that caliber. Moreover, there are other independent reasons why such long
thin projectiles are desirable. (3) The collapse of the cylinder could be made to _be
a true steady state process, and therefore more amenable to ana~is. This is dis-
tinctly advantageOUlil in correlating experimental results. (4) The modification of the
collapse process to incorporate a velocity gradient in the jet appears to be achievable
by a variety of means among which is simple ~~ave-shaping.

It is felt therefore that much more attention should be paid to cylilldrical liners
both because of their simplicity and because of their important potentialities in a
weapon. A comprehensive program has been under way ·for the past year. A1 though the 1 / ·
jet from .a cyl:lllder so far has not achieved a penetration as good as that obtainable Y
from the cone that can be put into the same body, the payoff will be big if it can be
accompliShed.

SUMMARY
The radiographs discussed represent the first ones obtained with the new wire
rotator and low voltage flash radiography techniques. They represent only a start on
the task of understanding the fine details of the effects of rotation on shaped charge
liners. ·

The attempts to interpret these radiographs 11111st also be considered preliminary.

' The large amount of visible detail in the radiographs seems to make them quite fruit:t'ul
in providing a firmer experimental basis for a theoretical approach. However,
additional radiographs are needed to remove some of the uncertainties which remain be-
cause these first radiographs are single exposures.

CONCLUSIONS
It is concluded that rotation of a conical copper shaped charge liner causes two
distinct defects tO appear in the jet. The first is the early break-up of the jet
axially into particles, which, therefore, even in- the absence of an,y other effects
11011ld teild to make copper behave· like steel and hence reduce its penetration. The
second effect is a radial break-up of the jet which appears to be assignable directly
to the action of the centrifugal forces resulting;'-from the rotation of the liner. It
takes the fol'lll of a bifUrcation of the inain jet followed by what may be successive bi-
turcations of each branch (polyturcation).

The understanding of these phenomena is progL"tlssing rapidly. It is believed that

this understanding is a prerequisite in the process of bringing new quantitative ideas
to bear on the problem of how to overcome the deleterious ei'fects of rotation.

Anacysis of the radiographs from rotated liners .suggests .the possible use of the
shaped charge, perhaps in simplest form with a cyl:llldrical liner, aa an experimental

149

&&Jib-. .
 Q2!LL£i 5

device for making accessible otherwise elusive phwsical properties of liner materials
subjected to high strain rates and 11Jl11Sual combinations of stresses such as torsion .
and tension.

The cylindrical liner is e:xpected to exhibit several very useful prop!3rties, among
Which are the ability to resist rotational deterioration, and the possibility of making
its performance depend on length rather than caliber of projectile.

150
 EXPENDABLE FLASH X-RAY TUBE
George Hauver

George Bryan

Ordnance Engineering i.aboratory, Ballistic Research Laboratories

Aberdeen Prov.l.ng Ground, Maryland

ABSTRACT

For the flash radiography of explosive phenomena, an expendable

x-rrq tube ma;v be used without protective shielding, .facilitating
flexible instrumentation, Experimental expendable tubes of the field
emission type are described, along with preliminary performance data,

INTRODUCTION

Upon preliminary considerations an expendable or inexpensive x-ra;v tube seems to

offer several advantages over the tubes. curr~tly used for the flash radiography of
explosive phenomena, As mentioned by Fleming , an expendable tube mrq be used without
shielding, Ordinarily, massive protective shielding is essential for- the radiographic
study of hea:vy charges and rearrangement of instrumentation then becomes a major con-
struction job, By ellminating the necessity for shielding (or at least reducing the
required shielding by a substantial amount), the great advantage of flexibility of
field instrumentation is attained,

The tubes under investigation have been design~ for use with the low voltage
x-rrq circuit developed by Paszek, Trqlor and Squier .for flash radiographic studies
at Aberdeen Proving Ground. A schematic diagram of the circuit and a 3-electrode
field emission type x-rrq tube appears in Figure 1. A potential, less than the
critical breakdown potential of the tube, is applied between the cathode and target.
A negative pulse is applied to the trigger electrode, brealdng dawn gap AB, and con-
sequently gap AC. It should be noted that, at present, the x-r83' tube is a major item
in the cost of a low voltage unit and that an inexpensive tube woul.d greatly reduce
the total cost.

CONSTRUCTION

Of the tube types .first constructed, some were made at the Ballistic Research
Laboratories while others were made to order by a c011111lercia1 manufacturer. The design
of those constructed in the Laboratories is shown in Figure 2, which represents a tube
with a Pyrex: envelope into which is sealed 40 11111 tungsten leads, The tungsten target,
S mils thick and l/4 inch in diameter1 is spot-welded to a IS 11111 nickel strip which
in turn is spot welded to the tungsten lead.

151
 19.2 MEG ~
'>

~ 5 MEG.
;: !== lp.F
.,.
"'
1_; ...

AB
5 MEG.•
:::: lp.F ">

15 KV I
-- .OIJI.F
3C45

IIll
,;
;:.--~ ~

~~
.
>
-;-\
P. s. :> 3.8
-
.,; .Oip.F • :: ,> MEG.
134KVI
P.S. ' 10;1

4P30i ;
Figure 1-Low voltage flash x-ray ·circuit.

~-------------4·----------------~

TUNGSTEN

CATHODE

"'- L' GAP

32

TRIGGER

Figure t---Ordnance Engineering Laboratory Tube.

152

d TST
,, _ __,.,__ T
 I ?§gg"•lf 'jj!P

Tubes were evacuated to pressures of from 10-2 to 10-.3 microns and tested lihile
still on the vacumn system. The vacuum system consisted of a three stage oil dif-
fusion pump backed by a Welch Duo-Seal forepump, liith a liquid nitrogen cold trap be-
tween the tube under evacuation and the diffusion pump. Pressure was measured with
National Research Corporation Type 507 ionization gauges. Tubes were not carefully
degassed but baked at .35o degrees Centigrade for periods of 12 hours.

The commercial tubes are of the types shown in Figures .3, 4, 5, and 6. These
types differ only in their trigger and cathode structures. Tube envelopes are of
Kovar sealing glass, into which are sealed leads of 6o mil Kovar. Tungsten targets,
42 mils thick and 1/2 inch in diameter, are spot welded and copper brazed to the
Kovar leads. The standard target-cathode spacing is 5/B inch, although for the tube
type of Figure 6 this spacing varies between limits of .3/B inch and 1 inch. The com-
mercial tubes were evacuated to pressures of about lo-.3 microns, but were not thoroughly
degassed .by standard procedure.

TESTING

Tubes were tested with the circuit shown in Figure 1, and x-rq outputs were
measured by Victoreen ionization chambers of the 0-10 milliroentgen size. Radiographs
of lead test patterns l-rere taken on Kodak Type F x-ra;y film, using Patterson Fluorazure
{zinc sulfide) intenSifyine screens. Previous tests indicated Fluorazure screens to
be the most efficient type for low voltage radiography. Photographs of the visible
flash in the tube were taken along w.ith pinhole radiographs of the target.

The highest consistent output from the BRL tubes, measured 20 inches from the
tube, was 0.4 milliroentgen. An image of specular density 0.12 was formed on x-rq
film placed at the same distance from the tube. After several flashes, tli.e interior
of the envelope between the cathode and target was observed to h:rve developed craze or
fine surface fractures, preswnabl¥ caused by heating. Optical pictures indicated a
concEntration of vapor along tube walls.

Values of some circuit components were changed during tests of the commercial
tubes. Tubes tested with the circuit as shown in Figure 1 gave, on the first nash,
an output of 1.0 milliroentgen at .36 inches. An image of specular density o.B was
formed on x-rq film placed at the same distance. Pinhole radiographs indicated
greater x-ray emission from that portion of the target nearest the cathode, and optical
pictures indicated a greater concentration of vapor near the same portion of the target.
With the main capacitance reduced to 0.25 microfarads and the trigger capacitance re-
duced to 0.01 microfarads, an output of 1.0 milliroentgen was still maintained for the
first flash. In all cases, the tubes became gassy after the first nash, and the
second nash yielded 0.5 milliroentgen measured .36 inches from the tubes. Further
flashing produced no output meai!urable at this distance.

DISCUSSION

The preliminary investigations lead to the obvious conclusion that some changes
in design and construction are necessary to obtain increased and consistent x-rq output.

15.3
 4"

'-I" ii; GAP

TRIGGER ELECTRODE

Figure 3---(;ommereial tube.

1--.,....--------4" ------~

f
I"

TRIGGER ELECTRODE

Figure 4-Commercial tube.

1 - - - - - - - - 4" - - - - - - - . a . l

'
I"

CATHODE

TRIGGER ELECTRODE

Figure ~onunercial tube.

'~-~·--,----4"-------t
.p\t'• I II
1
11
ji: GAP

TRIGGER ELECTR:lDE

Figure 6--Commercial tube.

154
Several possibilitie;~ for improving the output nre euggested by the mechanism presented
by Kingdon and Tanis"'. When the tube is flashed, the electrons initially fall through
the mtire applied potmtial. As the voltage .falls orr .from this maximum, the radi-
ation becomes sorter and is finally cut orr entirely by the .formation o.f' a low voltage
arc. This arc is caused by the presence or gas or metallic vapor which, when ionized,
neutralizes the space charge and allows large currents to flow. With large currents,
most o.f' the voltage drop occurs in the leads with little energy going into the produc-
tion o.f' x-rays. If gas or vapor is evolved too rapidly the x-ray output is cut o.f'.f'
prematurely.

Apparmtl.v there are three sources of gas or vapor. First, the commercial tubes
were not thoroughly degassed and became gassy a.f'ter the first .flash. When the main
condenser was slowly charged .f'or the second flash, the gas was cleaned up but was
rapidly re-evolved and .ionized by the second flash. Second, photographs o.f' the visible
flash and surface .fractures :In the glass indicate that vapor was evolved .from the bom-
barded walls, in wh:j.ch case the envelope should be enlarged. Third, i.f' vapor is .boiled
.from the target, a larger area should increase the maximum output. In .fact, pre-
liminary- tests indicate a dependmce upon target area.

l. R. · o. Fleml.ng, "A Simple Flash X-Ra;,y Circuit" 1 Progress Report 16, Technical
Memorand11111 Report 161 Ordnance Research D:l:dsion, Old and Barnes Inc., .Pasadena,
California. (1951)

2. Paszek, Ta;ylor, and Squier, "Low Voltage Flash Radiography''• Transactions o.f' the
Hollow Charge Symposi111111 Aberdeen Proving Ground. (November 1951)

). K. H. K:l.llgdon and H. E. Tanis, Physical Review 53, 128-1)4. (1938)

:11 £2!2B seeaq asa I

4 £LICIJllb
 THE KERR CELL CAMERA AND ITS APPLICATIONS

E. c. Mutschler

Department of Physics, Carnegie Institute of Technology, Pittsburgh, Pennsylvania

' ABSTRACT

The Kerr cell method of high speed photography has been used
by this group to photograph both luminous and non-luminous transient
phenomena at exposure times do1111 to 1/2 microsecond. In the case
of non.;.luminous phenomena, the opening of the Kerr cell is synchro-
nized electronically with the light flash obtained from an tlectri-
cally exploded wire having a peak intensity or about 5 x 101:1 candle
power. The Kerr cell is or sufficient size to allow the use of a
1 inch, f/2.5 lens at an effective aperture of r/4. To operate·
such a large Kerr cell requires a voltage pulse or amplitude 25KV.

This method is ideally suited to the photography or detonating

explosives, mstal jets in shaped charges, shock waves in transparent
liquids and solids, and other phenomena having propagation veloc-
ities up to 10,000 meters/sec.

INTRODUCTION

A nUIIIber of unusual difficulties must be overcome if the metal jets from lined
cavity charges are to be successfully photographed. The sub-microsecond exposure
time required to "stop" the jet ~es out the use of mechanical shutters for this
purpose. The intense luminosity of the exploding charge rules out "open flash"
methods in which the phenomenon is lighted by a flash or light of very short duration
from a spark or from a flash tube. Fig. 1, a Kerr cell photograph of the detonation
or a cylinder or pentolite, will serve to point up the unusual conditions encountered
in the photography of intensely self-luminous phenomena. Note that, in spite of the
short exposure time or one microsecond, there is sufficient self-luminosity available
for photography. Furthermore, close inspection of the picture reveals a blurring of
the image at the detonation front due to movement during the one microsecond the film
was exposed. Cavity charge jets are relatively non-luminous when compared with the
exploding charge. Consequently, a brilliant light source synchronized with a shutter
able to produce one half microsecond exposures is required to photograph these jets
by visible light. In the c.I.T. setup, a length or fine wire, exploded electrically,.
fulfills the light soFce requirement by giving a peak intensity or approximately 500
million candle power, and a Kerr cell electro-optic shutter determines the duration
or the exposure.

I. KERR CELL CAMERA AND LIGHT SOURCE

Fig. 2 shows a schematic diagram of the setup for Kerr Cell Photography. The
explosiOn takes place in a bombproof separated from the observation room by a thick

157

sa V
 Figure 1- Kerr cell photograph of the detonation of a cylinder of pentolite.

Qbser vat ion R oonf'

( explodin q w in• '

( tra ci nq paper)

PHENOMENON

POLAROID

CELL

Camera
BOMB PROOF

Figure 2-Schematic setup for Kerr cell photography.

158

c[ L
concrete wall. A window of bullet proof glass permits the camera to take pictures of
events occurring in the bombproof. The phenomenon is outlined against a light-diffusing
screen of tracing paper illuminated from behind by the exploded wire. The Kerr cell
shutter, consisting of a Kerr cell between crossed polaroide, transmits light only'
during the one half microsecond when a high voltage pulse is applied to the electrodes
of the Kerr cell.

For maximum Kerr effect, the front polaroid is oriented so that light entering
the cell ],a plane polarized at an angle of 45 degrees to the vertical. The back
polaroid is set at 90 degrees to the front polaroid. The electrodes are immersed in
nitrobenzene, a transparent fluid exhibiting the highest Kerr activity known.2

When no electric field exists between the electrodes, the polarized light passes
through the cell unchanged except for some absorption. However, when a potential is
applied to the electrodes, the polarized light is changed progressively to various
states. of elliptic polarization and, in an ideal ca13e, emerges from the cell plane
polarized at 90 degrees to its original direction. Because the Kerr effect depends
on wave-length, this 11 full-open" or maxillllllll transmission condition can be obtained
only with monochromatic light. For other wave lengths, only the component of the
elliptically polarized light parallel to the plane of polarization of the back polaroid
passes through the cell.

For 1;.he type of Kerr cell described here, it can be shown3 that the intensity I of
monochromatic light transmitted by the back polaroid is related to the original in-
tensity ! 0 from the front polaroid, by the equation given at the top of Fig. 3. Here
the term e-111d is the fraction of the light ·intensity rema.ining after absorption, m is
the absorption coefficient, d the length of the light path, B the Kerr constant, V the
instantaneous potential, and-a the spacing of the electrodes:" In addition to the-
absorption of light by the nitrobenzene, there are several other factors that reduce
the transmission, namely a non-uniform electric field and the lack of a monochromatic
light source.

On the bOttom left of this f.lgure is a line draWing of a two electrode cell showing
the electrode spacing a and the electrode length d. The aperture of the Kerr cell
shutter can be nearly doubled by using three electrodes as is shown at the bottom lett.
This arrangement is used in the C.I.T. camera.
Fig. 4 shows the disassembled Kerr cell shutter. At the top center is the
electrode assembly consisting of three copper screen electrodes and two supporting
t.'rames. The center electrode is suspended betwen tw glass rod insulators. The cell
bod;r, a rectangular box of glass plates cemented together, is directly below the
electrode assembly. At the bottom left is the light proof Kerr cell box with a needle
point spark gap mounted on ite top. The function of this needle gap will be given
later. To the right is one end of the Kerr cell box with a polaroid mounted on it.
For use in photograpb;)r, the glass cell boey must have end plates that are distortionless
and also free of strain. The latter requirement is necessary because glass that is
strained alters the polarized light causing light leakage through the closed shutter.
Likewise, internal reflection must be minimized since this also causes light leakage.
For this reason, screen mesh electrodes are used to reduce the reflection of light
striking their surfaces.

159
 2
I =I exp(- md) · sln 2 (TfdBV )
o a2

1= Light Transmit ted by Back Polaroid

10 = Light Transmitted by Front Polaroid
m • Light Absorption Coefficient
d • Length of Electrodes
a= Spacing of Electrodes
B • Kerr Constant
V • Voltage Applied to Cell

+ +

a l "'0
a a l
"'0

J Arrangement for
J
Doubling Aperture

Figure 3-Transmission of a Kerr cell.

Figure 4-Disassembled Kerr cell shutter.

160

K?'W?F!Pz , n
 ,,

The camera used with the Kerr cell shutter is made up of a 4 x 5 inch Graflex
camera back, a cloth bellows and an f/2.5 Aero Ektar lens having a seven inch focal
length. Because the aperture of the Ksrr cell covers less than half the area of the
lens, the effective speed of the combined lens and Kerr is reduced to approximately
f/4. At present the camera is being fitted With an f/2 Xenon lens which when combined
with the Kerr cell will give an effective speed of f/2.9. A long focal length lens
was selected to yield a large image without enlargement. The use of high speed film
and harsh development methods increases grain size so much that resolution is limited,
even for a J inch image obtained without enlargement.

High intensity light for thsKerr cell photography of jets is obtained by suddenly
discharging a J microfarad condenser charged to 24 kilovolts through a fine wire,
causing the wire to "explode" with an intense flash of light. This light source reaches
a peak intensity of about 500 million candle power in 7 microseconds and has a duration
of.about 25 to 50 microseconds.

Fig. 5 is a block diagram of the high speed photography system showing the
sequence of operations. The phenomenon initiates the synchronization circuit in one
of two ways. For non-rotated charges, a pair of enameled trigger leads are twisted
together and taped to the explosive. When the force of the explosion shorts these
leads, a thyratron is put into conduction forming a triggering pulse. For rotated
charges, a photocell is used to pick up the light from the luminous jet front.

A variable delay of 1 to 500 microseconds, is provided for in the delay unit. The
delayed pulse is then fed into the high voltage circuit, which is made up of the switch-
ing and synchronizing circuit, the exploding wire circuit, and the Kerr cell circuit.

Fig. 6 shows a drcuit diagram of the high voltage circuit. Drawn in heavy lines
is the exploding wire circuit consisting of condenser Cl, three-sphere spark gap SS
and the fine wire. The Ksrr cell circuit, drawn in medium width lines, is made up of
condensers C2 and CJ, and spark gap GG. The remaining components, drawn in fine lines,
serve as control elements. Both sections of the spark gap SS are adjusted to break
down at 15 kilovolts. When a trigger pulse is applied to the 5C22 hydrogen thyratron
thus firing it, the center electrode of gap SS momentarily goes to ground potential.
The full 24 kilovolt potential then appears across the upper section of gap SS causing
it to break down followed immediately by the breakdown of the loliBr section. With gap
SS ionized, a heavy surge of current from condenser Cl passes through the fine wire
exploding it. Referring again to the circuit diagram it is seen that spark gaps SS
and OG have a common air space. The ionization of this common gap serves to couple
together and to synchronize the exploclin& wire and Kerr cell. The time of breakdown
of gap GG can be varied through a range of about l-0 microseconds by adjusting the
spacing of its electrodes. Condensers C2 and CJ of the Kerr cell circuit discharge
through gap GG applying a voltage pulse to the Kerr cell shutter thereby making it
transparent. After a very short time, needle gap Ml breaks down rapidly discharging
the Kerr cell to its opaque state.

Fig. 7 demonstrates ~he operating characteristics of the Kerr cell shutter.

Shown at the upper left is an oscillogram of luminosity vs time for an exploding wire,
as seen through an "open" Kerr cell, that is, one with its polaroids parallel. In
this instance, transmission was reduced 90 per cent by the application of a voltage

161

5 I2&4iihL ..
 sa rr a 't t'sss t' s ·

LOW VOLTAGE HIGH VOLTAGE

TRIGGER ~ DELAY
.I SWITCHING
AND-
f-
EXPLODING
\WIRE
I SYNCHRONIZING

I
I
I
I KERR
' I
I CELL

Figure 6-Bloek diagram of Kerr eell camera system.

+24 KV

112 Cl
8
G

8
-----,...L 02
-r-

. .:·1
11021
Rll

-
EX I.ODING
WIRE CELL Nl

HIGH VOLTAGE CIRCUIT DIAGRAM

RI•IOOOO 01111 Cl• 3 l.lf SS • WIRE CIRCUIT SPARK OAP

R2•100 IIEOOHM CZ..I UF ( THREE SPHERES)
R3o 100 IIEOOHII CII•IOOO UF 110• KERR CELL CIRCUIT SPARK OAP
R4o I. 11180HII NO• KERR Ot:LL DI80HARGINO 8PARK GAP
t NUDLI POINT I )

Figure 6-High voltage cireuit diagram.

cyeymJDW'I
 Pa ra. lei

Figure 7-0scillograms of typical Kerr cell operation for various needle gap settings.
Each horizontal division corresponds to 0.5 microsecond.

163

o CO*WPFNTT%
pulse to the Kerr cell as is indicated by the dip at the middle of the light trace.
This corresponds to. a 90 per cent opening efficiency for a similar cell with polaroids
crossed. Synchronization, efficiency and exposure time can be checked with this type
of oscillogram. The other three pictures in Fig. 7 were obtained with the polaroids
crossed as for normal operation. These light intensity versus time oscillograms show
the exposures obtained for several needle gap settings. At the upper right is a one
microsecond exposure 'Which has nearly a square wave shape. At the lower left is a
three microsecond exposure, and at the lower right is a one half microsecond exposure.
In this last oscillogram, the decreased amplitude indicates that. the transmission
efficiency of the cell at this needle gap setting has decreased considerably. Exposure
times from one half microsecond to five microseconds are possible through adjustment
of the needle gap.
Before proceeding to applications of the Kerr cell photographic technique, some
of the advantages of the exploding wire as a light source should be pointed out. First,
it gives a peak luminosity that is much higher than that of any comparable light source.
Second, because of its negligible cost, it can be used in the bombproof 'Where it will
do the most good.
II. APPLICATIONS AND RESULTS
The photogra~ of jets traveling in air was the first application of the Kerr
cell technique attempted. Fig. 8 shows a jet picture from a typical charge. ~ile
a small portion of the actual jet can be seen near the top of the picture, most of
it is obscured by an opaque shroud. In fact the early pictures did not show any jet
at all until shadowgraph methods were adopted and the exposure time was made short
enough to stop motion.
The opaque shroud appeared to consist of a large number of high-speed particles
traveling nearly parallel to the jet and vaporizing continuously. This conclusion
was verified by taking a photograph of a jet passing through an evacuated glass tube.
Fig. 9 shows such a jet in a tube evacuated to a pressure of l0-.3 millimeters of
mercury. No shroud is seen in this picture, the particles being too small to be
visible in the photograph, and no metal vapor is formed at the jet tip in the absence
of air.
The shrouded jet show before in Fig. 8 was from a copper liner with a flat apex
as shown in cross-section at the left of Fig. 10. The lack of a pointed apex was
found to be responsible for most of the shroud, for when the pointed cones pictured
at the right of Fig. 10 were used, shroudless jets like that of Fig. 11 were obtained.
Note, however, that a si!Bll portion of the jet at the tip is obscured by the metallic
vapor produced by its reaction with the air.
In order to· determine how two simultaneously detonated cavity charges might
interfere, the setup given at the left of Fig. 12 was used. The two shaped charges
are detonated simultaneously by equal lengths of primacord running to a single detonator.
At the center is the action picture of this setup showing two perfectly formed jets.
However, when the detonation of the right hand charge is delayed 8 microseconds,
interference results as seen at the right of Fig. 12.

164
Figure 8-Jet from copper cone Figure 9-Jet in a tube evacuated to a
with fiat apex. pressure of lQ-3 mm Hg.
Figure 10-Copper liners with flat and pointed apexes.

Figure 11-Shroud\ess jet from pointed cone.

166

lbdt
Figure 12--lnterference of simultaneously or sequentially detonated cavity charges.
 Sho'Wil in Fig. 13 are pictures of jets formed from very small cavity charges
known as spit-back fuses. This charge is 1 inch long and 3/8 inch in diameter and
has a conical aluminum liner with an 80 degree apex angle, The jets shown in these
pictures are less than a millimeter in diameter and are probably the smallest jets
that have been photo~phed. Note that one of these jets is quite curved.

A jet penetrating a piece of plexiglas of dimensions 6 x 6 x 1-1/4 inches is

seen in Fig. 14• The plexiglas just above the jet tip has been shattered causing
this area to become opaque. A bow shaped shock wave is seen extending laterally
across the plexiglas block.

Fig. 15 shows a jet entering a glass target at an angle of 20 degrees to the

target surface. The jet inside the glass is surrounded by a transparent region
which in turn is bounded by a dark line. This dark line cannot be the result of
a shock wave since a shock wave would be conical rather than cylindrical, It has
been theorized4 that the transparent region consists of molten glass which has been
melted by the high pressure produced by the jet tip. Note that the jet is not de-
flected by the glass target even at this small angle,
Fig. 16 shows a steel jet in air which has separated into segments. This results
from the negative velocity gradient that exists along the jet lihich causes the jet to
lengthen and break up as it travels, This same effect can be shown indirectly by
pictures of the jet penetrating water. Fig. 17 shows two pictures of copper jets
with 2 and 12 inch standoffs respectively, t!lat is, the distance from the liner base
to the surface of the water, Although the jets are not seen in these pictures, their
nature can be inferred by the cavitation they produce. At the left the outline of the
cavitation for the two inch standoff case is smooth, indicating that the penetration
in this case is a continuous process, However, the cavitation becomes progressively
more irregular in outline as the standoff is increased, as shown in the other picture
in this figure •. Hence, the jet itself must become increasingly broken up as standoff
is increased.

Fig. 18 gives a similar comparison of jets penetrating water after penetrating

1-1/4 inches of steel, on the left, and 1-1/4 inches of glass, on the right. Here the
cavitation outline is much more irregular for the case of the jet that penetrated the
glass target.

Fig, 19 demonstrates the use of two Kerr cell cameras to obtain two successive
piotures of the same phenomenon. A jet from a brass liner is sho'WI! penetrating a 1/2
inch steel plate. The picture at the right was taken 4 microseconds after the one on
the left.
Fig. 20 is the setup used to photograph the collapse of a wedge liner. The liner
is extended beyond the edges of l;he charge to prevent explosive products from obscuring
the action taking place inside the wdge, Fig, 21 shows thf' collapsing liedge liner and
the resulting 11 jet" being formed. . ·

III. CONCLUSION
The Kerr cell photographic technique has proven to be applicable to a variety of
special problems in the investigation of jet phenomena. This technique can be used to

168
 Figure 13-Jets from small cavity charges in spit-back fuses.

Figure 14-Jet penetrating plexiglass. Figure 15-Jet entering a glass target at

an angle of 20• to the target surface.
...
&Ji Edt I21 ib 22& a

Figure 16-Steel jet in air.

170
Figure 17-Copper jets with 2 and 12 inch standoffs respectively.

Fi·gure 18-Jets penetrating water after penetrating 11.4 inches

of steel (left) and glass (right) .

171
 ltdil£1£. ._ . ._ __

Figure 19-Two successive pictures of a jet penetrating Y.! inch steel plate,
taken with two Kerr cell cameras.

172
 DE TONATE D
HERE

50-50 PENTOLITE
CHARGE

COPPER WE DGE LINE R

.0 4 3" THIC K

Figure 20-Setup for photographing t he Figure 21- Collapsing wedge liner.

collapse of a wedge.
supplement the flash X-ray technique in the stu~ of the fine structure of jets.
However, it is not applicable to the liner collapse problem except in the two di-
mensional case. Multiple Kerr cell cameras· can be used to take a number of pictures
of the same phenomenon. Available light intensity should be increased so that slow,
fine grain film can be used to improve the quality of the pictures. Also an exposure
time of 1/10 of a microsecond would give better time resolution and improve the
rendering of detail in the pictures.

1 Fundamentals of Shape4 Charges., CIT-ORD-21, p. B7, June 30, 1949.

2 Kingsbury, E. F., "The Kerr Electrostatic Effect , 11 Rev. of Scientific Instruments,
1122

3 Same as reference £•
4 CIT-ORD-31, Part III, P• 49, February 2B, 1951

174
 THE PJN TECHNIQUE FOR VELOCITY MEASUREMENTS

H, Dean Mallory~

U. S, Naval Ordnance Laboratory, White Oak, Silver Spring, Maryland

ABSTRACT

This is an electronic method for determining arrival times

of an event at various positions during its motion. The method
takes its name from the electrical switches which are charged,
pointed pins. The moving metal target or ionized shock wave is
electrically grounded so that on contact w1 th a pin, a pip is
generated on an oscilloscope screen, Pips from a se~ence of
pins are photographed ~ a still camera as they appear on the
screen. Time is determined from a crystal controlled sine wave
superimposed over the pips,

We believe that with this method, time can be resolved

better than can reasonably be expected of a photographic method
alone, Also, it is possible with the pin technique to investi-
gate the initial motion of a moving object over the interval
from zero to one millimeter with good reproducibility, It is
over this close-in range that photographic methods are at their
worst and the pin techni~e is at its best,

The pin technique is an electronic method for determining the arrival time of a
shock wave <ir moving metal target at various points, It consists of electrically
charged pins connected to one or more oscilloscopes through somewhat canplex circuits,
Pips from a se~ence of pins are photographed by a still camera as they appear on the
cathode ray tube, Time is determined from a crystal controlled sine wave superimposed
over the pips. Requirements on the rise time of the oscilloscope pips and the duration
of the sweep are sufficiently strict so that very few commercial circuits are ade~ate,

This paper will be concerned only with the technique as a whole, some experimental
details and results rather than electrical circuitry, Most of the re~ired schematics
can be found in the recent book, "Electronics", ~ Elmore and Sands,

The pin technique is being used at Naval Ordnance Laboratory in a program to

evaluate various explosives by measuring shock velocity through metals and free surface
velocities of targets propelled ~ exploding charges, We have for the most part
limited ourselves to the study of aluminum because its shock velocity and density
are such that the data obtained can more easily be related to the particle velocity
and pressure developed in the explosive than can similar data on other easily
available material,
If the detonation pressure is developed according to the von Neuman concept of a
high pressure spike which includes the reaction zone, followed ~ a s~are step wave

17~
behind the Cha'pman-Jouget point, it is apparent that thin targets will have an ab-
normally high velocity since they are propelled outward before the pressure-peak is
eroded. A free surface velocity versus target thickness graph should then show a
modified facsimile of the high pressure spike developed in the detonation wave. This
has been shown; subject to the uncertainties to be discussed. (Figure 1.) The experi-
ment has been designed. to enable 8-time readings to be made within a minimum of about
1 microsecond. The surface velocity readings are obtained from 8 sharp brass pins,
electrically charged and connected to an oscilloscope, and which are arranged in a
small circle within a few millimeters of the face of an aluminum target. This target
is an. electrically grounded flat disk, 6 inches in diameter and of any thickness tram
1-1/2 inches down to two mils. The target is propelled toward the charged pin con-
tactors by the explosion of a massive charge - the term massive is used in the sense
that all measurements are completed before edge effects become operative; actual
ctimensions of the charge are 6 x 6 x 5 inches. Since the pins have a circular con-
figuration, it is expected that the detonation wave :Impacting on the target plate be
planar and f'urtherniore, that it strike the target exactly normal. The plane waves are
generated in the explosive by so-called lense charges or plane·wave boosters. As the
target strikes a pin contactor, a conical shock wave will be set up in the target metal
with its apex at the point of the pin. In order that this disturbance does not in-
fluence the planarity of flow before successive pins are discharged by contact with the
grounded target, the pine must be spaced a certain distance apart. By consideration of
the shock cone arru.nd the pins, the target thickness and the· distance the target must
travel to contact all the pins, one can determine the minimum circle diAmeter on which
the pins can be placed. A minimum diameter is desirable inasmuch as the !latest portion
of the plane wave is likely to be near the charge axis. (Figures 2 and 3. )

The rate of propagation of a shock wave through metal is detennined in a manner

very similar to that used for tree surface velocity. In this case however, the charged
pin contactors are set in holes drilled to various depths in the target. Passage of a
shock causes the pins to short out. Considerations of shock cone and wave flatness
also apply to shock measurements.

Experimental difficulties m~ be classified into two broad groups - those which

can effect the magnitudes of measured velocities but not their reproducibility, and
those which can effect the reproducibility. They will be discussed briefly in this
order.

First of all, it is of prime importance to make measurements on the tree surface

of the target and not on the air shock which is built up in front of it. The intense
air shock contains ionized particles and is electrically conducting. It can therefore
effectively short the charged pin contactors and give pips on the oscilloscope which
are not a measure of target velocity. The problem here can be appreciated by a de-
scription of events occurring at the moment of explosion: The target plate is moving
toward the pins at a velocity upward of 1800 m/sec. It is a piston compressing air
ahead of it. Thus if the air shock is in same wq electrically grounded, the pins
will short out on arrival of the air shock front and the records will give only this
velocity instead of that of the target behind. Now the pin contactors, in sequence,
are charged to opposite polarities in order to facilitate reading the oscilloscope
records. Since the air shock has a finite and increasing depth - not the front but

176
 L ... 2 _L -- -- 1

4000
CA5T TNI- AI
FIRST OF!OER YELOC.ITY
OF FREE 5URFAC.E IN -'IR

']
... :mo
"'
~
0
>-
!:
'-'
0
-'
'" 200C
>

1000
I~

Figure 1-Cast TNT-AI. Fir.rt; order velocity of free surface in air.

P.tttl.

H. E.

SI'A.NnTI T r l l l - . - - - + -
laA" _,.,7' 1'111 ----it--

Figure 2-Contaet pin setup--Schematic.

177
 I I
Figure 3-Photograph of plate.

1.0

1.0 ALUMINUM- COMP . B

FRt:E 5URF"AC.E VELOCITY A':. FUNCTION
S EC.
292' _ _ __,_,,, / OF DISTANCE TRAVE:LE 0 BY TARGET
UNDE.R NOI'IMA\.. INC.ID"EONT Dt:TONATI ON
;3190

2 7 8 9 10 II
MII..I...IM£T!;.R

Figure 4-Aluminum- Comp. B. Free surface velocity as function of distance

traveled by target under normal incident detonation.

178
the mass of shocked air being moved ahead of the piston, it can therefore also short
a postively charged pin to a negative one even when it is not grounded. In this case
unintelligible records result. Insulation requirements are consequently twofold. The
air shock must be electrica~ separate4 from the moving, grounded, target and also
the pins must be insulated from each other in a manner sufficient to withstand erosion
by the hot wind in the shock but lightly enough to break do'Wll i111Jllediately on metallic
contact. We have used films of kerosene or vasoline on all 1110ving parts of the target
to keep the air shock from grounding. The pin contactors are painted along their
shafts with glyptal over which is slipped plastic spagetti tubing. The pin tips are
left bare except for a drop of light oil. It may be noted in passing that if the pins
are uninsulated, clean cut pips are observed on the oscilloscope on!y during approxi-
mately the first o. 7 nun of target motion. This has been taken to mean that the shocked
air has a depth of about 0.4 nun, which is the approximate distance between consecutive
pins, before the target (i.e. piston) has moved 1 nun.
Throughout this work we have neglected the air ahead of the target and considered
its free surface velocity as being that which would be observed in vacuum. This ap-
proximation appears reasonable in the .calculations and some substantive experiments
have been performed by firing into an atmosphere of butane. Despite the higher density
of butane compared with air, no detectible differences in target velocities wre noted.
Although the experiment is somewhat less tractable, firing into an atmosphere of
eydrocarbon gas has an advantage in that no inSulation\ is required on the surface
velocity pins. .

Among the experimental diff;l.culties which influence reproducibility of data, one

can mention such things 1 as,flatriess of target plates, planarity of the detonation ~ve,
variation in quality of the eXPlosive used, straightness of pin contactors, accuracy
of pin measurement and changes in pin placement due to handling and transportation to
the shooting site. Some comments will be given on these items.

The detonation ~ve striking the target should be plane since the pin contactors
are displaced laterally to avoid the effects of one interfering with measurements from
all the others. It is believed that the ~ve is flat to within 0.01 microseconds or
better so that this factor is under control. The presence of blow holes in the ex-
plosive along with variation in crystal sizes are factors which are difficult to con-
trol and w believe have been the most important contributors to nuctuations in the
data. We therefore used the procedure of firing nwnerous shots at each target thick-
ness in order to get a statistical average.

Targets less than 1/8 inch thick introduce their O'Wll peculiar problems, in part
due to flatness requirements which are especially trying when working with thin foils,
but principally the difficulties seem to be due to the very nature of the shock from a
detonation. It is in this thin target range that effects due to high pressure in the
reaction zone of the detonation show themselves. The reproducibility of free surface
velocities here, depends on reproducibility of reaction zone pressures which decay
rapi~ after the onset of reaction and consequently the slope of the free surface
velocity vs. target thickness line is very steep. One might therefore epect the large
velocity fluctuations of thin targets which sometimes occur. However, results are on
the whole gratifying.

179
 Insulation requirements for thin targets are somewhat more stringent than for
thick ones due to their higher velocities. As has already been mentioned, the heated
shocked air ahead of the moving target erodes the pin insulation. Due to the increasing
depth o£ the shocked air mass, those pins farthest removed £rom the target are subjected
to this erosion for the longest time. The later pins are prone to short out prematurely
unless completely covered with a film of oil. This difficulty is of course more pro-
nounced as the distance over which measurements are made is increased. The present
studies involve distances up to ·u DDD of target travel f'rom its rest position. We
believe that 3 orders of velocity have been observed f'or l/32" aluminum targets with
Camp. B explosive and that it has not yet reached its terminal velocity in this interval.
These velocity orders are thought to be due to reverberations of the shock wave in the
target. When the free surf'ace of the target begins to move, a backward traveling
rarefaction wave goes into the metal. When it reaches the explosive-metal interrace,
a new shock enters the target which on reaching the free surf'ace causes a sudden jump
in velocity. This process is repeated until terminal velocity is reached. It may be
seen f'rom the graph (Figure 4) that the let order (initial) velocity f'or l/32 11 aluminum
extends over onJ.y about lmm and that the target moves through this distance in less than
0.3 microseconds. In the interest of accurate measurements it is desirable to place
as many pin contactors as possible within the lst order range but in order to use 8
pins here, our electronic equipment would have to be faster by a factor of' three or
four. Thinner targets would require still faster response. At the present time, the
more interesting thin target velocities cannot be as accurately measured as we would
hope for. However, there are oscilloscopes in existence with which such short inter-
vals can be measured.

We believe that the pin technique is better than photographic methods for some
applications. CertainJ.y when the two are used together, more reliable information can
be obtained. Application of the technique to shaped charge work would require some
changes in procedure but there should be no inherent reason why it cannot be done.

180
 HIGH SPEED PHOTOGRAPHY WITH AN IMAGE CONVERTER TUBE

R. D. Drosd

T. P. Liddiard
B. N. Singleton, Jr.
u. s. Naval Ordnance Laboratory 'White Oak, Silver Spring, Maryland

ABSTRACT

The results of an investigation in the liSe of an image con-

verter tube as a high speed camera shutter and an image brightness
intensifier are given. With the 1P25A (sniperscope) image con-
verter tube it was found that exposure times of thirty niillimicro-
seconds with a gain in image brightness of two are easily attain-
able. The relative merits of the image converter and the Kerr cell
cameras are discussed. The construction of the 1P25A camera is
shown. '

All systems of high speed photography are compromises between exposure time,
light gathering power, optical resolution and field of view. Attempts to improve
these Sy"stems generally require a large expenditure of effort and money for rela-
tively small gains. However, there is one system, introduced by J. s. Courtney-
Pratt in 1949, that is basically different and offers, among other advantages, the
unique ability to produce an optical image that is brighter than the image which
is put into it. The device which makes this possible is the image converter tube.

Probably the most common example of an image converter tube in this country is
the 1P25A. This tube was used in the sniperscope and the snooperscope to convert
infra-red images to visible images. It was chosen for this investigation mainly be-
cause of its availability. It consists essentially of a 1.25 inch diameter photo-
cathode, a series of cylindrical electrodes !lnd a 0.63 inch diameter flourescent
screen. (F:i.g. 1)

The photocathode consists of a semitransparent film of silver and caesim oxide.

It is sensitive principally to light in the 1,000 to 12,000 angstrom range. Its peak
sensitivity is at 9,000 angstroms. 'When an optical image is focussed on the photocathode
a corresponding electron image is produced on the inner surface.

The electrons of this image are accelerated and focussed on the screen by the
various potentials on the electrodes. The ratios between the potentials must be constant
in order to maintain the proper focus. The overall potential may be varied over a wide
range without appreciably changing the focus or the image size. The maxi.IIIUI1l potential
that niay be applied statically is 4,500 volts but ·a:t least 14,000 volts may be applied

161
 NAVORD REPORT 1811

CATHODE SCREEN____.,.

Figure 1- 1P25A Image converter tube.

1 82
 PERMALOY SHIELOING

COPPER TUBING

F/!.9
IP25A

FII.B LENS z
ON 35 MM -+---11-- POLYSTYRENE ~
0
CAMERA ~
0
INFRARED ~
FILTER ,
IT1
0
SCALE • 1/2 SIZE
~
-1
Q)

COAXIAL CABLE

Figure 2-Axial-section of image converter camera and cross-section

showing the capacity divider.
~~--~;·~·~·---,--e===i·~·=·~B~/U ~=52A ...,
~~~TAG£ A LONG CABLE OPEN A RG-B/U
END
=

IN
... OUT

•
- z
a. b.
~
.... A
~==~il----------~+~~,
"'0
.
"'"'
RG-8/U 0
5- 30 ::r "'-1
.1.. l: 1.• INTERELECTROOE Q;

• ••T I' I· CAPACITIES

=
0 IBOH 6
OF THE IP25A

100(\. 1200
TO \IOLTAG£ DIVIDER

CAPACITY DIVIDER
SHORT LENGTH

c. d.

Figure 3--Circuits for three square-pulse generators and a capacity

voltage divider for the 1P25A.

OUrER HOUSING EXPLOSION CHAMBER

REMOTE
DETONATOR
PULSING
CIRCUIT

PROTECTIVE
~.I----WINDOW
z
!:0
IMAGE
"'
c

CIRCUIT
TUBE
PULSING
CIRCUIT .
"'"'
0
FRONT LENS
"'-t
iii

35 MM CAMERA

Figure 4--Typical set-up for taking instantaneous photographs of

detonation phenomena with the 1P26A image tube.

184
in short pulses without danger of breakdow. The duration of the pulse determines the
exposure time and is limited only by the transit time of the electrons. At 14,000
volts this is about 3 x lo-4 microsecond.

~n the electrons strike the screen the optical image is reproduced in the
visible spectrum. This image is one half the size of the original image.
According to the manufacturer•s specifications this tube has a gain of light of
as much as one when a potential of 4,000 volts is used. This means that the image on
the screen is four times as bright as the image on the photocathode since the same
amount of luminous energy is concentrated in an area only one fourth as large. With ,.
14,000 volts the calculated brightness relative to the image on the photocathode is (
about 12. If fl.9 lenses are used at both ends of the tube the overall gain in bright-
ness from object to the photographic film in the camera is approximately 2.4. In an
ordinary camera using the same lenses the relative brightness \iOuld be about 0.2.
(Figs. 2 & 3)

In the experimental investigation it was found that i t was necessary to shield

the 1P25A from stray magnetic fields. Even the presence of a magnetized screw driver
a foot away from the tube caused the image to shift several millimeters and produced
blurred images. Double permalloy shields proved satisfactory. A heavy copper tube
was placed over the tube to act as a shield against any surge magnetic fields that
might have been set-up by the high voltage pulses. (Fig. 4)

The f/1.9 lenses were all of 3.5 11 focal length. The camera used to record the
screen image was a Bolsey 35 mm with an f/1.8 lens.
The various electrode voltages were most easily obtained by means of a voltage
divider. A voltage divider made up of carbon resistors proved unsatisfactory for
pulses under one microsecond. Attempts to use a capacity compensated divider gave
no better results. A purely capacitive divider worked quite well. Apparently, the
carbon resistors acted like non-linear elements under the influence of the very short
pulses. With comparatively simple circuits it was found possible to obtain pictures
with a 0.03 microsecohd exposure time. The tube offers an electrical load of only five
micromicrofarad to the circuit.

Since both the photocathode and the screen are semi-transparent special precautions
are necessary to prevent visible light from leaking through to the photographic film.
This is easily done by placing a red filter in front of the image converter and using
orthochromatic film. The red filter is very efficient in the near infra red. It was
found that orthochromatic film was as efficient as panchromatic film as far as the pre-
dominately green light from the screen was concerned.

In detonation experiments only one special provision was required. The three inch
thick glass protective windows were found to pass only about 30% of the infra red light.
It was found that an equal thickness of lucite passed more than 90% of the light.

In the first detonation experiments a sinusoidal potential of 2,500 volts peak at

100,000 cycles per second was used. This permitted the tube to produce images at ten
microsecond intervals with exposure times of less than five microsecond. The explosive
was a stack of tetryl pellets encased in a jacket of wter to suppress the shock light.

185
 NAVORO REPORT 1811 NAVORO REPORT 1811

1/1
0
z
0
TETRYL 0
w
PELLETS 1/1
0
12. 1• OIAM . a:
'!f4 • LONG ~
~ a
t=3pSEC

0 . MULTIPLE EXPOSURE OF DETONATION WAVE IN TETRYL WOOD

ANO CORRESPONDING CHANGE SET-liP

b
12-TETRYL PELLETS
t • 4p SEC
5/e" DIAM., 3J4w LONG

( WRAPPED IN C£LLIJLOS£
AC£rAr£ FILM}

.~ ..
DETONATOR

t
c
=5p SEC
't
DETONATOR

b. ONE-HALF MICROSECOND EXPOSURE OF TETRYL

PELLETS EXPLODING IN AIR Figure &-Exploding t etryl pellets photo-
graphed at 0.03 J.l. sec. for various times
after initiation.
Figure 5
 NAVORD REPORT 1811 NAVORD REPORT 1811

a a
tz6p5EC
t • 2~SEC

b b
t z 3pSEC t F8pSEC

c d c d
t =3.$pSEC t =$pSEC t z/OpSEC t z/2p5EC

Figure 7-Exploding tetryl at various times Figure 8-Exploding tetryl for various
after initiation; one-tenth microsecond times after initiation; one-tenth
exposures. microsecond exposures.
Five exposures were obtained as the detonation proceded do1o111 the pellets. From this
picture the rate of detonation could be calculated. The quality of this picture is
poor due to the poor magnetic shielding which was used at that time. However, it does
indicate possible applications. (Fig. 5)

A half microsecond picture of a detonating stack of tetryl pellets was obtained

with a 5,000 volt pulse. In this picture the gas front is well defined. The wavy
pattern in the gas front is due to the gaps between the pellets and nonuniformity of
the density of the pellets. The detonation zone is vastly overexposed and halated.

All the ·remining pictures are of the detonations of single tetryl pellets. With
14,000 volts it was possible to obtain exposures down to 0.03 microsecond although
most of the pictures are 0.1 microsecond exposures.

The image converter camera has several advantages over a Kerr cell camera using
the same lenses. With the Kerr camera one must limit the field of view to maintain ap-
proximate parallelism of the light in the Kerr cell. Practical Kerr shutters transmit
less than 20% of the light. The relatively high capacitance of the Kerr cell makes
pulsing more difficult and high voltage and higher power becomes necessary. One
serious disadvantage of the Kerr shutter is that it depends on crossed polaroids to
create a "closed" shutter. Limitations of quality of the best polaroids result in a
closed shutter having a transmission of at least 0.1% of the transmission of an open
shutter. This leads to background exposure and the requirement for auxiliary shutters
for its control. The image converter shutter when properly designed has zero trans-
mission when closed.

In only one respect does the Kerr cell camera exceed the 1P25A camera~ The
resolution of the 1P25A ie about 450 lines per inch while the resolution of a Kerr cell
camera may be 1000 or more lines per inch. However, new and much better image con-
verter tubes will be soon developed which will have resolutions equal to the best the
Kerr cell cameras can offer.

lBB
 HIGH SPEED HIGH RESOLUTION STREAK PHOTOGRAPHY

c. T. Linder
Carnegie Institute of Technology Pittsburgh, Pennsylvania

ABSTRACT

A simple rotating mirror camera with high space and time

resolution, and great versatility in the photographing of various
explosive phenomena is described. The camera features a thin
single surface plane mirror so that the locus of the image is
circular to a high degree of accuracy, and employs film strips
'llhich are six inches in width. The selection of the appropriate
mirror width, speed of rotation, and radius of film track permit
the choice of a variety of fields of view and of writing speeds
up to 3 mm/IJ.Sec. The design employs a .single lens and an expendable
external slit placed at the location of the phenomenon to be photo-
graphed. An effective lens aperture of f/3.5 can be retained and a
1/1 magnification ratio is possible. Various applications to ~ich
the camera is particularly sui ted are discussed.

INTRODUCTION

There are tw types of cameras which are generally used in streak photograpey.
They are the rotating drum camera and the rotating mirror camera. There are many
contri~tors to the development of' theJ• Among these a~ MacDougall and Messerly, 1
Cairns, Payman, Shepard and Woodhead, Beams,4 Jacobs,!:> and numerous others. The
drum camera is limited to film speeds of .less than 300 meters per second because of
the mechanical strength of the drum materials. When higher writing speeds are desired
a rotating inirror camera is used. In this type of streak camera the image is renected
from a mirror, lihich is rotated at a known high speed, onto a st.ationary film strip.
In this way both a time and space record is present on a photograph,.of a rapidly pro-
gressing lUlllinous phenomenon. Such a camera was designed by Jacobs=' while he was at
the wartime Explosives Research Laboratory, Bruceton, Pennsylvania. When the c.r.T.
group moved to the Bruceton location, they were so fortunate as to obtain the use of
this camera. Writing speeds of the order of 1 mm/1JBeC and an effective aperture of
F/S were obtainable with it.
As our interests at c.r.T. shifted to fundamental details of shaped charges, the
limitations of the E.R.L. camera restricted the scope and the accuracy of such investi-
gations. The image size was limited by the 35 mm film width and the placement of the
optical components of the camera. Because of the loss of definition in the enlargement
of the negative, this permitted only partial compensation for the small image size.
We desired better time resolution and larger image size in a camera of simple design
so as to be easily adaptable to the photographing of various kinds of explosive phe-
nomena. Our new camera intends to increase the space and time resolution by using film

189
strips which are 6 inches in width and by increasing the radius of the film track. A
range of image sizes and writing speeds are available by employing three different s~ze
mirrors and three film tracks of different radii.

DESIGN DETAilS

The camera was not desi~ed for the photographing of a single type of explosive
phenomenon. Rather the attempt lias made to construct a multi-purpose camera which
could be easily modified for any type of investigation which is feasible with a streak
camera. Simplicity and dependability wre of prime importance. Also we desired to
keep the operation and maintenance such that sub-professional personnel could obtain
satisfactory results with it.

The dependability and the small amount of maintenance required by the E.R.L.
camera made it seem desirable to borrow many of its design features. Jacobs.:> has show
that the best effective aperwre obtainable with a two-lens optical system using com-
mercial lens-is about F/6. Since a larger effective aperture is desirable for certain
types of investigations, he chose a single-1ens system and used an expendable external
slit placed at the location of the phenomenon to be photographed. In order to make
·use of the full lens aperture we also have employed the single-lens system. The lens
is an Eastman Anastigmat with a focal length of 13 1/2 inches and an aperture of F/3.5.

The mirror is made from stainless steel plate 5 inches high, 3 1/4 inches wide and
1/4 inch thick. Two additional mirrors of widths 1 and 2 inches are in the process of
manufacture. When the full aperture of the lens is not desired the smaller mirrors
will be used thus permitting a higher rotational speed because of reduced air drag.
The mirror surface is obtained by polishing one side of the steel plate until optically
flat to within 1/2 wave length and coating it with aluminum. A protective coating of
quartz is then put over .the aluminum to reduce the abrasive action of dust particles.

Three film tracks have been constructed which accommodate a film strip 6 inches
in width. These tracks have radii of 7, ll, and 15 inches, thus giving a wide range
of writing speeds and image sizes.

Due to the fact tha-t the mirror is of finite thickness, the locus of the image is
not quite circular, hence the angular velocity of the image is not constant. In all
measurement.work with the camera, a correction for variations in image velocity is
taken into account. However, since the mirror thickness is much smaller than that of
the square mirror employed in the E.R.L. camera, the correcUon factor is consequently
much smaller.

The driving system consists of two 1/2 H.P. Dumore Electric Company universal
motors. It was not possible at the time of construction to obtain universal motors
rated higher than 1/2 H.P. with a speed of 8,000 r.p.m. Hence it was necessary to use
tw motors and to match them by the proper adjustment of the tensions in the driving
belts so that the load was equally shared. The speed of these motors is varied by
means of a variac.

190
c!Jih I St;

Figure 1

Figure 2

191
 200

- Film Track Radius Ro (in.)

-•
c
Track ( 1 ) - - - - - 7
Track 121-----n
u 16 0
..
c
0

..,•
Tro ck (3)-----15

... 12 0
...
0

..
0

•e
0 80
u

....
0

u (2)
...•
0
40
(3)

60

--
c

• 40
·-•
N

..
u

...•
0 20 (2)
(3}

0 2 3 4 5 6
Position of lens measured from the end of the foc11sino rail (ln.)

Figure 8--0bject size and distance from eamera port corresponding to

an uin focus" image which exactly covers the 6 in. wide ftlm track.

192
 The speed measuring system is essentially the same as that employed in the E.R.L.
camera. The bottom surface of the mirror shaft is half polished and half black. The
"chopped" light reflected from a light source to a phototube produces a signal which
is amplified and applied to the vertical plates of an oscillograph. The controls of
the oscillograph are set to synchronize the horizontal sweep to a full wave rectifier
operated from the 60 cycle supply voltage. The horizontal sweep frequency is 120
cycles with the accuracy of the a.c. supply. Stationary patterns of simple form are
obtained When the vertical input frequency corresponding to the camera speed is a
multiple of 120 cycles.
Since the camera features a thin one-sided mirror and a short film track, it is
necessary to predetermine the time of the event to be photographed relative to the time
when the mirror reflects light to the desired position of the film. This is' accomplish-
ed by firing detonators by the discharge of a 1.5 ~ condenser charged to approximately
lOKV. The condenser discharges through a mercury thyratron Which is triggered by means
of a light pulse reflected from the shaft of the rotating mirror.
Firing of the, detonator is accomplished by the following sequence of events: a
light source is turned on, the light from the source is reflected from a polished sec-
tion of a collar attached to the rotating mirror shaft and is picked up by a phototube,
the signal is amplified and triggers a type 884 thyratron, this reduces the grid bias
of a type FG-41 thyratron caueing it to fire, which permits the discharge of the 1.5
~ condenser, thus firing the detonator.

The position of the image on the film is determined by the orientation of the re-
flector collar relative to the main mirror. The consistency of picture position, how-
ever, depends on the reproducibility of the times required to fire detonators. The
firing circuit exclusive of the detonator is reproducible to within approximately
three micro-seconds.
A view of the camera proper is shown in Fig. 1. The bellows, lens barrel, aper-
ture stop holder, mirror housing, and a film track are visible. When it is desired a,
graded filter on a wire frame can be placed in front of the film track. Thus it is
possible to make full use of the F/3.5 aperture in photographing the less luminous,
slower moving portion of the phenomenon without losing the detail at the more intense
beginning. A light tight canopy fits over this section of the camera. The two driving
motors and the light sources for speed measurement and synchronization are shown in
Fig. 2 1 which is a view of the lower section of the camera.
OPERATING CHARACTERISTICS

The object size, corresponding to a 6 inch film width exactly covered by an in-
focus image, and the object distance measured from the camera port are shown in Fig. ·3
for the entire travel of the lens and for the three different film tracks. The mag-
nification ratio can be obtained simply from the object size curve. It is given by
1 1/6 the object size.

193
 I 7 ?FrTT5 'J 3 Pi??
' ',
The maximum image velocity attained with the 3 1/4 inch wide mirror, which permits
the use or· the full aperture of the lens; has been 2.3 mm/11 sec. This corresponds to
a mirror rotational speed of u80 r.p.s-. Higher llt"iting speeds will be obtained with
the smaller mirrors at the expense of a reduction in the effective aperture.

When smaller apertures are desired the appropriate one can be selected from a
group of -waterhouse aperture stops ranging from F/10 to F/128. An adapter iB located
at ~he rear of the lens barrel to accommodate them.

The film employed at_ the present time is Kodak linagraph ortho safety fi~. It
can be purchased in convenient 100 foot rolls which are 6 inches wide.

There is little or no trouble with synchronizing the image location with the time
of the event to be photographed. Once the reflector collar on the mirror shaft is
properly oriented the reproducibility is easily within the detonator variability of
20 ll sec or about 1 1/2 inches on the film strip at the most frequently used operating
speede.

Incorporated in the. circuits employed in the camera are the usual safety features.
The charging of the condenser, operation of the camera, and firing of the charge are
completely controlled from a single firing station. In accordance ·with the safety
requirements at Bruceton, the firing line circuit is completed by a second firing
station. The condenser and the firing line are normally shorted to ground except when
in actual use.

APPLICATIONS

In operation as a general purpose camera a wide variety of measurements such as

detonation rates, shock ~ve velocities, jet velocities, fragment velocities, and wave
shapes can be made. It is possible to make such a variety of measurement~;~ with accuracy
because of the 6 inch wide film, and the time and space resolution obtained with the
three different film tracks. It is useful for studies with small charges because of
the large magnification ratio. On the other hand, a wide field of view and large
aperture make mUltiple jet velocity and fragment velocity pictures possible.

As is al-ways the case with such laboratory equipment, improvements are continuall.T
in the making. It is now possible to obtain more powerful and faster uni'!ersal motors.
It is hoped that in conjunction with the mirrors of smaller width this will mean
greater writing speeds will be available. The use of a lens of 36 inch focal length
1lill permit the construction of a fourth film track and an image velocity in excess
of 6 imn/11 sec should be obtainable. These modifications will greatly facilitate wave
shape. studies.

The range of usefulness of the camera can be further extended by taking shadow-
graphs of relatively slow moving fragments of large size.

1. MacDougall, D. P. and Messerly, G. H., OSRD No. 682, July 8, 19u2.

2 Cairns, R. w., Ind. ~g. Chem.; 36, 79-85, 19uu.
3 Payman, Shepherd and Woodhead, Safety in Mines Research Bo!trd Paper No. 99, H. M.
Stationery Office, London 1937.
4 Beams, J. W., Rev. Sci. Instr. 1, 667, 1930J 6, 299, 1935.
5 Jacobs, s. J., OSRD No. 561U, January 2, 19u6.

194

+nw FEW'•
 EXPERIMENTS ON DETONATION PHENOMENA

G. R. Walker

Canadian Armanment Research and Development Establishment, Valcartier, Quebec

In an attempt to investigate some of the fundamental aspects of HE detonation, a

project was carried out at the University of Saskatche~ from 1944 to 1949 in which
photographs of detonating explosives were taken by a camera of the rotating-mirror
·type. The results obtained will not be dealt with at any length here because detailed
reports have been available for some time, and because none of the results have any
apparent practical value in the field of shaped charges. Instead, I shall briefly
describe our experimental set-up, and shall indicate what new facilities are being
planned for similar investigat~ons at GARDE.

Fig. 1 is a plan view of our camera.. The mirror is any one face of a 1 11 cube,
and is mounted on a Beams-type rotor turning at speeds up to 800 rev/sec for a maxilllUIII
writing speed of 5000 m/sec. The effective aperture is about f/20. At GARDE, con-
sideration is being given to the design of a similar camera with a comparable high
writing speed but with a much greater aperture. A detonation chamber designed for
photographing charges up to thirty pounds TNT is in the final stages of planning. It
is in the shape of a cylinder of length 40 feet and inside diameter 16 feet with light
doors at both ends which are opened before firing. The observation window is in the
middle of one side, and in the opposite side is a similar window for Schlieren photo-
graphy. It is plannecl. to collect structural data by means of strain gauges built into
the reinforced concrete.
Fig. ·2 is a photograph to illustrate the operation of the camera. The charge
consisted of a large number of 5-gram HE pressed pellets in tandem, contained in a glass
tube 32 DDn I.D., and initiated at one end by an electric detonator. The vertical charge
was imaged on the (vertical) slit by a field lens., and so vertical distance on the
photograph indicates distance on the cha•·ge. Horizontal dj 3tance indicates time, and
so the slope of the trace indicates velocity of detonation. Fig. 2 shows two and one
half ROX pellets and three tetryl pellets, and shows a sudden decrease in velocity at
the junction.

For fig. 2 the residual space within the glass tube was filled with water. Fig. 3
shows a similar charge with this space filled with propane. With the latter, there ie
strong luminosity as the shock wave hits the glass wall of the tube. In both, the
thin sloping trace is luminosity from the "hot" detonation zone. These photographS·
illustrate the well-known effect of water or propane· in quenching the luminosity of the
shock wave.

195
 '+--- Slit

Cardboard Tube -

S hutter Figure 2

F lamen1 Lamp
• ~ l nspectoon Plott:

503 8 mm

Figure !-Rotating Mirror Camera,

Horizontal Section
Figure 3
 CO!{pARATIVE EFFECTIVENESS OF ARMOR-DEFEATING AMMUNITION
A. Hurlich

watertown Arsenal Laboratory

Watertown Arsenal, Watertown, Massachusetts

ABSTRACT

Presentation of the comparative armor penetration performance

of the various types of kinetic-energy armor-piercing ammunition (AP,
AFC, HVAP, and discarding sabot shot) against single armor targets
as infiuenced by the type, thickness, mechanical properties, and
obliquity of armor. Performance of kinetic-energy armor-piercing
projectiles against spaced armor structures. Performance of HEP
ammunition against solid and spaced armor targets. Projectile de-
sign criteria and their relation to target characteristics. In-
fluence of tank design and construction upon selection and dev.,lop-
ment of armor defeating ammunition.·

There are,· at present, two major types of armor-defeating ammunition; projectiles

lihich depend upon their own kinetic energy to pierce or punch a hole through armor, and
explosive-loaded shell lihich, upon impact and detonation against armor, generate the
energy required for the defeat of the target.
/, 1{,_ •
&e. ,.unetic energy projectiles consist of more or less massive, cylindrical,
ogival-nosed inert shot made of hardened steel or tungsten carbide compacts designed
to have sufficient strength to remain substantially intact during the penetration
cycle. The armor penetration characteristics of such projectiles depend largely upon
their mass and velocity, consequently best armor penetration performance results wen
they are fired from high velocity guns. Kinetic energy projectiles may be further
subdivided into a number of types; monobloc (AP) and capped (AFC) steel shot, also (L.;
...,gemposite-rigid (HVAP) and discarding sabot (HVAPDS) tungsten carbide cored shot.
l'bnobloc steel shot are the most simple in design and least expensive of all types,
consisting of a solid steel bo~ to lihose nose a windshield, made of a thin steel
stamping, may be attached to improve its exterior ballistic performance. The steel
bodies of AP and AFC shot are made of alloy steel differentially heat treated so that
the nose sections have max:I.IIIUDI strength and hardness, with the ·hardness gradually de-
creasing towards the bases to provide tough fracture-resistant bo~ sections. The AFC
shot differ from the AP shot only in the possession of a steel cap placed over the nose
of the shot for the purpose of cushioning the forces on the nose ·of the shot resulti~~
upon impact against armor. The cap thus assists in keeping the point of the projectile
intact.
I'HVAP shot contain sub-caliber sized tungsten carbide cores fixed within light- ); ·
weight metallic carriers of gun bore diameter. The carrier accompanies the core to

197
the target, at which point the core breaks out of the carrier to effect the penetration•
By virtue of its high density and high strength, tungsten carbide is a more effective
penetrator than steel. In addition, because of the combination of a light-weight
carrier and a sub-caliber core, the HVAP shot weighs less than a full caliber steel
shot and thus achieves a higher IIIUZZle velocity when fired from the same gun. Since
the energy required to effect penetration is approximately .proportional, to the volUJIIe
of the hole produced in the arMOr, the greater arMOr penetration performance of the
HVAP projectile is, under ideal conditions, obvious.
Because of the lighter weight and consequent lower sectional density of the HVAP
shot, this type of projectile has poorer range-velocity characteristics than steel shot.
Although it may have a muzzle velocity initially 500 to 700 ft/sec higher than an AP
shot, the HVAP shot will have dissipated MOSt of this advantage within 2000 yards'
range. In order to overcome this deficiency, the HVAPDS shot liaS developed. The
tungsten carbide core is carried within a thin steel sheath and supported, during firing,
in a metallic or plastic carrier which is discarded shortly after the projectile eiiiBrges
from the muzzle of the gun. The sub-projectile of high sectional density proceeds to
,.
the target unencumbered U,r any useless mass. ,.
.
The explosive loaded or chemical energy arMOr-defeating allllll1l!lition consist of the
high explosive plastic (HEP) shell and the hollow charge (HEAT) shell. The HEP round
has a thin, helllispherical, deformable ogive and a base detonating fuze. Upon impact,
the forward portion of the shell collapses against the target and, upon detonation, .a
compressive shock •ve parallel to the plate surfaces travels through the armor, is
reflected as tensile "Waves, and produces a fracture parallel to the plate surfaces. A ·
disc, having a thickness of approximately 25-30% of the plate thickness, is detached
from the back of the arMOr at velocities of 500.1000 ft/sec and, within the narrow con-
fines of the interior of a tank, may produce considerable damage. The HEP shell rarely
perforates armor in the- true sense of the word, unless the armor is quite brittle, but
inflicts damage U,r a combination of disc formation and shock. The force of the deto-
nation of HEP shell may produce considerable damage of a secondary nature through
disruption of tank treads, detachment of fittings,, etc.

The hollow charge round produces a high velocity jet of discrete particles which
perforate armr by accelerating the plate material awy from the path of the jet. The
metal around the hole produced by the jet is compressed. This type of allllll1lllition is
too wll. knollll U,r this audience to justify any further discussion of its functiQning.

I will first describe the armor penetration performance of kinetic energy and
chemical energy projectiles against simple targets and then discuss their perforiiBllce
against more complex targets.

PERFORMANCE OF KINETIC ENERGY PROJECTILES AGAINST SIMPLE ARIDR TARGETS

I shall not attempt to present any equations to describe the armor penetration
.performance of kinetic energy projectiles; firstly because all equations which have
been proposed in the past are found to apply, with a good degree of accuracy, to onl;r-
a limited range of target conditions, and secondly, because a large number of geo-
metrical, metallurgical, and JIIBchanical property variables existing both in present

196
service projectiles and armor exert a profound influence upon the mechanisms of armor
penetration and projectile reaction.* Variables such as plate hardness may be intro-
duced into penetration formulae, but factors such as variations in projectile nose
shape, microstructure, toughness, and soundness of the projectile and armor steels
cannot be readily reduced to mathematical terms. '

An important consideration in the penetration of armor by kinetic energy pro-

jectiles is the ratio of armor thickness to projectile diameter (the e/d :ratio) since
the mechanisms of armor penetration and projectile behavior vary with the e/d ratio.
When the e/d ratio is greater than 1 (armor overmatches the projectile), the penetra-
tion tends to be effected by a ductile pushing-aside mechanism. Relative~ sharp
nosed shot are most effective, and the resistance of the armor generally increases as
its hardness increases. When the e/d ratio is less than 1 (armor undermatches the
projectile), the penetration tends to be effected by the punching or shearing out of
a plug of armor in front of the shot. Relatively ·blunt nosed shot are most effective
under this condition of attack, and the resistance of the armor generally increases as
its hardness decreases.

Data on the comparative armor penetration performance of kinetic energy projectiles

of the AP, AFC, and HVAP types are included in Table I. This table compares the pene-
tration performance of the 90MM AP T33, the 90MM AFC TSO, and the 90MM HVAP M304 shot
when fired at cast and rolled homogeneous and face-hardened armor from· 3" to 7 .6n in
thickness at obliquities of 300 tO 700. The comparative performance of these kinetic
energy projectiles against solid armor targets may be SUllllll&rized as folloliSI

a •..- Pbnobloc steel shot are more effectiVe than capped steel shot for the defeat
of undermatching armor at all obliquities of attack and are more effective than both
APC and HVAP shot for the defeat of moderate~ overmatching armor (up to at least 1-1/4

-
calibers thick) at all obliquities af·attack above approximately 450,

b. Capped steel shot are superior to monobloc steel shot for the defeat of great~
overmatching armor, (over 1-1/4 calibers in thickness) at obliquities in the range of
20° - 450, but both capped and monobloc shot are greatly inferior to HVAP ahot in the
low obliquity range against heavy armor targets.

c. HVAP and HVAPDS ahot are most effective against heavy armor targets at low ~\
and moderate obliquities of attack (the 90MM tungsten carbide cored shot can penetrate
10 to 12 inches at oo obliquity" and at short ranges) but their effectiveness is marke~
degraded at obliquities above approximately 4SO-- Soo.

The preceding statemnts regarding the comparative performance of AP and AFC shot
are well illustrated by Figures 1' and 2. Figure 1 represents data obtained from termi-
nal ballistic tests conducted at the Watertown Arsenal Laboratory in 'Which caliber .40"
sca],e models of the 90MM AP T33 and 90MM AFC T50 shot were fired at plates from 1/2 to

* .
This is not to imp~ that the factors which influence penetration are unknown, that
the performance of kinetic energy projectiles is very variable, or that penetration data
are either scanty or unreliable. As a matter of !act, i t is because penetration data
are so reliable and so extensive that we are not satisfied with equations that ;yield
only approximately correct estimates.

199
 APG St1QT SUP@!JR

"''
~ ... '
t..!.'
'
' '•
t·1~
•·"~";
"~''
~~
<K
:<t~
44,(

·~' '
.:""
> ••

•.
·= t ~ .~ ~
·~· ' ~ n~ :~
.'DC(' ',, ...
; *. ~ ~
.,'
- -
.~

~~I).
t"' •' t

•. • . ..• •• ,,
" ••
W A T t R T 0 WN AR S t NAL LAB0 RAT0 RV
AE S T A I C T E0

Figure !-Comparative performance of AP vs. APC proj,eetiles for the defeat of

homogeneous armor of varying thicknesses and obliquities, based upon data
involving scale model projectiles and armor.

. 200

,
 ·i·
r .. .i
!
... l
~ :' I
i
~
I...J.
ls:
It-
-·
1
I
1
I
I

:t·J
f
l
., '
'
...
•
-..- ~·.
'

91
--· .1.

''
~· 1
I
.• t ._ . "*·~ i
,_ I

., ' 'I
!

~ ·I

$ 1
I
5l "i
~
·g; i
~

II
...I .
~

~
' ~
l z
l

Figure 2-Comparative performance of 90 MM AP T33 vo. 90 MM APC T50 projeetilea

for the defeat of homogeneous armor ·Of varying thickneSses and obliquities.

201
 IITN.63C)-10,559
Figure 3-Schematic area of vulnerability curve.

202
Figures 4 & 6--Comparison of areas of vulnerability for various combinations of armor
thickness and obliquity having equal weight per unit of shielded volume
when attacked by 90 MM s~l projectiles.

203
2 calibers in thickness and at obliquities from 0° to 700 inclusive. The curves on
Figure 1 represent equal resistance curves; i.e. all plate thicknesses and obliquities
whose coordinates fall on the line designated 3000 have ballistic limits of 3000 f/s.
The lines furthermore represent the minimum ballistic limit for the target conditions,
whether the minimum ballistic limit was obtained with AP or APC projectiles. The
dashed line represents the boundary between target conditions where the AP shot was
superior and where the APC shot was superior. It will be noted that the areas of
superiority of the AP over the APC shot and vice versa are in accord with the previous
conclusions.

The data plotted in FigUre 1 represent very precisely determined ballistic limits
obtained over a wide range of target conditions. Similar data in full scale would
involve the expenditure of several million dollars, hundreds of tons of steel armor
and thousands of rounds of 90MM armor-piercing projectiles.

Figure 2 represents a similar treatment of data obtained in full scale tests con-
ducted at Aberdeen Proving Ground. These data are necessarily more limited in scope
than those used to obtain the curves shom in Figure 1 and hence the boundary condi-
tions of Figure 2 are considerable less reliable than those sho'llll in Figure 1. The
same general type of curve results, however.

A useful lillY' of presenting penetration data on kinetic energy projectiles is by

means of vulnerability diagrams of the type shown in Figure 3. A roughly elliptical
area exists for each gun-projectile-armor combination within which penetration of the
armor can be effected and beyond which the armor is invulnerable to the particular
attack. The gun must enter into this consideration since it influences the velocity
and hence the kinetic energy of the shot at all ranges.

Figures 4 and 5 show the use of vulnerability diagrams to illustrate the con..
parative performance of AP and APC projectiles against various thicknesses of armor
sloped at different obliquities. It will be noted that, for a fixed weight of armor
per unit vertical height, thinner plates sloped at higher obliquities (at least up to
53°) provide progressively more protection against APC shot. Against AP shot, however,
a given weight of armor sloped at 37° obliquity provided considerably more protection
than the same weight of armor in the form of a thinner plate sloped at 530 obliquity.
A comparison between the righthand curves of Figures 4 and 5 illustrates the improved
effectiveness of AP shot in attacking highly sloped armor targets.

Figure 6 illustrate!! the weights of 11teel armor required to protect against kinetic
energy projectiles at ranges of 1000, 2000, and 3000 yards as functions of the obliquity
of disposition of the armor and the caliber and type of armor-piercing projectile.
Note the steep dcwward slope indicating the marked degradation in performance of HVAP
projectiles with increasing obliquity of attack. It is also apparent that protection
against APC shot increases constantly as the obliquity increases, whereas armor ill most
effective against AP shot at about 300 obliquity, then becomes progressively more
vulnerable with increasing obliquity above 300.

204
 ' i\i I
i i 01\
: i \: I i- · ' viR r:- . :'\
:· . . ··> f-L fr -
" f--:-- \
-ci -, • . ' - - 1\- -
I ·i - -+-- \.
, ---
I

•:
• •• , j i +-c- --- -.. -" ·- -- ~-- -- ,, ..

§j 1 ---- L_: -~- 1-j--, -

•.
j
o: .~\.... ·~-t+- +
~~),· ~ l_;_ i ( ·---

&%i 1': :) .
I· - --~ ~1--;- .....r=.,_
Li! -- -.

i\.. - -1-'- -- -- --
-'11

-- '
i__:_ -- _'_

,.p.. '~',.--
, __ _
t-r--
,_
--f-- --
--- -~ wl!s'" '" .
-+

.........
r

• ...... ""

llfji '
'
.. 1-' r
.,.......
' ..
---

,. ' -,-,__ r -
'

;
.

- ' --

- .. i :_
l ' ·;
'
... . - --. .
I
' ... 1-:-
' --- 1---
1-;- -- 1- . - i--
'
i\
.----- - - 1-'- I-
1-- j---- 1-- I -
t-· 1-- ~ ' -:· ~""- ' -- !--·
i- 1- 1--- I-;- i -: i .. i
' ·. '"
ll. ' :· I - iT -:- :·- :- . .
;

I ~ es
.. -.
.
; . ,·
. '
'
'
'-'''
1-- !
: i
:
' i r- i ;
i '.,, '!' ~ : 'l'' :. i i .p
-...-.' fLGURf 171

Figure 6--Weights of rolled homogeneous armor at various obliquities required

to protect one unit of vertical area against various projectiles at selected ranges:

205
 Early in World War II, kinetic energy projectiles wre fired from guns with
muzzle velocities of 2000-2700 ft/sec. ~ the end of World war II, steel shot were
being fired at velocities up to )200 f/s and present guns are being built to fire
solid steel shot at velocities up to 3500 f/s and HVAP and HVAPDS shot at velocities
of 4000-4500 f/s. Coupled with these high velocities are good stability, high accuracy,
and high rate offire. These factors combine to yield a high probability of regis-
tering a damaging hit with kinetic energy shot. It is firmly believed that kinetic
energy shot will, at least in the foreseeable future, play an important role in tank
and anti-tank warfare.

PERFORMANCE OF CHEMICAL ENERGY PROJECTILES AGAINST SIMPLE ARMOR TARGETS

The available data on the armor-defeating performance of high explosive plastic

(IIEP) shell indicates that this round can cause the scabbing or spelling of armor up
to 1.3 calibers in thickness over a w.l.de range of obliquities. Unlike kinetic energy
armor-piercing projectiles, the performance of HEP shell is not greatly influenced b,y
obliquity of attack at least within the range of )0° to 600; the same thickness of
armor can be defeated over this whole range of obliquities. As a matter of fact, the
performance of HEP shell is worse in the range of 0° to )0° obliquity than at higher
obliquities due primarily to the fact that the explosive charge is not spread over the
face of the armor as effectively at ver,r low obliquities as it is at higher obliquities.
The IIEP shell is also degraded at obliquities of attack above approximately 6Qo.

HEP shell perform satisfactorily at striking velocities up to approximately 2500

f/s, but athigher impact velocities this shell is relatively ineffective because of
shock detonation of the explosive ~ch initiates at the nose of the shell at these
high velocities. To perform satisfactorily, detonation must be initiated at the base
of the shell to permit the generation and travel of a shock wave from the explosive to
the target.

It has been determined that brittle steel armor and unsound steel containing
laminations or segregations of inclusions are more readily defeated b,y IIEP shell than
are tough, sound steels. There is reason to believe that IIEP shell become increasingly
effective in cold climates since the toughness of steel armor decreases with decreasing
temperature, particularly i f the steel is insufficiently alloyed or poorly heat treated.

In view of the lower velocity of HEP shell as compared to kinetic energy pro-
jectiles, errors in range estimation assume more serious proportions than in the case
of kinetic energy shot. The probability of hitting the target, particularly at longer
ranges, is thus lower with IIEP shell than with kinetic energy shot.

Chemical energy armor-defeating ammunition do, however, have one ver,r great ad-
vantage over kinetic energy projectiles. Since they generate their destructive energy
upon impact against the target, chemical energy shell inflict as DNCh damage when
hitting from long ranges as from short ranges, whereas kinetic energy projectiles be-
come less and less effective as the range £rom which they are fired increases.
 The jet generated by the hollow charge (HEAT) shell continues in a relatively
straight line along the line of flight of the shell, consequently the armor penetration
performance of "this type of ammunition closely follows the cosine law. The penetration
performance of kinetic energy proje"ctiles follows the cosine law fairly well up to
approximately 300 obliquity, but at higher obliquities the deviation is very considerable
and is markedly influenced by the geometrical and metallurgical design of the shot.
Since the HEAT shell does follow the cosine law, a round which can penetrate 12 11 thick
plate at normal obliquity can defeat 6" thick plate inclined at 60° obliquity. For
comparison, the 90MM" HVAP M304 shot can defeat 12" thick plate at 0° obliquity at
ranges up to approximately 1300 yards, but cannot defeat even 4 11 thick plate at 60°
obliquity wben fired at point blank range.

The presently available data on the armor penetration performance of HEAT shell
indicate that the thickness of armor which can be penetrated 90% of the time is approxi-
mately 4 times the inside diameter of the cone. This behavior holds over a wide range
of obliquities up to possibly 70°. Thus a 90MM HEAT shell having a cone diameter of
approximately 3 11 should be able to penetrate about 12 11 of armor. The data contained in
Table II shows that the 90MM HEAT noBE15 shell can defeat 5" armor at 60° obliquity
and 411 armor at 68° obliquity.

In the case of defeat of armor by kinetic energy and chemical energy IIEP shell,
more or less massive pieces of metal flying at considerable velocities become available
to inflict damage behind the armor. In the case of HEAT shell, howver, only a thin
beam of tiny, incandescent particles emerges behind the a,rmor. Personnel or equip-
ment directly in the path of the jet will becoJDe casualties, but the damage may not
necessarily be serious. In order, therefore, to insure that the emerging jet will
possess a significant degree of lethality, it has recently been agreed that the jet
IQUSt have a residual penetrating ability of 211 of armor after defeat of the main armor
to be considered effectively lethal.

It li&S found early in World War II that spin stabilized HEAT shell fired from
rifled guns suffered a 30 to 50% loss in penetration performance as compared to non-
rotating shell. The centrifugal force of spin li&S su.i'ficient to cause the jet to cone
out and dissipate much of its energy-. This factor led to the intensive development of
fin-stabilized non-rotated HEAT shell. The depth of penetration of 4 times the cone
diameter applies only to non-rotated shell.

The depth of penetration by HEAT shell is inversely proportional to the square /J

root of the density of the material under attack, therefore the thickness of material f!
required to defeat the attack is also proportional to the square root of ita density. ·
Since the wight of material varies directly as its density, the weight of material
required to defeat HEAT shell varies directly as the square root of its density. Low
density DBterials are thus more resistant, on a weight basis, than are high density
materials. Thus aluminUJII and magnesiwn will offer better resistance to HEAT attack
than will the same weight or steel a,rmor. In their present s"tage of development,
bowver, aluminwn and magnesium alloys in section thicknesses comparable to 2" and
more or steel are significantly inferior to steel armor in resistance to attack by
kinetic energy- projectiles.

207
 In their present stage of development, fin-stabilized HEAT shell do not match the
accuracy of kinetic energy projectiles, and this, coupled with their lower velocities,
results in lower hit probabilities than are possible with kinetic energy projectiles.
The higher velocity and greater accuracy of kinetic energy projectiles make them con-
siderably more accurate than all present types bf chemical energy armor-defeating.
ammunition.

PERFORMANCE OF KINETIC ENERGY PROJECTILES .AGAINST COMPLEX ARMOR TARGETS

Kinetic energy projectiles have been tested against spaced armor targets consisting
of relatively thin (approximately 1/211) plates placed some distance in front of the
more massive main armor~ The function of the skirting plate is not to extract a sig-
nificant amount of energy from the attacking projectile, but to so affect it by yawing,
decappi.ng, or fracturing the shot that its performance against the main armor is de-
graded.

Results of firing various types of 57MM and 90MM kinetic energy projectiles
against spaced armor targets are shown in Table III. These tests were conducted at
Aberdeen Proving Ground under the technical supervision of the Watertown Arsenal.
Photographs of the projectiles were also taken as they emerged behind the skirting
plate in order to observe the effect of the skirting plate on the projectiles.

It was found that 57MM AP and APC shot were not fractured by passage through 1/211
thick skirting plate but were considerably yawed. In addition, the cap was always
removed from the APC shot. Surprisingly, the 90MM :A.P shot were found to be readily
fractured by passage through 1/211 thick skirting plate. Since the 57MM shot were not
fractured, parallel plate arrangements were found to be worse than the basic armor
since the shot were yawed in the direction of lower obliquity against the main armor.
Oppositely sloped spaced armor arrangements are indicated for cases whe.re the shot
cannot be fractured by the skirting plate.

Since the 90MM AP shot was broken by the skirting plate, both parallel and non-
parallel placement of the skirting plate were eq~y effective in degrading this shot.
The 90MM APC shot was not readily fractured by skirting armor, but its performance was
degraded against the target conditions shown in Table III because, once ita cap was
removed, it behaved essentially the same as monobloc shot, and the target conditions
chosen, namely 300 and 400 obli.quit;r, are those where monobloc shot are less effective
than capped shot.

Tungsten carbide cored projectiles may be seriously degraded by spaced armor

arrangements since the very brittle core my be rather easily fractured by skirting
plates. Once the core is fractured prior to impact against the main armor, the HVAP.
type of shot is rendered comparatively ineffective.

It is ess13ntial that spaced armor arrangements to defeat kinetic energy projectiles

be very carefully chosen, because it is entirel;r possible that some arrangements cause
ths projectile to penetrate far !IDre efficiently than it 110uld against the DBin al"IIIOl"
alone.

208
 Spaced armor arrangements can readily be designed to fracture capped shot. This
can be done by using two skirting plates separated from each other. The first skirt-
ing plate removes the cap and the second fractures the shot. The test described in
Table III, 'Where the 90MM APC T50 liaS fired against two 1/2" thick plates parallel to
and separated from each other and the main plate (3" at 55°) by 8 11 of space shows
llhat can be done with this type of plate arrangement. Thl.s arrangement could not be
defeated even at point blank range, whereas 4" armor at 550 obliquity can be defeated
by the same projectile at ranges up to600 yards and 3" armor at 55° at ranges up to
1600 yards.

PERFORMANCE OF CHEMICAL ENERGY PROJECTILES

AOArnST COMPLEX ARM:JR TARGETS

Since the IIEP shell defeats armor by the application of a severe shock 'Which
induces stress waves of high magnitude, it is obvious that the best way to cope with
the attack of thl.s type of ammunition is to prevent the shock wave from getting
started in the armor. It does not help liNCh to increase the thickness of the plate
since iarge increases in thickness are required to defeat IIEP shell. The British have
done an extensive amount of firing of IIEP shell against spaced armor structures and
have found that they could be readily defeated by spaced armor combination, by rubber
pads placed between armor sandwiches, etc. The skirting plate of spaced armor ar-
rangements designed to defeat IIEP shell must be supported well enough to prevent con-
tact with the main armor during detonation of the shell, since then the shock wave
will be transmitted to the main armor.
Table IV contains some data recently obtained at Aberdeen Proving Ground on the
performance of tbe 105MM IIEP T81E17 shell against 3" armor at 55° and then against
a spaced armor combination consisting of a 1/2" plate 8 11 in front of the 3" armor at
- 55°. This shell can normally defeat 5" plate at this obliquity. Its inability to
defeat a spaced armor combination consisting of a total of 3-1/2 11 thickness of steel
indicates how greatly this type of shell can be degraded by spaced armor.

There has been extensive work. on the development of defense against HEAT shell.
This t~1P8 of ammunition can penetrate such a great thickness of solid steel that other
means must be found to defend against it. Since it liaS apparent that low density
materials offered better protection against HEAT shell for a given weight of material,
a considerable amount. of effort liaS expended during World War II in developing low
density materials for thl.s purpose. The. best low density material developed li!LS HCR-2
~ch consists of a mixture of quartz gravel in a mastic base of 75% asphalt and 25%
wood flour. This material when placed behind a thin steel plate and attached to the
main armor of vehicles was found to give much better protection against HEAT shell
than a similar weight of steel or other materials with the exception of solid or
laminated glass.

Auother type of defense against the HEAT round ws provided by fi.xing 7• to 8"
long closely" spaced steel spikes to the surface of the main armor; the function of the
spikes being to break up the cone before initiation of the jet. Spiked armor structUres
have been found to be effective against several models of HEAT rourids.

209
 II 7"

10-··-·

:-:::.~ !~ .
AT n° M0 liAicEI! AR-'--, ~"..':":~
; ..

' ·
...

9
oeuaurrv. '" . _ .•••.ilAiiiio.:·;.;;;,~~~-~!i ••
:--- -----
: ~ .. I '- .
:-Tl ~-~-~ --e--1•" ••
.,
B'
: '! .: . i :
--;-
i i
i
1
. I
1
'
i .
"'iS 7
I I i. : I 7"
••
~
'
I
! • ••
" ii

I ...
II:
1\) '
~ 5 I"
0
.
"'z
"' ·"'
.... d'

.."'... I
0
r

Q IN>JllriMI11NllNN .~ UA~~UM.
w
Iii _'!: ftCHES
~
( . i
I
1 I
L
I
NUMBER OF TESTS

VATERTDVN ARIENAL LAIDRATDRY VTN.639-11 1 5DI

C0 N r I 0 £ N T I AL

Figure 7
 More recent developments sponsored by the Detroit Arsenal show that an arrangement
. of parallel angle irons, made of armor steel, placed on the surface of the main armor
offer considerably increased protection, particularly at high angles of attack, against
both HEAT and HEP shell. Although these angle irons did not defeat all the HEAT shell
fired at them, they were effective in signi.f'icantly reducing the probability of a per-
foration when hit by a HEAT shell.

It was found during World War II that spin stabilized HEAT shell could be fairly
readily defeated by spaced armor due to the degrading effect of the spin, particularly
when the standoff was increased by the spaced armor. More recent tests have sho'Wil that
spins even as low as 10 rev./sec. result in a 20% decrease in penetration performance
of l05MM HEAT shell against a spaced armor target consisting of a 3/4" plate 12 11 in
front of the main armor. This 20% decrease represented degradation in performance of
the same round compared to its performance against the same target ~en the shell was
not rota ted.
Very recently, tests of the 3.5" HEAT M28 rocket and the 90MM HEAT Tl06E20 shell
have been conducted against spaced armor targets. The basic armor consisted of 4 11
plate at 40° obliquity; firings were first conducted at this target, then at spaced
armor targets consisting of a 1/2" thick plate parallel to and 8 11 in front of the main
armor. In one case a 3" plate was placed behind the 4" armor, with 1/4" thick plates
stacked behind the 3" plate in order to measure the residual penetrations, but in all
other cases, a series of 1/4" thick plates wre stacked behind the main armor for this
purpose.

Figure 7 tabulates the data resulting from the tests described above. Starting
at the lefthand end of the chart, 14 rounds of the 3.5" rocket were fired against 4"
armor at 40° obliquity. Eleven rounds completely perforated the 4" plate and achieved
residual penetrations of 3 11 to 3-5/6 11 into the 3" and 1/411 thick plates behind the 4"
plate. Three rounds appeared to produce low order detonations; the 4" target, howver,
was perforated, but relatively little residual penetration into the 3 11 back-up plate
was achieved. ·

Nineteen rounds of the 3.5" rocket were then fired at spaced armor consisting of
the 1/2" skirting plate parallel to and 8 11 in front of the 4" plate. Of these nine•
teen rounds, four failed to perforate the 4" main armor, three perforated the 4 11
main armor but had no residual penetration ability, three more perforated, but had
residual penetration abilities of less than 1", three achieved residual penetrations
of 1 to 1-1/2 11 in depth, and the remaining six rounds achieved the same residual
penetrations as were obtained against. the solid armor target.

The performance of the 90MM HEAT fi08E20 as found to be very variable against a
simple armor target consisting of 4" plate at 40° obliquity. Oi' ten rounds fired, all
perforated the target., but the residual penetrations varied from 2" to 5-3/6". loJhen
tested against spaced armor with 8 11 spacing, two rounds of eight 90MM Tl06E20 shell
.fired failed to defeat the target:. Four rounds perforated the target and achieved
residual penetrations of the same order of magnitude as were obtained against the solid

2ll

t •
armor target, wile two rounds achieved even higher residual penetrations; 6-1/2 11 to 7"
in depth. This increase in residual penetration probably resulted from a more efficient
standoff caused by the 8" spacing. ·

A 1611 spacing between the skirting armor and the main plate greatly improved the
effectiveness of spaced armor against the 90MM HEAT Tl08E20 shell. Of ten rounds fired,
four were totally defeated, 4 perforated but achieved residual penetrations of but 1/4"
to 1 11 in depth, while only tw rounds performed as well as they did against the main
armor alone.

While the above tests are only elementary in nature, there appears good hope that
spaced armor combinations may be. devised Which will be even more successful against
HEAT shell. Spaced armor may be particularly effective at higher obliquities of attack.

Combinations of the main steel armor, low density materials, and spaced armr may
yet provide real defense against chemical energy armor-defeating ammunition

In conclusion, the thought should be expressed that both types of armor-defeating

ammunition should be brought to the field of battle. Any decision to adopt one type of
the kinetic and chemical energy anununition to the exclusion of the other type would
greatly simplify the enemy's armored vehicle design and construction problems. It is
possible to devise a reasonably simple defense against either type of ammunition alone,
but the problem of defense against both types together is an extremely complicated one.

212
 TABLE I
PENETRATION" PERFOHMANCE OF 90MM AP T33, 90 MM APC T50,
AND 90MM HVAP M304 vs ROLLED AND CAST HOMOGENEOUS
AND FACE-HARDENED ARMOR
Ballistic Performance of
Armor 90MM AP T33 90MM APC T50 90MM HVAP M304
- Bal. l-hx. Bal. Max. Bal. l-hx.
Thick• Obl. of Lim. Range Lilli. Range Lim. Range
ness 2 in. Type BIIN* Attack fLe zds fLe zds fLe zas
3 RH 280 450 1983 5000 2216 3975
3 RH 280 6QO 2629 2200 2853 1300 3145 1575
3 RH 280 65° 3026 625 3101 350 3543 675
3 RH 320 6QO 2645 2150 2795 1550 3246 1350
3 RH 320 6SO 2870 1225 3109 350 36ll 525
3 CH 26o 450 1979 5000 2437 3225

--- ----
---
3 GH 260 53° 2315 3550 2725 2550
3 GH 280 55° 2313 3550 2754 2475
3 GH 280 6QO 2586 2400 2683 2000
3 GH 280 65° 3ll2 325
3 GH 280 70° 307.3 375
3 FH 55°
6QO
2059 4725 2248 3825
) FH 2505 2725 2635 2175 2903 2125
3 FH 65° 2648 2125 2902 1100 3300 1225
4 RH 280 300 2054 4750 2149 4275
4 RH 280 450 2469 2875 2831 1400
4 RH 280 55° 2742 1750 3010 700 35n 625
4 RH 280 6oo-... 3079 450 3162 125 3638 475
4 RH 320 55° 2719 1850 3138 225 3503 775
4
4
RH
RH
320
36o
6oo ....
6oo,
3075
2943
450
915
--3097 375
3748
3680
225
375
4
4
CH
CH
240
240
5SO
6oo'
2785
2933
1575
1000
----
3208 Above MV 3800 100
4 CH 280 5~ 2620 2250 2744 1750
4 GH 280 3007 700 3135 250
4
4
GH
CH
280
320
6o '
6~
6oo, 2947
3129 250
950 ---- ----
3175 15 3669 400
4
4
FH
FH
---
--- ~
50° --- ---- 2391 3290 2868
316o
2225
1550
4 FH 55° 2765 1650 3397 1000
4
5
FH
RH 240
6oo-...
300
2763
2804
1675
1500
3069
2343
475
3400
---
5
5
RH
RH
240
320
450
300
3146
2967
200
875
2976
2461
825
2925
----
5 RH 320 450 3167 100 3177 15 ----

213

J£ r
 Sf -. L til :

TABLE I (Cont'd)

Ballistic Performance of
90MM AP T33 90!~ APC T~O 90MM HVAP !004
Armor Bal. Max. Bal. Max. Bal. Max.
Thick- Obl. of Lim. Range Lim. Range Lim. Range
ness a in. Type B!IN* Attack fLs zds fLs zds fLs ;lds
5 CH 240 30° 2543 2575 2234 3900
5
5
CH
FH
240
---
450
30°
2740
2475
1750
2850
2905
2394
1100
3175 2819
---- ----
2325
5 FH 45° 2866 1250 28J9 1200 3208 1425
6 RH 260 300 3214 Above MV 2750 1725 2562 2925
6 CH 240 30° 2907 1100 2632 2200 2487 3100
6
7.6
FH
RH 260
30°
30°
---- 2863
3182
1275
50
3333
2892
1150
2150

* BHN - Brinell Hardness RH - Rolled homogeneous armor

Grea tar than CH - Cast homogeneous armor
Less than FH - Face-hardened armor
mv - AP, APC - 3200 f/s
mv - IIVAP - 3850 f/s
TABLE II

PENETRATION PERFORMANCE OF HEAT SHELL AGAINST ARMOR

Obliquity No. of Rounds
Projectile· Thickness of Armor of Attack Fired Results
90MM HEAT* Approx. 5" 5 3 complete penetrations
Tl08ElS 2 partial penetrations
II
4.1" 1 Complete penetration
n 3.4" 2 Ricochet, no penetrations
II
4.5" 1 Ricochet, no penetration

* fired from range of 50 yards, 2400 f/s muzzle velocity

214

' ..
 TABLE III
PERFORMANCE OF KINETIC ENERGY PROJECTILES AGAINST SPACED AJU1JR COMBINATIONS
Ballistic Performance
M!xirmun
Mlin Armor* Thickness Ballistic Range for
;
Thick- Obli- Skirting Arrange- Limit f/s Penetration
ProJectile ness guitl Plate ment** S~cing (Protection) -yards
57MM AP KTO 2'' 40° None 1997 1850
II 211 40° 1/211 A 1611 1172 2500
II 211 40° 1/2 11 B 1611 2620,2638 175
57MM APC M86 211 40° None 2149 1500
II 2" 40° 1/211 A 1611 1943,1940 2000
II 211 40° 1/2 11 B 1611 2755 Above m.v.
90MM AP T33 3" 55° None 2505 2725
II 311 55° 1/2 11 A 811 2952 950
II
4" 30° None 2025 4875
II 411 30° 1/211 A 1611 2383 3250
II 411 30° 1/2 11 A 811 2368 3300
90MM APC T50 3" 40° None 2040 4800
II 311 40° 1/211 A 1611 2452 2950
II 311 40° 1/211 A 811 2501 2750
II
4" 30° None 2171 4200
II
4" 30° 1/2 11 A 1611 2681 2000
II 411 30° 1/211 B 1611 2657 2100
II 411 30° 1/2" A 811 2669 2050
n 3" 55° None 2777 1600
II 311 55° 3/411 B 20" 2657 2100
II
3" 55° two 1/2 11 A 8" partial at Above m.v.
plates 811 apart 3249
90MM HVAP ~04 3 11 55° None 3018 1850
II 311 55° 1/211 A 811 3832 50
II 4" 30° None 2226 3750
II
4" 30° 1/2 11 A 811 2690 2625
II 411 30° 1/211 A 1611 2780 2425
II 411 30° 1/211 B 1611 2702 2600
611 30° None 2590 2900
"II 611 30° 1/211 A 12 11 3689 350
M!.in armor is rolled homogeneous plate of 260-280 BHN.
* Arrangement:
** A - Skirting plate parallel to main armor
B - Skirting plate and main armor sloped towards each other,
each inclined at same angle but in opposite directions
from normal.

215
 SELL 2 2£ 5 If
TABLE IV

PERFORMANCE OF HEP SHEIL AGAINST SPACED ARMOR COMBINATIONS

Thickness
Main Armor Skirting Arrange-
Projectil.e Thickness oDriguitz Pl.ate ment Sl?!cin~ Result on-Tar~et

l.05MM HEP 3" 55° None Average spal.l. size

T8l.El7 1.0 11 X 7 11

" 3" 5~ 1/211 paral.lel. 8" No spall:i.ng, target

not defeated

The l.05MM HEP TBl.El. 7 shel.l. is normally capabl.e

of spalling 5" thick roll.e d holliOgeneous armor.

216

'R£££2lf
J

W. E. Soper

Armament Research Establishment, Fort Halstead, Kent

'
Introduction

During the years following 1945, the attention devoted to armament research was
naturally severely curtalled, and, in this process, research on the hollow charge and
squash head virtual1y ceased, apart fr0111 a few sporadic attempts at ad hoc work. This
state of affairs is now ceasing, but the process of re-establishing a good research
team equipped with modern apparatus is a comparative1y slow process. For this reason,
this paper will be devoted large1y to a survey o£ the field as it appears in Great
Britain, and the directions in which the now growing research effort wlll be expected
to turn. It cannot be a record o£ considerable achievement, although an account will
be given o£ such ad hoc work as has been carried out.

1. Hollow Charge
Cone Production

It was obvious from the American results during the war years that, were deep
penetration is required, copper must replace steel as the material for the conical
liner.

A limited amount of work of a rather ad hoc nature was undertaken to try to formu-
late the requirements to specify for given cases. The first results, obtained with
cones with 30° and 450 apex angles, 2.45 inches in diameter, pressed from 1/16 inch
sheet metal were rather disappointing, although in the main s0111ewhat better than with
steel cones. These first copper cones were, however, rather poor and had obvious ribs,
although the variation in vall thickness was not large.

It vas therefore decided to attempt to set a standard by making cones which were
not pressed. These cones were all to have apex angles of 45°. Two methods were tried.
The first vas electrodeposition on to a graphited steel former. This produced a cone,
the thickness of vhich.diminished in a regular manner from apex to base, the mean
thickness being about 0,;06 inch. Under certain conditions, the surface vas fairly
good and free fran pits, and these cones were made up into charges without any ma-
chining, apart from cutting off the skirt to leave the cone 2 inches in diameter at the
base. In otDer cases the outside vas machined. The difficulty was to obtain repro-
ducible specimens.

The charges made .t'rom tDese cones perforated 200mm of hanogeneous armour and this
vas a considerable :Improvement in performance over the pressed cones. The second
course vas to machine cones from 2 inch drawn bar. This method is, of course, much
more fiexible and probably the best method to use for cones for research purposes.

217
 - '~.~ ... ~ - -,; "' . ,., ...

With this method cones of different wall. thickness were studied including sane
with wall thickness tapering either from base to apex or from apex to base •.

The results of these tests are rather interesting.

~ Result

Electrodeposited thickness 0.041 Perforated 200mm homogeneous armour

"' o.o65 200mm

Turned thickness 0.035 Failed at 200mm n 11

o.o45 Perforated 23Qmm homogeneous armour
o.o66 23Qmm n n

Turned tapered thickness Apex 0.08 Base 0.05 Perforated 200mm

Apex o.o5 Base o.oa " •
Apex 0.066 Base 0.032 Failed to perf. 200mm
Experiments were then abandoned on the tapered wall.. This does not necessarily
imply that the tapered wall thickness is abandoned for all time; it merely means that
continuation on an ad hoc baeia was not thought to be profitable, in view of the pran-
ieing results obtained from cones o£ uniform wall thickness.

Fran this basis, a contract was given to a firm, Messrs. Pl.eesey, to develop a
pressing technique; and, in fact they prodllced, after sane months of experiment, sane
very eatie£actory cones of 2.6 inches diameter and nominally o.o6o inch thick. These
cones turned out surprisingly uniform. The apex was thin being .of the o:Mer 0.050 to
0.055, but below this the thickness measured at four places at four different heights
did not dif£er by more than 0.003 inch. This cone perforated 300mm of armour. The
cones were produced by seven successive pressing operations, each step being followed
by annealing. Encouraged by this the Armament Design Establiebment put an order on to
the Royal Ordnance Factories to produce a large 5-inch diameter cone 0.1 inch thick.
These did not turn out eo well, but after machining on the outside a more or lees uirl.-
form thickness of o.o96 inch was obtained. These cones hwe not so far been fired
statically, but dynamically, in the 5 incb rocket, a perforation of 10 inches at 6o 0
from normal (equivalent 500mm) was obtained by one shot in the A.D.E. trial.

Lethality

This discussion leads directly to a question on which a great deal of interest now
centres. When assessing the performance o£ a gi-v:en hollow charge, it has been the
custom to base it upon the thickness of armour which is perforated irrespective of
what damage is done behind the armour plate. This, however, does not appear to be
completely logical because the primary aim ie. to cause lethal d11111a.ge behind the struc-
ture which the hollow charge has to penetrate.

It has almost alwa,ye been the custom in Ordnance Board trials against armour to.
place behind the rear face o:t' the armour, two mild steel plates 1/lt:i inch thick,
separated by a distance of 6 inches from one another, and situated so that there is a

r·
218
 distance a1' 3 feet between the rear face of' the armour plate and the first surface of'
the target plate structure. The evidence of' damage which would occur within a struc-
·ture protected by the armour is thus obtained. It has been a common experience to find
: that at the extended limit a1' ita performance a hollow charge will do little or no
damage to the target plate structure.

This then leads to the question of' what criterion should be adopted by the de-
signer when c.alled upon to defeat a specified target. Should he design on the basis of'
!
perforation or should he allow some overmatching coefficient in order to do lethal
damage bshind this specified target? This question is now being hotl;y debated in Great
Britain. It is obvious that, if' an overmatching coefficient is introduced, it will
1118an an increase of' calibre, and, thus, a1' w.eight of' the .round designed; and it is
important then to know a precise value to put to this coe£ficient.

Experiments to this end are now in progress and, at this moment, it is impossible
to give any very clear or emphatic decision. This work has been pr:lmarily sponsored by
the Ordnance Beard, ani a programme has been arranged to fire, statical.ly, in order to
eliminate some of' the firing variables, a number of Service IIIUnitions against plate at
and around the critical penetrating thickness with targets of wood, A.P. shot complete.
with filled cartridge cases and tanks containing Diesel on. The hollow charges to be
used are the Energa Grenade, the 3-1/2 inch rocket head both of which have copper liners,
the 95mm shell with steel liner to be followed later b:y the 5 inch and 4-l/2 inch
rocket heads with copper liners.

Sane of' this programme has been fired and in giving the present trends, it is
with the understanding that. as the trial proceeds these trends m~ be profoundly
altered. What appears to be emerging at the moment is that, at the critical thickness/
of' plate, round to round variations are considerable both in penetration and damage;
but if a point is taken for the perfonnance where a series of' 9/JJJ successll!s against
the plate is achieved, then at this level of performance considerable lethal damage
m~ be inflicted inside the armour protected structure.

This conclnsion is, naturall:y, of little help to the designer who wants to know
whether to add a constant thickness to his specified target, or whether it is better
calculated ·on the basis of a percentage of armour thickness.

Later experiments vi th the larger 4-1/2 inch and 5 inch heads, none of which have\
yet been fired, ma:y help to resolve this scaling question. \
I
Meanwhile a short series of fuings of the standard 2 inch experimental charge
using various thicknesses of liner are to be fired to ascertain how nearly the maxi":"
IliUm penetration and maximum damage criteria can be reconciled.

Regu].arity

It is obvious from what has been stated in the prev:l.ous passage that one of the
greatest illlprovements that could be made to the comrentional hollow charge W'ould be
to increase its regularity-. In the United States a very great volume a1' painstaking

219

•
work has been dcme on this theme. The regularity and &ylllllletry or the liner has been
studied from the point or view or thickness and axial symmetry; the unitormity and
homogeneity or the high e:zplosive charge has also been studied together with the errect
or orr-axial initiation. It has been shown that caretul attention to these details has
mads a considerable imprOV'elllent to the range of spread but has not elim:1Jlated it.
Particular attention is nov being directed in Britain to the question o! symmetrical
initiation. There is some evidence to show that with the ordinary method or initiation
with a detonator and booster pellet it cannot be assumed that a By~~~~~Strical detonation
wave will resalt from the most carefUl geometrical set-up, end an investigation was
started by Evans and his group on :Improved methode of initiation. This work has nov
been taken over by the new hollow charge group. It is as yet too early to give ar:ry
reeul te but the purpose or the p~sent preliminary investigation is to see 1! there is
lillY improvement in regularity with a true point initiation and a symmetrical detonation
wave.
Lines or FUture Investigation

There is stUl a chance that some advantage lll!ey' be obtailled from the liner with a
tapered wall thiclmess 1! it can be adjusted to the optimum collapse velocity possibll-
with the help of a shaped detonation wave. There is a very considerable field !or en-
quiry here aDd it may be necessary to depart f'rom the conical shaped liner. The
prlmary object will naturalll" be to increase the mass or the jet without decreasing the
velocity gradient. There is a large source or untapped energy in the explosive 1! onll"
we can devise the means to obtain it.

Not much attention has been given at the present time to the shock waves which are
involved in the various steps of cone collapse, jet formation and target penetration.

Consider the collapsing conical liner, the detonation wave coming through the ex-
plosive starts a shock wave through the thickness or the liner. At the other side a
shock wave is transmitted into the air and a shock wave or rarefaction wave returns
through the liner. Work an this problem is now being started and the simple case or
transmission through a nat plate is first being studied. This work acts as a link
between the hollow charge and the squash head which is the responsibility or the same
group at the Armament Research Establishment.

Work is also being started on 1:he study or the penetration phenomenon by the
hollow charge jet with special reference to the shock wave which must precede the head
or the jet.

It is unfortunate that this work is not yet in a stage where some concrete results
could be exhibited.

These problems are being tackled by a combination or methods including oscUlo-

grapey1 Kerr-Cell cameras and rotating mirror cameras. We are also very interested in
the promising technique being developed by Stanton at the N.O.T.s. using tracers to
record the paths or collapsing elements or the cone.

220
 . The Filling of Holl011 Charges

The standard British filling !or hollow charges is RDX/TNT 00/40. There is, how-
ever a certain amount of di!!icul.ty attendant upon this as British mamrl'actured RDX/
TNT l,oj4f) tends to be a rather thick '1DDJanageable mass. It is thus very di!ficul.t to
avoid the formation of cavities in the filling and to get the filling down to the
fiange of the cone without bridging. Consequent]¥ the filling factory organisation
has often extracted a reluctant concession to reduce the proportions to 50/50. This
is a bad and retrograde step !or two reasons, the more obvious being the loss of powsr
inherent in the charge between the proportions {:J)/40 to 50/50. The second reason is
that the tilling is now very IIDlch more fiuid and the RDX tends to segregate, and zonal
examination of the segregated tilling shows that the settling is not ~etrical and
the case ~ arise where there ~ be for example a zons of ~5/35 RDXfTNT on ons side
of the charge with perhaps 55/45 RDX/TNT on the .other side at the same level. The
· manufacture of RDX/TNT IIDlst be pursued in the zone of physical chemistry.
Rotation

The problem of compensating !or the effects of rotation on the performance of the
holl011 charge jet has proved to be very tl'OublesoiiiS. One of the first di!ficul.ties is
to ascertain what exact]¥ is the mechani8111 by which the jet is perturbed. Quite simply,
it Jll.ey' be due to tangential velocity effects during cons collapse or it ~ be due en-
tirely or in part to centi-i.f'ugal effects acting against the forces causing collapse as
visualised by Schumann. The fonner effect seems to be unlikely i f the collapsing liner
is conceived to be coherent. But in either of these cases the effect of rotation is to
prevent the complete !ormation of the jet by reducing the forces causing collapse.

The other mechanism by which the jet ~ be perturbed after fomation is by centri-
fUgal spreading due to the conserved aneular momentum. Shell trials against homogeneous
armour, and armour protected by a thin (1/4 inch) skirting plate at 5 inches distance
suggest from the obvious instability of the jet that this effect at least is operating.

With regard to caupensation it is considered that with the wealth of experience

which has been gathered in the last few years on the fiuted cone it would be unprofit-
able !or us in Great Britain to enter this field on the research side. But we are not /
satisfied that the last word has been said on behal£ of the cylindrical liner. This j/'
type of liner has obvious advantages i f it can be made to work efficiently, the chief
of which is that it will have a smaller angular momentum, and it might be possible to
use it in practice without compensation.

There is obviously quite a big field of work on the unrotated cylindrical liner
before aey rotation work is begun. There are maey obvious theoretical di!ficul.ties,
but it is intended as soon as possible to survey the ground with a view to increasing
/
the collapse forces while del~ the collapse process-by suitable shaping of the
detonation wave and adjustment of the thickness of the cylindrical liner.

Weapon Design

The endeavour of our designers is lll8i.n:cy" concentrated on obtaining a high level of

performance at large angles of incidence. The electric nose switch developed by the
·'

221
 _, £€REI SJCMii§ !Iii&ll

Design Establishment has established that functioning can be obtained up to angles of

6o 0 from normal with the rocket head at striking velocities of the order of' 700 rt/sec.
It is hoped that with introduction of' the new A.R.E. detonator which will .function with
very small de~ when actuated by a very low current, a perfol'lllance comparable with that
obtained static~ will be possible at high angle of incidence and high rate of
striking.

Trials with an electric switch in shell have shown that some re-design may be
necessary in order to get good results.
2. Squash Head

The squash head shell has developed on very different lines from the hollow charge.
It has been very largely a matter of empirical development without aey pure research.
As a consequeney a great rnazv trials have been carried out and it is often quite diffi-
cult i f not impossible to correlate them.

The principle is fil'lllly established that a thin walled shell filled with a plastic
or semi-plastic H.E. filling and fitted with a fuze with a small delay will under
certain conditions release a 'scab 1 from the back of the plate on detonation.
It has been assumed that during the short time of dela;y between the impact of the
shell on the plate a.nd the functioning of the fuze that the thin nose of' the shell is
squashed out on to the plate giving a fairly close contact between a comparatively
large section of the plate and the H.E. filling.

Kille camera records have been taken of the impact of shell on plates but the actual
mechanism of collapse and spreading is to a large extent obscured by nash, smoke and
debris.

This type of shell works reasonably well at low velocities of impact up to about
15'00 rt/sec and at angles from nomal up to about 6o0 from nol'lllal.

At higher velocities a general deterioration in performance takes place especially

at angles near nol'lllal.

At higher velocities still, a poor per!Ol'lllance is marked at angles from normal up

to 30° and again at larger angles of the order soo to 6QO. This poor performance appears
to be associated with premature detonation or explosion of the H.E. filling by impact
shock.

This study of this phenomenon is the major imrestigation on squash head shell at
the present time.

It has already been remarked that our knowledge of the exact mechanism of squashing
is extremely sparse and our knowledge of the behaviour of explosives under high rates
of strain is very poor.

222
 Apart from obvious weakness in design o:r Jlllii'JY squash head shell, there are several
possibilities :for the cause o:f this premature detonation, They will be listed: -

1) Aeration o:r the :filling leading to detonation by adiabatic compression

on :impact,

2) Heat o:f deformation o:f case leading to ignition o:f the :filling,

3) Heat generated by the shook wave running back into the :filling.

The subject of (1) is under investigation and the scope o:f the investigation is to
extract the a:lr :from the plastic explosive in a vacuum incorporator and then to :fill by
extrusion into an evacuated shell, It is thought that this is a practical proposition
and a pilot plant is being constructed,

Under (2) it is proposed to line the shell, more especially the nose, with an inert
material ldlich will absorb or del.lv the transmission o:f heat to the high explosive.
Trials o:f this nature have not yet been carried out,

The heat generated by the shock (3) is thought to be the most likely cause o:f pre-
mature ignition. or detonation.

From the theoretical side some calculations have been carried out by Thornhill in
the A.R,E. on the possible shock temperatures in the :filling at different striking
velocities, This calculation can only be very approximate because the equation o:f
state o:f plastic e:xplosive is entirely unknown and it was necessary to assume data by
analogy- with certain organic compound :for which data is evailable.

It is clear from these results, without ascribing any accuracy to the actual
temperatures involved, that at higher striking velocities the shock temperature rises
very sharply and from this point of view it is obvious that there is a maximum striking
velocity beyond which i t is impossible to go,

Again using the actual temperatures which Thornhill obtained and combining them
with the heat o:f formation o:f R.D.X. the time of survival of the e:xplosive and differ-
ent striking velocities was obtained, It happens that between 2000 :rt/sec and 2500
:rtfsec striking velocity the time of s.urvival works out at 1 millisecond and 1 micro-
second respectively. It is not claimed that this agreement with practice is anything
more than :fortuitous but it does demonstrate that there is a limiting velocity at
Striking,

Attempts to damp out this shock weve h!EVB been made and suggested. The simplest
is to fill a quantity of inert plastic in the nose of the shell, This has been tried
out in the new l20mm and gwe some successi'ul shots at 2400 :ft/sec. This resuJ.t· needs
confirmation with a large munber of shots. A suggestion by Thornhill vas to encourage
streamline now of the plastic H.E. by putting in the nose o:f the shell along the axis
a streamlined plug,

223
 Another problem which has not yet been resolved is the fiJ.ling of squash bead
shell. There are two possible fillings, Plastic Explosive and pourable RDX/Paraffin
Wax. Owing to the design of the British squash bead shell it is difficult to fiJ.l
with Plastic Explosive except by hand stanuning. The competitive pourable RDX/Paraffin
Wax is not alw;zys pourable without trouble and there are varying reports that it is
not so efficient in bringing off a •scab' as P.E.

Research on the mechanism of squashing is new being studied pbotograpbical~ on

the small scale and it is hoped that the results of these studies will assist in de-
signing a more efficient shell, the problem being to get the largest spread of H.E. in
the shortest possible time at all angles of attack.

224
 HELLER WARHEAD DEVELOPMENT*
~t·'''
R. W. Foster
Ganadian Annament Research and Development Establishment, Valcartier1 Quebec

1. INffiCDUCTION

The Heller is a platoon anti-tank weapon o£ 3.2 inch or 81-mm calibre undergoing
· final engineering trials. Canadian User trials are expected to take place in December
1951. The rocket burns all propellant within the launching tube so that it is non-
propulsive during flight. This method was chosen as it provides more regular external
ballistics than the orthodax: rocket which m~ burn and provide thrust after launching.

The rocket is fin-stabilized with a chamfer on the leading edge o£ the fins which
imparts a slow spin in the launcher to compensate mis-alignment or dissymetr.y obtained
in manufacture. (Fig. 1).

The shaped charge warhead is initiated by a spit-back fuze. However, it is in-

tended to replace this at a later date by an electric fuze under development.

2. WARHEAD DESIDN

2.1 The Copper Cone Liner

When the design of the warhead was first considered, it was decided to use a
copper cone liner with a solid angle of 42°. The warhead.was filled with RDX/TNT
6of4D. These decisions were made on the basis of work on shaped charges in other
countries. An independent investigation into these variables would have invo1ved more
, work than C.A.R.D.E, could complete within a reasonable length of time.

Tools were designed and manufactured for producing the copper cone liner, and
300 warheads were manufactured for full-scale static penetration trials. These pene-
tration trials were fired to find the optimum stand-off and cone liner thickness for
the shaped charge, For stand-off, suitable wood stands were used to give a range of
stand-offs from 1.5 cone diameters to 2,25 cone diameters. The range of cone thick-
nesses was obtained by pressing cones of ,09 inches wall thickness and machining the
outside surface o£ the cone. In this way, wall thicknesses from .05 to .09 inches
were obtained. (Fig. 2)

Briefly, the results from these first trials showed that the optimum stand-off
was 1. 75 cone diameters and the wall thickness should be greater than .09 inch.
A new set of tools was then made to produce cones up to .125 inches wall
thickness and it was possible to fire trials with oone thickness of .100, .ll5 and .125
inches, again machining the outs:i.de surface as required - (Fig. 3). Penetration de-
creased with these greater thicknesses and i t wa.s decided from the curve o£ penetration
versus cone wall thickness that .095 should be the thickness for the cone,

*Project S/N-17-1; Reference N-17-1; Tech. Letter N-17-1-5.

225
cMhll£ 7
7

Figure !-Heller Round.

226

II I h£2£ z; L
a nm!TJ$1 5 I

Figure 2- Cone Liner.

Figure 3- Cone Liner.

227

SJtD
 mru 8 •r I ssti 7

. However, considerable difficulties have been encountered with the CODeS

mBilllfactured from the second set of tools. The tools are simple in design and use un-
guided punches. The cones are drawn in ll operations and two ironing operations in-
stead of four operations and two ironing operations used with out first set of tools.
The new tools produced cones with better control over metal now but it was difficuJ.t
to keep the inside surface of the cone concentric with the outside. The variation in
wall thickness around any circumference of some cones was as !llllch as .005 of an inch.
It was found that this -cype of variation caused an unacceptable degradation in pene-
trative performance and a considerable number of double jets were formed.

Tools are at present being designed at C.A.R.D.E. to produce cones to toler-

ances which are considered to be necessary fran trial results. The wall thickness will
be .095 + .005 but within this range, the wall thickness must not vary more than .0015
around aiiy circmnference. They must not vary more than .005 along the slant surface.
Hardness of the inside surface must be between 70 and 75 on the Rockwell l5T scale.
Runout of the cone flange with respect to the longitudinal axis must not exceed .002
as this locates the cone in the warhead. The temperature and time of the final
annealing treatment before ironing operations ha:ve not yet been decided; it is be-
lieved that a temperature between 5oo0Ji'. to 6oo<IF. for one hour will be suitable.
This should give a uniform grain size from cone to cone.

2.2 WARHEAD CASING

The warhead casing material is SAE 4130 steel heat treated to give 30-32 on
the Rockwell C Scale. It varies in wall thickness from .115 to .16o.

Two types of warhead have been used for static trials. The first t,ype (Fig.
4A) is 7 inches in length and has a parallel section of .16 inches at the front end.
However, in flight trials the warhead (Fig. 4B) was found to ha:ve a better ballistic
shape and was much easier to load into the launcher. The warhead with the parallel
front section had a tendency to jam during loading in the launcher unless the rocket
was very carefully aligned by the loader •
•
Static penetration trials were carried out with the warhead design as shown
in Fig. 4B, and it was found that penetration had improved about 15% although the H.E.
content had been reduced from 1 pound 6 ounces to 1 pound 1 ounce. It is believed
that this increase in penetration is due to the length between apex of the cone and
the exploder pellet approaching the optimum length for this warhead design. This
variable is being investigated b:y firing trials but the:y will not be completed in
time to affect the present design.

3. FILLING
The use of a low viscosity RDX/TNT 60/40 has eliminated !llllch of the trouble
from entrapPed air ·and shrinkage originally encountered when the standard filling was
used.

The following steps are used for the present method of filling: - '

(a) The warhead is filled to just below the required height of fUling '!fith
low viscosity RDX/TNT 6o/4D which is poured at a temperature of B5°C.

228

sa BBEiti £La
r $I Snrmtr
· •x •' ~
zd&n

229

CGIU lbi£4 BilL

 &&41 IS&! ILL 22222) 26222 JW!

(b) A special. riser is preheated to 9S°C and is set onto the neck o£ the war-
head, This riser is filled with H.E.

(c) A heat insulating cardboard cover is then placed over the riser and the
whole assembly is set aside to cool for 3/4 of an hour.

(d) As the filling cools, shrinkage cavities are prevented from forming by
the molten H,E. in the riser. When the filling has solidified the riser
is removed by inserting a bar through holes in the upper riser wall. and
giving a sharp twise. This breaks the filling and allows the riser to
be removed,

The special. riser consists of a hollow aluminum cylinder having a thin per-
forated base, The height of the filling is controlled by the depth to which the riser
is allowed to set into the warhead. By this method, the height of filling is controlled
by riser design and no scraping of the filling surface is required. The method has
given perfect fillings in 9S% of the warheads filled,

High and low viscosity RDX/TNT 60/40 have been canpared in penetration triaJ.s
but there was no difference in performance.

Trials have aJ.so been fired to compare SO/SO and 70/30 with the 60/40 RDX/TNT,
These triaJ.s have shown 60/40 to be the most suitable for the present design of warhead.
4. STAND-OFF

The static penetration trials have shown that the optimum stand-off for the
warhead is approximately 1.7S cone diameters, The baJ.listic cap and fuze provide a
stand-off of 2 cone diameters; with a fuze functioning time of between So to 60 micro-
seconds, the effective.stand-off after fuze functioning is very close to 1.7S.

S. FUZING

A modified American MS2 fuze for the rocket"is used,

The modifications consist of: -

(a) An all W&)"S mechanism, which will work up to 70° from nonnaJ., to replace
the original. striker.

(b) A spit back charge liner and housing to replace the booster magazine.

This fuze has functioned extremely well in all firings to date.

An electric fuze is also under development and it is hoped to hold the first
firing trials with this fuze within the next two or three weeks, A few varheads have
been fired statically with a mock-up of the electric fuze to check the functioning of
the detonation train and this has been satisfactory.

230 ,
emu IBm: mi.
66NZIBEN BXE sua s LMh

Cm!flWTJAl.
CARDE
'::'!oc3.J .J.1.

Figure 5-Test Fixture.

Figure 6-Target plate for Spit-back Fuze.

231
a t3NilibkN iiAOZseea: s z

-A
-·' '\
.....
<ll

~
E-<

0
~
00

I
t-
,..
<ll

~ .'
~

23 2
6. SHAPED CHARGE LINER FOR SPIT-BACK FUZE

A fixture was made to investigate the perfomance of the shaped charge used
in the spitback fUze (Fig. 5).
The fi,xture is arranged so that the charge liner can be aligned on a cross
scribed on the target plate of l/4 inch mild steel. The target plate is set at the
same distance from the spit-back pellet as the exploder pellet is in the actual round.
(Approximately 7-1/2 inches.) The efficiency of the charge is assessed by the damage
caused to the target plate and by the distance from the cross on the target plate to
the jet centre. (Fig. 6.)

From these trials, materials and shapes for the charge liner and housing,
and methods of filling the charge liner can be investigated. At present a charge
liner of aluminum with hemispherical shape fitted in an aluminum housing is used but
this may be changed i f other types prove more suitable.

It was noticed during these trials that the jet from the spit-back pellet did
not always strike the cross on the target plate. Under actual conditions, this would
result in off centre initiation of the shaped charge. The effect of this was investi-
gated in penetration trials by placing the detonator 1/4 inch off-centre and results
showed that penetration was decreased 30%.

To reduce this possibility of off-centre initiation it is proposed to use a

cone liner stem with tapered bore, the diameter of the bore at the exploder pellet to
be .25 to .3 inches. With this design, any mal-alignment of the spit-back jet is cor-
rected and the C.E. exploder pellet initiated on centre.

The guiding properties of the tapered bore have been checked with the fixture
used for trials with the small shaped charge liners. A small shaped charge for the
spit-back fUze was mounted in the fixture and aligned on the cross scribed on the tar-
get plate. A copper cone liner was then mounted between the shaped charge and the
target plate with the stem of this cone liner aligned on a point 1/4 inch and in later
trials 1/2 inch from the cross on the target plate. In both trials, the tapered stem
iid deflect the jet from the spit-back pellet.

SUMMARY OF TRIALS

The Heller warhead is 3.2 inches calibre and in static penetration trials has
given 13 inches penetration in homo armour plate.

In llight- trials against 21bmrn of armour at nomal angle of attack 8o% of the
rounds fired have defeated the armour.

In flight trials against 120mm of armour at 6o 0 to normal 8o% of the round

fired have been successful.

233
 CURRENT PROGRAM ON HEAT AND HEP ARTILLERY PROJECTU.ES

Lt. Col. R. E. R~le

Office, Chief of Ordnance, Washington, D. C•

.,
ABSTRACT

The current program for fin-stabilized HEAT artillery projec-

tiles is discussed. Status of development is given in each of the
various calibers. Principal development has been in 90nln and l05rmn
sizes, and developments in these calibers are being extended to
other sizes. Highlights of these developments are discussed, in-
cluding the electric i'uze, the effect of standoff and slow spin,
and the effect of cone variables on penetration. A definition of
K factor as employed in a recent tripartite conference is given.
A K factor of 4 appears descriptive of current HEAT rounds. Accu-
racy of fin-stabilized HEAT rounds is discussed. The accuracy
appears to be about .35 mil horizontal or vertical probable error,
compared to about .15 for spin-stabilized projectiles.

The current program for HEP shell is discussed. A niD1!ber of

designs have been released for production in 75nun, 76nun, 90mm and
lD5mm sizes. Principal development has been in 75mm and 105mm
., sizes. Most significant recent improvements are use of one-piece
type shell construction, and use of a new plastic filler. An
electric fUze is under development, which should i'urther improve
performance. Present HEP shell are shown to be capable of defeat-
ing about 1.3 calibers of plate, regardless of obliquity, at
velocities up to 26oO fps.

1. HEAT Projectiles

a. Introduction
The current program on HEAT fin-stabilized artillery projectiles will provide .
rounds for recoilless rifies in 57nun, 75mm, 90mm and l05mm sizes, for tank guns in 76nm,
90mm, and l20mm sizes, for howitzers in 75mm and 105mm sizes. Programs for the immedi-
ate fUture will cover such rounds for 105non and 155non tank guns and for llOnun howitzers.

The principal development efforts during the past two years have been on the
90mm Tl.OB round for tank guns, the lD5mm recoilless rifie rounds under the BAT program,
and the 105mm Tl31 round for the standard howitzers. From these programs, a number of
features which either improve accuracy or penetration, or which permit simpler fabrica-
tion hBVe been detemined, and are being applied across the board to other rounds wher-
ever possible.

r
• 235
CONFIDENTIAL Jr 4 0 II 12

.... - ...
Figure !-Sectioned round, T-108, E-15, 90 MM R&D TMl-1451.

c.....,yzery '!:!W.,, !t!" , ,..,, 44) 'H ' tf(t•

C"""&WLb_'!'p . . .A n] WP Cr.._e I ....

l'I.<LS<r. - .zq ~>

.s~ ,o<. - ·~c-.
~
~a.,L •-.e\mw• It~ oy
.
a·~ ,),;) BPYNQ
t!9,t ......,...,., W!!I!.+P .,. owq.,NK• Me.h ""'*"'"' p• fM4aSl!\.l ),,,.
a-... ~ ... ,..,.,.c.......,_e'"P-•....-•• •·~ · <.A,.~,.c.""""""' ...
.,. . .
,....,",.__ ..,..,..,...,...,
Ot\"'w
c-.•""'••
t ••••,.,.,.;• M"Mbh " ,.._. ' Of&'! ""+J••Otr "'"' awu.,a,:;,

Figure 2-Fiight test round, 105 MM HEAT TMl-1526.

 The 'Figures illustrate a few typical new HEAT rounds,

First, Figure 1 shows the fin-stabilized, 90rnm, Tl08 HEAJ round for tank guns,
Note the long boCIII extending back into the case, the guide rails in the case, the shot-
pull device, Several thousand of these spit-back. fll.zed 90mm Tl08 rounds are now in
field service stocks. Production is currently being switched over to electric fused
rounds, Next, in Figure 2 is the :J,P5mm Tl31 FS HEAJ round for howitzers, Note the
boom extending beyond the case, Production' of the lOSmm Tl31 round for howitzers wi11
begin in December at the rate of SOOO per month, No BAT rounds have yet been released
for procurement, but substantial R&D quantities are on order, including SOOO lOSmm Tl84
rounds and 3000 T118 rounds which are similar in general to the 90nn Tl08 round, and -
1000 lOSmm 'n3..8. potato-masher shaped rounds for the Firestone BAT weapon. The latter
is the only HEJrir l'Qll.nd of current development be:!,.ng procured in substantial R&D
quantities, which is not fin-stabilized, Figure 3 is a drawing of this shell, The
Tl38 spins slowly at about 2_s-rps,a.nir:i.s stabilized through its unusual nose shape.

b, Penetration of HEAT rounds

(1} Fuzing

Under the 90mm Tl08 program, an electric fuze has been developed lolhich, due
to its fast and more reliable action, is replacing the spit-back nose fuze. The time of
functioning, from first nose contact to jet reaching the plate i!l of the order of CiJ
microseconds compared to about 120 microseconds for the spit-back fuze, The electric
fuze has the added advantage of simplicity (the fuze being essentially a crystal in the
nose which generates its own current on impact}, the ability to function properly up to
71 degrees, and by use of inner and outer ogives, a more difficult fuze to defeat by
such devices as spikes and angle irons,

(2} Stand-off

Tests in 90mm and lOSmm sizes have indicated optimum stand-off occurs at
4 to S cone-diameters, Built in stand-ott for most of the current rounds is 2 to 3
cone diameters, and in this region, about 3/411 in penetration is gained for every inch
increase in stand-off for 90Dm and lOSnun rounds. Figure 4 is a curve showing penetration
vs, stand-off for lDSmm BAT rounds. The lower curve here illustrate_s the· fact that
cones with a simple conical apex have less penetration than with a spit-back tube, This
result has been obtained in both 90mm and lOSIIDU sizes, The first electric-fll.zed 90mm
Tl08 round employed a cone located deeper in the shell body at about 2,8 cone diameters
stand-off, instead of ·2,4 as used in the spit-back type, The penetration increased from
about 911 to about 12 11 for this 311 cone wen electric fuzed, Of this 3" improvement,
about 2 11 can be attributed to increased stand-off due to faster action of the electric
fuze and about 1 11 to the increased built-in stand-off. Recent static 3.5" rocket tests
have shown that the explosive behind the cone can be. considerably reduced without
significantly degrading penetration where constant stand-off was used, indicating cones
should be as far back in the shell as practical.

237

s er
I
-U.CIT'I! I'LU5 ON& RI!OU
~
v E.1
E--t

Go/'0 5""- BEFIOR£ ~ ....,.ER Ulo.J(o .....

SCJ'IlE. : & -o
•g~ "•~;, BUC!!i L AI'TJft TliHTtNINS ~
SCALZ: 2-1

Figure 3-Assembly T-138 Projectile for initial trial firing.

 STANDOFF VS PENETRATION
'-
-....
w
w
IOS1111 T 138 COIE-.100 IICH WALL
t; 25
....a
!- 20 ~
......
............... LIN ~ WITH SPITBAC TU

,~ "'
~-""'"

I ~
•w
:I:

-•:=
15

10
UNERS-51
""
~PLE APEX)- ~ ....
I

0
5
~
•....w
..
•w 0
10 20 30
STANDOFF INCHES
Figure 4--Standolf vs. Penetration, 105 MM T-138 Cone-.100 inch wall.
40 50

SPIN RATE VS PENETRATION

IOS11M Tl38 COllE -.100 IIICH WALL
::i
w 20
w 71tz INCH STANDOFF

=u ~
t;
11:11

-
a
0
w
~10
•-I
'\. ~
Q•
~ 5
•....w
•w
.. 0
50 100 150 200
SPIN RATE REVS/SEC
r ·. Figure 5-Spin rate vs. Penetration, 106 MM T-138 Con&----.100 inch wall.

239

a:;;:&
 ACCURACY FIRINGS OF SHELL
HEAT-FS 90MM TI08
EFFECT OF CONE VARIABLES OF 90mm HEAT
100
(2800 FPS - 1000 IDS)
j"''lll ' 10171''
~l ~ 115-& c:oME DIO TRAtr.NERSE Ell STD

~1!1..
VARIATION
90 "RS-5 CONE .010 WAU. WAVINE!iS 100'1 Ell STD

~~ SUPERIMPOSED ON T34 TANK

80 ll
,
"
RS·6 I ,

70
REFERhcr, ACWMTE CONES Cl8 ROUNDS,

~..,' •..1... ~.J .. '...J

• ••
I 0

0
"" RS-5 DIAMET£RS STAND OFF

• •
0

30 l
' \
~ •
• •
•
• ••
~

I' ~ ••
' ,_
20

10 ~
!'r-..
0
B 9 I II 12 13 14 13 16 17 18
PENETRATION INCI£S 10FT - - - - - - -

Figure 6-Elfect of cone variables

of 90 MM HEAT. Figure 7-Aecuraey firings of shell
HEAT-FS 90 MM T-108.
 (3) Rotation

Current fin-stabilized HEAT rounds rotate about 10 rps, this spin having
been found necessary for best accUracy. The potato-masher shaped, l05mm Tl38 round
previously shown, rotates about 25 rps. Fig11re 5 is a curve showing the effect of spin
on penetration as determined for 105mm projectiles. It is noted that about 1 11 pene-
tration is lost for 10 rps, 3" for 25 rps, and for 200 rps (as encountered in the M67 V
105mm standard rotated HEAT rounds), penetration drops from 1811 to 6 11 • The scaling
law m~ be applied to these results qy plotting penetration in cone diameters vs. V/N
(instead of RPS .. v/ND), but this scaling law is not accurate for moderate changes in
scale. V is muzzle velocity, N is twist of rifling in calibers per turn.

(4) Cone Variables

A number of cone variables have been examined in 90mm and 105mm sizes,
such as eccentricity, waviness, wall thickness, cone angle and other factors. Figure 6
is a curve showing the results of studies of cone variables of the 90mm Tl08 round.
The data is plotted in terms of percent of rounds fired versus penetration in inches.
For the 88 reference, accurately made cones, it is seen that 50% defeat 13-1/2 11 , 90%
defeat just under 12" 1 at 2 cone-diameters stand-off, against al'l!lor plate. The. lower
curves show that 2'1 to 3" penetration is lost i f wall thickness in a transverse plane
varies .oo5, or i f the exterior cone elements are sinusoidal waves of .010 amplitude.
Similar curves, not shown here, indicated that other factors investigated did not
appreciably degrade penetration, such as having thicker waJ.ls at top than bottom of
cone, or having ths shell cavity as forged, which resulted in a certain amount of
ovality. other tests in 105mm size have indicated that provision of a wide flat at
the base of the cone is harmful, that the cone angle should be about 45° 1 and that the
cone wall thiclp]ess should be about 3% of cone diameter. In the 105mm Tl31, various
cone thicknesses have been tried, and penetration was 311 better with .100 inch waJ.ls
than with .o65 inch walls.
'
(5) K Factor
In a recent tripartite tank gun conference it was agreed that the K
factor should be defined as that depth into homogeneous armor measured in cone diameters
to which 90% of the rounds will penetrate, and that a K factor of 4 on this basis /
.closezy describes actual performance of current HEAT (FS) rounds. This does not allow
for overmatch required to give full lethality, which appears to be about 2 inches,
based on both u.s. and U.K. tests. Incidentally cone diameter is taken as the ineide
diameter at the base of the cone.
c. Accuracy

Spin-stabilized shell average ab®t .15 mil horizontal am vertical probable

error. A primary f!Cctor in development of fin-stabilized HEAT rounds has been the
problem of achieving acceptable accuracy. It has been necessary to use high-alloy
aluminum fine and booms, close tolerances, unshrouded tails and a slow spin (of the
order of 10 rps) to achieve the current accuracy of about .35 mil horizontal and
vertical probable error. Figure 1 shows a chart of the hits of a typical group of 90mm
noB rounds fired at 2800 .fps against a vertical target at 1000 yds. The silhouette of

241
a Russian T34 tank is shown to provide a sense of the relative· scale. A typical spin-
stabilized projectile would show a little under half the spread obtained here but it
is observed that all of the shots fall within the silhouette of the tank. A total of
239 TlOB rounds fired have averaged .35 mil. other tests mvolving shrouded f'inB,
uncemented joints, and relaxed tolerances have shown poorer accura~, and therefore
such designs have been discarded. The first production lots averaged .42 mil. The
latest production lot from. a new marrufacturer gave .20 mil for a 9 round group. Re- ·
coilless rifle HEAT rounds, incJnding the potato-masher type Tl38, have average about
1/3 mil probable errcr. Smooth-bore firings of non-spinning rounds gave poor accura~.
If fins were canted enough to achieve accura~, spin built up to over 50 rps at 1000
ytls, reducing penetration. Hence, best results are achieved by obtaining about 10 rps
at the muzzle by simply making use of friction bet"~."een the projectile body and rifling
within the bore. Use of .full twist and depth of rifling appears satisfactory, ani has
the advantage of permitting the firing of cheaper, more accurate spin-stabilized HE and
WP rounds. Loss of gases through the rifling around the projectile has little effect
on tube life or muzzle velocity.

2. HEP Projectiles

a. Introduction

The HEP round is another annor defeating type which does not depend on kinetic
energy to defeat the plate. The plate defeating ability of such a round depends on the
use of a soft thin-walled nose containing a plastic explosive which is deformed on
striking plate, to the general shape of the plate. In Figure 8 we have a diagram,
somewhat exagerated, showing its principle of operation in spalling the back side of
the plate by means of a shock wave. A number of individuals have written papers covering
the nature of the wave, and the distribution of pressure within the wave.- Essentially,
a ccmpression wave is reflected as a tension _wave strong enough to cause tensile failure.

Over 100,000 105111111 HEP shell have been produced. The basic shell was developed
for howitzers; by splining the rotating band, it has been adapted to the standard 105mm
recoilless rifle. HEP shell designs have also been released for procurement in 75mm
size for rifle, howitzer and tank gun, in 76mm size for tank guns. other development
type llEP shell range in caliber from 57111111 to 165111111.

If' the nose of a HEP shell strikes the plate at too high a velocity-1 shock deto-
nation occurs, and the resulting explosion occurs before sufficient explosive is in
contact with the plate, and the explosive wave moves away from the plate. Failure to
spall is the net result. The HEP shell designer attempts to get a high-capacity shell
by using walls as thin as set-back forces and stability calculations will permit. He
chooses a powerful, plastic explosive, with as little impact sensitivit,y to shock as
possible. He tries to select a base fuze which will set off the explosive wave after
just the right amount of defonnation, or squash. The delay- time required is a function
of striking velocity, plate obliquity and caliber. A new electric fuze utilizing a
probe extending up toward the shell nose is now in the planning stage, which, i.f
successful, should pennit more nearly ideal fuze action. The length of probe will be
adjusted to give optimum crush-up.

242
 .,

ARMOR PLATE DEFEATED BY HEP SHELL

( HIQh Charpy, Hlp Obllqultr Plate)

ACT lOll 10

/
'

SQUASH HEAD TILE 9

SLO p ·~/d•l.3 J
v
8
•• ••
DE FOR PLATE )! !• 0
!

SHELL 7
/ v
;;; 6
/
\ v v'j
1 . .- .

LIMIT
~
..
0

!<i 0
..
~
'\
c
'
/ --
DIRECT 1'5
"'
4
/
0" ASSURED
SUCCESS

I v v
!1i
SHOCK
WAVE
3
v
2
/
I E E E
.,...E E
0
0>
E
2

// ~BACKSPALLS
~. 0
2 3 4 0 8 7 8

~ CALIBER (INS.) (d)

Figure -Armor plate defeated

Figure 8-Action of squashhead projeetile. by HEP shell.
 :"'
i
~

I
...
0

..
4
f
,;
4
••
~

•r !:
""
..
T
~

2 b
 b. Plate Defeating Ability

Considerable experience has been obtained in four sizes, 15lDI11; 90mm, 105mm and
120mm. The performance of these rounds are plotted in Figure 9 against ability to
defeat plate at high obliquity. The points are plotted as one-inch bands, based, for
example, on the ability of the 10511111 HEP to defeat 511 plate but not 6" plate. Thus,
the perfOl'lllallce is bracketed, and the line represents expected average perf'onnance.
However, the lOJoJer point for each caliber has been used in the chart since this is a
known or demonstrated capability. It is noted that in 751D111 size, the line runs close
to the 4-inch point. A few 75mm rounds have actually defeated 4-inch plate, but at
only one velocity, 1700 fps. Also, in 105mm size early firings against very low charpy
611 plate were successfUl. When it was learned that plate toughness, or charpy value,
strongly influenced results, all firings subsequently have been against tough, high
charpy plate. The charpy test measures toughness by the ft. lbs. of energy required
to break a notched bar. Low charpy is considered 5 to 15 ft. lbs., high charpy above
35 ft. lbs. This plot indicates tbe HEP round to be capable of defeating about 1.3
calibers of tough, high-obliquity plate, regardless of caliber. It is not known i f
the slope of the· curve is actually a constant, but :f'i'om the firings to date it is not
far from constant for the range 75mm to 120mm. For very large . calibers, one would
expect the curve to droop, or for HEP to become less effective. However, the line is
fairly close to the plate thickness of 8-inches defeated when Aberdeen Proving Ground
fired UK 6.511 rounds.

There have been three principal methods of fabricating a soft-nosed HEP shell.
(1) The UK HEP shell used butt welding. (2) The US was wary of production diffi-
culties of welding1 and first used an overlap, brazed joint to produce the two-piece
HEP shell. (3) More recently1 production methods have been asveloped to produce a
one-piece HEP shell with inner walls gradualJ;y tapering outward toward the front, then
closing in to the nose with no joint. In Figure 10 are cross-section drawings of the
105mm one and two-piece designs. The technique of making the one-piece type involves
a series of drawing operations followed by a final nosing and spinning operation to
close in the nose. The method is cheap, and many private facilities have the type
presses and other machines required. The advantage of the one-piece type shell has
been that the walls can be gradually thickened toward the rear, being only thick enough
to withstand set-back forces, which permit maximum nose capacity and insures a better
squashing action of the shell on striking plate. The principal effect of the one-piece
·design on plate performance has been to raise the effective velocity range, permitting
better hit probability. Aey HEP shell has an upper and lower striking velocity limit
of effectiveness, the upper limit set by wall strength, and deflagration on short range
impact, the lower limit is set by the walls being so. strong that proper crush-up, or
squash, does not occur. For example, the latest 105DBD HEP shell far howitzers has an
effective striking velocity range from 2100 down to 1400 fps 1 compared to 1700 down to
900 fps for the original two-piece HEP shell. HEP shell for tank gims now appears to
have the upper effective limit near 2600 fps. Up to recently the HEP shell filler has
been the new plastic explosive, Camp C-4. Tests now indicate certain superiar features
of Comp A-3, an explosive which is press-loaded (in granular fom) to produce a firm,
solid filler, but which nowe plastically on impact. Use of A3 has pemitted the
velocities to go up as high as 2600 fps vithont shock detonation, and has permitted de-
feat of plate at norinal impact of thickness equal to that defeated at high angles of
obliquity. For eXample, with 105IIDII shell loaded with C4, the shell defeats 511 oblique
plate and 3" norinal plate, but with A3, 58 plate is defeated at any angle of attack.

'.

......
I
I

I
-···

.. . ....
. ~ ..· ;.
 §£Bfi §p ·p 7

Nose shape has been varied, between hemispherical and ogival with no significant
difference noted. The o~val nose was chosen because of reduced drag. It is con-
sidered that future improvement of HEP shell performance is more likely to come frau
use of the electric fuze previously mentioned, than from further changes in shell
geometry,

c. Accuracy

In general the average rotated shell, whether HE or HEP, has an acruracy of

about .15 mils HPE or VPE. The HEP shell differs from the HE shell in that the walls
are thinner, resulting in lower stability factors but results to date indicate accuracy
at 1000 yds. of HEP shell is about the same as for HE shell. In the fabrication of' HEP
shell, special attention has been paid to keeping wall eccentricities to a minimum. At
long ranges, there is some evidence that HEP shell accuracy is less than that of' other
spin-stabilized shell. The actual acruracy figures for tank guns range from .12 to .16
mil HPE and VPE, based on firing results in 75mm, 76mm and 90mm tank guns. Accuracy of
HEP shell fired from howitzers has averaged about .2 mil HPE or VPE, and from recoiless
rifles about .25 mil HPE or VPE.

246
 I
I I
I:
DEVEIDPMENT OF SHOULDER FlRED SHAPED CHARGE RbCKET HEADS

H. s. Weintraub
s. Fleischnick
I. B. Gluclanan

Picatinqy Arsenal, New Jersey

ABSTRACT

In order to evaluate the design features to be incorporated in-

to a lightweight T205 Head, this arsenal conducted a shaped charge
investigation by varying different parameters in the standard 3.5°
M28A2 Head. Results obtained through this investigation and perti-
nent to designing the prototype T205 Head were as follows:

a. The opt:l.mwn cone thickness for the 3.5" M28A2 Rocket

is o075" t .003°.
b. The HE charge canbe reduced to 1.3 lbs. without
affecting penetration, ;: , - · ,.
(
c. The optimum standoff distance for the 3.5" M28A2
Rocket is 9 11 •

d. The booster pellet ma,y be placed as close to or atop

., the apex of ·the copper cone without aey significant difference in
penetration being noted,
Prelllllinary tests with a. modified 3.5 11 M28 type rocket head
which eliminated the nat at the base of the cone (the M28A2 Head
has a flat of approx:imate],y .1" at the cone base) indicated that an
increase in penetration of approximately 2 inches might be expected,

Studies to improve the perfomance of the 2.36 11 T59E3 Rocket

Head (1) bY the use of explosives having higher rate of detonation
than Composition B, such as 70/30 Qyclotol and 75/25 Cyclotol, (2)
by sh~~Joling the detonation wave front traveling through the charge
and (3J bY use of a tandem cone arrangement., have resulted in no
marked improvement ·in penetration to date •
•
Investigations are being conducted with the T2016 and T2017
Rocket Heads to determine whether increased standoff and corre-
sponding],y decreased explosive charge (overall length of head
being COilllltant) would result in inci'eased penetration.

247
 Dl'TR.ODUCTION

This paper is divided into four parts. The first and major part covers the develop-
ment of a 3.511 T205 lightweight Rocket Head. The second part deals briefly with the
development of a 2 .36" T59E3 Rocket Head. The third part gives a .very brief description
of the development of the 2. 75 11 T2016 and T2017 Rocket Heads. Although the T2016 and
T2017 Rocket Heads are not shoulder fired, they have been included in this repor~ be-
cause the design problems are closely allied to those encountered in the development of
the 3.5" T205 Rocket Head. The fourth part gives a very brief description of an inde-
pendent investigation conducted for the 105mm T43 (M324) HEAT Shell, The results of
this investigation are now being reviewed to determine the feasibility of incorporating
the data :Into a 3.5 11 Rocket Head design.

DEVELOPMi>NT OF A 3.5" T205 LIGHTWEIGHT HEAT ROCKET HEAD .

BY: H. S, WEINTRAUB

1. A conference, which was attended by representative·s of ORDTU, Redstone Arsenal

Division of the .RolJm and Haas Company, Ordnance Rocket Center, Redstone Arsenal,
Aberdeen Proving Grrund and Picatinny Arsenal, was held at this Arsenal in April 1951.
The object of this conference was to establish a high priority program for the develop-
ment of a 3.5" Rocket having greater accurac;y-, range, penetration and velocity than the
now Standard 3.5" M28A2 Rocket (Figs. Ia, Ib). The new rocket is designated as the 3.5"
T205 Rocket.

2. The respons:lbilities for the development of the 3.5" T205 Rocket were fixed as
follows:

a. Picatincy Arsenal -
(1) Coordination and technical supervision.

(2) Design and development of a lightweight HEAT Head for the T205 Rocket.

(3) Design am development of an alternate T205 Motor design, employing

trap-type construction.

b. Rohm and Haas -

Design and development of the T205 Motor, employing head-end suspension.

c. National Bureau of Standards -

Design and development of the Piezo Electric TYPe Fuze for the T205 Rocket.

3. The characteristics to be achieved by the T205 Rocket, as put forth by OllDTU,

especially those pertinent to the head design are listed as follows:

248
 Figure !a- Rocket, HEAT, 3.5", M28A2.

ROCKET RE AT 3.5-lN M2BA2

.......... , ...
TENP u..i=eo'r'ro ,2~D-F-~"' I
I l 1
I '
I

I
« TAIL tJ/.tfAI.(INf
A • ia#N~ ,., ~~~
I · ~Wn (_,...,_,.D W . . .U/IIMI£1J

....... n 1 ...- ?11r( W t'flll l• o. U IOC ,..t' ,H ...: I oot l )IIO(C ....,,. C ' '
" ( f r .. .....

tJI't c,...,...,..,..,...
"' ' """' ""'"" · • , , . ..... -u .. •Tif'"" r~r ~ r s -·~ .....,.,
•• u ~~t ~r.o .,. , " "'lll""'"

t.!

"

. .=? , . - ~#,_..,,_,/711(~ .....-........

,.,...~=
---- .n.SI~

...,...._r .1 ,.""'<
I/IC,.... f •

~
,..,_.,, _...,.IIW ' """I M #" l'fl ' llrC
.oiiCII'kii~ . . . ..-N. . .#.IIT~

~~----·~ diiU~

, ;:;;;~~,y-I'C~ _ ..,,_..,....-.; ..,..,

I .... k #I"'CC~ ~ .,~ .. ~eJ J"ur,.·~~~ lA~ ~
.,.,,Ju.l _,_. . I f

Figure ! b-Rocket, HEAT, 3.5", M28A2, Assembly.

2L9
 a. Penetration to be achieved by 90% of the rockets:

(1) 6•• penetration of amor plate at 60° obliquity is required.

(2) 5" penetration of amor plate at 70° obliquity is desired,

b, Lethal Fragmentation:

Fragmentation equal to that obtained with the M28A2 Heads is desired,

c. Accuracy:

(1) 80% probability of the first round hitting a target 7.5 1 x 7 .5• at 500
yards is desired, with the rocket having a probable error of 1 mil.

(2) The rocket has to exhibit equal or better aerodynamic stability when
compared to the. M28A2 Rocket.

d, Initial Velocity:

500 ft. per second is required,

e, Weight:

(1) A weight of 7,5 pounds for the complete T205 Rocket is required,

(2) A weight of 7 pounds is desired i f possible of attainment.

f. Length:

A maximum overall length of 23.55 11 (same as M28A2) is required, In addition,

this round must be able to be fired from the automatic launcher as well as the M20
shoulder-fired launcher.

g. Fuze:

The PI, BD, '1'2030 Electric Fuze is required,

4. The Rohm and Haas Company- selected for their motor development a design (Fig.
II) that had the same metal parts weight (3.3 lbs.) as the M28A2 Rocket Motor, This
motor, however, contains approximately 30% more propellant than the M28A2 Motor. The
National Bureau of Standards has under design a fuze that weighs ,86 lb, less than the
M404Al Fuze of the M28A2 Rocket. The increase in weight of propellant and the decrease
in weight of the fuze did not meet the velocity requirement· of 500 fps and the maximum
weight requirement of 7.5 lbs. for the '1'205 Rocket, Thus, it became necessary to pro-
vide a lliUCh lighter head than that of the M28A2 Rocket, To meet these requirements, a
head weight of approximately 3.5 lbs. or 25% less than that of the M28A2 Rocket was
required (Fig. III). Thus, the main problem for this Arsenal was to design a lliUCh
lighter 3.5" HEAT Head that would have better performance characteristics with respect
to penetration than those of the 3.5" M28A2 HEAT Head,

250

a
I
, :HIC
 .....
.....

251
 ---~•.

ri
•

:' --·-
'
.
I j .•·. ,;.-

•
l
•

252
 ...,J/t

Figure IV-Head, Rocket, HEAT, 3.5", Figure V- Rocket, HEAT, 3.5", T205,
T205 Pc Mk P-83152A Assembled with design No.6 as shown on Dwg. PX-8-720
tetryl pellet (See photograph M-39684) assembly with booster detonator assembly
Prior to assembly with blasting cap. & DuPont No.6 Blasting Cap.
Plates are hot-rolled (RHN B-70)
FS 1015 Spec QQ-S-633.
 ~------~n

Figure VI-3.5" M28Al Rocket head of ammunition Lot PA-40-12.

Figure VII-Head, Rocket, 3.5" modified T205 head to contain

1.01 pound H.E. as shown on piece marked P-82795F.

254

GIL! EELS! hilL

 \j)\JI ~.2 ••
5. Development of the lightweight head was investigated by conducting static
penetration tests for a variety of modified heads, The following is a list of the more
important variables that were investigated!

a, Reductions in the weight of the HE charge,

b. Changes in the body contour to detennine the best distribution for the re-
duced HE charge,

c. Reductions in cane thickness to save weight.

4. Decreases in the distance between the apex of the cone and the booster
pellet to szve weight without sacrificing penetration performance,

e. Changes in standoff distance to increase penetration,

f. Reductions in thickness of head wall to save weight,

g, Changes in ogive shape to obtain more stable fiight characteristics,

6. In order to expedite the test program, most of the initial static firings were
made with modified M28A2 Heads, The firings were conducted by placing the modified
heads perpendicularly on top of a stack of l-inch mild steel plates 16 to 20 inches
high. As shown in Fig, IV, the stack was usually arranged with three 5" x 5" x 1"
steel plates at the top and 411 x 4" x 1" plates for the remaining height. These plates
were of FS 1015 hot rolled steel, in accordance with Specification QQ-5-633, and had an
zverage Brinell hardness of lhO (Rockwell B-70). Detonation was accomplished so as to
simulate, wherever possible, the conditions that take place when the head is assembled
to the rocket. To accomplish this, the M28A2 Heads were asse~~~bled with booster deto-
nator assemblies, and the detonation of the heads initiated by a duPont No, 6 blasting
cap, Although this method of obtaining the static penetration data does not exact~
duplicate firing against homogeneous armor plate, it does afford a good evaluation of
the perfonnance of a design relative to other designs tested in the same manner. In
addition, it has been shown that the penetration of armor plate is approximately 60%
of the value obtained with mild eteel.plates,
7, The following tentative conclnsions may be drawn from the testsa

a, The reduction in the weight of the HE charge from 1.9 lbs. (Fig, VI) to
1.4 lbs. and 1.01 lbs, (Fig. VII) does not indicate aey significant difference in
penetration, These heads were fired at a standoff of 4,2n, vhich is the standoff ob-
tained by static firing of a 3.5n M28A2 Head,
b, The body contour (Fig. VIII) may be modified within limits without adverse
effect, provided that the portion of the head body extending f'rom the base to the apex
of the cone is approximately cylindrical and maintains a diameter approximately equal to
the base of the cone, The portion of the head 'bocl7 for the M28A2 and '1'205 Heads, which

255

~i{i,i_SLLUL
 ;

.' " ' ....

. ' ,., '" t
• t

....... '··~ ·

c:=
:iii!: ____ .., :ill:
,......
C"':)

.. f
~· W•

...,. ,....
C":»
~ 1- -
~ ..,.\ - "'
:Ia
f\)
C/.)
~
'\:· \ C'-)
\Jl.
(]'\
- l -+ __...._ __ _
-+- ~
--
...,.,
/ -~
... IT\
, s ,.,,
1=2Q5 ttMR • " * : ( -.~,.,

-=

~ [~$~~-(
.
..,..a.,.tr

Figure VIII-Rocket, HEAT, 3.5", T-205, Shaped Charge, head design study.
 COl~
llNCLASSiflfU
....~~....--.~~'-sm.sit~·

&....lrtJIUI
l ·582
!-->
..
•

5
.5'1

-·
......
·"'
·295

- - .•.,.
CON£ NALL TN/(KNESS (INCN£SI
.... . .. ... 7-lJ-51
.... 7-10-Sl
7 J!ll, 1951
JILl· I I -
-
Figure IX~Penetration vs. cone thickness for the 3.5" M28A2 head.

FUtT?.J.rtOIII Vi
filii til£ STllrll.iltl ),5"
lf2!M.2 fiOCi\!T

Mo. l'o:md
?olr:t .........f1.E:1

~·,~----~r------r-----1-----6·1-~~~--~--+------+--~3 ,;
lS
1:~
,)9
4 lS .71
s lS .ss
-6 lS .60
~~,~----~r-----1-~~~------+------+------+------+--: tl
• l5
lD ll
ll l5
1.2 . 13
13 l5

0
'·· .... 6.0
STANDOFF (INCHES)
'U ...
~1gure X-Penetration vs. standoff for the standard 3.5" M28A2 rocket.

~JNCf.5ASSJfJED
/

- ~-5-131 B OR
82-5-131 H

DISTANCE FROM
UEAD "TYPE APPROXII\IAIE 001'£ Al-EX TO BASE
AECE MARK WHG1. CF H.E". Of" OOOSlfR CAYil'l' COM~RISON Ck \£AD CON10UR5
82-5-131 1.83 2.24
P-82911 N03 1.55 l. 'T I
P-8281 l N03A 1.35 :c.4
F~gure Xlla-Comparison of head contours.

r.

~ 2
0
~

~
~
m
~
F Xllb

0 .zs .50 7$ /.() 1.25 /.S'O I. u 2.0 2../!S zs

otsrANce <2Faoo.s~ PELLET l""lt'()M c.:w.e .APe;f O~.o!\f)
Figur~ XIII>-Pene:~~tf!)(i~ttit~ from cone apex --
25R
szmxt
SWWQfjri 01
£

259

CbitPIPEWSX 11
 19.U- . ~ · - -- - _ _ ..,

--- 1

- - I :: ;>;t .~ ·

1
-- · - JIMR -·---
-.t
.CtOWI/JC
l
'1(.0 'J()~£LP
\ ~
'-1.
,,
\ ....
GROO~ ', -:~ I

\_...1EE VIEW A - .J. •I.!Z -.ot:r

- SPHERICAL
j ·

______
____ _____ , .,....
--OITIII--LAa1ftUa,

'
I. L C., S(CTIOIII 1tl 4110 1U Till , . - .
- 011 Till MVII.A~ Of ITS COIIT'IInt • M•
FII.DE - TO •• u•uur••o•uu o •t.- ot - ' " '
-~ nuw

Figure XIV-Rocket, HEAT, 3.5", T205, Head design #4.

Wind tunnel model, dimensioning.

13.2.4 -:02.
FOR FURTHER INFDRNATION
SEE. VIEW A7 SEE DRAWING PX 8-723
- 4.87 ---·~1 ~o.--;-- 64'2.;02 I
-11'1
I /

2.1R
I
i
D

+If)
(I')

SPt-ERICAL
.04+ .02

.20 R - -
BLOCK NO.7 PX· 8-76S A
VIQ/1/ A
SCALE1 +
Figure XV-Block No. 7.

260

II EFDJW!t¥
 z3 7 · ·· k'snr r ..
runs for the height of the cone, should have a diameter of approximately 3 .5 inches.
The copper cones of the heads have an apex angle of 42° and a diameter of approx:lmately
3.5 inches. ·
c. In the M28A2 Rocket, a maximum static penetration occurs with a cone ·
thickness of .075 inch as compared to a thickness of .093 inch for the M28A2 Cone (Fig.
IX). This result was obtained by firings conducted with groups of M28A2 Heads. Each
group of heads was assembled with different cones having the following thicknesses:

.o5o"t .003"
.o6o 11 t .003"
o070 11i, o003"
.080 11 t. .003 11
.09311 t. .003"
d. The static penetration of the mild steel plates by the 3.511 '1'205 Rocket
increases almost linearally with standoff (Fig. X), until a max:!Jmlm penetration of
about 18 inches occurs at about a 911 standoff. The static penetration of the M28A2
Rockets at full ogive standoff (4.2 11 ) is about 14.5 inches.
e. A head wall thickness of .050 11 may be used instead of the .090° thickness
of the M28A2 Head, bu.t with no significant difference in static penetration. Since the
use of a thinner wall head and a smaller HE Charge will undoubtedly affect the frag-
mentation qualities of the head, this problem is now under investigation.

f. Reduction of the distance between the booster pellet and cone apex in the
M28A2 Rocket does not seem to reduce the static penetration of the head (Figs. XI!a
and XIIb). The tests leading to this conclusion were conducted with groups of M28A2
Heads, each group of which had the booster pellet placed at a different distance from
the apex of the cone. Ths distance ranged from 0 inch to 2 .2 inches (M2 8A2) in in-
crements of 1/4 inch. This was accomplished by increasing the depth of the bore of the
HE cavit,r into which the booster pellet was inserted.

B. The above work indicates that acceptable penetration characteristics can proba-
bly be achieved with a head of MUCh less weight than the M28A2 Head. The weight re-
duction, together with the necessary head contour changes, moved the center of gravity
of the '1'205 Rocket to the rear (closer to the center of pressure). This change in ·
center of gravity was sufficient to decrease the flight stability of the rocket when
an ogive of standard contour is used. It was, therefore, found necessary to conduct
an aerodynamic investigation of ogives having different contours, so as to develop an
ogive shape that would improve the fiight stability without adversely affecting pene-
tration. Wind tunnel teets showed that the head designs Nos. 3, 4, 7 and 8 (Figs. XIII,
IIV, IV and XVI) had satisfactory fiight stability characteristics. However, head de-
sign No. 4 gave unsatisfacto:cy penetration because of interference of the ogive with the
jet. As a result, an ogive design having longitudinal clearance at the base of the cone
to avoid interference with cone collapse and jet penetration was then designed. This
design gave satisfactory results when tested for penetration (Fig. XVII). Aerodynamic
tests for difi'erent type head contours are still being conducted at NACA.

261

?? ana
 FOR FLRTHER 11f'"0~1lotl
~------------1a24~oe------------~
SEE VEIN A SEE OAA\'l ltiG PX &· 724
~---~+~+02--~~

L__ 131~01 I=
1---4-.tfl----'"1
.......
I
I

Sf'tE~L

BLOCK NO 8 PX-e-767A

Figure XVI-Bloek No. 8.

[)£SIGN NQ8

vET CLEARANCE

Figure XVII-Jet interference in T-206 heed expressed on basis of

two dimensional cone collapse.

262
Figure XVIII-Head, rocket, H E AT, 3.5" T205 Pc Mk P-83152A with 1.45"xlh" tetryl
pellet and #6 Du Pont blasting cap prior to assembly for static penetration test firing.
See photograph #M-39685.

Figure XIX-Rocket, HEAT, 3.5", T205, Design No.3 with propellant charge & igniter.

263
 ,. -,__.,, ~- .

9. Using the data obtained in the various tests, the lightl,reight prototype '1'205
Head design No. 3 (Figs. XVIIT and XIII) has been established. This head design
differs from the M28A2 in the following manner:

a. The HE content is reduced to 1.4 lbs.

b. The cone thickness is reduced to .o6on.

c. The distance between the booster pellet and the cone apex is reduced to
1.711 •
d. The outside contour including the ogive is changed to accomodate the re-
duced charge and to provide satisfactory flight stability.

e. The total. weight has been reduced .to 3.5 lbs. The '1'205 Heads have been
manufactured and are about to undergo static firing tests. Flif!)lt tests of the '1'205
Rocket (Fig. XIX) employing the Picatinny Arsenal head are scheduled to be conducted in
the near future. It is expected that the characteristics of the rocket will be a~?
follows:
(1) Length - 23.55" max.
(2) Weight- 7.2 lbs.

(3) Velocity - 500 ft. per second.

(4) Maximum range - 1400 yards.

(S) Expected penetration - 14-inch hanogeneous armor plate.

DEVELOPMENT OF A 2.36•• TS9E3 HEAT ROCKET HEAD

Br: s. FLEISCHNICK AND I. B. GLUCKMAN

1. The 2.3611 T59E3 Rocket (Fig. 1) is a completely redesigned rocket intended to

replace the 2.36n M6A6 Rocket. The T59E3 Rocket weighs 6.5 lbs., has a velocity of 420
ft. per second at 7oor, and has a penetration of approximately 8 inches of hanogeneous
armor plate.

2. Due to the occurrence of some erratic penetration performance of the T59E3

Head, work was initiated and is still being conducted at this Arsenal to improve both
the depth and the unii'o:mity of armor penetration. The investigations conducted and
those still under investigation include the following:

a. Tests conducted with explosives having higher rates of detonation than

Composition B, such as 70/30 Qyclotol and 75/25 Cyclotol, have not indicated that acy
imprOV'8JIIent in the penetration performance of the T59E3 Rocket Head can be expected.

264

a '32
Figure 1- Rocket, HEAT, 2.36", T59E3 with fuze, rocket, T2000 El (Mod A).

NOTEd

/ _..___
~ Pk8-o3JA

Figure 2-Rocket, HEAT, 2.36", T59E3, Loading Assembly.

265
 b. Tests, in which it was attempted to shape the detonation wave front
travelling through the charge, were conducted with the T59E3 Heads. However, no im-
provement in penetration was noted,

c, The use of' a tandem cone arrangement (Fig. 2) indicated some marked im-
provement in the penetration of' the T59E3 Head. However, several basic difficulties
encountered with this type of head design must first be overcome, The first problem is
to design a supplementary charge (located behing the smaller cone) to give a very fast
.jet with little or no slug, which would produce a large hole in the armor through which
the main jet could then pass, The second and more difficult problem has been to delay
the fonnation of' the main jet so as not to destroy the jet trcm the supplementary charge.
Means of overcoming these problems are still under investigation.

d. An investigation to determine the optimum standoff for the T59E3 Head has
indicated, in static tests, that an increase in the existing standoff ~ approximately
one inch (to offset for the nose crush of the fuze on ballistic firing) may result in
an additional inch of penetration, It is felt that this change can be made without
affecting overall rocket length, since it may only require relocating the cone closer
to the rear of the head, A corresponding reduction in the length of the e:xplosive
charge, without change in standoff, did not indicate any reduction in penetration.

e. Tests are now being conducted with TS9E3 Heads having different cone thick-
nesses to determine if the .059" thickness of cone now used in the T59E3 Rocket is
optilmlm.
f. An investigation is being conducted of the effects of non-uni£onnity and
eccentricity of the head metal p!ll'ts, While this investigation is not as yet complete,
preliminary results indicate that the amount and uniformity of penetration can be
significantly improved ~ closer control of eccentricities of the head metal parts.

DEVELOPMENT OF A 2.75" T2016 AND T2017 HEAT HEAD

BY: I. B, GLUCKMAN

A development program is now in progress at this Arsenal to design a 2. 75" T2016

Heat Rocket Head using a mechanical fuze and a 2. 75 11 T2017 HEAT Head using an electrical
tuze. The T2016 (Fig. 3) will have an approximate nose crush of 2 inches while the
T2017 will have approximately 1 inch of nose crush. Tests to determine optilmlm para-
meters of these heads are in progress. The present tendency is to increase the standoff
of these heads ~ moving the cone to the rear; however, this would result in a smaller
amount of explosive and the limiting amount in this regard has not yet been determined,
The optimum cone thickness for these heads was found to be .076• in static tests.

BY: S, FLEISCHNICK

As a result of a recent investigation conducted ~ this Arsenal, to determine an

optimum design of cone for the 105111111 T43 (M324) HEAT Shell, it was found that improved

266

JL£12 anrt..
Figure 3-Head, Rocket, HEAT, 2.75", T2016 metal parts.
 - ----- ·-~

\
, I

I \'
'
penetration was ootainek, \n static tests, by removal of the nnatn (and/or radius) of
explosive at the base of the cone between the slant surface of the cone and the in-
terior walJ. of the shell. A 11fiat 11 of approx:imately 0.1 inch exists at the base of the
cone as assembled in the M28A2 Rocket Head. It was considered desirable to detemine
whether improved performance could also be obtained in this head by use of a cone which
pemitted no nat. The results of exploratory tests indicate that as much as two (2)
inches more of penetration can be obtained by eliminating the nat.

-,

268 ./

J 'em'!'S'T
 1..
THE PRESENT PERFORMANCE AND PROBLEMS OF THE 105MM BAT RIFLE

H. P, Manning
c. w. Musser
H. w. Euker

Frankford Arsenal, Philadelphia, Pennsylvania

ABSTRACT

As candidates for "The BAT Rifle 11 there are three l0511UU re-
coilless rifles, four mounts, four spotting rifles, one set of
fire control, four types of HEAT rounds and five types of spotting
anmunition including tracer. These will be assembled intc four
"weapon systemsll ,

The n,36 · Rifle 1 TllB Armmmi tion and 1'149 Mount have given
accuracy at 1 1 000 yards represented by HPE of 0,29 mils, and VPE
of 0.44 mils. Static penetration tests have given 16.9 inches
penetration against homogeneous armor. These data represent
ammunition already superseded by improved designs. Caliber .50
tracer and spotting ammunition have shown reasonably good fUnction-
.ing and matehing perfomance.
The major problems ourrent:cy- being encountered with the Tl36
Rifle, nlB Ammunition, 1'149 Mount, T43 spotting rifle combination
are: recoil balance seems to be undu:cy- sensitive to loading density)
the method of spinning the long boom, fixed fin T118 type projec-
. tile; the amount of spin .to meet accuracy and peDStration require-
ments; variations in measured jump ·under different conditions have
been encountered) questions of overall evaluation of the spotting
device and techniques for its use are yet unanswered,

One of the candidates for the BAT Rifle is the fi36 l05mm Rifle on the 1'149 Mount
and incorporating the caliber .50 Spotting Rifle, direct and indirect sights. In Fig.
1 you will note the spotting rifle is the T43 and the sight is the interim type. On
the 1'149 Mount the elevating hand wheel is in the vertical plane and the traversing
hand wheel in the horizontal plane. The firing button (pull out to fire spotting
rifle - push in to fire major caliber) is located at the center of the elevating hand.
wheel. The handl~ for operating the breech mechaniBm for. 105 rifle can be inserted
from either the tcp or the bottom.

The 1111JlJ111Ilition for the n,36 Rifle is shown in Fig. 2 1 the projectile being the
long boom, cme caliber fixed fin, T118 type and the cartridge case being the T30
fabricated case, The case is cylindrical and there is no requirement for gt1ide rails.

C!Mh£ m
 ' ··- .
1\)
-.J
0

..

Figure 1-105 mm rifle T136, 105 mm mount

T149, caliber .50 rifle T43 and
the interim sight.

Figure 2-Cartridge case (Dwg. FD15561,

rev. 1) and projectile, 105 mm, T118E2.

J
 Associated with the development program on the Tl36 Rifle and Tll8 ammunition
and technical~ inseparable from that development are the development of fin stabi-
lized ammunition Tl84 type for the M27 Rifle and the development of a Tl70 Rifle for
,' the 11 0NTOS 11 infantry vehicle. Fig. 3 shows the characteristics of these three rifles.
The main points to note - that the M27 Rifle has an 800 cu, in. chamber volume and the
other two rifles have 500 cu. in. chamber volume, This results in the loading density
of .3 for the M27 and ,5 for the other two rifles. The barrels on the M27 and the Tl70
are identical on the inside 1 the Tl70 having a thinner wall than the M27. Note also
that the M27 was designed for a 70 1 000 psi working stress, the Tl36 far 100,000 psi
working stress and the Tl70 for 85 1 000 psi working stress, The barrels of the M27 and
the Tl70 are rifled one turn in 20 caliber for the stabilization of the thirty pound
spin stabilized projectile, The Tl36 rifle 1 being designed specifical~ for fin
stabilized projectiles, is rifled one turn in 480 calibers to give it 2,18° per foot
roll to the projectile, in order to average out asymmetries which otherwise would
cause large deviations from the average trajectory.

One of the major features of the Tl36 Rifle is the incorporation of the "strain
compensation" principle in the barrel design, In order to achieve minimum weight for
the rifle, it has been designed so that the working stress of the steel is 100 1 000 psi
at rated maximum pressure. On conventional design, the bore diameter would have been
4.134 11 + ,002 and bourrelet diameter of the projectile would have been 4.128 11 - ,005,
With the very thin barrel of the Tl36 Rifle, elastic strain in the steel would give
rise to an increase in diameter of ,013 11 • The diametral clearance between the pro-
jectile and the bore during firing would be at least ,019 11 and with normal manufacturing
tolerances co1lld be as high as ,026", This condition is illustrated in the upper
sketch in Fig. 4. Several bits of experimental evidence support the contention that
all this clearance appears on one side of the projectile, On the "strain compensated"
principle of barrel design, the bourrelet diameter is made sufficient to give a small
clearance in the bore dur·ing projectile firing, while the bore is elastical~ expanded
by the gas pressure, as illustrated in the lower sketch in Fig, 4. In the Tl36 Rifle
the bore diameter is 4.134" and bourrelet diameter. is 4.145". During firing, clearance
between the projectile and the bore lies between ,002 11 and .007 11 at the extremes of
manufacturing tolerances. The T136 is the second recoilless rifle designed on this
principle. It has been shown ·that, with spin stabilized projectiles, accuracy is
greatly :Improved in this manner. However, sufficient experience has not been gained
to be certain difficulties will not be encountered in rifles designed with these high
stress levels.
Figure 5 shows long boom fixed fin 105mm HEAT projectile, For thE! present pur-
poses this can be either the Tl84 or the Tll8 projectile since the features which dis-
tinguish between. them are not shown, These projectiles weigh 16,8 pounds, both will be
fired at a muzzle velocity of 1700 f/s. These projectiles contain a 3.6 11 diameter cone
incorporating a spit back tube. The fuze is of the T208 series, actuated by the barium
titanate piezo electric crystal, Neither the crystal nor the base element are shown in
the sectional view, Stand off is 8-1/2 11 and fuze t:!me is expected to reduce this by
not more than 1/8". The explosive charge is two pounds of composition B with a head of
charge of 7 o3". The boom is perforated and contains the pr:!mer charge for the cartridge.
The projectile is fastened into the case by a retainer which acts as a shot start device
rupturing in tension at 4 1 000 pounds pul.l. Spin rates at the muzzle are of the order
of ten to fifteen rps. Penetration tests have been conducted with Tl84 type projectiles

271

Jl£12 !SEC ih&

 105 mm RIFLE CHARACTERISTICS

Rlfle no.. nclature T136 M27 T170

Projectile welpt ( lb) 16.8 16.8 16.8
lllua•le velocity (f/a) 1700 1700 1700
Rated mea preaaure (pal) 10,000 9,500 10,000
Chamber wol (cu in) 500 800 500
Char1e welpt (lb) 8.8 8.6 8,8
Loadin& density 0.5 o. 3 . 0.5
ProJectile ~ravel (in) 98.7 98.7 98.7
(orl1lnal deal1n) (ln) 88
Min yield atrenath of ateel (pal) 150,000 125,000 125,000
Worklna atreaa (pal) 100,000 70,000 85,000
Ratio bore to throat areaa 1.35 1.45 1,37
Bore diameter (In) 4.134 4.134 4,134
Bourrelet diameter (In) 4. 145 4.134 4.134
Rlfllna depth (ln) 0.005 0.037 0,037
.~
I I
( turna/eal) Tn' n
Twht of rlflln1 }
(de1/ft.) 2.18
Bare rifle welpt (lb) 185 340 220
Wt of complete mounted ayatem (lb) 383 702

Figure 3-105 mm rifte characteristics.

NON- COMPENSATED SYSTEM

--....
____.,
COMPENSATED
-SYSTEM

Figure 4----Principle of strain compensation.

272

1serwnrxm' ¥
 Figure 6-105 mm T118E18 projectile body
Figure 5 showing plastic rotating band on bourrelet. Figure 7
 ACCURACY- 1118 AID TIS' ROUIDS

••deJ ... ....., s,; ..

defflt.
at ••••'•
,,.
SpJn at tOO yd•
def!lt. ,,. •r•
•U•
....
•Jh
,.,h ...,.
pd

.u

.
TUIB:I Teet Rlald <2 <1 (.&3). 1.9 (.:1&). 10 (9). 900
.,,,d
Tll8Bl0
Tll8BU
Teet
Teet
...
•••••,.
<2
<2
<I
,
• &I (.24)•
.29 ...
.94 ( ••• , • 10 (8).

•
900
900
TIUBU TU6 <2 >u

,..
...••• .10 12

•
900

,.
Tli8BU Tl31 T14t·Gnd ... <2 >41 .17 900

,.
TllUU TUO Tl4t•J•ep <2 .37 .25 7 900
Tll8E13
TIIIBU
Tl36
TUI
TI41-Jeep
Tl4t•J . .p
<2
<2
•••
•• 9
.29 (.24)••
.et (.Ia)••
10
10 ....
1000

Tl18El3
888
Te•t
t::t ••••• <2 >41

• ...
••• •••
••
8 900

...
TliBBll Rlald 2.1 10 1.1 7 900
•

...... ...•••
1175-Jeep c• , ..2 900

..
TU4E2 ••7
,
.•••
TI14B2 1175-J•ep C4 1000
Tl84B4•:: ••••• 2.5 12 2.3
••• .18 7 900

..,.. ····· .. ....,.......

,
•• c.......... , ........ ,,,. ..... ,.,, •••.
. . . , . . . . . . . , ........ , • • • , 0,140 h .

• • • Rifled 0,10' ie deep at taht af ••• tera h 410 tell•••••

Figure S-Aceuraey-T118 and T184 rounds.

SPIN RATES

TI8~E3 Projectile -Plastic Rings

Ring widths Spin rate near muzzle

(inch) (rps) (deg(ft.)

none 3 to 4.5 .64 to • 95

I
T 9, 10 1.9, 2.1
I
if 18, 19 3.8, 4.0
I
2 30, 30 6.4, 6.4
I
2 (improperly positioned) 143, 159 30. 3, 33.7
I
if (in machined seat) 11 to 13 2. 3, 2.8
Figure 9-Spin rates, T184E8 projectile-plastic rings.

274

sbt:ll£££ aa
 a nPT?YW ?T S ·· I ' m Mw

having straight fiDs and fired from the M27 .Rifie so that the muzzle spin is appro:xi•
~~&tely 4 rps. F:l.'" rounds fired against 15 11 of stacked armor at 400 feet gave caaplete
perforations at normal. :Impact. Five rounds f1red at 911 or stacked plate at 6o 0 obli·
quit;y at 400 rt. range gave an average penetration of J.h.6n 1 penetration ranging from
14-l/4" to 15-l/4".
· The special features of the nJ.8El8 projectile, shown in Fig. 6 1 are the step
bourrelet and molded plastic rotating band. The step bourrelet is approximately l/4"
wide and appears at the rear edge of the ogive. Immediately shead of the step
bourrelet has been machined a shallow band seat and a plastic rotating band molded
into place. This rotating band is not pre-engraved.

The nBW£3 projectile, to be. fired either in the B27 or n70 rifles, is character-
ized by a plastic -ring located around the tapered portion of the body- to the rear of
the bourrelet. This may be seen in Fig. 7. During firing the band. apparently mOV"es
fozward, is engraved by the rifling, and imparts spin to the projectile by friction
and slippage between the band and the projectile. ·

· Figo 8 shows. the accuracy: 11hich has been obtained in the development of the n18
and n84 type projectiles during the program. The top two lines show the fairly poor
accuracy obtained with straight fin projectiles fired from a smooth bore rifle. The
accuracy targets showed vhat appeared to be a normally distributed group with occasional
· projectiles .well outside of this gr-oup. The figures in parenthesis show the target
sizes when only those rounds exhibiting no mal distribution were considered. The
n18El3 type of projectile bad a 4° cant to the flipper at the rear of the fin. The
average of horizontal and vertical probable errors for· all of these rounds, fired OV"Br
900 of 1:.000 yards range, iS •335 mils. nJ.BEll design is similar to the Tl18El8 except
that the body metal iS al1111linWIIJ thus the last line of nJ.8 1S represent firing from a
barrel rifled one turn in 480 calibers. The n811E2 rounds had fins with the rear
flippers canted 4o. Tile n8l!E4•s in the last line of the table were fired with l/4"
plastic bands in a machined seatj as shown in a previous .f':igilre 1 in an M27 with a barrel
rifled one in 20 calibers. Spins imparted to the projectile were approximately 12 rps
at·the mnzzle and 8-l/2 rps at the target.
Fig. 9 shows spin rates observed near the muzzle for various widths of plastic
rings. Note the l/2" rings although cemented to the body were \mocked loose in
chamberiilg the ra1nd into the gun and were fired in this condition. Note the very
high rates of spin obtainsd. Note that l/4" bands placed on the projectiles without
a band seat and those which were held in place in a machined seat give entirely
different rates of spin. The ·l/8•, l/4" 1 and l/2" bands were properly positioned al-
though not held there b;y a machined seat. Results such as these increased the mistrust
of plastic rings sufficiently so that other methods of inducing spin were designed and
imrestigated.

Fig. 1D shows a design made in accordance with principles laid out in a BRL Rsport
by Mr. Vinte. This has been variously described as the reaction engine or broached
bourrelet. The slots operate as no_zzles and the passage of the chamber gas through
these slots gives rise to thrust, the tangential component of which applies a torque

275
Figure 10-105 mm T184 test slug body
showing eight "reaction engines"
machined in bourrelet. Figure 11
 fW --1:1!& &ad i§ 12 I d It /1

·,-

SPOTTING RIFLE ANMUIITIOM

ProJectile type: Tracer Tracer Spottln1

No•encleture: Tl39 Tl39E3 T140E17
Ballistic Coeff, (Cs>= .765 • 6611 • 423
Mu••le velocity (f/a): 1739 1825 1975
Elevation Blea (•ll): 0 -1 -1
Mi•••atch at 200 yd (•il ): .13 -.ss . -. 22
• • 400 yd (•ll ): .20 -.25 • 38
• • 600 yd (•ll): .19 -.02 ,59
• • 800 yd (•ll ): .09 .06 .40
• • 1000 yd (•i 1 ): -. 12 -.02 -. 38

Ballia&ie eealf, el Tlll ow Tl84, c1 • 1,41 (Cs • - .tl)

Figure 12.-Spotting rille ammunition.

277
to the projectile. The one shown in the figure incorporates eight such engines around
the bourrelet. No appreciable degrading of performance has been observed in the
firing of 'these projectiles. Brief tests in the M27 Rifle showed that the desired
spin can be :Imparted by this method.

A second method for :IJilparting spin in the M27 is shown in Fig. ll. This is
described as the Jmurled bourrelet or the turbine type. This type operates because
the gases flowing past the bourrelet inside the grooves of the barrel expand into the
knurl and impact angular momentlllll to the projectile as they strike the surfaces of the
knurl. Analytical studies of this method have resulted in several additional designs
intended to give a range of spin rates. The design shown here has given a one point
check of the analyBis.
Experiments with both of these methods of inducing spin within the gun are in the
ear:cy stages and will be continued until a definite conclusion can be reached. The
minimum rate of spin which nmst be imparted to the projectile within the gun in order
to give good accuraa,y has not been definite:cy determined, although present indications
are that the spin of 2.18° per foot (10 rps) will give good accuracy up to 1,000 yards.
Whether the spin will decay sufficiently to give poor accuracy at 2,000 yards has not
been determined.

A portion of the BAT development program is the spotting device, consisting of a

caliber .50 semi-automatic rifle, tracer cartridge, spotting cartridge, and combined
spotting-tracer cartridge. A tracer cartridge and several spotting cartridges have
been developed. Fig. 12 shows the matching characteristics of several bullets. The
first column represents a tracer with insufficient length of trace. The second column
is a satisfactory tracer bullet. The third column is one of the spotting bullets.
The low ballistic coe.fficient is undoubted:cy due to large yaw of the bullet in flight.
It is believed that the design can be made to eliminate yaw and maintain the good
functioning characteristics upon :Impact. A combination tracer-spotting bullet is
under development and will probably supplant the separate bullets. The effect of such
large mismatch as shown for the Tl.h0El7 bullet is not now completely determined, but
may be sufficiently large that this bullet will be useless at this stage of the
development.

278
 DESIGN VARIABLES AFFECTING THE PEII.R)RMANCE OF THE BAT· HEAT ROUND*

Earl W, Ford

Defense Research Division, The Firestone Tire and Rubber Company, Akron, Ohio

ABSTRACT

This paper discusses same of the factors investigated in the

development of a HEAT round for the BAT project,

Some of the factors investigated are: cone angle, stand-off

flange effect, target material, effect of interference, cone shape,
wires through cone, base element location, temperature effect, spit
back tube length, cone ·wal.l thickness, cone material.• constant head
versus constant volume, and confinement, The more pertinent factors
will be discussed,

Firestone has adopted a 42° cone, pressed from sheet copper,

having no flange and a short spit back rube, with a wall thickness
of .100 inch for use in the T-138 HEAT rounds, This round is to be
fired at 25 r,p,s, The design of this round is explained in detail,

Time has not permitted carrying each investigation to the ulti-

mate and hence the data are in many cases only indicative of the
trends, Work is continuing on certain features in order to be more
specific as to the optimum conditions.

As part of the development of a BAT weapon it has been necessary to study many
variables which would effect the performance of a HEAT round. The effect of spin rate
on the several. variables was unlalown and hence had to be studied, It is felt that the
results of the various tests can best be 8Ullllll8ri.zed in tables and graphs. In practi-
cal.ly alJ. cases the data given represents the average of at least five tests,

The variables will be discussed in an order which is not necessarily that of

importance, Certain data are canmon to al.l rounds to be reported: The liners were
turned from copper bar stock purchased to government specification QQo.C-50la, All
rounds were fired from cylindrical. bodies, In part of the program the cylinders were
T-138 bodies and later a slightly less complicated body was developed for use as a test
body. Data will be presented to show that the results obtained with the two body types
were comparable.

* Data Obtained at ffiL and ERIE

279
.. ---- ---- ---- -----~. t:"::WO ' __,_- - - - -
--, ,, -- - -- - -
~
~

ri e--
•= ------
/--- \ '
---- --- --0-- ~--
-- ...,
-~

~~
Rl f!E\IS/!1
1 "
STAM:
-----;.\ 0-- ~

- '"~...
00
I
...
P----- ----7 --
,.,,./
'i'
-
. IITAICIOPP
'
0-
0-- --- - --~
-

•
It l
. 00 .. 40 .. . .
00 E Alia.! ( 8EES)
eo 0
SroANDOFF -INCH£5
'

Figure !-Penetration as function of Figure 2-Penetration as function

cone angle. of standoff.

..

e
ov / r---.., .. AT Z!RO

••' "
~
~
~·

·~
Ii -- ... AT 4& APS

,.u, f'"'
•

'"""" '
0
~ • ''t " "t
9TAIIIDOI"P-IIIOIII!S
27
' ..
FigureS
Figure 4--Penetration as a function
of standoff.

280
 TARGET MATERIALS
MATERIAL SPIN RATE PENETRATION
BOILCR PLATE 0 /8.0
~ENOEEIV 711ROET PLATE ALTERIV.tliT£0 nq
0
WITH Bon. Eli' PLA r £

AsERotEN TANGer PLATE 0 115

A8£1i'OE£N 7Aii>G£TfiAr£ ALr.t:;<o?- 0 I 7. 0
;\-AT£.0 tt11TH BotLER PI..AT£

Bo!L£RR.ArE 25 14,'1
H01¥00CNEOV.:S ARMOR P.LA TE zs /4.1

ABEROE:EN 771Ru£r Po~.Ar£

0 /8.1

HtGH OvAJ.ITY A.vMaR RArE 0 /6.0

ABE HOE EN TARGET P<..ATE 45" /0.7

IIJGJ.< OuAI..ITY ARMOR 46" t/.4-

Figure 5-- Target Materials.

' '
281
 ORC 224 ORC 230
(A) I Bl

ORC 314
(C l

Figure 6

20
!-._

-~ •coNE -NO ill.

15
- --:r- ~ ~
DRC 224 .TEE

~
- DRC 230' TEE "'-......:

~
-
-
-
...
<n1o
:r -
~ -
z
Q
.....
<l5
a:
...z
.....
-
...
a.
-
-

0
-
10 20 30 40 50
SPIN RATE (REV/SEC)

Figure 7

282
:r ,.
 I &EM 1J 22 tJ\
It is obvious that cone angle would affect the penetration obtained. Consequently
a program to study the effect of cone angle on penetration at various spin rates was
carried out. Figure 1 shows a summary of the data. It should be remarked here that
30° and 25° cones have been made but have not been tired. From the data available it
is apparent that a cone angle of approximately 42° is opt:l.mnm for the 105mm shell of
the T-136 type. A consideration not readily apparent is the overall length of the cone
and hence the shell body. It is fortunate that the optimum cone angle was 42° since
the T-136 projectile body as finally developed would not have accanodated a much
longer cone. The errect of cone angle apparently is much less pronounced at higher
spin rates.
Another important consideration in the development was the effect of stand-off dis-
tance on penetration. Figure 2 shows the effect or practical point detonated fuze
standoff distance on penetration for two cone angles at two spin rates, when using
copper cones. Figure 3 shows the results ror 45° steel cones. ~igure 4 shows the
penetration obtainable at higher standoff distances. These standoff distances would
require something otlier than a point detonated f'uze. The data shows that the standoff
distance has an effect on penetration which becomes quite pronounced for 450 cone angle
at the lower spin rates. It is also evident that extremes of standoff distance do not
result in greatly increased penetration.
A comparison of penetration through mild steel plate and homogeneous armor plate
was made. Figure 5 shows the results obtained. It appears that for slowly spinning
projectiles the difference in penetration is in the neighborhood of 5% whereas ror
zero spin .rate the difference is about 15%.

Since the T-138 projectile has a large mass or steel in front or the shaped charge
it was necessary to rind whether the tee would adversely affect the penetration. Some
preliminary work at the Ballistics Laboratories using aluminum tees having 1/4" and 1/2 11
diameter holes had shown that the hole had to be at least 1/2" in diameter. The ques-
tion of the relief :Immediately in front of the cone became of importance in terms of
the location or the center of gravity or the rounds. After preliminary firings had
shown the necessit,y of considerable relief the interior tee configurations shown in
Figure 6a, b 1 were tested. As a result of the data !'ran these tests the interior tee
configurat,ion chosen is that shown in Fig. 6c. The data for the two tees tested and
for rounds having no tee are shown in Fig. 7. It is seen that with DRC-224 tee con-
figuration the penetration is essentially as gcod as with no tee. The design being
adopted Fig. 6c is certain to be no worse than DRc-224.

Certain projectiles used by foreign cotmtries were known to have liner shapes
other than the conical shape that is standard in this co1.mtr,r. Certain special liner
shapes as shown in Figure 8 were tested. Figure 9 shows in tabular fo:m the results of
these various rounds as compared to the 45° cone. The data are incomplete as to spin
rate and standoff distance but clearly indicate that the conical sh~~pe is superior to
aey others tested.

The errect or the booster being buried beneath the rear surface or the Canp B was
investigated. In tbll original test rOl.mds the teteyl booster vas placed in the metal
plug just tangent to the rear surface of the Comp. B. In the actual round, it is

- 263
 w
. '

LINER SHAPE DATA

LINER TYPE SPIN RATE STANDOFF PENETRATION
IR.P.S.I UN.l UN.l

SEMI- CYLINDRICAL 0 7t 7.25

2.5 IN. O.D.

SEMI-CYLINDRICAL 0 7i 11.86

l
1.75 IN. O.D.

SEIII- CYLINDRICAL 0 9.55

7*
I. 00 IN. O.D.

HEMISPHERE 0 7~. 6.81

NO SPITBAal !0 7'11 7.118
0
0 ""•
7'11
9.60
9.72
TRUMPET !0 7h a. 7!
0
0 ""•
7'11
8.64
19.4
48" CONE !0 7~. 1!.4
45 71't 9.4

.Figure 9-Liner shape data.

Figure 8
 ?PEP 6 a

EFFECT OF BOOSTER LOCATION

~v_v-~ tt:": R STANDOFF PE~ETR~JION CONDITIONS
LOCATION {inches) 1nches

In b8s11 plu1 7~ /1.33

Mepl~ plu9 in
In Comp.B 7~ 17.16 bBse plu9
Steel in b.vse
InCDmp. B 7~ 1144 plu9

BoD.5f:er S'"
£mbedcleJ8 7~ /7.5S S lmp/11 Cone

BDoster m 7~
base plu!J ''·~1 Stmple Conti

Figure 1()-Elfeet of booster location.

-·

-
I( 0
ll

15
IP .~
-e·
r.

.t:'l
0
~"'
f--.... t:::--..

fi>
~
~
~

~ 10
fi
~
Ill
zIll
IL
1-
5
1-
PL :TE T ~MP. I Fl X
1- CH ~, ~MP. f Fl 1!1
r AI TEN P. I Fl 0
t-
0
10 20 30 40 50 60 70 80 90 100
Figure 11
 ---------··

SPITBACK TUBE LENGTH

/ ~
CYLINDRICAL TEST BODIES ORC "·6
NON·ROTATED, 7t IN. STANDOFF

'
~
SPITBACK TUBE LENGTH PENETRATION
liN.) ttN_l

0 17.33
'
l2S 17.50

2.+0 11.80
•

.050 .100 .ISO .too

WALL TKICKNESS (IN.)

Figure 12-Spitback tube length.

Figure 13

I
'
2

' m
~·
8
'
m.
"' z
' ,_Qo 7 "''"
' ,_,
"'""
' .."'"',•
z
0
' 14
'
' •0000 IJfJflO ..... .D040 QJ. .oo60
MAXIMUM DIFRM:flfCE IN WAL.L THICKNESS (IN.)
.005 -0010 .0015
STANDARD DEVIATION IN WALL THICKNESS IJN.)
.0020 002.5

Figure 14 Figure 16

286
 2

<f---_
5
-......._
rtJ.-
~•>-,
''
' ', Copper_?&) ·~
~

•
0 :sj1"; '~
4.5i s.o.
fiAiuminu' -bronze 7.5 lllS.O. s.o..
••;r ~~5"
9!; II
s.o.
~'WI'•uml ~um cones 7. 5in.s.o.
5

0
0 10 20.30 40 50
ROTATION IREVS/SECl
Figure 16

CONFINEMENT- TEST BODIES

SPIN STANDOFF
BODY TYPE CONE ANGLE PENETRATION
~~~T~~ (In l (in l
T·/38 0 7~ 4-5'0 17.4-
T-138 0 7~ 4-S /8.7
T-138 ~5 7~ 4-S I O. 7
Test Cq/mJ-, 0 7~ 4-S I 7.3
, ., 4-5 7~ 4-S I 0.0
T-138 0 7~ 60 14-. 4-
T-138 ~5" 7~ 60 Cl.6
r.st Cv/,d'er. 0 7~ 60 /3.B
.. , 4-S 7-;l 60 9.8
Figure 17-Confinement-Test Bodlea.

287

a a 11.
 with pi/of rintj /
removed ,. ~/
w1th pilot Yinlj S/z.t1_1.VS rill'/
LINER REGISTER

LINER SPIN RATE STAND\OFF PENETRATION

r.o.s. I 1n. (in.)
Sro. +s• CtJ. 0 7~ 1'1. Z6
~5" MoP1FIEO 0 712 19,Q8

Figure 18--Liner Register.

288

n
necessary to have the tetryl booster buried in the Comp B due to the size or the fuze
base element. Several experiments were run under slightly dii'ferent conditions which
indicate that the booster can ee ahead of the rear surface of the Comp. B by as much
as 1.0 inch without serious impairment of the penetration. Figure 10 shows the results
in tabular fom,

The effect of the temperature of the high explosive, the target material and the
ambient temperature was investigated at considerable length. Figure 11 shows the re-
sults, The general conclusions are, that over the temperature range imrestigated, the
penetration is not a function of the temperature of the Comp. B of the target plate or
of the ambient temperature,

At the outset of the program a spitback type fuze was anticipated in order to get
superquick action to take advantage of as much standoff distance as possible. Conse-
quently, the cones were btlilt with a spitback tube extending back from the apex af the
cone to the location of the tetryl booster. When other types of fuzes became available
the suggestion was made that the spitback tube be eliminated, Certain data had indi-
cated that some type af inert material back of the apex of a simple cone produced an
improvement in penetration, Hence, a series of tests using dii'ferent lengths of spit-
back tube were carried out. Figure 12 shows the data from this test, Within the
experimental error there is only slight reason for using a spitback tube.
The effect of cone wall thickness has been investigated in two ways. First from
rather drastic changes in nominal wall thickness • using .OS0 11 • ,100 11 • and .2oon.
Rounds of each thickness were fired at each of three standoff distances for zero spin
and, at one standoff distance, at 4S rps. The data are shown in Figure 13 and show
only that there is a maximum between .oson and .200 11 • Using data from other tests on
cones of .lSO" wall thickness it is apparent that the optimum penetration is obtained
with a wall thickness of ,100 inch, The Second study on cone wall thickness involved
the effects af variations in the wall thickness of a given cone, Figure 14 shows a
plot af the penetration as a function of the maximum difference in wall thickness and
Figure lS shows penetration as a function of standard deviation in wall thickness. It
is evident from these data that within the limits of variation involved the penetration
is independent of either variable,

Investigations were made to study the effect of various liner materials. Figure
16 shows the data available, It is quite evident that copper cones are definitely
superior to any other materials used.
Another factor which has been investigated is the effect of the wall thickness
of the test body or projectile on penetration. The available data are shown in
Figure 17. As far as the comparison af T-138 test projectile bodies as compared to
the regular penetration test body - the performance is essentially equal. Further
work is in progress.

The effect of non-concentricity of the cone axis and projectile body has been the
subject of repeated discussions. A series of rounds were modified in that the register
on the cone was removed. This resulted in the axes of the cone and test body being
non-coincident by as much as .003 inches, For this amount of eccentricit,y the data
shows little or no effect. Figure 18.

289

'A I
 SESE( 2 tsp

SUMMARY

1. A cone angle of approximate~ 42° has been foUJid to be optimum.

2. A standoff distance of 7 1/2 inches is found to be practical both from the

penetration and from the projectile design viewpoints.

3. At spin rates in the vicinity of 2~ rps the difference between penetration in

armor plate and mild steel is about 5%. At zero spin the difference is about 15%.

4. A tee design for the T-138 projectile has been found which satisfies the
flight characteristics and the penetration requirements.
'). A. conicu shaped liner has been foUJid to be the best for use in the project.

6. The fact. that the booster charge is buried beneath the surface of the Comp. B
has not been. found to be deleterious.

7. The temperature of the Comp. B, target or the air does not affect the pene-
tration over the range investigated.

B. The use of' a spitback tube on the cone appears entirely optional.

9. A cone wall thickness of .100 inch appears to be optimum.

10. Copper has been proved to be the optimum liner material. v

ll. The cone must be fairly concentric with the body, but variations of' .003 inches
does not mater~ affect the penetration.

290

a a aa
 p a s Tr?·E •·

APPLICATION OF THEORY TO DESIGN

Norman Rostoker

Carnegie Institute of' Technology, Pittsburgh, Pennsylvania

ABSTRACT

The scope of' the discussion will be limited to the design of'
hollow charges with metal cone liners~ Design parameters that will
be considered are material, geometry, and confinement of' the explos-
ive, and material and geometry of' the liner. Performance will be
assessed on the basis of' penetration depth, hole volume, and hole
shape.

Several specific charge designs and their observed per-

:f'onnance will be described. On the basis of' these examples an
attempt will be made to evaluate the guidance afforded by theo:cy,
in its present status to the design of' shaped charges. The ex-
tent to which further developments in theory ~ assist in pre-
dicting per:f'omance from design parameters will be considered for
the purpose of' clari:f'ying research objectives.

At the outset, it is appropriate to consider to what extent physical theory based

on representations may be of' assistance .in the design of' shaped charges. There are
limitations to the accuracy with which liners and charges can be manufactured to de-
sign speci:f'ications. Such limitations are hardly unique to the shaped charge problem,
but perhaps unique in relative. importance. A good illustration of' the e:f':f'ects.qf'
these limitations is provided by the comparison of' penetration-standof':f' curves:!/ in
Fig. 1. The two CUITes were obtained for two lots of' cones of' identical nominal de-
sign specifications cast into supposedly identical charges of' 50/50 pentolite. The
reasonable agreement of' penetrations for low standoff's lead one to believe that such
penetrations could be predicted from the nominal charge design. However for large
standoff's it is unlikely that ~ theory will be of' much assistance in predicting
. even the average penetration. In I1I8.Izy aspects shaped charge performance, charges that
are supposed to be. stereotyped turn aut to be iridividuals. The approach to the prob-
lem of' shaped charge design is consequently twofold.

(1) By means of' physical theory based on representations which embrace

nominal charge design parameters we should attempt to predict those aspects of' per-
:f'omance which exper:lments iDdicate to be determined by nominal design.

(2) Many aspects of' shaped charge performance are observable, but not pre-·
dictable from charge design. It is necessary to track down the causes of' charge
individuality in order to minimize them and increase the class of' predictable vari-
ables in performance.

291
 • ,....... 1-
I- +-
,....... 1-
1- !--
!A B
I
c 0
I
I

I
------ ---
.0.
~.--

,.,. l.l_ ;,.

•
.~. •0 r
Ol l
' -~·
--7n
~ /1
I

t'' ~\\~,
;

T'
0
ID
~
I
1<,
,,' ,, \~

...
8 [ _ _ _ JI
1_ ' II 1\

Q ..; I
"'.; I
• L.-, --
·. q
I J."
4 Mild

I I I "' I Sleelfoce
plate.
"'"' j
/i6
z
0
u 3"
14
I I J
\
z \
Q ~"
.... \
\ 74
a!.... \
Case (I I Cho'lfe \

"'z bose a free '\ ~"

cr boundary. \
\
94

'
N
;s Picatinny Lot ( I ) .
Picotinny Lot l2) ,
• (1947)
. (19~01 ..... ____ _
~=
2 4 6 8 10

STANDOFF INCHES

Figure 1-Comparison of two lots of eonea

of identical nominal design. Case (2) Ctuuge ba.se
in contact with a I" thick
mild steel plate drilled throuoh
the center ,
CHARGES Cost 50/50 pan10lite with tettyt boosteR.

LINERS hi9AI copper ; thickness D37" t apex angle 44• \ \

\
base diameter 1.6 3". \
\
\
\
\

Figure 2--Penetration produced by M9Al

cone-section charges in 2-S aluminum.
 77777'
7

3 Ed!& iil&d&t

In this discussion we shall be concerned principally with the present status of

the first approach. As yet the theory or shaped charges contributes little quantita-
tive assistance to design. The main reason for this is that we do not yet have an
adequate quantitative understanding of the dependence of the velocities and directions
of liner elements upon the geometry of the explosive. However, we do have an apparentl,y
sound qualitative understanding that has developed from extensive experimental studies
and theoretical investigations of over-simplified representations such as have been
carried out by the group at N. Y. u. This claim best be supported by considering
several specific charge designs.

1. Cone Section ChargeJ/

When the charge base is a free boundary as in Case (1) or Fig. 2, the penetrations
produ.ced by charges B and C are considerabl,y larger than we should expect similar
sections to produce when they are part of the full cone. The velocity gradient in a
jet is principall,y due to the rarefactions from the charge boundaries which reduce the
pressure acting on the liner and consequently the ultimate collapse velocity attained
by liner elements. The ultimate collapse velocity attained by any liner element de-
pends on the mass per unit area of the element and its location with respect to the
boundaries of the charge. Elements near the cone base for a standard charge design
such as charge D, are closer to a charge boundary and attain an ultimate collapse
velocity that is less than the ultimate collapse velocity of elements near the cone
apex. The relative magnitUdes of velocities of the corresponding jet elements are
similarl,y disposed. In tenus of these ideas the behaviour of cone section charges may
be understood. ·

-· We shall compare the cone section in charge B with an identical section of the
liner in charge D. In charge B the rarefaction from the charge base reduces the
pressure acting on the liner. The pressure dec~ is more rapid for basal elements.
From the point of view of the same cone section in charge D, the rarefaction from the
base has. been postponed by the presence or the ~est or the charge so that the pressure
on the basal liner elements of the cone section should decay less rapidl,y. Therefore
we expect the cone section in charge B to have a greater velocity gradient and produce
greater penetration than its counterpart in charge D. Similar remarks appl,y to the
cone section in charge C.

The essential correctness of this explanation may be tested by placing a mild

steel plate in contact with the free bou!ldary at the base of charge B. Then a weak
compression shock wlll reflect back into the detonation products, followed by a
rarefaction which begins when the steel plate starts to move. In general the pressure
near the base will dec~ much more slowl,y than it would if the boundary were free, be-
cause the steel plate moves slowl,y comp,ared to the escape velocity or the detonation
products. The result is that the jet from charge B has a smaller velocity gradient
and produces less penetration as m~ be seen in case (2) of Fig. 2.

293
 &Ittl&ii22 3 a

T
I"

1.7 5" 1.88" 2.00" 2.2 5"

.
5

/\ ,'fo;, ~ ,lflt, h~\ ,rv,.,

tl
,,,~' '\\,, ,, ,,~~· ~,, ,,,, ,,,, ,f \\
.
'l ,.,I! ''t, '
//' ~~
, ,,
,.,,
I'
~~,, •'~
\\
,, 'I \\ ~,

·~~•
" '}•\
II '• t : ·',, ,:' ~.
·~,
,, ~
,;1 ·~
.,,, ,,
,,
//
l' ., ,.,
ll
/1
\'
•'
'\ ",,,, ,,/' ,, '•"
I•
/
1
I•

'' ,1
'
'•'•
,, ,,,,
'•
,,\• ,,
• ~
'l.
i'!.'
'
'~·,.
,,,,
' I '

I"

14.5 19 26 29
c.c. c. c. c. c. c. c.
2"

-~---

5"
-----
6"
----------------------- --- __.. _-------------------------------

Figure 3-Proflle diagram of charges of various diameters and penetrations

produeed in mild steel targets.

CHARGES: CAST 50-50 PENTOLITE WITH CYCLOTOL BOOSTER

L.INERS: M9AI STEEL; TtiiCKNESS .036"; APEX ANGLE 44°;

BASE DIAMETER I. 6 3"

294

66f!ii IL..
 ?R?WJWWU 2 !&din.

,
T..

1.75" I. 8 8" 2.00" 2.25"

5"
t//i\,,
,,,,
II
,,,.
\\

,, ,,
,.
,.
,,
1\

~~ ·~

..1"
7 19.5 21.5 24.5 32
c. c. c. c. c. c. c. c.
1:_ ____ cc. _______ _ 1-------- 1------ >-------

2"
------~ f---- ----- e..------- 1-------- ------ '------

~: __ --- - --------· 1---- ----- ~------- f----

..4.':._ _____ 1---------- ~---·---- I---------- --------- f---

..?:. ________ ---------------------------- " v

\
6"
-------------------------------------------- --- ----------------

Figure 4--Profi.le diagram of charges of various diameters and penetrations

produced in mild steel targets.

CHARGES! Cost 50-50 pentolite with cyclotol ooosters

LINERS: M-6 steel; thickness .062"; apex angle 42",

base diameter 1.63".

295

san r
 .._ ________.
2. ·Control of Penetration and Hole V~lume by Explosive Diameter

Consider first the case of a 44° steel cone with an oOj~ in. wall thickness for
which exper:1Jnental results obtained by duPont illVestigators.V are shown in Fig • .3·
The penetration as a function of charge diameter has a maximum for a diameter of 1.75
in. which is slightl;r greater than the 1.6.3 in. base diameter of the liner. For
charge diameters less than 1. 75 in. the basal elements of the cone receive insufficient
impulse from the explosive to form a useful jet. By increasing charge diameter the
basal elements of the jet are not wasted and the penetration increases up to an optimum
charge diameter when the cone is efficientl;r employed and the velocity gradient of the
jet is a maximum. Further increase in charge diameter reduces the penetration because
the resultant increase in impulse delivered to the basal elements decreases the velocity
gradient in the jet. The hole volume produced is a monotone ·increasing function of
charge diameter.

Additional results of duPont investigators for 42° steel cones of wall thickness
of .062 in. are shown in Fig. 4. The interpretation is essentiall;r the same. In this
case the opt:lmum charge diameter is 1.88 in. instead of 1.75 in. in which case the
maximum penetration is the same as for .0.37 in. cones. The reason for this is that
the .062 in. liner has a greater mass per unit area and requires a greater impulse to
achieve the same velocity. The hole volume is once again a monotone increasing func-
tion of charge diameter.·

The performance of the above charge designs can easil;r be understood from Q1ll'
qualitative notions about the rarefactions from the charge boundaries. Rarefactions
proceed in from the free boundaries of the charge and reduce the pressure acting on
·the liner. If the belt of explosive around a section of the liner is narrow the
pressure· decays rapidl;r and the impulse imparted to the liner is correspondingl;r law.
Optimum penetration is achieved when the cone is employed efficientl;r in that the basal
elements are not wasted, and when the jet has the maximum velocity gradient. The hole
volllme is a monotone function of the total kinetic energy of the jet so that it in-
creases monotonel;r with the impulse imparted to the liner •
\
.3. Confined Charges

If the free boundaries of the liner are covered with aiJiY confinement material that
provides inertia, the confinement material moves slower than the escape velocity of the
gas. The expansion and consequent decay of pressure at the liner by rarefaction waves
is retarded. For design purposes, the addition of confinement tubing is essentiall;r
equivalent to increasing the explosive diameter. An increase or a decrease of pene-
tration may be produced depending on whether the original charge diameter is less or
greater than the optimwn diameter. The hole volllme of course increases with confine-
ment.

TP~ effects of confinement are illustrated in Fig. 5 which is a reproduction of

duPont!V data. The charge diameter is 1.6.3 in. which is slightl;r less than optimum
for .0.37 in. cones. The o75 in. confinement tube is equ;l.val!lnt to an increase in
charge diameter of more than .25 in. so that a decrease in penetration results :from

296

l
 tGWN'TVW' '' 2 3 22Ii&El&..

-"'
LLI
:J:
(,)

-
;!!: 4

0 Charges confined in .75 in. thick

seamless steel tubing

A Unconfined charges

40

"'2 30
(,)

LLI

~
~ 20
LLI
d
I

10

.0 2 .04 .06 .08

CoNE WALL THICKNEss I INCHES l

Figure 5-Effects of confinement on penetration and hole volume produced

by 44 • cones and 50/50 pentolite charges in mild steel.

297
 confinement. Charges for the thicker cones are also 1.63 in. in diameter which is
considerably less than optimum. The confinement tube is equivalent to an increase in
charge diameter which inc~ases the penetration.

Although the present status of the theory is such that it can only provide
qualitative assistance in the design of shaped charges, it is clear from the above
examples that even this much assistance permits us to exercise a considerable degree
_ of control over penetration and hole volume by suitable charge designs.

tl The hole volume can be increased to any desired magnitude by increasing the
( amount of e:xplosive 1 adding confinement, choosing an explosive of higher detonation
pressure, or a combination of all three. The penetration at a given standoff and for
1 a given liner material is limited by an apt:IJnum that is achieved when the basal ele-
1 ments of the cone are efficiently Eimployed 1 and the jet has a maximum velocity gradient.
The optimum penetration .may be achieved by adjusting the charge diameter, adding con•
finement, varying liner thickness, or by a combination.

It is anticipated that current work on the dependence of velocities and directions

of liner elements upon e:xplosive design will eventually make possible quantitative
application of theory to design. Quantitative success in the application of theory
to design will be always limited by the accuracy with which design specifications can
be met. However, there is much to be gained in the appreciation of the possibilities
and limitations of shaped charge applications by attaining the quantitative under-
standing of the collapse process that mathematical analysis of realistic representations
can provide.

BIBLIOGRAPHY

"}j CIT-ORD-32 April 30 1 1951

'lj CIT-ORD-23 October 311 1949

l/ Theory and Application of the Cavity Effect; May, 1943, Contract W-670-0RD-4331;
E, I. duPont de Nemours and Co., Inc,

W Theory and Application of the Cavity Effect; March 1943 1 Contract W-670-0RD-4331;
E. I. duPont de Nemours and Co,, Inc.

298 ,.

spy
 a 22. SSC2L 5

FOREIGN DEVELOPMENTS IN. SHAPED CHARGES

Harry Bechtol

Develop!l!ent and Proof Services, Aberdeen Proving Ground, Maryland

ABSTRACT

This presentation deal.s with recent shaped charge deve1op-

ments from the following sources 1

1. European - France and Be1gium

2, Russia and China

The French 7.3mm Rocket and the Be1gian. Energa Rifle Grenade
have been sel.ected as being representative of recent European
advancements in shaped charge tecl:mo1ogy, Both of these devices
have many exce1l.ent features among them being 1ightness of overall
weight, improved mechanical. types of fuzes and shaped charges using
RDX type filler and copper 1iners,

ibe Energa Grenade emp1oys a 2 .711 diameter 1 45° copper cone

simUar to 1ate mode1s of U. s. shaped charge ammunition. Pane-
. trations of 1011 or 3.7 cal.:lllers have been observed, Full standoff
is 1.3 calibers,
The French 7.3mm Rocket is unique in that it emp1oys a trumpet
shaped or doubl.e ang1e cone of 2.73 11 diameter. The cone starts
at. an ang1e of approximately" 220 at the apex and finishes at an
angle of appro:x:lmatel.y" 580 at the base, Cone thickness increases
from apex to base (approximately" .055 11 to ,084"), Penetrations of
12 ,on or 4.4 cone calibers have been observed, Fu11 standoff is
1.5 calibers.

Russian and Chinese designs of shaped charges Observed thus

far are general.l.y" considered to be primitive by current standards,

Russian 76mm and l22mm shaped charge artUl.ery projectiles have

been examined and tested. Except for spitback fuzes, they resemb1e
very early" U, S, designs, which used stee1 cones. Penetration is
roughly one caliber for either round.

The Chinese 8711111l Rocket of very recent manufacture is now

under examination. The Rocket Launcher is a c1ose copy of the
U, s. 3.~ 11 Rocket Launcher, The 87mm Rocket, strangely" enough,
is a spinner. The shaped charge for this round has TNT. filler
and a fabricated (wrapped and· seamed) cone. Performance details
are not ava11able at the time of preparation of this abstract.

299
 INTRODUCTION

This presentation deals with recent shaped charge developments from the
following sources:

1. European - (France and Belgium)

2. Russia and China

The description of the French and Belgian developments is not an attempt to

enumerate all of the various calibers and types of shaped charges now under considera-
tion. Instead, we have selected two developments which we consider outstanding and
about which we have a reasonable amount of knowledge.

The information to be given on Soviet and Chinese hollow charge ammunition is

probably the limit of our knciwledge of their present application of shaped charge
teclmology.

If our information is aceurate, it might be considered comforting, for it shows

the Soviets and Chinese to be considerably behind us and our allies in development of
shaped charges.

EUROPEAN DEVELOPMENTS

The French 7Jmm Rocket and the Belgian Energa Rifle Grenade haYe been selected as
being representative of recent European advancements in shaped charge munitions.

The Energa Rifle Grenade is a development of a Belgian firm known as Mec!ir*. The
characteristics of this grenade are as follows:

Caliber 2.8711
Weight 1.32 lb
Velocity 165 fps
Fuze Type Spitback
Design Standoff lo) Cone Diameters

Explosive Pressed RDX - TNT (81.9% - 18.1%)

Liner Material Copper - Tin ·Coated
Liner Diameter 2.715 11
Liner Angle 16'
Liner Thickness .o4811
Liner Manufacture Probabl,y Cold Formed and Annealed

* ~ Societe Anozzyme Belge de Mecanique et d 1Armement

300

s a £L ,.
 Armor Perfonnance* - Defeats 10 11 at 0° consistently. Also defeats 5 11 at
65° (ll.811) consistently. Has defeated 10" at 300
(11.5") but not consistently. These data may seem to
be contradictory at first glance. However, when we
consider that these are the results of firings at actual
ranges, the explanation is obvious. A low velocity
missile mtch as a grenade has an appreciable trajectory
at any range. Hence, it has a noticeable angle of fall
at the target. When striking a high obliquity plate, a
small angle of fall makes a considerable difference in
the amount of armour to bs defeated. When striking a
low obliquity target ()0°) a small angle of fall makes
little difference in the amount of armor to be defeated.

In view of the above, we consider 10 11 at )0° to be

the upper limit of the Energa Grenade.

Advice received from London indicates that the

Energa Grenade has recently been improved. The cone is
supposedly the same, but the weight of explosive and the
standoff have been increased. This new round is supposed
to defeat 325mm of annor (12.8 11 ) at oo obliquity.

The French 73mm Rocket is a development of S'l'RIM**• The Rocket is designed for a
shoulder fired launcher much like the 3.511 Bazooka.
A quantity of these 73mm Rockets have been tested at Aberdeen Proving Ground with
excellent results. The folloWing summarizes the important characteristics observed:

Caliber 73mm
Weight 3.0 lb
Velocity 5'25 fps

Explosive Cast 50/50 Cyclotol

Fuze Type Base Inertia

Design Standoff 1.5 Cone Diameters

Liner Material. Copper

Liner Diameter 2.7)11

Liner Angle Double 22° and 58°

Liner Thickness Varies .055 " Apex to .08411 Base

* .. As reported by Board No. 3, Fort Benning, Georgia.

** • Societe Technique de Recherches Industrielles et Mechaniques.

301

%1!!6 'iii
 Liner Manufacture Cast*

Armor Performance - Local tests indicate that the round defeats 12 11 at 0°

consistently. other targets defeated were 611 at 600
(12 11 ) and 411 at 70° (11. 711 ). The upper limit of per-
formance was not determined, but it is no doubt slightly
greater than 12 11 • In SUJlU11Bry, it can be said that the
French and Belgian designers of shaped charges have made
excellent use of the latest techniques and have come forth
with service rounds which give performance comparable to
the best u.s. rounds to be developed to date.

The French liner is especially interesting since it

embraces a design which has been given very little con-
sideration in this country. In fact, u.s. experiments
along this line with 105mm trumpets showed the trumpet to
be no better than a cone.

RUSSIAN SHAPED CHARGES

Ttlo shaped charge rounds of Russian designs and manufacture have been examined am
tested in this country. The following are the characteristics of these rounds:
Caliber 122mm 76mm
Weight 29,2 lbs 11.93
M, V. (Approx) 1093 i'ps 850 i'ps
Fuze Type Spitback Spitback

Explosive • Cyclotol Cyclotol

Liner Material Steel Steel

Liner Angle 420 280 40 I

Liner Thickness .080 11 to ,08911 ,017 11 to .090 11

Armor Performance - The 122mm Projectile has been subjected to static test
and gave a max:IJ1lum penetration of 4". No dynamic test
results are available, but Soviet Ammunition Manuals indi-
cate that the round is gGod for about 4".

The 76mm Projectile has been tested at Aberdeen

Proving Ground recently, Statically, it defeated 2.75"
to 3,0 11 • In dynamic firing tests, the performance was
about the same, the best being 2 11 at 470 (2 .9").

* "' The liner examined locally appears to have been cast, machined inside aru:l aJUlealed.
However, advice received from London indicates that similar cones for the STRIM
Grenade are formed from Jmm plate am finish machined.

302

'
 a 5 223 22 ldn

CHINESE SHAPED CHARGES

Only one example of a Chinese shaped charge has been captured and examined to
date. This. is an 87nun shaped charge rocket projectile that is fired from a weapon
which is evidently a copy of the u.s. 3.5" Rocket Launcher. While the Chinese Launcher
bears close resemblance to the u.s. Launcher, the rocket itself is entirely different
from the u.s. 3·5" rounds. It appears that the Chinese made no attempt to copy the
U.s. Rocket nor any of its good features. In fact, it appears that the Chinese round
might have been designed independently of the u.s. model.

The Chinese 87mm Rocket is spin stabilized, resembling somewhat, the u.s. 4.5 11 M16
design. A small quantity of the Chinese Rockets have been captured and are now at
Aberdeen Provine Ground for examination. None have been fired to date, however. The
following characteristics have been observed:

C~iber 8~

Weight ll. 75 lbs

Velocity Unknown

Fuze Type Spitback

Explosive TNT

Liner Material Steel

Liner Angle 55°

Armor Performance - Unknown, however, it is doubtful that the penetration
will be much greater than 1 caliber, based on u.s.
experience with rounds of similar design (75miu M66).

303
s ,..
 "bEE mz seem g r n

ROTATED-NON-ROTATED
REP<RT ON l20MM SPIN STABILIZED PROJECTILE WITH NON-ROTATING SHAPED CHARGE

Stanley Dubroff

Artillery Ammunition Department, Frankford Arsenal, Philadelphia, Pennsylvania

ABSTRACT

In order to secure an armor penetrating round 1n the l20mm

caliber which will have greater penetration than the AP shot, this
"Shaped Charge" carrying round is being developed. Considerations
leading to this design are: compactness of round, due to its spin
stabilization; improvement of annor penetration of the IIShaped
Charge" by reduction of spin to approx:lmatezy zero.

The proposed round has an exterior member which is rotated at

the i'ul.l spin rate for this weapon. Because of the large moment of
inertia provided by the cylindrical section, stability is secured.

The inner member carrying the "Shaped Charge" cone is carried

on ball bearings. Load pressure distribution on the inner and outer
member reduces the load on the ball bearings to a very small valne
during 11 set backn.

The above is an abstract of the paper given at the Sympos;I.Ulll. The ma-
terial discussed in this paper is covered essential.J.y by the Frankford Arsenal
RNR Report from which the following is taken.

THE DEVELOPMENT

The development of the subject shell divides itself into four main parts:

1. !Jivestigation of all possible methods of mounting the inner member of the

shell 1n the outer member.

2. The design of first engineering samples.

3. The manufacture of engineering samples.

Testing the completed r011nd.

305

8 9!WW?
 'A"i'fPFMPI 'fr we 'd$
1
F'ss f

~UIREMENTS

The following is a tabulation of requirements for the l20DD11 HEAT T230 spin-
stabilized projectile with nonrotating charge:

1. The projectile should be launched with the least possible rotation im-
parted to the shape charge.

2. The projectile should be lalUlched without any damage to the bearings and
the bearing races.

3. .After lalUlching, a method must be devised to prevent rotation of the

inner member of the projectile caused by bearing drag.

4. Design a projectile with the largest poss:lble diameter copper cone in

order to improve on the performance of the best available 12DDDI1 AP shot.

DISCUSSION

HISTORICAL

The following characteristics will exist in all designs of the shell:

a. The projectile· will be stabilized by rotating the outside member, which

should have heavy walls, for high axial moment, and should be as short as practicable,
resulting in relatively high stability.
b. The inner portion of the shell containing the shape charge should be
restrained from rotating and should be as light as practicable in order to minimize the
overturning maaent.

Since rotation decreases penetration, the idea of a rotating shell with a non-
rotating shape charge is a natural outgrowth. It was discussed in Artillery Division
long before it was given any serious consideration. Finally, in December of 1950, the
above idea was discussed in a conference at Office, Chief of Ordnance, ORDTA. During
this conference many ideas for effecting the design were discussed. Franlci'ord Arsenal
received from the Ordnance Office &lo.ooo and was informed that this Arsenal should
proceed with design studies on the 120IUIII rotating-non-rotating shell (RliR). The
initial approach considered the use of a thrust bearing for the rear bearing of the
projectile and a needle bearing for the forward. After the initial design studies were
completed, an investigation was made to detennine the stability of this t;n:>e ot pro-
jectile. Preliminary calculations indicated that this type of projectile was very stable
and, i f launched without damage to the bearings, had a very good chance of pertoming
in the proper manner.

306 r:,
I

eo lit 1BL24 2L !1.

 SCWWMW!. I s m I I ••

The follow.ing approaches were reviewed as possible solutions of this design prob-
lem.:
1. The use of a fluid lubricant to be used with mating parts to form an
ammlar orifice between the outer and inner members.
2. The use of special rifling to engage the inner member while the outer
member is being rotated by the normal rifi:i.ng of the gun.

3. The use of orifices in the base and ogive to produce a turbine-like

effect which will prevent the rotation of the inner member.

4. Setting the inner member in bearings which will. prevent the inner member
from picldng up full spin on firing of the gun, while the ou.ter member does pick up
full spin and provides the stabilizing infl.uence for the projectile.

Of the above mentioned possibilities, the fourth was chosen. The first was re-
jected because of sealing problems, and heat dissipation through the fluid at high
rotational speeds. The second was discarded because of the need of a special gun to
fire the round; and the third dropped because of the implied difficulty and expense of
manufacture. Therefore, the fourth possibility, which seemed to be the most logical,
was chosen for development, since it does away with fluid seals, special type rotating
bands or grooves and turbine blading to go with an orifice system.

Assumming that the projectile would be fired at qo,ooo psi the design was initiated.
The first design incorporated the use of a thrust bearing at the rear of the proje!)tile
to take the setback forces. (See lllust. No. ll}. Considering the weight of the pro-
jectile to be approximately forty (40} pounds, the Jll!IJdmum acceleration of this pro-
jectile being fired at a pressure of 40 000 psi would be in the vicinity of 20 1 000 g 1s.
Were the inner member to weigh eight (8~ pounds, the setback forces on a bearing placed
at the rear of the projectile would be l6o,OOO pounds. A thrust bearing to take this
load would have to be far larger than that which could be contained in a l20mm shell.

At this time it may be mentioned that a Government contractor had an Ordnance con-
tract to develop a projectile of this type, and upon testing the first designs found
that failure occurred at the bearing. A thrust bearing was utilized in the above de-
sign, and failure was caused by "brinelling". Personnel at this installation did not
have the adv11ntage of aJV :f'urther infomation than that which told of bearing fallure.
However, it was quickly perceived that the crux of the entire problem was the rear
bearing. That is, the setback forces caused the balls to imbed in the races and cause
depressions which prevented the balls from rolling freely.

There was no apparent ~ to solve this problem of so great a setback force until
the personnel working on this project came up with a possible solution: - to limit the
setback force by allowing the gases propelling the projectile to act on the inner
member, causing little or no relative motion between inner and outer members.

With this solution in mind the first design came off the drafting board with an
opening in the back of the projectile. The bearing at the rear of the projectile vas
a thrust bearing, and a needle bearing was put just to the rear of the cone near the
front of the projectile. (See Illust. No. ll). .

!S!il ibill42illf:
 It was then a problem to see if bearings were available to i'it in the spaces set
up i'or them in this design. Representatives were consulted !'rom New Departure Division
of General Motors, SKF Industries, Inc., The Torrington Camp~, Nice Ball Bearing Co.,
and Fafnir Bearing Co. to get their separate rec011111endations on the problem. Nice and
Fai'nir Bearing Companies could not supply the bearings and The Torrington Comp~ pre-
sented a needle bearing whicb would i'it, but advised that a needle bearing of' the type
presented would not take the bigb rotational speeds of' this projectile. The New De-
parture Division of' General Motors and SKF Industries, Inc., recommended a plain single
row, deep groove radial type ball bearing, as they were asked to i'Urnisb a bearing to
take the bigb rotation of' 20,000 rpm, and a thrust of' 10,000 potmds (with a life of'
less than a minute).

The New Departure Division of' General Motors and SKF Industries, Inc., advised of
the availability of' certain sizes of' bearings and !'rom this point on the projectile was
designed around the bearings.

Since the projectile_ would eventually be used with the new NBS 11 Lucky'1 i'uze, it
was decided upon to utilize this arrangement. This immediately presented the oppor-
tunity to place the front bearing i'arther i'orward than in the first design, thus
eliminating most of' the i'ront overhang in the projectile and ai'i'ording support !'or the
inner member close to both extremeties.

Due to space limitations, the largest cone which can be utilized at present is
105DDII. in size. The rei'ore, the same cone desil!1led !'or the 105DDII !'in-stabilized pro-
jectile will be used !'or the i'irst group of' engineering samples of' this projectile.

Upon the completion of' this design, the local representatives from the New De-
parture Division and one of' the f'actory engineers were consulted on the design and
possible problems of' lubrication. At the present, storage problems are not coneidered,
therei'ore, it was concluded. that a thin i'ilm of' oil on the bearing would be enough
lubrication i'or this i'irst test. There were also some alternatives in bearings oi'i'ered
in the event that the deep groove, radial type ball bearing oi'i'ered too much resistance,
and this drag caused undue rotation of' the inner member. (The i'riction exerted on the
balls by the high velocity of' the outer race revolving m~ cause additional drag on the
inner race causing the inner member to rotate. )

The bearing and bearing seat tolerances chosen were A. B. E. C class three (3).
At the present tble it is not known wether this tolerance is needed, but the problem
at hand is to establish a principle, while the production difi'iculties encountered may
be overcome at a later date. (It must h9W"Ver, be mentioned, that~ of' these
difi'iculties have been noted and possible production solutions are already available.)

Ten l20mm. Tl47El target practice rounds have been modii'ied i'or use as proof' slugs
!'or determining a propellant system !'or the projectile.
The i'irst five (5) projectiles have been inert loaded and will be 1'1red on the
transonic range at Aberdeen Proving Ground, Maryland. The remaining i'ive (5) pro-
jectilee will be utilized in another test, possibl:y against armor plate i'or pene-
tration, upon the successi'ul conclusion of' the first test,

308

W Ulillll Bll I 'IL

 66A11E&4 III • a ·it I ' atim .,..

Stability computations were performed on the first design and appeared satisfactor,r.
Information pertinent to flight characteristics will be obtained frOlll the first test of
the present design.

In order to be able to tell whether there is any relative motion between the outer
member and the inner member, the aluminum nose of the projectile (attached to the inner
member) will be anodized and then sand blasted with a spiral in order that it may be
photographed in flight. An alternate method for determining the rotation of the inner
member is to insert a magnet into the nose of the projectile perpendicular to the axis
of the projectile. There would then be set up wires running longitudinally in the range.
The projectile when fired would then set up impulses in this field by any rotation of
the magnet. These impulses would be recorded, and the rotational speed would be readily
available.
To have an absolute check on rotation, two pins are attached to the inner (non-
rotated) member and protrude past the base of the rotated member. One pin is cylindrical
in shape and the other is conical. The projectile is photographed in vertical and
horizontal planes along its trajectory and from these photographs the rotation 1118\V" be
computed.
TECHNICAL
In order that a basis for design would exist, and since no special requirements
were announced for this RNR (Rotated~Non-Rotated) projectile, the pressure of 40,000 psi
was chosen as a design pressure (based on the pressure at which the l20mm AP Shot is
fired).

From the elementary e~ation

(1) P x A • m a where P ., pressure, psi

A '" exposed area (sq.in.)

M • weight (lbs.)

A • acceleration (g•s)

it is easily seen that for any set conditions where the area and weight are given, the
pressure is directl:y proportional to the acceleration or

(la) ; • ~ • constant

To prevent too great a shock load from acting upon the bearings (see Ulust. No. 2) 1
or mare exactly, upon the rear bearing, the design problem existed whereb;r the exposed
rear area of the outer member divided by its weight be equal to the exposed rear area of
innsr member divided by its weight. When this condition exists, there will be no
tendency for relative motion between inner nember and outer member upon firing.
Under design conditions, the pressure being 40 1 000 psi, the acceleration of the pro-
jectile computed to be 181 951 g 1s. The setback forces (ma) for the inner member,

309
 seeypy;p

weighing seven (7) lbs., came to 132 1 657 lbs. This force had to be balanced by the
pressure of the gun multiplied by the exposed area of the inner member. From equation
(1) the exposed area was co111p11ted and the diameter determined.

For the given pressure and acceleration:

Al A2
m·M2'
where A is the rear area and M is the weight; the subscripts stand for the inner and
outer members. This is the condition which 11111st exist in order that there would be no
tendency for rel&tiva motion between inner and outer members upon launching.

The consultant from New Departure Division recommended an angular contact bearing
in the event that there was too much. rotation of the inner member due to drag on the
inner race. In this event, it is not desirable to leave the forces equal, for i f mry
forces are set up in flight that might cause the thrust to be forward on the inner
member, the angular contact bearing would not be suitable. However, i f the area of the
inner member was so controlled as to permit a set-back force of say 1000 lbs. the
bearing would function as designed. From this point of view, and to be sure that the
direction of the force be known, the exposed area of inner member was designed to allow
for a net force back, of the magnitude of 1000 lbs. This will be useful regardless of
whether the single row deep groove radial type or the angular contact ball bearing is
used.

One design idea for limiting the rotation of the inner member is to mill a spiral
~oove in the nose 1 with the spiral opposite to the rotation of the projectile •a outer
(stabilizing) member. If the rotation is excessive, this method for limiting rotation
may be tried. Another alternative, straight grooves in the nose, may be sufficient t.p
stop rotation.

Ano1:Jler problem existed in the design - concerning aerodynamic forces on the pro-
jectile. The first design had the inner member utilizing a nose which took up the en-
tire ogival area. With this arrangement, a large portion of the drag of the shell was
taken by the inner member. It is not known whether or not this would set up an;y undue
forces on the bearings. However, to do way with this possibility, the present design
utilized frontal exposed areas approximately proportioned to the masses of the outer
and inner members. The drag on the inner member is proportional to the area exposed.
The drag on the outer member is proportional to its exposed area. The above voo.ld ·
minimize the tendency for relative motion .between inner and outer member.

310

canwwwre
 a esrwnm

CONTRIBUTING PERSONNEL

ARTilLERY AlolMUNITION DEPARTMKNT

PHASE DIVISION ~

Engineering & Design Projectile & Case Mr. w. s. Scheingol.d

n II Mr. H. s. Lipinski
II II Mr. s. Dubrofi'
n n Mr. G. Callum
II n Mr. J. Brady

Design Mr. W. F. Kauffman

II Mr. J. J. Luczak
II Mr. s. s. Bloom
II Miss M. Bostwick
Manufacturing Experimental Shop Mr. J. King
n· II Mr. R. Ganser
II n Mr. F. Goepel
n It Mr. F. Majke

311

a a
 ' : I h T?tiBE

~'U
en
LW
::c
(._)
:::::2:

-e
en

co

-r--

CD

-&..n

-r

-(""')

- C"\..1

Figure 1-120 MM, T230 projectile metal parts asserr.bly.

312
 7

313
 •— oo
LU
— — oo nz
—— oo

•05
•O5

CO
•CO

•CO •CO
•CO

•to •to
•in

•CO
-co
•CNJ
•OJ
•oj

-c=,

-•-****
Figure 3—Inner body assembly Figure 4—Inner body assembly
(before machining). (after machining).
Figure 5—Inner body assembly,
nose in place.
 s; a

Figure 6--Inner body assembly, bearings, outer body and ogive assembly.

315
 cmmsrnw!, I a I

tin

Figure 7-Rear bearing seat. Figure 8-Forward bearing seat.

Figure 9-Labyrinth seal (base plug & rear lock nut).

 Figure 10--Projectile base (arrow indicates exposed inner member) .

..
317
I ?P'E' ·p
 22£& &EM 16 IIIIG!iiiZL

< Ill
j Q
.ifj
...,
a;
"'·a 3
·a
~

1...... I...
[ eg,
~
~

318
 sa: as sua£) £2II&IIl£ a

STUDY OF THE EFFECTS OF ROTATION UPON THE PENETRATION

OF JSTS FROM 10,5MM SHAPED CHARGE:S
L. Zernow
J. Regan
J. S:lmon
I. Liebennan

Ordriance Engineering Laboratory, Ballistic Research Laboratories

Aberdeen Proving Ground, Maryland

ABSTRACT
An analysis of the targets into which rotated shaped charges
were fired was carried out prior to .t'lash radiographic studies
previous~ reported.
Clear evidence was found for the bifurcation· and subsequent
apparent po~cation of the jet at increasin~ higher rota-
tional frequencies. Photographs of sample targets illustrating
this effect are shown.

A rotational standoff effect is established, which shows

that the. depth of penetration of a r~~pidly rotated shaped charge
decre~es sharply with increasing standoff.

Additional evidence for the :Importance of shocks and com-

pressibility in the penetration process is found in peculiar holes
of essentially square cross .section associated with the jets from
the rotated projectiles. It is noted that these effects may be
i.ruportant in lethality studies.
A ~ic model of a rotating hollow liquid cylinder
which is unstable under rotation is proposed as a basis for
understanding the bifurcation process.

INTRODUCTION
The need for a systematic study of the effects of rotation on large cal:lb8r(sbaPed
charges stimulated the development in 1950 of the wire driven projectile rotator l) -
which gl"Elatly a:lmplitiad the e:xper:lmental problems. This system was :Immediately put to

(l) So Kronman, L. Zernow - "A W;t.re Driven Projectile Rotator for Hollow Charge
Studiasn - IllL Report No. 798, March 1952 •

.319

anmsnwu
 20

...
"' 14
:1:
u
z

z 10
0
!i
"'ti
z
.
•
•
7 1/2 STANDOFF

le
• 1.0

•
•o,L---•---~--~.~--.~--~,~~~.~.~-.~47--~,.~0--~,.~0~72~
20 40 0 0 00 0 0 00~~.~7~240
20 oL-----~,±o----~2~0~----~~~----~4~0----~50~-­
FREQUENCY OF ROTATION (RPS) STANDOFF (INCHES)

Figure 1-Etreet of Rotation on Penetration Figure 2-Effect of standoff at 240 RPS.

at various standoffs. 45 • cu. cone 45 • cu. cone with spitback tube, 105 nun.
with spitback tube, 105 mm.
 qpn

use in studying the effects af rotation upon 105mm projectiles. This information was
urgently needed in connection with the development of a 105mm weapon by t.he Firestone
Tire and Rubber Company as well as by several government agencies. As it turned out,
predictions based on the scaled performance of 5.1mm projectiles were badly in error.
The actual experimentally observed deterioration was about twice that predicted on the
basis of the 57mm scaling calculations. These calculations have not yet been critic~
examined so that this deviation from the scaling law m'IY, at least in part be
attributed to causes such as non-similarity of the liners.

During the course of these experiments careful attention was paid to the targets
which showed several very interesting and reproducible characteristics that could be
correlated with the rotational frequency and the standoff. A study of these targets
proved Uluminating, and it was possible to predict some of the details regarding jet
breakup(!bich were later observed in the flash radiographs of the jets from rotating
liners. J

THE EXFERIMENTAL FROGRAM

The initial experiments which were carried out covered the range of rotational
frequencies from 0 to 6o rps. The standoff used in these experiments was fixed at
about 2 cone diameters, since this was approximately the standoff at which the service( )
weapon would be expected to function. The data collected in these initial experiments 3
were au~ented to include the effects of higher rotational frequencies and of variation
of the standoff.

EXPERIMENTAL RESULTS

Figs. 1 and 2 contain a summary of the experimental results. Penetration is

plotted against rotational frequency in Fig. 1 and against standoff in Fig. 2.

The points of interest on these plots are:

(1) The crossover of the curves for 9 3/4", 7 J./2 11 and 4 3/4" standoff,
(2) The clearly evident rotational standoff effect at 240 rps.

The crossover frequency was not established very accurately because it is not of
significant importance at present. However, the reality of the crossover is established
beyond doubt, and it is understandable simply as the resuJ.t of the competition between
increasing penetration due to standoff and decreasing penetration due to rotation.

Simply stated, the experimental range of standoffs for a copper liner is such that
at zero spin, the penetration is still increasing with standoff while at high-spin

(2) L. Zernow, s. Kronman, F. Rayfield, J. Paszek and B. Ta.yl.or - "Flash Radiographic

Study of Jets from Rotating J.05mm Shaped Charges" - Shaped Charge Symposium Trans-
actions - November 195J..
(3) Most of the data obtained in the early firings appears in the J.950 and J.95J. reports
of the development of the B.A.T. weapon by the Firestone Tire & Rubber Compaey.

32].

c& ILL
 6o r~ r~

Figure 3-Bifurcation and Polyfurcation in the bottom target plates

resulting from rotation at various frequencies.

322
 814£± S "tt N ••
(2qo rps) the rotational dispersion effect is such as to decrease the penetration with
increasing standoff. These two end conditions necessitate a crossover such as is ob-
served.
Sane evidence for this rotational standoff effect was observed by the British (4) in
1943 in their original exper:iments on 3" charges. However 1 the trend since that t:ime
to carry out the exper:imental work with small eharges has resulted in the crossover
point be:i.ng shifted to higher frequencies and hence, the e:xtJer:imental data reported with
small charges does not go out far enough in rotational frequencies to f~ establish
the rotational standoff effect at high spins. This can be made clear by the following
reasoning:-

If 1 as postulated by Birkhoff 1 the rotational effects scale with the parameter

c.i> 1 where ro is the rotation frequency and D is the caliber, then a cone of 1/2 the
105nm diameter would require a rotational frequency of 480 rps to show effects that we
can observe at 240 rps. Very few exper:iments have been carried out at this high
rotational frequency, which is much greater than that attained by H.E.A.T. projectiles
in aey weapons. As an example, the 57mm H.E.A.T. projectile is fired at 215 rps. It
is likely that we could still have observed evidence of the rotational standoff effect
at frequency of rotation lower than 240 rps.
A semi-empirical correlation of rotational scale effects proposed by Winn(5) of
the Firestone Tire and Rubber Company does not take this rotational standoff effect into
account. Therefore, its reliability must be considered limited to the region of
rotational frequencies belOW" the c::ossover frequency.

TARGET STUDIES - EFFECT OF INCREASING ROTATIONJ\1, F'RF;;tUENCY AT CONSTANT STANDOFF

As the rotational frequency was increased during these e:xtJeriments a series of

striking changes could be observed on the target plates. The reduction in penetration
was found to be accompanied at first by definite enlargement of the hole bottom and
finally by a bifurcation in the last plate. The spacing between the two holes in-
creased as the rotational frequency increased at a given standoff. Finally at the
very high frequencies, the bifurcation gave way to a series of many holes spread around
a circle.
Figure 3 shows pictures of the last plates at 3 different rotational frequencies.
These photographs clearly show the increasing spread of the bifurcation and the ulti-
~te polyturcation.

The data obtained are summarized in Fig. 4 which is a plot of the distance between
the centers of the two holes (or the mean diameter of the circle for the polyfurcations)
against the rotational frequency.

(4) AG-3987 - Hollow Charge Rotated Projectiles.

(S) Firestone Tire and Rubber Compan,y - Progress Report - June ll, 1951.

qpcypf)$''1
 DISTANCE BETWEEN CENTERS OF HOLES
VI. FREQUENCY OF ROTATION

--en
11.1
:z:
u
....z
en
z 11.1 Z.O
11.1 ..1
""
1\)
~
11.1 0
3: :z:
1-
11.1
Ill II.

11.1
en
u It:
z 11.1
0
L5

LO

(
;..........-
,._.
,... --c -- ~

c 0.5
1- 1-
en z11.1
-
Q u 0
20 40 60 110 100 120 140 160 160 200 220 240

FREQUENCY . OF ROTATION (RPS)

Figure 4-Distance between centers of hole& va. frequency of rotation.

' ·.
rum 5

I§

Figure 5-Square hole from a rotated charge.

Figure 6-Square hole from a rotated charge.

32.5

&Nfl 1££14 I I I -
 0

Figure 7-Hypothetieal steps in the process of bifurcation

of a rotating hollow liquid cylinder.
/

326
 Among the other interesting observations are those shown in Figures 5 and 6. Here
apparently a hole with approximl!ltely square cross section has resulted from the rotation*
or the charge!

It is quite evident that processes other than the simple incompressible hydro-
dynamic ones are needed to explain these results. The need for considering in detail
the tar~t compressibility during penetration and the study(qf shock propagation in the
target (and perhaps in the jet as well) is clear. Birkhoff b) has alre~ considered
the effects of compressibilit,y on penetration and concluded that target. compressibility
decreases the penetration. Further effort in extending this is believed justified be-
cause such studies may make important contributions to the better understanding of
lethality or the H.E.A.T. round.
HIDRODYNAMIC MOOEL OF THE BIFURCATION PROCESS

In the course of the observations of bi.t'urcation in the target plate it was natural
that one should attempt to find a hydrodynamic. model which could explain the process.
Following Birkhoff's workt 6 J and extending the reasoning to non-ste~ state hydro-
dynamic models, the investigation led to art examination of the stability of a rotating
holloW liquid cylinder. This appears to be an extremely difficult problem to solve
analytically, although one could make very plausible a posteriori "intuitive" guesses
regarding the process of bi.t'urcation of a rotating hollow liquid cylinder, examples or
which are· sketched in Figure 7.

During the investigation of this problem** it was learned that Sir. James Jeans had
carried out a theoretical investigation "On the Equilibrium of Rotating L:l,quid
Cylinders". (7) He was interested in the three dimensional problem connected with the
behavior of nebulae !lnd the formation of planetary systems, but in this work he tre·ated
the simpler two dimensional problem or a cylindrical mass of fluid held together by a
gravitational potential, using a very speciali~ed approach. The essential conclusions
he reached were:

(1) There was a critical rotational frequency below which the system would
be stable,
(2) Above the critical frequency the fluid cylinder would break into two
parts.

It must be noted first that Jeans· treated the full ·cylinder, whereas we are
really interested in the stability or a rotating hollow cylinder. Unfortunately it
appears that h:l;s specialized treatment will not work ilr.a hollow cylinder. In 8.f11"
case, it is not presently known whether the stability of a hollow liquid cylinder is
markedly different from that of a solid cylinder. Intuitively one might guess that
there would be considerable siinilarit,y~ Jeans predicts a necking down and finally a

* Dr. Soper indicated that similar observations had been made by the British.
(6) Garrett Birkhoff - "Hollow Charge Anti-Tank Projectiles" - BRL Report No. 623 1
February 1947.
(7) Phil. Trans A cc 67 (1902).
** Pointed out to one or the authors (LZ) by Prof. T. H. Berlin or the Physics Dept.,
Johns Hopkins University •

• 3~7

RAJ EM. rz
Figure 8 {Water analog of a shaped charge} 8-Unrotated.
Figure 9 Jet & Slug 9-Rotated.

328
break-up into two smaller cylinders of unequal diameters 1 whereas the experimental
observations indicate that the break-up is into two approximately equal portions. One
could, of course, postulate successive bifurcation of each of the two cylinders re-
sulting from the first bifurcation in order to account for the experimentally ob-
served polyfurcation at high rotation frequencies,

WORKING HYDROD~~IC ANALOG OF JET FORMATION AllD THE EFFECTS OF ROTATION

Early in 1950, two pictures were cbtained from the laboratory of Dr, Schardin in
France,* These pictures which are reproduced in Figures 8 and 9 represent a converging
conical surfaca of water fonning a jet and a slug as predicted by hydrodynmnic theory.
In Fig. 81 no rotation was :Imparted to the water. In Fig. 9, rotation was deliberately
imparted to the water. The fonnation of a hollow region in the interior of the jet and
slug can plainly be seen. If one examines the jet in the rotated case, one can
imagine that there is incipient bifurcation, although a comparison with the unrotated
case leaves one somewhat in doubt, since there appear to be extraneous striations in
the unrotated jet as well.

This was an extremely striking demonstration and it was an easy matter to dupli-
cate** the unrotated jet with locally fabricated equipment, It has been more difficult
to duplicate the rotated jet, although the attempts to do so are still in progress,
In addition, a rotating hollow li:J.uid cylinder analog has been constructed. It appears
at present that extraneous physical properties of the water, such as surface tension,
may mask the e'ffects one is seeking, Steps are being taken to eliminate these and thus
make the analog more realistically duplicate the system i t is meant to simulate.
One can occasionally obtain definite bifurcation in the water analog of the
rotating hollow cylinder, but it is not at present a satisfactorily reproducible
simulator. This work is being pursued further and the plans include treatment of the
fluid and possibly choice of fluid to eliminate the factors that are believed to be
masking the results,

This short discussion of the hydraulic analog has been included because it is
indicative of the diversity of the possible methods of attack upon the problem of
understanding the rotational deterioration of the jet.

SUMMARY AND CONCLUSIONS

The process of deterioration of a jet by rotation is· shown to involve a bi-
furcation of the jet, at the lower rotational frequencies. Higher rotational fre-
quencies appear to result in polyfurcation,

A clear rotational standoff effect is found at 240 rps which shows the pene-
tration falling off monotonically with increased standoff.

* Through th.e courtesy of Professor Garrett Birkhoff of Harvard University.

** This work was done by Mr. s. Kronman and Mr. F. Rayfield.

329

LL&ifL$
 --------------

. cezrpyNTI 2I 2 I

Evidence is fOUIJd for processes other than the simple incompressible hydro-
dynamical ones normally considered in target penetration. The need for extension of
the analytical treatment has been recognized by others before. These observations
accentuate the need for an extension.

The postulated instability of a rotating hollow liquid cylinder is proposed as a

hydrodynamic model on which to base the underst~nding of the bifurcation process.

/
 S'SP7I ::; 1!116ILI 244Siu

SPIN COMPENSATION

E. L. Litchfield

Carnegie Institute of Technology, Pittsburgh, Pennsylvania

ABSTRACT

Techniques and methods for the manufacture of fluted cones as

developed in cooperation with the National Bureau of Standards are
discussed. Techniques used in producing C.I.T. laboratory size
samples are essentially those required for large scale production
of liners for weapons use. It is shown that the problem of ob-
taining good fluted liners is of the same magnitude as that obtain-
ing good smooth liners.

The results obtained from tests with various fluted liner de-
signs are presented. Several fluted liner designs have produced
100% compensation (i.e. static smooth liner performance) at spin
rates as high as 15'0 to 180 r.p.s.; other liner designs have pro-
duced ._as much as 75'% compensation at spin rates as high as 300 to
330 r.p.s. Additional fluted liner groups which may be of impor-
tance in future weapons design are also discussed.

The need of compensation for the effects of spin degradation upon the performance
of shaped charge ammunition is well known. A typical demonstration of such degradation
is given in Fig. 1 1 which represents the penetration depth vs. rotational frequency
data for the copper cone used in the 5'7mm HEAT shell. The penetration at 210 r.p.s.,
the spin frequency of the shell fired from the 5'7mm recoiless rifle, is only 2.8 in.
although the liner will give 9.3 in. penetration at 0 r.p.s.

The concept of compensation for t~~ effects of spin through the use of fluted
liners first appears in E.R.L. reports1/ Apparently, the idea was advanced initially
by Pauling; and it was tested by E.R.L. with small numbers or relatively crude liners.
The first laboratory size group of identical fluted liners was obtained by E.R.L. just
before the termination of NDRC at the war's end.

Thus, at the time this investigation was undertaken rr,y C.I.T., neither fluted
liners nor methods for the manufacture of such liners were available to permit further
testing of the idea. The problem was standardized with the 5'7mm copper cone (1.69 in.
't)ase diameter, )j2° apex angle and 0.045 in. wall) because of the higher potential
penetration of copper liners; and was to include the development of methods of manu-
facture applicable to mass production of fluted liners.

]/ National Defense Research Committee of th_e Office of Scientific Research and Develop-
ment, "Target Penetration rr,y the Jet from a Rotating Cone Charge" 1 Report OSRD No.
3874, July 10, 1944.
National Defense Research Committee of the Office of Scientific Research and Develop-
ment, "Target Penetration by Rotating Cavity Charges" 1 Report OSilD No. 5'5'98 1
November 5', 1945'.
331
 10 SMOOTH 57 MM CONES

9 !! f-COMPOSITION

~-50/50
B
PENTOLITE
8
.
z 7
-I
I 6
z
0
!
I
~
\-.>
\-.>
1\J
~
a:
.....
LIJ
z
5

4 t:
!
LIJ
Q. 3

2 •
0 ~~-----+----~------1------+-----~----~----~-----+----~----
0 50 100 150 200 250 300 350 400 450
ROTATIONAL FREQUENCY-- R.P.S.

Figure 1-Smooth &7 MM cones.

 j 2 t' I r
.,.
Responsibility for the development of applicable methods of manufacture was
asSU!IIed by the National Bureau of Standards; and a IlUlllber of linear flute designs have
been manufactured. A representation of a typical product is given in Fig. 2. This
figure shows several cross sections of a fluted cone, The canted surfaces, which are
usuaJ.ly segments of circles canted with respect to the pitch circle of the smooth cone,
and the flute or offset surface joining the extremities of two canted segments may be
observed.

The first cones manufactured by N.B.S. were formed with a·· fluted punch and a
rubber pressure pad. The fluted punch is formed by machining the fluted cone design
into a steel replica of the cone to be fluted, This punch and a rubber padded smooth
die are then used to form flute contours into the copper cone, The rubber acts as a
hydraulic medium, forcing the cone against the punch and causing flute contours to
appear in the exterior surface of the cone as well as the interior, Typical flutes
formed in this manner are sho1m in Fig. 3. The contours formed on the interior of the
cone are reasonable but not perfect reproductions of the punch; but the contours formed
on the exterior of the cone bear ·little resemblance to the contours actually present on
the punch.

Fig. 4 shows profiles which were produced when N.B.S, developed a method of pro-
ducing fluted dies, The dies are cobalt-phosphorous alloy replicas of the fluted punch.
Such a replica, prepared by electro-deposition and seated in a tool steel block, may be
used as a fluted die, The contours in Fig. 4 were produced by such .a die used in con-
junction with a rubber padded smooth punch.

It would seem that the next step in development of fluted liners would be the use
of matched metal dies and punches. Fig. 5 shows typical contours for such liners.
Little experimental information is available for designs of this type and they will not
be discussed fUrther here,

Examples of typicaJ. results from fluted liners which are of interest in the 57mm
HEAT problem are given in Figs. 6, 1, and 8. Fig. 6 shows data which gives evidence of
100% coli!Pensation at 150 r.p.s. and a curve shape which is a reasonable facsimile of
that of the smooth liners, except for the shift in the frequency at which best pene-
tration performance was obtained. (The solid curve represents performance data for the
smooth 57mm copper liners.) These cones had sixteen linear flutes with a nominal maxi-
mum depth (at the cone base) of 0.025 in. formed by the method illustrated in Fig. 3.
It should be noted that, with few exceptions, the reproducibility of results is as good
as with smooth liners. Penetrations as high as 7.4 in. at 210 r.p.s, should be compared
with the 2 ,8 in. to be expected from smooth liners at the same frequency.

Fig. 7 shows penetration depth vs. rotational frequency data for a group of liners
having 16 linear flutes of 0,035 ino nominal max:lmum depth, also fo:nned by the method
illustrated in Fig, 3. These liners show COII!Pensation at a DDlch higher frequency than
those of Fig. 6; but they do not show as high a degr-ee of COII!Pensation. The best per-
formance is about 7 in. penetration, or 75% COII!Pensation, at 330 to .36o r.p.s. This
penetration is nearly .3 times the penetration of the smooth liners at the same frequency.
Even at 210 r.p.s., where only 6o% COIIIPBnsation is shown, the penetration of the fluted
liners is twice that of the smooth liners •

.3.3.3

L &&. ..
Figure 2

334

ft?WF
 wsnzz :Sll

Figure 3-Typical RS flute contours formed with fluted punch and rubber padded die.

[_:--JJ

Figure 4-Typical SR flute contours formed with fluted die and rubber padded punch.

Figure 5-Typical SS flute c:ontours formed with matehing steel dies.

335
 rsnrm 6
I

3
7

;Ifill

NBS LOT 5
10

9
18 X.028
••
a • • • I I
!i 1 •• I • • •
I

z•' 6 ••
S! • I
.~•. 8 •
... 4

a
•• • ••
• •
•
0
0 80 120 180 240 aoo a80
ROTATIONAL FREQUENCY--R.P.S.

Figure 6-NBS Lot 3, 16 x ,025 RS,

10
NBS LQ:T 4
9 t8X.OS5 •AS
a
I• 1 •
I
•;:: •
0 • •
... •
c
a:
z
. 4
• • •• • • •
a • •
• • •
0
•
0 eo 120 180 240 aoo 360 420 480

ROTATIONAL PMQUENCY--R.P.S.

Figure 7-NBS Lot 4, 16 X ,085 RS,

NBS LOT 8
10

e
181.018
••
• •
..
!

~
I

!Ca:
1

5
•
•
•
•
•
•
•
..
:..
4 • •
a •
•
•
•
0
0 eo 120
ROTATIONAL
180 240
FREQUENCY- -R .P.8.

Figure 6-NBS Lot 9, 16 x ,016 SR,

- lHIO 420

.,,

336

It
'•J·"
 a ..?P sc:.
Fig. 8 show penetration-frequency data for another group of liners having about
the same dynamic characteristics as those shown in Fig. 6. However, the data repre-
sented in this figure show quite clear~ the variability that some groups of fluted
liners have exhibited. The best performance to ·be expected fran these liners at 210
r.p.s. would be about 7.5 in. or about 3 times the penetration of the smooth liners at
that frequency. These liners had 16 linear flutes of o.ol6 in. maximum depth formed
by the method illustrated in Fig. 4. The opti11111111 frequency of the liners of Fig. 8 is
more than twice that of the liners having the same nominal specifications formed by the
method illustrated in Fig. 3.

The accUIIIUlation of data from several lots of fluted liners has indicated certain
correlations between design parameters of fluted cones and their performance. For cones
having sixteen flutes, the optimum frequency is essentialzy a linear function of the
flute depth. This linear correlation holds for flutes formed as illustrated by either
Fig. 3 or Fig. 4, but, for the same nominal flute depth, the liner having flutes formed
sharp on the exterior has an optimum frequency which may be as JIIUch as 3 times as large
as that of the liner fluted with a punch and rubber padded smooth die.

Difficulty has been encountered in obtaining camplete compensation at optimum fre-

quencies above 150 r.p.s. If flute depth is increased beyond that required to give an
optimum frequency of 150 r.p.s., the performance at the optimum frequency begins to
degrade. This is apparently due to mechanical failure of the liners along the corners
of the flutes, which in turn is likely due to the fact that the simple linear flutes
used thus far are not the ideal. Empirical design of non-linear flutes from data ob-
tained in tests with the linear flutes is contemplated. The interpretation of degrada-
tion with increasing flute depth is strengthened by results obtained with heavier
bllllllcs. Much deeper flutes can be tolerated when heavier blanks are used but, as would
be expected, the optimum frequency for a given flute design decreases with increasing
wall thickness.

For constant flute depth, as the number ·or flutes is increased from 16 to 36, the
direction in which the liner must be rotated in order to achieve compensation is re-
versed i f the flute orientation is kept constant. This means that i.f a 16 flute liner
has a positive compensation frequency, the corresponding 36 flute liner will show a
small negative compensation frequency. This reversal in direction of compensation was
entire:~¥ unexpected at the time of the initial observation but has since been established
unquestionabq by experimental obsel"'ations. As the flute number is increased from 36
to 6o, the magnitude of the negative compensation frequency increases slightq. With
flute numbers in excess of 6o, no further increase in compensation frequency with flute
n11111ber has been observed. These experimental obsel"'ations provide the basis for re-
jection or modification of early theories of compensation and will provide a critical
test of future proposals for such a theory.

There is no apparent reason that a fluted liner design which will compensate at
210 r.p.s. and have good reproducibility CBilDOt be developed. The proper compensation
frequency has been closely approximated and offers no real problem. The more trouble-
some problem of variability is now being investigated through gaugings of each fluted
cone before firing. These data require tedious and laborious statistical treatment; but
significant results are being obtained.

337
 · snrr esmlifii ·wer mar
There are two general fields for theoretical work in connection with spin de-
gradation am spin compensation. The first of these, dealing with spin degradation,
implies a theory to explain the effects of rotation upon the collapse and penetration
of smooth liners. The second general field, spin compensation, means here a theory
for the mechanism of compensation by fluted liners.

The eff~c;ts of spin upon smooth shaped charge liners have been treated briefly
by Birkhoff .£./ These equations were derived for steady state collapse geometry but may
be adapted readily to non-steady state collapse geometries. No definitive experi!llent has
yet been carried out; but the equations appear applicable to such data as ace available.
Theoretical treatment of the mechanism of compensation by fluted cones has been
attempted by several groups. The initial version of the E.R.L. theory assumed that
only the infiuence of the canted surface collapsing toward ita center was important.
It was determined, however, that the observed direction of compensation was opposite
to that predicted by the theory. A later theory developed by E.R.L. predicted com-
pensation in the proper direction (for small mmbera of flutes) but does not give
predictions in quantitative agreement with observation.

Since that time, theoretical work, primarily by C.I.T., has progressed until now
it is believed that three effects may influence the behavior of fluted liners. The
first of these is the variation in impulse received by a liner from a detonation wave
with the thickness of the liner. The second is the variation in impulse received by a
liner with the inclination of the liner to the detonation· wave. The third effect (the
only one of the three considered by E.R.L.) is the tangential component of velocity
which (possibly) can be obtained from the collapse of the fluted liner. Of these three
effects, it is believed that the first is by far the most important. Each of the first
two effects may be experimentally separated and studied by means of plate liners.
Experimental and theoretical investigations with these simple models has been accom-
plished; but application of the results to a theory of fluted liners is difficult.

A theoretical paper to be given later by L. H. Thomas (Watson Scientific Computing

Laboratory) treats by shock wave techniques the first effect mentioned for liner de-
signs that actually can be manufactured but which are different from those that have
been tested.

I In conclusion, although no quantitative theory is yet available, compensation by

fluted liners seems to be the simplest and most promising way to counteract spin degra-
dation in shaped charge !11l11111lJlition. There is no apparent reason that liners m81111factured
to reasonable industrial tolerances and designed to compensate at service frequencies
cannot be developed.

Y Ballistic Research Laboratory, Report No. 623, "Hollow Charge Anti-Tank Projectiles",
Aberdeen Proving Ground, Maryland, February 10 1 1947.

338
 ..1n
., - 2 C.aL~
i!!. Z'-
YtCl.Jau

MINIMIZING THE EFFECT OF ROTATIOII UPON THE PERFORNANCE OF LINED CAVITY CHARGES

Hugh Wi.nn

Defense Research Division, The Firestone Tire and Rubber Company, Akron, Ohio

ABSTRACT

The effect of rotation on the penetration of shaped charges

is described. An empirical correlation, useful for design pur-
poses, is presented, which permits one to estimate the spin rate
penetration curve for shaped charges with reasonable reliability.

In an effort to overcome the deleterious effect of rotation

both fluted liners and double body projectiles have been studied.
It is shown that a degree of spin compensation has been obtained
by each method and that the prospect for an ultimate solution to
the problem is bright.

It is well lalown that the penetration of lined cavity charges is greatly reduced
by the rotation required for spin stabilized shell. The extent to which penetration is
reduced by rotation is shown in Figure 1. These curves have certain very interesting
features. The penetrations at 0 and 240 rps are proportional to the base diameter of
the liners. This means that the penetration efficiency-spin rate curves for these two
liners coincide at 0 and 240 rps, but at intennediate spin rates, the efficiency of
the 10511111 liner degrades with increasing spin much mere rapidly than does the 57mm
liner. As a first approximation it appears that the rate of loss of penetration at
stated spin rates is proportional to the non rotated penetration, that is,

L" FSP (1)

0

and Ps • P (1-FSP ) (2)

0 0

where L .. 1 - ~
0

S = .spin rate, rps

Ps = penetration at S rps
P .. penetration at 0 rps
0

Values of the FS vs S function are shown in Figure 2. The curves drawn represent the
composite data for several hundred rounds. It is very significant that the curves for
the 57mm and 105mm liners coincide up to 45 rps. Thus, aver this range equations (1)
and (2) hold for both 51 and 105mm liners, and also presumably for a wide variety of

. 339

q sr:
 zo

1\
~~
-----
/OS" 1'1/H

~. 4'7 HH
1--- -a__

CIT· Oil/) ""--4

---e
If/~

-
0 so /00 200 26"0

Rt:vs/st:c
Figure I
L 3 •FSP.
I?• R(I-FSI?
• •)
.10

.08

~ --- - 157 Hl"f

C.t.T.-ORD - /i/8

~
.06
v
'·~ .o~
I.JI
/
b. /OSHM

•o.z
)
w-.
:· ..

0
0
7,
Figure 2

340
~~gp;;
;;. i-)
 20

REVS/SEC
25
~
~
20
~ IS

-z-
~
.•
·w
z"'
;:: i::::
f;. ~
0
rr: ~ /0
z
z
0

l
10
"~
!5
~ ~
··~ 5
'Z
"'... s

5 10 II! 20

PENETRATION- ROTATING ( IN.l

Figure 8-Eft'ect of rotathm upm

0
penetration' 0 .1() 20 .JO 40 so
/?EV'.J/SEC
Figure 4
 zrsrrr s
£,. ..
1

a
other shapes ani types. Equations (1) and (2) and the correlation of Figure 2 have
been used to evaluate the penetration-spin rate behavior for series of liners of vary-
ing non-rotating penetration. These data are shown in Figure 3. Predictions based on
these curves are remarkably sound. In some seventeen comparisons, totaling some 200
rounds and including such variations as 450 to 6Qo conical liners both with and without
spit back tuhes1 trumpets, hemispheres, mild steel and armor plate targets, 45 and 105mm
diameters, standoff, and charge shape, the estimated values were within the standard
deviation of the experimental measurements. Based on Figure 3, Figure 4 shows the
estimated penetration curves for liners whose non-rotated penetrations are 10, 15, 18,
20 and 25 inches. It is significant that vecy little advantage results by using
larger liners, or more perfectly made liners, if they 111\lSt be spun faster than about
40 rps.
Since rotation tends to cancel out the improvements gained by using larger or more
perfectly made liners, in order to achieve penetrations in the range of 15 or 20 inches
one must either design a round in which the rotation is kept below 30 or 40 rps, or
one IIIUst modif.r the liner and/or charge so as to shift the axis of optimum perfomance.
Fin stabilized rounds, "slow spin" rounds of the T-138 type, and double body projec-
tiles, which will be described presently, are ex8J11Ples of methods by which the rota-
tional rate may- be held at an acceptably low level. The use of fluted or serrated
liners is an example of an attempt to shift the axis of optimum performance from zero
to the higher spin rate where spin ·stabilized projectiles are feasible.

During the past year, Firestone, working with the groups at Carnegie Institute of
Technology and BRL, has manufactured and tested ten series of nuted liners of varying
design. At the present time a certain degree of success has been achieved, BIJi
certain design factors have been at least qualitatively established.

The penetration-spin rate curves for the ten aeries of liners are shown in
Figures 5 - 14, inclusive. The nute profile for each design is also shown in these
figures. All liners were conventional 105mrn copper liners machined from hard drawn
electrolytic copper bar meeting Federal· Specification QQ-C-50la. All rounds were
assembled in cylindrical mild steel test bodies, loaded with Comp. B at Picati.rley'
Arsenal, and fired at a 7 1/2-inch standoff against a mild steel target. The design
relationships noted will be described after the various curves have been examined.
Figures 5 - 9 show the behavior of 16 flute liners. The performance of the DRD 17-1
and 2 liners shown iii Figure 5 is very disappointing. The best performance was o~
12 inches penetration at 0 rps. Figure 6 shows the behavior of the DRD 17-6 liners.
This series, with a heavier wall, but being otherwise similar to the DRD 17-1 and 2
liners, has a penetration of 16 inches at 15 rps. Figure 7 shows· the perfo~ance of
the DRD 32-3 liners. These liners, differing from those of the DRD 17-1 and 2 series
by having a curved rather than a flat flute profile, penetrated 8.5 inches at 35 rps.
·Figure 8 is for the DRD 34-3 series. This series • having a deeper flute than the
DRD-17-6 series! but being s:lmilar otqerwise, has a penetration of 9.6 inches at 55
rps. The DRD 7!:1-2 liners shown in Figure 9 were the o~ liners with matching internal
and external flutes. The best penetration was 17.0 inches at 5 rps.

Fignres 10 - 12 show the behavior of three series of 36 flute liners. It shou1d

be noted that compensation in the positive direction ie obtained with 36 or more £1\ltes
 ~ \

~
">
1!0
\ I
~

~"' IS I
">
~
I..J I
~
~
' 10
I
h: .;oo J.-VALI..
z CONrRDL
~
~ s
~
0.:

0 L--------L--------L--------L--------L--------L--------J
-so 0 50 !00 150 zoo 250

REvs/sEc
Figure 5

~
~
">
~
20
i
16 -.028- .ISO

\
"'~ I

I
I

~
~
~• 10

~ $ z
i::
~
"~ 5

~
Q
.Jl:) 0 so 100 /.Sn i!SO

R.cv.fsEc.
"" Figure 6

3113
 PST 2 lSi!

;-.1
~ 20

"
VJ
<:::)
I

16-.028 -.IDQ
::::!
~
I
IS

~ CONTROL 0.100 H/ALL I

~
~ I
~
10

K
~~
5
~

~~ J

I
-so 0 so 100 ISO zoo ,;:so

REV SEC

Figure 7

<?<-----~

;S"f----+~-
I

oL-______J-_ _ _ _-L------~--------L-------~----~
..so 0 so
___. 100 zoo
REVS/SEC

Figure 8

344
LLLI
 .. weygy

\
zo~----,--

3
C'a TI?OL 0. tOO J4-'4LL
I

oL-----~~----~------~------~-------L------~
so 0 so /00 150 zoo
R£vjs£c
Figure 9

~·~------r-------~

I
I
36 -.010 -.100 I
'

\ 1

$-~------~------+-------+-------+-------+--------

oL-------~-------L------~--------L-------~------~
-.5£7 0 .5b 250

Figure 10

J45
 ~~------~------~

~~-----+------~----~~------+-----~------~

0~------L-----~L-----~------~L------J------~
-41:> 0 100 ISO 200 250

REVs/ SEC

Figure 11

~ 2D
'
~
..,
~
\-1 '
l
'
~
~ /S
l ~ ' ,.-.o.eo~~;t

(~ f\ I
I

~ 1,..-"""D .!l.l-3\ I
~
~ H) ., I }
' I
~
p

CONTROL O.I.SO w-"''

~ t--..
~
~ s

~
~

0
-so 0 KIO /SO .zoo
.li'EVSI SEC.
Figure 12

346

3 iii
 3 7 §

10

~
~
~ :s
~
~
I(

0
:So 0 /OD 2SO

RE.V~JEC.
Figure 18

zo
I
'

I r 1
~r-;·~·
I I
·~ "~~
d:O - .0.20 R .ISO

P.ltO tS".I-~
·,
10
.
~~~ ~

(CDNT,IfO,/. oJsi>·-u

0 SO /OD lSD 2DO ,ZSD

HE.v.sj.sEc

Figure 14

34~

~.......... .
 ..
S£42£$2.,,. 3 7 ( ,.
only when the flute orientation is the reverse of that used in the 16 flute series,
There are two series of DRD-20 liners, representing different manufacturing tolerances,
shown in Figure 10. The close tolerance liners penetrated 18,2 inches at 10 rps,
while the second series did not show a.ey compensation, The DRD-20 liners, historically
the second series tested in Firestone's program, showed the importance of carefUl
manufacture and all subse~ent series met the close tolerance specification. Figure
11, for the DRD 35-3 series, shows a max:lmum penetration of 15,0 inches at 10 rps.
Both flute depth and wall thickness are greater than for the first 36 flute series
(Figure 10), Figure 12 shows the DRD 33-3 series having curved flutes instead of flat
flutes, A penetration of 15,2 inches was measured at 15 rps.

Figures 13 and 14 show the behavior of two DRD 63-3 6J flute series of liners
which differ only in wall thickness, The thinner wall series liners penetrated 14,0
inches at 65 rps and the thicker wall series penetrated 15.3 inches at 25 rps,

The perfomance of these ten series are SU!IIIIlarized in Table I. Let us now con-
sider the effect of design variables upon optimum SPin rate and, independently, upon
penetration, A comparison of item 10 with 11 shows that for a given flute contour,
the optimum spin rate .is certainly a function of wall thickness, the thinner wall
thickness, the thinner wall compensating at the higher SPin rate, A comparison of
items 3 and 5 shows that, other things being equal, the optimum SPin rate increases
with increasing depth of flute, A comparison of item 8 with 11 shows that the
optimum spill rate depends upon the number of flutes, As indicated by the negative
sign in brackets after the 16-flute series, a given flute orientation causes· a 16
flute liner to compensate in a direction opposite to that of a 36 or 6J flute liller,
It appears that the inversion poillt is near 25 flutes,

A comparison of items 8 and 9 shows that with 36 fllltes the curved and nat f'lute
designs cause similar per.t'ormance, but as shown in items 2 and 4, with only 16 flutes
a curved flute is better than a flat flute,
I
Let us now consider the penetration obtailled at the optimum SPill rate, In each
case where the milllmwn wall thickness of the fluted liner is not less than .090 inch
the penetration is as good as the non rotatiilg penetration of a smooth liller having a
uniform wall thickness equal to the max:l.mwn wall thickness of the fluted liller. When
the milrlmum wall thickness is less than about ,090 illch the penetration is proportion-
ately reduced, Therefore, a minimum wall thickness of about ,090-olOO illch must be
maintained if fUll penetrating efficiency is to be obtained, Now, ill general the
optimum wall thickness of smooth 105mm liners is about .100 illch, Thicker walls re-
sult ill :reduced penetration, If it is desired to design a liller for operation at high
SPin rates relatively deep flutes will be required, and i f a millimum wall of .090 illch
is to be mailltained the nominal wall thickness will be fairly large, Under these con-.
ditions the. fluted .liner may be e:x;pecte.d to show a total penetration at the high SPill
rate equal to a non rotated smooth liner of equivalent maximum wall thickness (100% ·
compensation) j but the penetration will be somewhat less than the non rotated pene-
tration of a smooth liner of optimum wall thickness,

The IIRD 78-2 liners showed a rather anomalous behavior, This series did not show
an appreciable compensation. All series which do show appreciable compensation have a
flute profile with a unifomly illcreasing wall thickness, Because of the particular

348

s&Ziif .i"'
 azz s rim k'gwati~m

geometry of the llRD 78-2 liners, the wall thickness is approximately constant aver the
major portion of the flute profile. As a result no appreciable compensation is ob-
served even though a portion of the mass of the liner should have been compensated.
From this we conclude that compensation arises mainly as a result of a synnnetrical
variation in wall thickness, and that flute contour is of importance only in so far as
it ce:uses a variation in wall thickness. This statement agrees qualitatively with the
11 thick-thin effect 11 described by Dr. E. Pugh and his co-workers at Carnegie Institute

of Technology and tends to substantiate the recent predictions of Dr. L. H•. Thanas of
Watson Scientific Computing Laboratory as set forth in BRL Report No. 765. If further
studies confim that a symmetrical variation in wall thickness is a primary and con-
trolling pre-requisite for compensation, then a liner fluted on the inside only should
be as effective a.s one fluted on the outside only, except that the direction of com-
pensation should be reversed. Experiments now in progress should help to clarify this
question.

Another different approach to the question of spin compensation which has received
some attention in this laboratory is one employing a double body projectile. An ex-
terior portion has a conventional rotating band and spins at the usual stabilizing
rate while the charge carrying portion mounted on ball bearings is constrained by its
own inertia to rotate at a relatively lower rate. Figure 15 shows one type of such a
double body projectile. There are two major problems involved in the development of
this type of roUild. In the first place, either the bearings lliUSt be capable of
accepting the thrust load caused by setback without developing too large a frictional
torque, or both parts of the projectile must be subjected to the same acceleration so
that only a relatively low unbalanced thrust load need be sustained by the bearings.
It is possible to balance the acceleration by the relatively simple expedient of ad-
justing the base areas of the two parts so that the area of each exposed to the pro-
pellant gases is proportional to its respective mass. This required that the non-
rotating body fit inside the stabilizing section and, for a given bore diameter limits
the size of the shaped charge. On the other hand, i f the total thrust load can be
carried by the bearing system it is possible to use a larger size liner. Since the
thrust loads developed in the BAT weapon are not unreasonable and because penetration
is of supreme importance, the studies in this laboratory have thus far been directed
toward the development of a bearing system capable of accepting the full thrust load.
A simple amru1.ar pivot bE)aring of the type shown in Figure 15 has demonstrated con-
siderable promise. Lubricated with "molykote", a test slug with such a bearing
developed only 40 rps in the non-rotated section, even thou~ the stabilizing section
was spinning at about 240 rps. A ball bearing thrust system is under development.

The second basic problem with double body projectiles concerns their flight
stability and accuracy. If the round .functions properly~ and the charge carrying part
is indeed rotating at a relatively low rate, how large must the rotating portion be to
stabilize the round? Unfortunately we do not now have an answer to this question.

The effect of rotation upon penetration be shaped charges has been des.cribed and
a chart ha.S been presented which permits the designer to estimate the spin rate pene-
tration curve of a specific liner with reasonable accuracy •

.3~9

1!1! '
 -- ·-DRC-215 TE.E. ONE Rr~'t·-

\ DR.4'-27<4 THRUST BU..GiN6 O"'E. FU.O'D

\ \
"'N~20206 NEY..." D~PARTURE
eAt..l. BE:P.,.._.ING TWO REO'D

ROTATING BAI\OO,(ONE IN 20TURNS). ONE RE.Q'D.

1"-a L.H. NC-2 .JAM NUT,

1
ONE Rt.Q 0 - - STAKE
IN 3 PL-ACES AFTER FINAL A'::>SY

I
I
ONE. RE.Q'D LoRB·Ze9 NOSE ONE REQ'C.

! 307·~P.:,L: :U:,:G_._;O:,N:,:E: . .;~.! E,9='D,

ORA• •

Figure 16-Double body projectile assembly,

 nrcnFT 9 · r • ~., . .,.
·----- • H CWPSU?P

Certain of the design variables governing the perfomance of fluted liners have
been presented. It appears that 1D5mm liners of the type described here show can-
pensation only when there is a significant variation in waJ.l thickness across a flute
profile, that a minimum wall thickness of about .090 inch must be maintained in order
to secure full penetrating efficiency, and that the spin rate at which the best
average penetration ocaurs is increased by: (1) deepening the flutes, (2) reducing the
minimum wall thickness, and (3) by increasing the number of flutes when 36 or more
flutes are used, or by reducing the IIU!Ilber of flutes when 16 or fewer flutes are used.

TABLE I

Summary: Behavior of Serrated Liners

Item Drawing No. Fig. Geanetry vo p p

vo 0
No.

1. DRD17-l 5 1.6-.02 8-.100

n
0 12 18
2. DRD17-2 5 0 lO 18
3. DRD17-6 6 16-.028-.150 (-) 15 16.0 15.0
4. DRD32-3 7 16C-.028-.100(-) 35 6.5 18.0
5. DRD34-3 8 16-.070-.150 (-) 55 9.6 15.0
6.
1·
a.
DRD78-2
DRD20
DRD35-3
9
lO
ll
16cP-.034-.100 (-)
36-.010-.100
36-.020-0.150
'
lD
lD
17.0
18.2
15.0
18.0
18.0
14.3
9. DRD33-3 12 36C-.020-.150 15 15.2 14o3
10. DRD63-3 13 6o-.020-.100 65 14.0 18.0
n. DRD63-3 14 60-.020-.150 25 15.3 15.0

3,51
ntrr
 MCB!T SST¢¥ htn ,, ·~.

A ZERO CliDER THEORY OF THE INITIAL MOTION OF FLUTED HOLLOW CHARGE LINERS*

L. H, Thomas.

Watson Scientific Computing Laborator,y, New York, New York

ABSTRACT

'When a cased charge detonates we mq divide the early motion

of the casing into three parte: the initial state of rest or
steady motion, a confused regime of shock waves and renected
rarefactions, am a more or less steady motion under the continuing
pressure of the explosion gases.

For a light casing a plausible zero order theory treats the

whole intermediate stage as the motion of a single shock wave
forming a refracted extension of the detonation wave into the cas-
ing, which sets the casing impulsively into motion. The momentum
impulsively given in this approximation to a certain area of the
casing is proportional to its thickness, the physical reason for
this being the longer time that the material is confined by that
adjoining it when it is thicker.

This picture makes it possible to estimate the angular momen-

tum of a zone of a hollow charge liner immediately after the
passage of the detonation wave, and the angular momentum is not
likely to change much later as eq~ pressure on the outer surface
of the liner would produce no change. The assumption that nearly
zero angular m~entum is a necess~ condition for the form~tion
of a good jet nmi leads to a prediction of the effect of aqy given
nuting in counteracting initial rotation which is in qualitative,
and perhaps rough quantitative agl'Sement with observation .•

I f this zero order theory is borne out by further comparison

with observation, it ~ assist the design of efficient liners for
rotating projectiles,

1. The impulse ,wave in a light casing. (COIIQ)a're BRL Report 475 (1944) § 7.)
Consider first a shock front travelling with velocity U into metal of density p
0
at rest, and propelled by pressure P• If the metal density at pressure p is pl' and i f
the subsequent velocity of the metal relative to the shock tront, in the opposite di-
rection to u, is u , conservation of mass and momentum gives
1
P1 - u1 !-u, p0 (velocity relative to the front)

p Io
- ----
* Also published as Ballistic Research Laboratories Report No. 765.
353

t::s 1¥.
 ,____ v
EXPLOSION GASES UNDISTURBED- EXPLOSIVE
e
UNDISTURBED CASING

Figure 1

z z
GENERATOR

CONE MIDDLE
PROFILE

CONE AXIS

Figure 2

Figure 3

?F 1
 :rsen s 2

Po u • Pl ul

Po u2 • p1·Ul2 + P

The :Impulse per unit volume, p0 (U- u ) • pjrJ, in the direction oi' motion oi' the
1
shock i'ront o1' velocity U • { (pfp {1- pj~}) into the undisturbed casing.
0

We picture the detonation wave travelling through the explosive at a dei'inite

speed V with a definite pressure jump P extended through the casing by a 'virtual shock
wave' in the casing, with a pressure jump varying !'rom P inside to 0 outside, and a
mean oi' p " ~ P. We may suppose that this 'virtual shock wave 1 has a configuration
determined by its mean speed u, a property oi' the casing material, and that this
material is set in motion illlpulsive:cy with impulse pjrJ per unit volume peiJ>endicular
to the virtual shock. (Figure l.)

It the angle or incidence or the detonation wave on the undisturbed casing is e

and the angle oi' refraction oi' the virtual shock is It 1 as usual

v u
sin e .. sin jiJ •

2. Geometry ror a fiuted cailical casing.

We suppose the detonation wave in the explosive peiJ>endicular to the cone axis.
We rei'er the cone to its middle proi'il.e, hali'way between the :lnner and outer suri'aces,
drawn in a plane peiJ>endicular to its axis. At an;r point 0 in the proi'ile take
rectangular axes with Ox parallel to the cone axis, Oz produced negative:cy passing
through the axis. (Figilre 2.)

I1' the halt angle of the cone is II\ , the cone generator through 0 has direction
cosines

cos II\ 1 0, sin II\

and i f the tangent to the profile through 0 makes angle y with the y-axis, this tangent
has direction cosines
o; cos y, sin y.
The profile suri'ace normal then has direction ratios

-cos y sin a 1 -sin y cos II\ 1 cos a cos y

and the direction oi' progress oi' the waves along the surface has direction ratios
. 2
cos II\ 1 -sin a sin y cos y, sin a cos y

3SS

. 73?2 f'
Thus the angle of incidence of the detonation wave, a, is given b;y

cosa• cosxsina • sin a - -;:=:;::=co=s=a=:==::::::::==- •

J coi a + coi x sin a
2
J
coi a + cos2 x sin2 a
and the angle rj that the virtual shock makes with the cone surface is given b;y

sin rj " ¥cos J

a/ coi a + cos
2 2
x sin a ,
Impulse pjU, per unit volume, perpendicular to the virtual shock bas components
(pjU) sin rj in the direction of progress of the waves along the surface and -(p/U)
cos rJ perpendicular to the surface, so that its component in the y-direction is

_Usin rj sin a sin r cos r + tr cos rj sin X cos a

Jcoi a + coi r sin2 a Jcoi a + coi r 2
sin a

• - §i cos a sin a cos r sin r

2 2 2 + § 2 sin X 2
a 2
cos J 2 2 2 if-
cos a + cos r si.Il 11- ~os~
cos a+cos rsill a cos a+cos rsin a ~

per unit wlume,

If t is the thickness of the cone in the y-z plane perpelldicul.ar to the m1dd1e
profile, the volume per unit area or the x-y plane is t/cos r, so that the ~lse in
the y-direction per unit area of the x-y plane is, i f we abbreviate (Figure 3,)
sin a COB
cos a
X to e
( ~..---~-+-~-2 - ¥~} 1 - ~
t sin r ll_ lt v- t sin r
P. -,--==...,--r-.,._ _
cosrtJ l+~2 • coer'!JJ J2
2 11
.. 1-7+~ +v~

To get the angular moJDentwa given a zone of the cone, per unit distance along the
cone axis, we DDlst DDlltjply the above e:xpression b;y the distance hom the cone axis, r
~~ variations in which around the cone we ma,y reasonably neglect, and integrate with
respect to y completely around the developed profile.

3. Results.
The integrand found in § 2 has the same sign as X• If we integrate with respect
to II separately for parts of the profile with positive X and parts with negative y,

356
Figure 4

Figure 5

357

R&kli ·."
 Suppose now that t is constant; then the integt>and continually increases as y in-
creases and so e diminishes. Hence in a profile with t constant and y on the average
larger·when positive than when negative(: the total angular momentum impa_rted by the
virtual shock is in the y-direction. Figure 4.)

Suppose on the other hand that the thickness in the z-direction is constant rather
than the thicknei!IS perpendicular to the middle profile. Then t • T cos y 1 where T is
constant, and we have integt>and

c:-cl-
r?- >
rTp_cosa 7
lT sin a

This expression continuall,y decreases as y increases and so~ decreases. Hence in a

profile with T constant and y on the average larger when positive than when negative,
the total angular momentum imparted by the virtual shock is in the negative y-direction.
This seems to e:xplain qualitatively the difference in direction observed for different
profiles. (Fi~ 5.)

Taking p • 7 x 1010 dyne/cm2 , U • 5 x lo5 em/sec, and a density p • 9 gm/cm3,

~ • 1.5 x 104 em/sec. This iB aba.lt. 25 times the observed peripheral velocity, and 1/25

: not unreasonable for the mean of tan y (1 - ~)/ /1 -~ + e 2

+ ; c:- •

A more accurate theory obtained by setting up an artificial shock .f'ront· of con-

stant velocity in actual casing as a re.f'raction of the detonation wave might be worth
while.

358 ;

5 c.
---------- ·- ---- ---
 E CZC& G • IL: -

SHAPED CHARGE DAMAGE BEYOND ARMOR

D. R. Kennedy

u.s. Naval Ordnance Test station, Inyokern, China Lake, California

ABSTRACT

The development of shaped charge weapons has been concentrated

primarily on achieving max:iJnum penetration capabilities of the
charge, and little attention has been given to the effects of the
jets beyond the defeated armor, or, more specifically, within the
enemy tank.

Realizing the need for such information, a program was

initiated at the Naval Ordnance Test Station early this year t6
determine the comparative effects of various shaped. charge liner
materials and cone angles in producing damage beyond defeated
armor.

The charg~:~s :Investigated were identi.cal to the 6.5 in ATAR in

size and load. Copper, almninmn, and steel were utilized as cone
1 .
materials, with cone angles of 422• 60, 90, and 120 deg.
Three series of tests were conducted to determine (l) the
mass, size, nmnber, and spatial distribution of fragments ejected
beyond armor, (2) the distribution of jet fragments impinging on
angle plates beyond the primary armor, and (3) the pressures and
temperatures generated within confined steel targets by the various
shaped charge jets.

Almninmn liners were discovered to have unique damaging

properties when fired at extremely close stand-off distances
against defeatable armor. Indications are that copper liners,
although superior in depth of penetration capability, are de-
cidedly inferior to steel and aluminmn liners as damaging agents.

Shortly after the onset of the Korean War, the Naval Ordnance Test Station at.
Inyokern, California was asked to develop a shaped charge warhead adaptable for use
on the 5.0 in. High Velocity Aircraft Rocket (HVAR) motor. The design specifications
for the warhead required a 100% defeat of 13 l/2 in. of armor since it was thought that
this penetration would be required to stop the North Korean armor.

Accordingly, a warhead was developed which incorporated electric fuzing, a 6.125

in. 60o low carbon steel liner, and a 20-lb. Composition B explosive load. The weapon

359

(&i 77ITT2 f:O

 as issued to the service was designated the 6'!5 A.TAR (Anti-tank Aircraft Rocket) Mk 1
Mod o. During the cou.rse of the Station's evaiuation of the ATAR a test was conducted
in which several ATAR's were fired through 12 in. of Class B armor into a confined
steel target box encompassing a vol'lllllll or 240 cubic feet, or about the vol'lllllll or a
typical tank, It was concluded fran these tests that lethal fragments ejected beyond
the defeated armor were confined to a narrow cone and that the fragment ricochet was
negligible.

As a result of the obs.ervations made on these tests an exploratory program was

initiated to seek eane means of increasing the beyond armor damage by shaped charges
of the ATAR caliber. The effects of varying the liner material, thickness, and apex
angle were studied in a test program composed of three parts, (1) the determination
of armor damage, spatial distribution of· material beyond armor, and size and mass of
the material ejected beyond the armor; {2) the distribution and velocities of frag-
ments impinging on angle plates beyond the primary armor; and (3) the pressures and
temperatures generated within· confined steel targets (i.e., tanks) by the defeating
shaped charge jet.

All of the charges investigated were similar in size and explosive load to the
production 6V5 ATAR. Cone materials investigated included copper, steel, and
aluminum. Apex angles included 42 1/2°1 60°, 90° and 120°. All liners of a given apex
angle were identical in geometry and had a liner thickness of 6%f this thickness being
used in preference to the nonnal 3% (for copper and steel liners} since one of the
objectives of this program was to increase the production of material beyond the de-
feated armor. .

The first part or the test program utilized a large ice target structure designed
to receive and retain all of the material ejected beyond a massive plate of 6 in. Class
B homogeneous armor. A l/6 in. X 50 in. X 6o in. sheet or aluminum was placed 2 rt.
behind the armor plate to ~cord the spatial distribution of the jet and spall frag-
ments;. A 7000 lb. ice mass arranged roughly in the fonn of a cube was placed behind
the spatial distribution plate to stop and retain the ejected material.

In addition to the various configurations of 6 in charges discussed above, a few

service issue shaped charge heads were also fired for determination of the their com-
parative effect beyond the 6 in. armor. The sel"V'ice charges included the 3.5 in.
· Rocket HEAT·. M28A2 (Superbazooka), the 5VO Rocket Head Mk 25 Mod 1 (a .Nav,r head for the
HVAR), and the 6'!5 ATAR.
Scme of the conclusions reached as a result of this series of tests were1 (1) no
strict rel!ltionship between the hole dimension in the target plate and the amount of
material produced beyond armor were noted, although as a rule the larger the hole in
the armor, the gt'E!ater was the mass of ejected materialJ (2) doubling the liner thick-
ness for a given si11e of liner generally more than doubled the mass and number of
fragments produced beyond armor; (.3) steel cones produced greater llWDbers of lethal
fragmente beyond armor than did any or the other materials tested;. (4) the smaller cone
angles (i.e. 42 l/2° and 6o 0 ) produce ~ater llWDbers. of fragments (as well as being
capable of greater penetrations)J and (5) incJ:"easing the cone diameter increased thE
output of lethal material gt'E!atly in excess of the ratio of diameter increase. The
spatial distribution of fragments beyond armor revealed that the small angle steel
liners had a maximum spreild or approximately 110° with the area of heaviest fragment·
concentration falling in a spread of 6QO.

LWh£ w
 iOWP'Wl'I 2 .. I •••

. The second part of' the program, that of detennining the distribution and velocities
of' f'ra~ents impinging on a 45° angle plate placed behind the primary armor plate (6
in. homogeneous armor) vas on:cy- partially successf'ul. Attempts to record fragment
veloc"ities vera .unsuccessi'ul with the instrumentation than available. Photographic
records indicated, however, that steel liners vera superior to identical gecimetry copper
liners in the production of large masses of ricocheting fra~ents off the imgle plate.

The third part of the program, the determination of pressure and tenq>eratures
generated within confined structures by various shaped charge jets consisted of
statically firing each charge through 6 in. armor plate into a welded steel target
box 5 1/2 ft. deep, 5 1/2 ft. high and 8 ft. vide (a volume of 240 cubic feet). The
interior of' the target box vas painted white before each shot to record the incendiary
flash and fra~ent impact pattern. The target interior was instrumented to record
transient pressure and temperature rises caused by the entrance of the shaped charge
jets.

Same of the observations noted from this test were that (1) steel cone charges
produced a:verage temperature rises for all charges fired of 1028'? over a recorded
1 second duration as compared to an average rise-of only 3180 F for copper cone
charges including the 3.5 HEAT M2BA2 1 the 5'JO Mk 25, and the various 6 in. charges; (2)
the transient pressure rises a:veraged 18.8 PSI above ambient for the steel cone charges
as against 5.3 PSI for copper cone charges; (3) the temperature and pressure rises were
independent of the charge size in the range tested; and (4) the scorched surface on the
vall of' the target box opposite the jet entrance hole was about three times as large in
area for the steel cone charges as for the copper cone charges.

Based on the results of the first three parts of this program for the determination
of' jet effect beyond defeated armor, ve submit that as far as possible, the weapons
designer should consider the use of soft steel or cast iron for a liner material in-
stead of' the COIIIllOnly used but highly critical (as regards supp:cy-) copper. The loss
in maximum penetration ability (a criterion which we feel is great:cy- over-rated) "is
more than offset by the increased beyond armor effectiveness •.

)61

'(!@;Iii £3 L2! Z22i£ ~ •

.'
 I

THE DAMAGE EFFECTIVENESS OF SHAPED CHARGES AGAmBT TANKS

F. I. Hill

Terminal Ballistics Laboratory, Ballistic Research Laboratories,

Aberdeen Proving Ground·, Maryland

A:BSWACT

A SUIIIIIIary of the damage fran firing over one hundred 3.5 in.
rockets and l:lmited nwnbe:rsof 90mm Tl08 projectiles ve. the T26E4
tank is made. From these data, it is tentatively concluded that
interior damage after a perforation U,r both these projectiles is
comparable. However, the conditions to achieve a perforation with
the better fuzed projectile are not so rigorous. A comparison of
terminal ballistic data for the 90mm HEAT and kinetic energy pro-
jectiles is made indicating that general damage from perforating
kinetic energy projectiles is cons:l.derabl,y gl'eater than that . .from
HEAT rounds. A further anal,ysis is made in which the vulnerable
areas ·of these rounds are compared. This analysis indicates that
the vulnerable areas to mobility, firepower and total destruction
are not cons:l.derabl,y greater for the kinetic energy projectiles.
An explanation of this apparent anomal,y lies in the fact that the
principal targets inside a tank are anmunition, fuel and the
engine, and that the personnel are not the most important targets.
A brief estimate is made of the relative overall effectiveness of
90mm HVAP and HEA'l' rounds to show their expected relative prob-
abilities of a hit being a kill for these projectiles against the
M26 tank.

At its present stage of development the shaped charge round from a terminal
ballistic point of view is a cornparativel,y effective one against tanks. This statement
in no way implies, however, that a perl"oration by this round will necessarily destroy
a tank. It has somet:lmes been desirable from a terndnal ballistic standpoint to
approach the 100% level of e:r:t"ectiveness in a weapon for destroying its target once it
has. hit. Perhaps this would be desirable in antitank weapons provided the attendant
sacrifices in rate of fire, weight and possibl,y hit probability were not too severe.
With present weapons· such sacrifices are probably too severe. The e:r:t"ort to achieVe
even a perforation on some surfaces of tank armor has been great enough that the ob-
jective of just perforating the armor has been one that was a sufficientl,y high goal.
We may well ~k: ''What lies behind such an objective?" The answer is that a perfora-
tion of the armor usuall;r allows the projectile or jet to enter the tank and initiate
the forces of destruction that a tank carries within itself. These forces are
primariJ:y resultant from ignition of the ammunition and i'uel although they are not the
onl;r ones. A shot into the engine compartment ~ also destroy the tank's effective-
ness in a like manner. A. shot into the radiator or water jacket. will cause the engine
to bu.ri'I itself up or a .sh.ot into an oil line or cooler will eventuall,y cause the
engine and transmission to destroy themselves. At the present time, we have no anti-
tank weapon that is capable of inflicting. comparable damage on a tank independent of
363

cs m«
these forces. Some projectiles do more than others, but, still, those projectiles
capable of initiating a tank's potential of self destruction compare closely in overall
effectiveness caliber for caliber.

At the request of BRL, the Development and Proof Services here at Aberdeen have
been conducting extensive firings against some T26E4 tanks (essentially the Pershing)
with holl011 charge· projectiles to determine their terminal ballistic effectiveness.
Most of these firings have been perfomed with the 3.5 in. Rocket. Some have been per-
formed with some early 90mm noB projectiles and some static firings have been made
with the 6.5 in. ATAR. The tanks for the most part have been stowed with radio and
fire. control equipment, wooden dummy personnel, inert ammunition and in sane cases·
fuel. As long as possible the engine has been operating. Of course, after considerable
firing the tank .is more damaged and many of these things can no longer be simply re-
paired 'to put the tank back in operation.

Assesmnent · of damage is made on the basis of whether the tank is destroyed com-
pletely ( 11 K11 damage) or whether firepower ( 11 F11 ) or mobility ( 11M") damage have been
done. Since the tanks are usually fired on without IIIIIIIIUllition or f\lel stowed auxiliary
experiments have been performed to determine their susceptibility to attack by HC
ammunition. Table I lists the resUlts obtained from ~iring on gasoline and diesel
fuel. To obtain these data firings were made through armor at fuel in small cans
(5 gal.) 1 55 gal. drums and actual fuel tanks. The jets were aimed both above and
below the fuel level.
TABLE I

ltesUlts or Firing vs. Gasoline and Diesel Fuel

Diesel Fuel Uasoline

Armor Container No. of % Annor Container No. of %

Penetration Size Shots Fires Penetration Size Shots Fires
ins gal ins gal

7o25 5 8 0 7.25 5 3 100

2.75 5 5 0 3.67 5 5 100

7.25 ss 28 36 7.25 55 1 0

2.75 55 24 33 4.25 6 6 100

2o75 55 1!?. .!!Q. 3o00 M26 Fuel 1l .Jl2.

Tank
Gross Ave. 75 29 28 90%
 o £LIIS££Li2! I a,..
Examination of the above table shows the very high susceptibility of gasoline to fuel
fires as compared to Diesel fuel. It is notable that small containers of Diesel fuel
were not ignited. Of further interest is that the ignition of the fuel in the gasoline
tanks appeared to be independent of whether the fuel tanks were full or empty. Data on
the effect of fragments against ammunition are very extensive and will not be gone into
here. HEAT rounds will explode ammunition i f the jet passes through the case. The
small high velocity fral?)llents cannot be expected to ignite a case i f the fragment does
not strike the primer or the black powder booster. Our data on firing at ammunition
are limited. However it i.s adequate for the illustrations to be shown today.

Table II is a summary of the damage obtained in firing on the T26E4 tank with
various types of HEAT ammunition. As a gross summary it is of some value in comparing
just how much damage can be expected for these rounds. The data fran these tests are,
of course, used in a different manner to assess overall tank vulnerability because the
shots are distributed over the tank. Most of the damage obtained in the tests was
determined by the location of the shot and the shot was located at the places where
the greatest doubt of the result would be. For instance repeated shots into the
ammunition stowage from the side are of little interest.
TABLE II

Summary of Damage from Hollow Charge Rounds

Round Direction of No. No. of Per Perforation Per Perforation

Attack of Perforations Crew Co!!JEt
Shots Crew Eng. Killed Wounded M F K
Co!J!2t Comi!t
3.5 Ground Attack 70 32 7 .30 .69 .32 .25 .18
Rocket (45° from
Front & Side)

3.5 45° Air 16 8 6 .25 .75 .64 .53 .43

Rocket Attack

90mm Ground 10 8 0 .37 .63 .45 .54 .12

noa
6.5 45° Air 6 5 0 2.40 .6o .96 .85 .&J
Rocket Attack

90mm 100 yds 19 9 0 l.&l .55 .87 .82 .58

HVAP 1500 yds 26 15 2 1.00 2.00 .85 .73 .47

90mm 100 yds 23 17 0 1.12 .aa .90 .81 .58

AP

&!i . ti:
 Examination of the detailed data as well as Table II shows that i f there is a
significant difference in the damage from the 90mm TJ.08 round and the .).~ inch Rocket
is has not been shown with the first 10 rounds fired. However, the 6.5 in rocket does
vecy much more damage. The nature of the damage for all the rounds is such that some
wounding of personnel is possible a small distance from the direct path of the jet but
little other damage outside the jet path is obtained. This is borne out by the small
number of persons killed and wounded by the jet. At the bottom of Table II for the.
sake of comparison there is shown the damage for the 90mm HVAP and AP rounds. Note
that damage after perforation is greater for these rounds.

Because the above shots were selected and nonperforating shots were not even
tried the data of Table II are not a fair comparison between the projectiles. A
compariS>n of the average vulnerable area of the 3.5 in. Rocket and the 90mm M304 HVAP
proje<ltile has been made in which the vulnerable area of the M26 tank to the two pro-
jectiles has been computed. Vulnerable area is the product of the presented area and
the kill probability of a hit on that area. Table III lists these data:
TABLE III

Average Vulnerable Area of M26 Tank

Round Vulnerable Area
M F K
3.5" Rocket
90 HVAP MJ04 (500-1000 yds) 26.1 8.6
The reason for the similarity of the vulnerable areas is attributable to two th:!.ngs:
The first is that the chief targets in a tank are the engine, ammunition and fuel
rather than the people, and the second is that the· 3.5 Rocket can perforate a few
surfaces that the 90mm round cannot.

If it is assumed that the damage after a perforation is essentially the same for
both the 3.5 Rocket and the 90mm Tl08 round, an overall comparison can be drawn con-
sidering the expected ranges and angles of attack. This has been done using a range
and angular distribution function, f(R 1 9), obtained-from data on battles fought .Ln NW
Europe. The kill probability, PK.(9 1 R), considered as a function of range and angle
can be computed from the chance '&hat a hit will be obtained on a vulnerable area at
angle of attack 9 and range R. The comparison can then be based on the expectation of
a kill, PE• averaged over all ranges and angles of attack. This can be eJq~ressed as

PE .. J J
0
2000

0
2n
PK(a,R) f(O,R) d& dR.

The integration for range is taken to 2000 yds. because only a small % of shots occur
at ranges greater than 2000 yds. The values of PE computed for a single shot fired at
an M26 Tank by a 90mm M3 gun without a range fi!:Jder are as follows 1
 a; !Ldiiattm.

TABLE IV

PE
Type Kill 90mm HVAP 90mm HEAT
M304 TJ.08

M .83 .&J
F .79 .54
K .73 .45
The numbers of this tabla are satisfactory for a comparison but their value tends to
be high because the vulnerable area is asswnad to be concentrated in a given area.
This overall comparison shows the importance of having a high velocity and good
accuracy as well as considerable terminal ballistic damage effectiveness in a round.
WhUe sane sacrifice in penetration perfonnance of the Tl08 might be desirable to
gain ballistic damage on an M26 tank it probably wUl not be so for the same caliber
round were it to be used against a more heavily armored vehicle such as the Stalin III
and it is doubtful if any of its velocity or accuracy characteristics should be
sacrificed to such an end.

r
,..,WttlDEIC lilts
 SHAPED CHARGE DAMAGE TO AmCRAFT STRUCTuRES

G. c. Throner

u. s. Naval Ordnance Teat Station, Irvokern, China Lake, California

ABSTRACT

The search for an ideal warhead for ground-to-air guided

missiles has been centered upon four different types of heads,
the moat promising of which, according to some investigators, is
one using the shaped charge principle.

Since 1950, the Naval Ordnance Test Station has been engaged
in studies of the effects of shaped charges fired at aircraft
structures from long standoff distances. Tests have shown that
the hypervelocity shaped charge jet fra~ents can be directed
against aircraft targets 100 to 150 feet from the point of deto-
nation and produce K-kill damage. The type of damage incurred is
referred to as "vaporif'ic11 damage, and is characterized by a
brilliant nash which envelopes the target and produces severe
structural and skin damage to the aircraft.

Present investigations lead to the hypothesis that 11 vaporific 11

damage is the result of target material combining explosivezy with
ita surrounding atmosphere because of the high energy imparted to
the material by impacting jet fragments. Single hypervelocity
(8 1 000 to 15,000 f'ps) pellets ha:ve been fired from IIPugh chargesll
against enclosed, multiple-plate aluminum targets containing both
inert and reactive a'bnospheres such as helium, nitrogen, oxygen
and engine exhaust gas. It was shown that llvaporific" flashing
and the accompanying damage can occur only in an a'bnoaphere which
will react with the target material when the latter is produced
in finely divided fom by high velocity impact. The possibility
of protection by purging the interior of aircraft wings with
engine exhaust gas is cited.

Since ground-to-air guided missiles at the present stage of development are unable
to approach an aircraft target sufficiently close for effectiveness of a blast-frag-
mentation warhead, it is intended that a shaped charge warhead be gimbalsd within the
missile and guided so that it will track the target and detonate at the point of
closest approach. If the jet misses, this warhead would be no less effective than a
blast-fr~ntation warhead because a heavy, cont.rolled-fra~entation case for the
shaped charge will be incorporated.

mwnmmr·e.
 G?T liS a·( :":•

K-kill damage is a necessary requirement ~or this warhead because in operation

against flights o~ more than one aircraft, each individual target must ~all immediate]Jr
when hit so that succeeding missiles will not be wasted on the same target.

A variety o~ shaped charges ranging ~roni 4.6 in. to 8.5 in. in diameter and with
aluminum, steel and copper cones have been ~ired against small ~ighter aircraft at long
stando~~ distances o~ 100 ft. and greater in a cursory and random exploration o~ their
destructive capabilities in this application; most o~ these charges having been origi-
nally designed ~or use against armor targets at close stando~~ distances.

In the near future, a preliminary model o~ the above described guided missile
warhead will be tested against B-29 bomber air~ame sections. This preliminary test
model has an aluminum cone 15 in. in diameter with a 120 lb, high exp~sive load, much
larger than any o~ the charges tested against aircrafi so ~ar.

The specliications o~ same o~ the more successful char~s ~ired against .aircraft
are given in the ~allowing table:

Chg. Type Wt, o~ Length Cone Cone Cone Cone Wall Cone
lSSU!~•· Expl, Expl, o~ Che; Dia. Wt, Angle Thiclmess Met 11
A B 23.6 lb 15 in 6.125 in 6o9 ~ 120° 6% Aluminum
*B B 20.0 15 6.125 1375 6oo 3% Steel
c C-3 69.0 22 8.5 3493 800
3% Copper

*The Type B charge is a production model 6'5 ATAR head stripped 0~ ogiva and fuze.
This charge has a cast, unmachined steel cone,

The target airplanes were strippad o~ guns, ammunition, fuel, instruments, radio
and in same cases the engine, but were structurally sound, They were plac.ed in
appropriate positions on the ground, usually with nose down and tail elevated to ex-
pose the greatest wing area, Charges were supported statically at the indicated
stando~~ distances, ·

'l',ype A Charges: Four 6 in. aluminum cone charges were ~ired against F6F fighter
aircrafi at 100-ft, stando~~ distance,
No, 1 The wing was sheared o!~ at a point approximately 6 ~t. inboard ~wingtip,
The remaining part o~ the wing sustained removal o~ large areas o! skin and severe
damage to internal structure. Assessed as K-kill.
No. 2 Aimed at wing-"to-fuselage juncture just afi o~ engine compartment skin at root
o! wing blown wide open, ~ire set to empty fuel tank, main spar cut; electrical lines,
hydraulic lines and oil cooler des.troyed, engine accessories damaged, K-kill.
NQ_._) All o~ wing structure removed in :impact area except ~or main spar and leading
eage: Main .spar was cracked, K-kill,
No, 4 Aimed just below horizontal stabilizer shooting perpendicular to axis o! fuse-
lage. Severe damage to tail o~ fuselage, all controls destroyed, Skin o~ horizontal
stabilizer bulged outward by internal pressure. Assessed as K..;kill.

370
 ~e B Charges: Four of the 6V5 ATAR heads using cast, urunachined steel cones
were li'ed against F6F fighter aircraft at 100 ft. standoff distance.
No. 1 was a:lmed at a wing. The gun and ammunition compartments of that wing were burned
out and the main spar was cracked. Assessed as A-Possible Kill. Failure dependent
upon presence of amnruniti.on in wing.
No. 2 also a:imed at a wing. Approximately 16 ft2 of skin was removed from both top and
bottom surfaces of wing. Three ribs and some stringers were destroyed, however the
wing retained most of its strength since the main spar was undamaged. A-Possible Kill.
No. 3 was aimed at side of fuselage 3 ft. aft of canopy. A peppered area about l ft.
high and 2 ft. wide was produced on the entrance side. On the exit side a hole 3 ft.
in diameter was completely blown out. Fuselage structure was weakened to point of
failure in fiight. K~kill.
No. 4 a:!Jned at side of fuselage 1 ft. aft and 2 ft. below canopy. All cockpit canopy
glass. was blown out except for heavY, bulletproof windshield. Fires were set in cock-
pit and along fabric-covered ailerons. · Inside of fuselage was buckled and bulged and
heavY damage was sustained on the side away from impact. K-kill.

~e C Charge: One 8.5 in. copper cone charge loaded with 69 lb. of composition
C-3 pastic explosive was fired against an F6F .fighter at 100 ft. standoff distance •
.This charge was aimed at the folding wing juncture and removed all the wing structure
in a path 2 ft. wide except for the spar. Debris from the wing was found 150 ft • .from
the plane. This round was assessed as K-kill.

CONCLUSIONS

It has been demonstrated that shaped charges fired at long standoff distances are
capable of producing K-kill damage to aircraft.

Experimentation to date indicates that the following conditions are necessary for
innicting K-kill damage to aircraft of the F6F type construction:

a. The charge IIIUSt be aimed so that the jet will strike the aircraft.

b. The charge d:Lmneter IIIUSt be 6 in. or greater. This fixes the explosive
load at a minjmum of 20 lb.

c. The standoff distance IIIUSt be 125 ft. or less.

d. The shaped charge cone material should be aluminum, steel, or copper in

order of decreasing effectiveness.

e. Extensive 11 vaporific 11 dam~ can occur only in an atmosphere which will

react with the target material when it is produced in finely divided form by high
velocity impact.

371

117
 -- ---------------

I
~

'·
 Mi~r
7

' ' '

DEFENSES AGAINST HOLLOW CHARGES

R, J. Eichelberger

Carnegie Institute o! Technology, Pittsburgh, Pennsylvania

ABSTRACT

A resume o! the principles set 1'ortb in recent reports on

this subject are given, together with some more recent results
on the use o1' glass as a protective device. (See bibliography)

Glass blocks or plates fastened on the outside o1' homoplate

and protected on the outside surface by a relatively thin homo-
plate are believed to provide the most practical protection.
Aluminum armor fastened to the outside o1' basic homoplate pro-
vides several advantages over glass but would add more weight !or
the same protection.
Under certain circumstances patterns o1' spikes or layers o1'
explosives between steel plates may provide very low weight pro-
tection.
In principle any low denei ty material can be used !or pro-
tection, since the weight that must be added !or a given pro-
tection is proportional to the square root o1' the density o1' the
protecting material. The protective qualities o1' aluminum are due
to this relation but the protective qualities o1' glass are much
greater than this relation predicts, Glass is the one known
exception.

This discussion o1' means o1' defense against shaped charges will involve only the
phase o1' preventing perforation o1' a vehicle that is struck by such a weapon. Such
aspects o1' the problem as reduction o1' the probability o1' a hit by increasing maneuver-
ability or decreasing damage by charges that do defeat the vehicle by appropriate dis-
position o1' components, personnel, and contents will be completely neglected.

The problem o1' protecting a vehicle against shaped charges is complicated by the
!act that such charges are not exclusively artillery weaponSJ they are at least equally
important as infantry weapons, Consequently, a tank must have its entire surface pro-
tected, insofar as this is possible, instead o1' having only selected vulnerable areas
heavily protected while the remainder is only lightly armored, The weight oi' pro-
tection required is nmch greater on this account, In !act, especially !or the lighter
vehicles, the weight o1' armor that nmst be added is prohibitive unless some active
means o1' defense is used, He.avier vehicles with thicker basic armor could possibly be
protected adequately by passive means.

313
 ? ppppzt: 7

-•
.• II.

- m
~
~
I
I
. --..,
t I
I
I II.
.......
~ I ...
0

..-•....., I
I
I
.."'
a
Ill

0 • I
I
/

/
I
I
I
/
I

I
I
I
I

/
/ ....
..
/
/ ...
II.
~
k

!:.

D..~

ct+...
.
0..
•
"'
D..
II

"'
D..

-rn

374
C&CZ&Elt£ a
 As indicated above, the means of protecting an armored vehicle naturally divide
into two categories: passive armors that simply absorb the energy of a shaped charge
jet, and active protective devices that induce malfunctioning of the charge or destroy
the jet after its formation, We shall consider the passive types first,
The protection to be expected from ordinary materials added t.o a reasonable
thickness of steel basic armor can be predicted to a fair degree of accuracy by means
of the residual penetration theory devised by Fireman and Pugh. Very briefly, the re-
duction in the residual penetration into the steel backing is computed as illustrated
in Plate I, The data that must be available are the penetration-standoff curve for the
weapon under consideration, fired into material similar to the basic armor to be pro-
tected, and the density of the proposed protective material, The reductions in pene-
tration are in the same ratio as the square roots of the densities of the protective
materials,
If an air space is left behind a layer of protective material, the residual pene-
tration is computed from the ratio of the residual penetration immediately behind the
protective layer to the penetration at the same distance from the charge with no inter-
vening protection, and the corresponding point on the penetration-standoff curve, That
is, as shown on Plate I,
PR
psI = p s • "'PR~+:-li.,.....P~p

With this addition, the theory can be a:(>plied to layers of various materials and to
spaced armor (including skirting plates) as well as to homogeneous layers of protection.
The relative ability of some typical materials to protect mild steel backing is
shown by Plate II. The thickness of almninurn, steel, or lead required to protect any
given thickness of backing against the standard laboratory charge can be taken directly
from the plot.
Very roughly, the theory yields the result that the thickness of material needed
to provide a given degree of protection varies inversely as the square root of the
density of the material, and the weight varies directly as the square root of the
density.
The manner in which skirting plates and spaced armor provide protection is illus-
trated by means of the residual penetration theory in Plate III, Against the labora-
tory charge fired at 2 in, standoff, a 1/4 in, steel skirting plate placed 9 in, fran
the basic armor will reduce the penetrati·on fran 5 1/4 in. to less than 3 inches,
Similarly, a series of 12 - 1/4 in. plates spaced 1/2 in. apart will reduce the pene-
tration from 5 1/4 in, to 1 1/4 inches, These figures sound rather good, but it must
be observed that a considerable part of the effectiveness of the skirting plate and of
spaeed armor lies in the rather rapid fall in the penetration-standoff curve at large
standoffs. Against really well constructed charges, the skirting plate would be
practically useless and the spaced armor would have little or no advantage on a weight
basis over solid armor.

315

qpp
 rnsww u s r ., I ,,

Steel Cones
6

5
"'-~
•
:!
~
.••
•
;; 3
•
~
•
"' 2

2 3 4 6
'
Tartet Thlckn'ass
0 2 3 4 5 6 7 8
Stondotr lin. I

Figure 2

S1ee\ Cones
6

5r .rV's.o. Curve
~
1/4 in. plate
~

~
~
0

~
4
------
.
;;
~

""
3
;;
...:;;"
..
a:
2

2 in. Air
0 3 4 5 6 7 ', B 9
Target thickness (in) ..... ,
o,L_____+-----*2----~3----~4~----5~---,6~~~7~---,st---~g~--~R----.ITI~
Standoff (in.)

Figure 3

376
SSi s;zpq.
RESIDUAL PENETRATION (IN.)
ro u *• 01

<
a>
o CO
<O
<D .
1<=
«_ m
ro ^XdlZ' "
O "" — a
(0
7
O
fl»
1

Ol
 !?GPEE'7'
7

25 r
:;;
~
E
h
:::: 20 r
""'c
,J
Q,

,J
c
15 r
""
co~
;:
Ci
!!!
~
0
,. 10 r I STEEL
"zu;
f---- ~·78-'- --------
""'
.
"z
"'f
Q,
5 r-

c 1------- -f!!·!..,--- - - - - - -
I _[_ I _j I I I
0
2 3 4
POSITION OF PLATE IN STACK
5
• 7 8

Figu'i-e 6

280

260

240

220 .o=l2

200 JO= 10

180

..
,.
•t
€120 1"•4 Spac•d A.r•or
•
~100
1/4 in. piGtel IPOCed 1/2 in •

Aluminum-
..
ii 80
~

60 G Sloss
II= 2

40

20

0~----~----~----~----~----~~----~----~--~~--___j
1 2 3 4 ~ 6 7 8 9

Tllickneu Protection Un.l

Figure 7

378
 Furthermore, the spaced armor is under no circumstances as good protection as an .¥
equal weight of homogeneous material. In Plate III, the spaced armor has been so 1
arranged that it is generally equal in weight to the same thickness of solid aluminum. '
The plot shows that the aluminum provides considerably better protection than the same
weight and thi-ckness of spaced steel armor.

It should be noted that, of the ordinary passive protections, aluminum probably ~~·
constitutes the most practical compromise. It is ·effective against weapons other than ,
shaped charges, presents little difficulty in application, and is not exhorbitantly I
expensive. In order to be efficiently used, however, it nrust be backed by a reas.onable
thickness of stall armor which will stop the relatively slow tail of the jet. 1

The only passive protection that has exhibited any extraordinary stopping power
against shaped charge jets is glass. A typical exper:lment demonstrating the effective-
ness of glass is illustrated in Plate IV. In the specific exper:lment, against copper
M9Al cones, a 2 1/2 in. brick of glass allowed only 2. 7 in. average residual penetration
where as the same thickness of steel allowed 5.3 inches.
A more general. picture of the effectiveness of glass is shown in Plate V.
According to the curve for glass, which is based on experimental data (a total of 100
shots), glass is more effective than steel on a thickness basis in thicknesses less than
4 1/4 inches. Considering its density alone, glass should provide the same degree of
protection as aluminum. The plot shows ho~r far more effective glass really is. As a
matter of fact, in thicknesses less than 3 1/4 in., glass is even more effective than
an equal thickness of lead.
A still more graphic illustration of the extradordinary stopping power of glass.
can be obtained by comparing individual layers of the protective material il'l the tar-
get. This has been done by using the available observations and computing from the
observed reductions in residual penetration the density of a protective material of
the same thickness and used in the same position in the target that would provide the
same reduction in residual penetration. The result is referred to as the apparent
density of the plate of glass at that position in the target. In plate VI, graphs are
shown indicating the apparent density of 1/2 in. plates of glass tested against M9Al
copper and M9Al steel cones. It is of special interest to note that the glass is con-
siderably more effective against the copper jet than against the steel jet, and it
should be remembered that previous comparisons of glass with other target materials
have been made on the basis of tests with steel jets rather than copper. The plots
in Plate VI show that for copper jets or -steel jets the glass is much more effective
at the front of the target (that is, against the front of the jet) than toward the
rear. The apparent density at the front of the target when tested against a copper
jet is greater than 20 and against steel the apparent density is greater than 10, as
compared with the real density of only 2.6 gmfom. It should also be noted that even
the 8th plate in the test has an apparent density greater than the real. density 2 .6.
It nrust be remembered that comparisons of this sort show the advantages of glass
on a thickness basis and that the disadvantages usually inherent in the use of pro-
tective materials of high density are not encountered in the case of glass. As a
matter of fact, a considerable weight saving is obtained, as has been mentioned before. )
Weight for weight, at the front of the target a given amount of glass will provide a
protection against copper jets equivalent to that of more than 5 times its weight of
steel. The advantage is slightly less when used against steel jets but is still quite
considerable.
319
 nit:r t'crm [
1

The relative effectiveness of the various kinds of passive protection is illus-

trated in Plate VII. Here, a specific example has been considered - the protection of
1 in. of basic armor against the CIT standard laboratory charge lined with an 1~9Al
steel cone - and the thickness of protection plotted against the weight. The points
on the curve occupied by 11 plastic 11 homogeneous materials of various densities are
indicated, and points lying well off the curve illustrate ~~e low efficiency of spaced
armor and the very high efficiency of glass.
Plate VII shows clearly the need for a compromise between high density protection,
of which only a relatively small thickness would be needed, and low density material
which would allow a saving of weight. As a rough estimate o:f requirements :for defense
against a 3 1/2 in. Bazooka, one could multiply both thickness and weight qy a factor
of 31 remembering that the basic armor protected also scales and is equal to 3 inches.
It is evident that, o:f the so-called "plastic" materials represented by the curve,
aluminum represents about t.lte ~ost practical compromise, and that a thickness of about
18 1/2 in. weighing 255 lbs/ft 1 neglecting any necessary :fastenings, would be re-
quired. Mr. F. I. Hill of BRL has estimated that about 35 to 38 tons o:f aluminum armor
would be required to adequately protect the TL2 and T43 tanks against the 3.5 Bazooka.

If glass were to be used, on the other hand, Mr. Hill has estimated that only 16
tons oi' added armor would be needed to protect the M42 and 17 tons to protect the T43.
In other words, the use of glass in place of aluminum would permit saving of weight of
. mare than a factor o:f 2,

It is evident that, in spite of practical difficulties in its application, the

advantages of glass as a protective material cannot be neglected, It-provides, in the
ex!llllple chosen, a thickness advantage equal to that of a material of density about 7
gm/cc, and the weight advantage of a material of density less than 2 gm/cc. In order
to realize these potential benefits, however, certain prec!tlltions must be taken: a
metal face plate must be used to detonate the charge before it damages the glass, and
the glass must be used in blocks of fairly small area, with separators of shock absorb-
ing material between, to prevent damage over a large area due to a single strike.
Same of the advantage of glass is lost because of the need for these extra accoutre-
ments, but rather detailed engineering on these aspects indicates that only a small
fraction of the advantage is lost.
In practice, it is likely that only a limited thickness of glass is really desira-
ble, since aluminum gives comparable shaped charge protection at the greater depth in
the target and would provide far better protection against A.P. and similar "Kinetic
energy" projectiles, It should be a simple matter to design a combination of glass and
aluminum which, together ·with the basic steel armor, would provide complete protection
against A,P, projectiles as well as against shaped charges and squash head,

It should be noted that the advantages described have been observed only in solid

I
:.

1
· glass. Glass products such as fiber-glass and Doron provide only the protection to be
1
expected from plastic materials of the same density; they exhibit none of the superiority
of solid glass.
;

)80
 Several active means of defeating shaped charges have been tested to some extent,
The one that has been most extensively tested consists of a system of heavy spikes
fastened to the basic armor, Plate VIII shows a small test plate made for testing
against M6 Bazookas. The spikes act ~perforating the windshield of the projectile
and defonning the liner before the detonation wave reaches it, thus destroying the
symmetry of the jet fonnation process, Obviously, several factors are vital to the
success of such a method, The spikes 11ll1St be able to enter the windshield with a
minimum of resistance and lliUSt be strong enough to resist bending or fracture during
the process.

In. laboratory tests with dynamically fired M6 Bazookas, the criteria for good
spike shape, spacing, and length were found and, ~ observing the simple design llrules••
so established, it was possible to achieve 100% effectiveness in protecting 1 1/2 in.
of steel. The test plate of Plate VIII, after the test, is shown in Plate IX, The
penetration in the plate was negligible and the damage to the spikes themselves can be
seen.

The M6 Bazooka, with its low flight veloc:l.ty and slow fusing, was obviously a
relatively simple shell to defeat ~ this means. L:l.mited field tests against a variety
of larger American and Ge~an weapons indicated a surprising degree of effectiveness
in defeating even such shells as the high velocity M67 and the point-detonated
Racketenpanzerbuchse, On1,y the latter weapon failed to be consistently defeated ~
the spikes designed in ac.cordance with the very simple 11 rules of thll!ll.b 11 ,

vnxtle it would seem that spikes could eas~ be defeated ~ the fast modern fuses,
the very considerable saving in weight that might possibly be gained makes this con-
sideration still worthwhile, Mr. Hill has estimated that the added weight needed to
protect a T42 tank would be about 2 1/2 tons (compared with at least 16 tons of

I
passive armor), or somewhat less than 4 tons for the T43 (compared with almost 17 tons),

Another means of active protection tested consists of thin layers of explosive be-
tween layers of inert material (aluminum or steel). The jet from the CIT laboratory
charge is completely defeated by five layers of Pentolite 1/10 in. thick interspersed
with 1/4 in. steel plates. The practicality of such protection is qu~stionable, but
its evidently general effectiveness makes it appear worthy of further consideration,

At one t:l.me, it was thought that the efi"ectiveness or a jet could be reduced ~
using oxidizing agents, Thorough testing has demonstrated that there is no benefit to
be gained from such materials,

The possibility of arranging small shaped charges on. the surface of a tank, to b03 ..V
detonated ~ the attacking charge and destr'oy its perfonn.ance in much the same WI!¥ as
spikes, has also been suggested. No tests have been made on this idea,

In general, it appears that the most certai.n:cy practical protection against shaped
charges would consist of a combination of glass and aluminum_, of appropriate thick-
nesses, together with the basic steel amor. Such armor necessarily weighs a good deal,
however, and a decision is necessary as to whether the resultant loss in maneuverability
of the vehicle is serious, Protection can possibly be obtained with 11lllch less added

381
..• .. -c:::-
. ....
Figure 8

Figure 9

382

sgrwsnn a
weight by using spikes, but their usetulnes.s is evidently limitedJ they may be entirely
ineffective against the newer ruses. Layers of explosive give generally good protection
and woul.d be effective against any type of weapon, It may be possible to devise means
of making this technique practical, in which case explosive layers may well afford the
lightest and thinnest defense.
--.------ -

BIBLIOORAPHY
BRL Technical Note 281 by F. I. Hill, "Proposed Program for Development of Armor to
· Defeat HEAT and Squash-Head Projectile s 11 , August 1950,

CIT Special Report, Contract W-36-061-cmD-2879, Eme:rson M. Pugh, 11 Survey of Devices

for Protec.ting !mor against Shaped Charges", June 30, 1949.

OSRD No. 6384, Final Report, "Protection Against Sbaped Charges", November 1945. /

383

d awns
&ILL I
0
5
I ,.

1
 APPENDIX ~
BIBLIOGRAPHY AS DISTRIBUTED BY EMERSON M. PUGH AT THE SYMPOSIUM

BRIEF HISTCRY AND BIBLIOGRAPHY OF CLASSIFn.:D REPORTS LEADING TO THE PRESENT

THEORY OF METALLIC JET CHARGES

A number of papers have now been released and published on the general subject of
the theory of the action of lined cavity charges. The principle applied in releasing
this information to the general public has been the desire to make fundamental in-
formation available for the use of the industries that might have need for it and to
stimulate research. It should be remembered, however, that material which has been
released in this manner is usually out of date by the time it is published and ma;y
con~in some serious omissions of material which, if published in unclassified papers,
woul.d jeopardize national defense. It is, therefore, highly desirable that individu-
al.s working in this field, who are cleared for access to confidential and secret docu-
ments, consult.the original classified sources.

Those -who are just entering this field will find that there is a large agregate
of useful and reliable information available to them. The information is so extensive
that it is believed to be worthwhile to pick out reports that represent peaks of pro-
gress in the understanding of the phenomena, and to list them in one document for the
use of the workers in the field.

The list that follows represents a first attempt at a compilation of these peaks
of progress reports. It is recognized that this list is quite incomplete because the
time for compiling a complete list has not been available. Also, other. individuals who
have worked in the field will have different ideas concerning which reports represent
the peaks of progress, and the author would greatly appreciate receiving comments and
suggestions for modification of the list. If the list can be made to represent the
views of those scientists who are most familiar with the field, it should be worthwhile
publishing it in a special classified document. Such a document should be useful not
only to new people starting in the work, but should also be useful to refresh the
memories of those of long experience.
The author has had so little access to German documents that the following list
is certain to be deficient in Geman references. It contains chiefly references from
workers in England and the United States and the English references are known to be
incomplete.

OTIB-U68, "Substantiating Material in Support

of Eval.uation of Compensation in
Favor of Explosive - Experimental
Company and F. R. Thomanek."

This report mentions a secret patent (404/39) issued August 14, 1936 to captain
Wimmer based on work carried out from September to November 1935. This work is
summarized in a report but the author did not have access to it and therefore no
description of the patent can be given here.

385
 prr 2 •. If ,. n.

The Westralisch-Anhaltische Sprengstoff-Actien-Gesellschart of Berlin patented in

Germany an explosive charge having a cavity of conical or other suitable shape either
hollow, filled with an inert material, or lined with sheet metal. The effect of the
liner was probably not realized as it was not stressed.

Experiments in the CTR in 1924 showed that the lined charge produced a substantial-
ly different effect than the unlined charge. However, a basic knowledge of the pane-
. tration of lined hollow charges was not reported until February 15, 1939 by WaPruf' 5.
In a German Army Ordnance Dept. (WaF) report of June 10, 1937, the value of a
space or standoff between charge and target was recognized in tests using hemispherical
liners. A 7 .5c:m aluminum lined .shaped charge was ordered in l~ovember 1938 and adopted
in July 1939 after trial by the troops. Hollow Charge 15 was ready by the end of 19b0
and tested on the M~t~inot line in March 1941. Oil June 15, 19b0 the Waf reported the
superiority of sheet steel over cast iron for hemispherical liners. A report of August
3,1940 indicates that conical liners with apex angles between 20° and 4SO were investi-
gated. Best results were obtained with 30° cones of 1 to 1.5mm wall thickness at 25
to 30mm standoff.
An evacuated cavity was suggested by F. R. Thananek and investigations were made
along these lines until May 1938, but no improvement in penetration was obtained. In
19b0 Thananek also suggested the use of a conical liner whose wall thickness increased,
from apex to base.

R. w. Wood, "Optical and Physical Effects

of High Explosives," Proceedings
of the Royal Society, LOndon 1936,
Band 157, s. 249. (1. Plastisches
FliBen von Metall)

Various el.-periments performed with fuzes with a copper lined indentation in the
bottom are reported. These fuzes upon collapse form a single slug rather than a jet
and slug as do the conical lined cavity charges. Concentric rings were etched on the
liners of some fuzes and their slugs collected in water or cotton. A theory for the
formation of the "pellet" was developed and substantiated by firing fuzes with smaller
charges to obtain various stages of formation. Wood states that the pellet has a
velocity of 1800 m/sec., but he does not indicate haw the velocity was measured. Al-
though the fuzes used in this report were quite different f'ran the metallic lined
shaped charges of tod~, Woo4 1s investigation is interesting and the results agree
remarkably well with observations made considerably later.

AC 1591, Advisor,y Council on Scientific

Research and Technical Develop-
ment, "The Mechanism o{ Jet
Formation in the Munroe Effect",
(Pcys/Ex. 216; OSRD Liaison
Office WA-87-13) Jan. 12, 1942.

386

ana
AC 1749, Advisar.y Council on Scientific
Research and Technical Develop-
ment, "Supplementary Experiments
on Munroe Jets in Air", (Phys/Ex.
234; OSRD Liaison Office W-156-26)
Feb, 25, 1942.
AC 2461, Advisory Council on Scientific
Research and Technical Develop-
ment, "The Comparative Properties
of Munroe Jets Formed by Various
Lined Hollow Charges" (Phys/Ex.
303; SC 4/25; OSRD Liaison Office
WA-556-20B) Aug, 4, 1942.

Above three reports by W, M•. Evans and A, R, Ubbelohde,

These reports show penetration versus standoff curves with coordinates in terms
of the diameter of the charge. Evans and Ubbelohde (probably independently) dis-
covered the increase in penetration with standoff and implied that a linear scaling law
is applicable, They made estimates of the velocities of these jets and started measure-
ments of these velocities, They also introduced the notion of optimum thickness of
the liner and made some determinations of this quantity, Spherical caps and wide angle
cones were generally used by this group. This work furnished the inspiration for the
beginning of the Explosives Research Laboratory at Bruceton, Pennsylvania in 1942.
Work was touched off in the United States by patents of H, H, Mohaupt (a Swiss
national: U.S. Patent No. 2,419,414; applied for on October 31 1941) and W, Blackinton
and John J, Calhoun (U,S, Patent No, 2,413,680, applied for November 21, 1942). It is
pJCObably because these patents used conical liners having small apex angles (rou~
450) that most of the experimental work in the United States was concentrated on
conical liners of small apex angle,

E, I, duPont de Nemours and Co,

(OEMsr-764, W-670-0RD-4331,
and w-672-0RD-5723)
The first research work in the United States appears to have been started by C, 0,
Davis and his group at the Eastern Laboratory of the duPont Company in 1941, although
the earliest report known to tile author is dated March 1942, This group furnished
most of the original information for the design of weapons. They determined optimum
values for the various parameters used in making shaped charge weapons; liner material,
liner thickness, liner apex angle, standoff, length and diameter of charge, effects of
confinement 1 etc.
OSRD 682, Explosives Research Laborator,r
Bruceton, Pa. Set up Wlder Div.
8 of N.D.R.C. with the joint
sponsorship of the Carnegie
Institute of Technology and the
Pittsburgh Station of the Bureau
of Mines. Hereaf'ter simply E.R.L.
( OEMsr-202 ) II A Rotating Drum
Camera for the Qptical Study of
Detonations,'' by G. H. ~lesserly,
Progress Report July 8, 1942
(NDRC Div. 8).

This report describes the construction and operation of a rotating drum camera
designed to measure the velocities of explosive phenomena.

AC 3518, Armament Research Dept. (Great

Britain) 11 The Effect of Lateral
Charge Confinement on Munroe
Jets" (Part I), by W. M. Evana
(ARD EJq>losives report No.
60/43; SC.5; 0SRD Liaison
Office WA-528-14), Feb. 22, 1943.

Linear scaling was experimentally determined for depth of penetration, entry

diameter and volume of the hole. These quantities are shown to depend upon the con-
finement, but as the thickness of the confining wall is increased, very little change
in performance was noted.
Penetration

AC 3596, Advisory Council on Scientific

Research and Technical Develop-
ment, 11 A Note on the Theory of
tiB Munroe Effect" by J. L. Tuck,
Feb. 27, 1943, Pbys/Ex. 393;
SC.3; 0SRD Liaison Office
WA-638-24). ·

A theory of jet formation based on a hydrodynamic mechanism is presented.

AC 3654, Advisory Council on. Scientific
Research and Technical Develop-
ment, 11 Studies of Shaped Charges
by Flash Radiography 1. Pre-
liminary" by J. L. Tuck (Phys/Ex.
399; SC.6; OSRD Liaison Office
WA-693-ll or II 5-4944) March
15, 1943.

388 .
.•

QQ'WF
 ew;nrmr'T Ssmwir[ Izf ~· r ..

Shows nash X-rays of collapsing cones and describes the equipment and procedure
used.

OSRD 1338, E.R.L. Progress Report on

liThe Mechanism of Action
of Cavity Chargesn to March
15, 1943, by G, B. Ki.stiakawsky,
D. P. MacDougall and G, H.
Messerly, ~losives Research
Laboratory, Bruceton, Pa,
(NDRC Div. 8) dated
April 12, 1943.
The most ilnportant early work on the understanding of the fundamental prop-
erties of these charges was d~ne at the Explosives Research Laboratory, Bruceton,
Pennsylvania, under D. P. MacDougall. Much of this is SI.D1ll1larized in OSRD 1338.
The experimental data furnishing the background for this report is found in Div,
8 NDRC Interim reports rnuubered CF-1 to 7. These reports contain the first evidence
that the penetrating medium is a high velocity jet of liner material followed by a
low velocity slug containing most of the liner material but having little effect upon
the target. It was discovered that the front of the jet travels at high velocities
and the rear at much lower velocities. Rough, but correct, values for the diameter
of the jet were determined. An equation was developed for determining the velocity of
penetration of jets into target materials which showed that this velocity is dependent
primarily upon the density of the target materials, Experiments were performed to
verify this result, It is pointed out that the process of penetration into targets
consists of using up the jet from the front end and that penetration stops when the
total effective length of the jet has been used up. It is postulated that the
velocity gradient in the jet accounts for the improvement in penetration with stand-
. off. The reduction in penetration of the jet after optimum standoff is reached is
·. presumed to be due to breaking up of the jet into particles and subsequent radial
l
spreading. This report is a remarkable source of information concerning the funda-
mental processes of jet penetration.

SC-1, E .R.L,, "Studies of Shaped Chargesn

by D. P. MacDougall, M, A.
Paul and G. H. Messerly
(NDRC Div, 8 Interim
report) Aug. 15-Sept. 15,
1943.
Conical lined charges were fired into ice blocks and liner fragments recovered
for study of jet distribution in the target. The fragments recovered were not
correlated with their original location in the uncollapsed liner,

389

II' ana
 It is pointed out that while the high velocity front part af the jet should pene-
trate practically all target materials in accordance with the s:lmple theory of pene-
tration, the penetration of the slower rear part of the jet should depend more upon
the target material. A residual penetration theory is presented in which the reduction
in penetration for a given thickness of material is shown to be approximately proportion-
al to the square root or the density of that material. This theory is shown to hold
exper:lmentally for lead, steel, aluminum, water and mai\Y' other materials,· This consti-
tutes one of the best experimental proofs of the soundness of the fundamental concepts
upon which the penetration theory is based.

OSRD 4357f, Carnegie Institute of

Technology ( OEMsr-950
"Fundamentals of Jet
Penetration; Theory of
Jet Penetration", by E,
M. Pugh and E. L. Fireman,
Nov. 15, 1944, (Monthly
Report OT&-4£1 NDRC Div •
g). -
OSRD 5462e, Carnegie Institute of
Technology, "Fundamentals
of Je•t Penetration" by
E. M. Pugh and E. L.
Fireman, Aug. 15, 1945
(Monthly re~t. OT&-13e 1
NDRC Div. _g_).

These reports contain an integration of the penetration theory and verifies the
fact that exper:lmental penetration versus standoff curves cannot be fitted without
assuming that the metals are drawn out ductilly as they travel along the axis.

AC 8355, Armament Research Dept.,

Fort Halstead, England,
"Penetration by Munroe
Jets; Secondary Penetra-
tion and the Effect of Target
Strength," Part I by D. C,
Pack and w. M. Evans, April 1945.
(ARD Explosives rept. no. 51/45
and ARD Theoretical Research
rapt. no. 11/45; encl. 1 to
MA London rapt. R3812-45; Pcys/Ex.
656J _sc 151; OSRD Liaison Office
wA:.4586-ll).

A relation for determining the effect of target strength upon penetration is derived,

392
.-

£Q? $w
05RD 6384, Carnegie Institute oi'
Technology, "Protection
Against Shaped Chargesu,
Final Report Aug. 1,
1943, Nov. 10, 1945, by
Pugh and W, H. Bessey
rept. no, A-384 Div. 2).
This report summarizes the simple penetration theory, the integrated penetration
theory and the residual penetration theory with corrections obtained by integration.

Page 64 discussed penetration theories used in England and the United States.
Page 76, the anomalous behavior of glass in residual penetration exper:illlents is
ascribed to a strength effect nearly equal to the theoretical strength of the glass.

Page 85 gives theoretical and exper:illlental proof of the detrimental effects of

spaced armor for use as protection against well-made charges. It shows why spaced
armor has been effective against poorly made charges. It is shown that theoretically
and experimentally if a given thickness and weight can be added to a tank, it is best
that the material be of uniform dens! ty rather than lumped as it would be in spaced
protection.

CIT-ORD-3, Carnegie Institute of

Technology 11 FundamentaJ.s
of Shaped Charges; Part E.
Theory - An Attempt at a
Theory of Hole Volume"
by E. L. Fireman and G.
H. Winslaw: W-36-061-0RD-
2773, Third Bimonthly Report;
June 1, 1946.

CIT-ORD-4, Carnegie Institute of

Technology, "Fundamental.s
of Shaped Charges 1 Part
E. Theory - Cal.cul.ation
of Hole Profilesu by G.
H. Winslaw, W-36-o61-
0RD-2773, Fourth Bi-
monthly Report•
W-36-o6l-ORD-2A79 First
Bimonth:cy Report; Aug.
31, 1946.
A theory developed by H. A. Bethe ( 11 An Attempt at a Theory of Armor Penetration",
Ordnance Lab., Frankford Arsenal, May 1941) for the penetration of armor by projectiles
is adapted to metallic lined shaped charges. This adaptation states that the work per
unit volume of hoJ.e is equal. to the sum of the work required to penetrate a unit volume
in an indefinitely long period of time (static work) plus a term proportional to the
average oi' the squares of the velocities of penetration. Curves are calculated for the

393

a .&
hole volume in mild steel and armor plate as a function of standoff and compared with
exper:lmental data. A correlation between the total energy delivered by the jet and
the average square penetration velocit,y is given by plotting each against standoff.

Jet Formation

AC 3724, Advisory Council on Scientific

Research and Technical Develop-
ment, 11 A Formulation of Mr.
Tuck's Conception of Munroe
Jets11 by G. I. Taylor, (Phys/Ex.
427; OSRD Liaison Office
WA-638-32) May 27, 1943•
This report gives a fundamental hydrodynamic steady state theory of jet formation
based upon the flash X-rays of AC 3654.

BRL 368, Ballistic Research

Laboratories,DHigh
Speed Radiographic
Studies of Controlled
Fra~entation. r. The
Collapse of Steel
Cavity Charge Liners"
by L. B. Seely and J. C.
Clark, June 16, 1943.

Excellent flash radiographs of the collapse process in conical liners are shown.
These pictures are very similar to those of Tuck (AC 3596) and were probably taken about
the same time.

BRL 370, Ballistics Research

Laboratories, "Mathematical
Jet Theory of Lined Hollow
Charges" by Garrett Birkhoff,
June 18, 1943.
This theory is very similar to that of AC 3724 and was fonnulated entirely inde-
pendently. In addition it furnished a mathematical theory of the heating pra:luced during
the liner collapse and jet formation processes.

RC 193, "The Explosion of a Long

Cylindrical Bomb 11 by G.
r. Taylor.
In this paper Taylor derives his theorem showing that the motion of a liner that
is struck by a constant velocity detonation wave must collapse in the direction parallel
to the bisector of the angle between the original liner and the collapsing liner. Thi.s
was added to his collapse theory for wedge shaped and conical liners •

. 394
 ?EWU s .. I
5

BRL 623, "Hollow Charge Anti-

Tank Projectiles" by-
Garrett Birkhoff,
Februar,y 10, 1947
This report contains an excellent summary of the knowledge of the phenomena up to
that date. It also contains additional material on scaling effects in rotated and
non-rotated liners (some of which had previously been distributed in the form of pri-
vate letters from Dr. Birkhoff). It also contains results of some metallurgical
examinations of sectioned slugs.

CIT-ORD-21, Carnegie Institute of

Technology (W36-061-0RD-
2879) Eighteenth Bimonthly
Repart, Annual Summary,
"Fundamentals of Shaped
Charges", June .)) , 1949,
TIP C3310, by Emerson M.
Pugh, R. J. Eichelberger
and Norman Rostoker; Part
II by R. v. Heine-Geldern
s. Foner and E. c. Mutschler.
Part I presents a non-steady state hydrodynamic theory of jet formation and experi-
mental verification of this theory.
Part II contains a discussion of charge imperfections and their effects and a
description of the Kerr cell shutter and the associated circuits used to operate the
exploding wire light source and the slmtter. Photographs of explosive phenomena are
shONil.

CIT-ORD-31, Fourth Bimonthly Report,

"Fundamentals of Shaped
Charges 11 , Feb. 28, 1951.
Part I Theory by Emerson
M. Pugh, R. J. Eichelberger
and Norman Rostoker; and
an experimental verifica-
tion by R. J. Eichelberger
and E; M. Pugh; Part III
by R. v. Heine-Geldern, s.
Foner am E. c. Mutschler.

Part I is a restatement of Part I of CIT-aiD-:21, and Parts II and III summarize

much of the important work done by the Carnegie Institute of Technology group since
late 1945.

Part II contains further remarks on the generalized theory of jet formation, a

discussion of target characteristics, and a discussion of charge imperfections and
their effects.

395

91 !Iii&
 ______. ...

Part III shows many photogt"aphs obtained with the Kerr cell camera setup des-
cribed in Part TI oi" CIT-ORD-21. A discussion of protection against shaped charges
is also given.
 DISTRIBUTION LIST

No. of No, of
Copies Copies

6 Chief or Ordnance 2 Canmander

Department or the Army Naval Proving Ground
Washington 25, D. C. Dahlgren, Virginia
Attn: ORDTB - Bal Sec
ORDTX-AR l Chief, Bureau or Ordnance
CRD~ Department of the Navy
CIIDFA - C.V, Parker Washington 25, D. c.
ORDTA - Col. W.L. Bell,Jr. Attn: Mr. H, K:lmble, Re2c

10 British - OHD'IB for d:l..stribution l Director

2 cys - British Joint Services Naval Research Laborator.y
Mission, Attn: Miss Washington 25, D. c.
·Mary G. Scott Attn: Technical Infonnation Div.
l cy - A.R.E., Of Interest to:
W. E. Soper l Canmander
u.s. Naval Air Development Center
4 Canadian Joint Staff - ORDTB for John~ille, Penns,ylvania
distribution Attn: Aviation Armament Lab.
or Interest to: R. Walker
Ro Foster, C.A.R.D,E,, 2 Canmandant
ValCartier u. s. Marine Corps
Washington 25, D. c.
4 Chief, Bureau of Ordnance Attn: Division or Aviation
Department or the Navy Plans and Policies Division
Washington 25, D. c. Maj. w. Lo Flake
Attn:· Re3
2 Chief or s~rr
2 Chief of Naval ~esearch u. s. Air Force
Technical Information Division Washington 25, D. c.
Librar.y of Congress Attn: AFDRD - AR
Washington 25, D. C.
2 Canmanding General
3 Canmander Wright Air Deirelopment Center
Naval Ordnance Laboratory Wright-Patterson Air Force Base
White Oak Ohio
Sliver Spring 19, Maryland Attn: WCEGO
or Interest to:
Mr. S, J. Jacobs 2 National Advisory Committee for
Dr, G, Ko Hartman Aeronautics
1724 F Street, N. w.
Canmander Washington 25, D. C.
u.s. Naval Ordnance Test Station or Interest to:
llzy-okern Mr. Coleman Donaldson
P. o. China Lake, California
Attn: Technical Librar.y & 2 Atanic Energy Canmission
Editorial Section Division of Militar.y Applications
or Interest to: Mr. J.s. Rinehart Wash4tgton 25, D. c.
Mr. G.c. Throner· Attn: Col. R. G. Butler
Mr. A,L, Bennett
397
 0 £22 d£:0 I!i!Sii!&!Ld

DISTRIBUTION LIST

No, of No, or
Copies Copies

2 Director 3 Commanding Officer

National Bureau of Standards Office of Ordnance Research
Washington 25, D. C, Duke University
Attn: Ordnance Development Lab, 2127 Myrtle Drive
Durham, North Carolina
2 Chief of Engineers
Department of the Arrrry 1 Commanding Officer
Washington 25, D. c. Engineer Research & Development
Attn: ENGNE Laboratory
Structure Dev, Branch. Ft. Belvoir, Virginia
Attn: Technicil.l. Intelligence Br.
1 Canmanding General
Detroit Arsenal 1 Professor Garrett Birkhoff
Centerline, Michigan 21 Vanser Building
Harvard University
3 Conunandine Officer Cambridge 38, Massachusetts
Frankford Arsenal
Philadelphia 37, Pennsylvania 1 Dr, L, H, ThQIIas
Attn: Mr. H, S, Lipinski Watson Scientific Computing Lab,
612 w. ll6th Street
J Canmanding Officer New York 27, New York
PicatinQy Arsenal
Dover, New Jersey 1 Professor G. B. Kis:tiakowsky
Attn: Technical Division Chemistry Department
Harvard University
1 Canmanding Officer Cambridge 38, Massachusetts
Watertown Ar11enal
Watertown, Massachusette 1 Professor N. F. Ramse.y, Jr.
Department of Pnysics
1 Canmanding Off.icer Harvard University
Chemical Corps Chemical & Radio- Cambridge 38, Massachuset te
logical Laboratories
Arnw~emicalCenter, Muyland 1 Professor John von Neumann
Department of Mathematics
2 Commanding General The Institute for Advanced Study
Redstone Arsenal Princeton, New Jerse.y
Huhtsville·, Alabama
Attn: Technical Library 2 Professor Richard Courant
Institute for Mathematics and
1 Los Alamos Scientific Laboratory Mechanics
P, 0, Box 1663 New York University
Los Alamos, New Mexico New York, New York
Attn: D. P. MacDougall
1 Dr. A. W. Hull
1 Commanding Officer Research Laboratory
Holston Ordnance Works General Electric COIIIPa.JV
Kingsport, Tennessee Schenectady, New York

398

mrrttmm 'Y
 RES'IT!'l 5 ••
'
'• I

DISTRIBUTION LIST

No. o! No. o!
Copies Copies

2 Dr. w. H. Prager 2 Firestone Tire and Rubber Co.

Graduate Division or Applied Akron 17 1 Ohio
Mechanics Attn: E. W. Fcrd
Brown. University H. Winn
Providence l2 1 Rhode Island THRU: District Chief
Cleveland Ord District
2 Chief, Explosive & Physical 1367 East Sixth Street
Sciences Division Cleveland 4, Ohio
Bureau or Mines
4800 Forbes Street l Budd Ma.nuf'acturing Co,
Pittsburgh 13, Pennsylvania Red Lion Plant
Attn: Dr. Charles M. Mason Philadelphia, Pennsylvania

2 Director THRU: District Chief

Applied Physics Laboratory Philadelphia Ord District
8621 Georgia Avenue 1500 Chestnut Street
SUver Spring, }ia.ryland Philadelphia 21, Pa.
Attn: Mr. H. S. Morton
l Professor Melvin Cook
THRU: Naval Inspector o! Ordnance Department or Physics
Applied Physics Laboratory University or Utah
8621 Georgia Avenue Salt Lake City, Utah
Silver Spring, Maryland
THRU: Chief or Naval Research
l Institute !or Air Weapons Research Tech. Information Div.
University o! Chicago Library or Congress
Chicago 37 1 Illinois ~/as~ington 25 1 D. c.
Attn: Paul E. Shanahan
l Dr. c. o. Davis
THRU: Conunanding O!ficer Eastern Laboratories
Mid Central Air Procurement E. I. duPont de Nemours & Co., Inc.
District Gibbstown, New Jersey
165 N. Canal Street
Chicago J5 1 Illinois THRU: District Chief
Philadelphia Ord District
Carnegie Institute o! Technology 1500 Chestnut street
Pittsburgh, Pennsylvania Philadelphia 211 Pa.
Attn: Dr. E. M. Pugh
l Dr, T. C. Poulter
THRU: District Chief Stanford Research Institute
Pittsburgh Ord District Palo Alto, California
200 Fourth Avenue
Pittsburf?P 22 1 Pa. THRU: District Chief
San Francisco Ord District
P. 0. Box 1829
Oakland, California

.199

------- -----~
 "" '-·.
DISTRIBUTION LIST...._ · •

No. of
Copies

1 Annour Research Foundation

Technology Center
Chicago, Illinois

THRU: District Chief

Chicago Ordnance District
209 W. Jackson Boulevard
Chicago 61 .Illinois

1 Arthur D. Little, Inc ..

30 Memorial Drive
Cambridge 42 1 Massachusetts
Attm Dr. w. c. Lothrop
TBIW: District Chief
Boston Ordnance District
Boston Al'll\l' Base
Boston 10 1 Massachusetts

.,

ARIIIIY·RARITAM ARII.E.AUAAPDl,MEfUCMEN,1U•51·474