Low power and high performance detff using common

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
__________________________________________________________________________________________
Volume: 02 Issue: 10 | Oct-2013, Available @ http://www.ijret.org 317
LOW POWER AND HIGH PERFORMANCE DETFF USING COMMON
FEEDBACK INVERTER LOGIC
Y. Yaswanth Kumar1
, A.Varalakshmi2
1
II M. Tech VLSI System Design, 2
Asst. Professor, Department of ECE, SITAMS, JNTU Ananthapur, Andhra Pradesh,
India, yaswanth.yadalam@gmail.com, varalakshmi.a@gmail.com
Abstract
The power consumption of a system is crucial parameter in modern VLSI circuits especially for low power applications. In this
project, a low power Double Edge Triggered D-Flip Flop (DETFF) design is proposed in 65nm CMOS technology. The proposed
DETFF is having less number of transistors than earlier designs. Simulations are carried out using HSPICE tool with different clock
frequencies ranging from 400MHz to 2GHz and with different supply voltages ranging from 0.8V to 1.2V. In general, a power delay
product (PDP)-based comparison is appropriate for low power portable systems. At nominal condition, the PDP of the proposed
DETFF is improved by 65.48% and 44.85% over earlier designs DETFF1 and DETFF2 respectively. Simulation results show lowest
power dissipation and least delay than existing designs, which claims that the proposed DETFF is suitable for low power and high
speed applications.
Keywords: CMOS, flip-flops, Double-edge triggered, power dissipation, delay and PDP
--------------------------------------------------------------------***--------------------------------------------------------------------------
1. INTRODUCTION
In recent years, consumer's attitudes are gearing towards better
accessibility and mobility that caused a demand for an ever-
increasing number of portable applications, requiring low-
power dissipation and high throughput. For example,
notebook, hand-held, palm-top computers are portable
versions of microprocessors which require low power
dissipation designs. In these applications, not only voice, but
data as well as video are transmitted via wireless links. The
weight and size of these portable devices is determined by the
amount of power required. Low power design is very
important to support these devices, particularly with respect to
the integrated circuits, in which excessive power consumption
will make the chip over heat and cause error or even
permanent damage. The seriousness of this problem increases
with the level of integration and thus we need to consider the
power consumption of the chips for every component in the
device. Hence, the major concerns of the VLSI circuit
designer were low power dissipation, high speed, small silicon
area, low cost and reliability. The battery lifetime for these
products is crucial; hence, a well-planned low-energy
consumption design strategy must be necessary.
The design for low power issues can’t be overcome without
précised power prediction and optimization tools. Therefore,
there is a critical need for certain tools to calculate power
dissipation during the design to meet the power constraints to
ignore the costly redesign effort. Power dissipation is given by
the equation, P = α CV2f. Voltage scaling is the most effective
way to decrease power consumption, since power is
proportional to the square of the voltage. However, voltage
scaling is associated with threshold voltage scaling which can
cause the leakage power to increase exponentially. In order to
reduce the complexity of circuit design, a large proportion of
digital circuits are designed to be synchronous circuits. The
power dissipation in synchronous VLSI circuits is contributed
by several factors: i.e. I/O, logic and clock power dissipation.
Clock power dissipation is divided into three major
contributions: clock wires, clock buffers and flip-flops.
Studies show that 40-45% of total power dissipated in
integrated system is due to clock distribution network and
90% of which is consumed by the flip-flops and the last
branches of the clock distribution network. D-type flip-flop’s
(DFF’s) are one of the most fundamental building blocks in
modern VLSI systems and it contributes a significant part of
the total power dissipation of the system.
The most popular synchronous digital circuits are edge
triggered flip-flops. Single-edge triggered (SET) flip-flops and
Double-edge triggered (DET) are edge-sensitive devices, that
is, data storage in these flip-flops occurs at specific edges of
the clock signal. During each clock period, single-edge
triggered flip-flops latch data at only one edge, either rising or
falling edge of the clock signal. Double-edge triggered flip-
flops latch data at both rising and falling edges of the clock
signal during each period of the clock signal. Hence by using
double-edge triggered flip-flops (DETFFs), the clock
frequency can be significantly reduced - ideally, to half while
preserving the rate of data processing. Using lower clock
frequency may translate into considerable power savings for
the clocked portions of a circuit.

_______________________________________________________________________________________
The DETFF design aims at saving energy both on the clock
distribution network (by halving the clock frequency) and flip-
flops. It is preferable to reduce circuits’ clock loads by
minimizing the number of clocked transistors. A simple
DETFF is implemented with about 50% extra transistors than
the traditional SET flip flop, however, this issue is being
resolved in recent intensive researches. It motivates to use
DETFF for larger and complicated designs, especially for
application specific integrated circuits (ASIC) design. Several
DET flop-flops have been proposed in the literature. The main
attention has been in improving the performance of the
circuits.
This paper is organized as follows: Section II explains the
DETFF structures consist of conventional and proposed
DETFF circuits. The nominal simulation conditions along
with transient analysis and optimization performed during
simulation are discussed in Section III. Section IV consists of
Results and performance of proposed design and conventional
designs comparisons in terms of average power dissipation,
delay and PDP. Paper ends with the conclusion.
2. DOUBLE EDGE TRIGGERED FLIP FLOP
STRUCTURES
There are several ways to implement a DETFF; In general
they can be categorized into two ways. The first idea is to
insert additional circuitry to generate internal pulse signals on
each clock edge. The second idea is to duplicate the pathway
to enable the flip-flop to sample data on every clock edge. The
same data throughput can be achieved with half of the clock
frequency by using DETFF. In other words double edge
clocking can be used to save half of the power on the clock
distribution network.
2.1 Conventional Double-Edge Triggered Flip-Flops
The DET flip flop given in [1] is shown in figure 1. This flip-
flop is basically a Master Slave flip-flop structure and has two
data paths. The upper data path consists of a Single Edge
Triggered flip-flop (SETFF) implemented using transmission
gates. It employs positive edge. The lower data path consists
of a negative edge triggered flip-flop implemented using
transmission gates. Both the data paths have feedback loops
connected such that whenever the clock is stopped, the logic
level at the output is retained. This flip-flop has 22 transistors
excluding clock driver. In these 22 transistors, 12 transistors
are clocked transistors.
Fig -1: DETFF1 given in [1]
The figure 1 shows a pair of parallel loops. When the clock
pulse changes from low to high, the upper loop holds data and
the down loop samples data. But when the clock pulse changes
from high to low, the top loop switches to sample data and
then the down loop switches to hold data.
The DET flip flop given in [2] is shown in figure 2. This flip-
flop is also basically a Master Slave flip-flop structure and
consists of two data paths. The transmission gates (TG) in
both the data path are clocked such that the upper data path
works as positive edge triggered flip flop and lower data path
works as negative edge triggered flip flop. The upper data path
consists of transmission gates TG1, TG2 and inverter I1. The
lower data path consists of transmission gates TG3, TG4 and
inverter I3.
The input data is connected to TG1 and TG3 and the output is
taken from inverter I5 whose input is connected with TG2 and
TG4. Both the data paths have feedback loops connected such
that whenever the clock is stopped, the logic level at the
output is retained so as to maintain the static functionality. The
feedback in upper data path consists of an inverter I2 and a
pass transistor P1. The feedback in lower data path consists of
an inverter I4 and a pass transistor P2. This design is identical
to figure 1 except feedback path. The feedback transmission
gates of figure 1 are not on critical paths and replaced with
pass transistors. This flip-flop has 20 transistors excluding
clock driver. In these 20 transistors, 10 transistors are clocked
transistors.

_______________________________________________________________________________________
Fig -2: DETFF2 given in [2]
2.2 Proposed Double-Edge Triggered Flip-Flop
The proposed DET Flip-Flop design is shown in figure 3. This
flip-flop is also basically Master Slave flip-flop structure and
it consists of two data paths. The upper data path consists of
transmission gate TG1, pass transistors (P1, P2) and an
inverter I1. The lower data path consists of transmission gate
TG2, pass transistors (P3, P4) and an inverter I2. The
transmission gate and pass transistors in both the data paths
are clocked such that the upper data path works as positive
edge triggered flip flop and lower data path works as negative
edge triggered flip flop. The inverter I3 serves as a common
feedback inverter and the output of the inverter I4 forms the
output of the present invention.
This design is identical to figure 1 except feedback path has
been changed and the number of clocked transistors is reduced
by using n-type pass transistors instead of transmission gates.
The transmission gates TG2, TG3, TG5, and TG6 of figure 1
are replaced with pass transistors P1, P2, P3 and P4
respectively. Basically n-type pass transistors give weak high
but in figure 3, the n-type pass transistors are followed by an
inverter, which results in strong high. Thus the proposed
DETFF in figure 3 is free from threshold voltage loss problem
of pass transistors. The feedback network of figure 2 is altered
by placing the n-type pass transistor instead of p-type pass
transistor. Since, the area incurred by NMOS is less than that
of PMOS transistor in order to compensate the mobility
constraint of NMOS and PMOS transistors.
According to the DETFF1 and/or DETFF2, one of two loops
always samples the data and the other loop always holds the
data. Therefore, only a single feedback loop is required to hold
the data. The proposed DETFF design eliminates one feedback
loop and reduces the number of transistors compared to the
existing designs. The two inverters I2 and I4 of figure 1 are
used for two feedback paths, but here a single feedback
inverter I3 is used as a common feedback inverter for two
feedback paths as shown in figure 3.
The input data is supplied to one end each of two transmission
gates (TG1, TG2). The other ends of the two transmission
gates (TG1, TG2) are separately connected to the input of two
inverters (I1, I2). The outputs of the two inverters (I1, I2) are
separately connected to the inputs of another two inverters (I3,
I4) via the pass transistors (P1, P3). The output of inverter I3
goes through two pass transistors (P2, P4) to connect to one
end of the above mentioned two transmission gates (TG1,
TG2) separately, and the output of the inverter I4 forms the
output of the present invention. Both the data paths have
feedback loops connected such that whenever the clock is
stopped, the logic level at the output is retained so as to
maintain the static functionality.
Fig -3: Proposed DETFF
The novelty of the proposed flip flop lies in the feedback
strategy using common feedback inverter and pass transistors
to make the design static. This improves the power efficiency
of the proposed flip flop. This flip flop has 16 transistors
excluding clock driver. In these 16 transistors, 8 transistors are
clocked transistors. It is preferred to reduce number of clocked
transistors to achieve low power dissipation. So, the main
advantages of the proposed design are low power dissipation
and increased performance with low transistor count. Thus the
proposed Double Edge Triggered Flip-Flop (DETFF) has
become more efficient in terms of area, power and speed
which claim for better performance than conventional designs.

_______________________________________________________________________________________
3. SIMULATIONS AND TRANSIENT ANALYSIS
Simulation parameters used for comparison are shown in
Table 1. For a fair comparison we simulated all flip-flops on
same simulation parameters.
Table -1: CMOS Simulation parameters
Technology 65nm CMOS
MOSFET Model BSIM 4
Nominal Conditions Vdd =1V, T=25° C
Duty Cycle 50 %
Nominal Clock Frequency 400MHz
Data Rate at Nominal
Clock Frequency
800Mbps
All simulations are carried out using HSPICE simulation tool
at nominal condition with different range of frequencies from
400MHz to 2GHz and with different range of supply voltages
from 0.8V to 1.2V. The simulated waveform for the proposed
DET flip-flop at nominal condition is shown in figure 4. This
waveform shows that the DETFF is storing the data in both
rising edge and falling edge of the clock signal.
Fig -4: Output waveform for proposed DETFF
We have optimized the transistor sizes. The transistors, that
are not located on critical path, are implemented with
minimum size. We have improved the feedback path in our
design. We simulated the design to achieve minimum power-
delay (clk-q) product (PDP). Transistor count is also included
to maintain a fair level of comparisons. In DETFF1 there are
12 clocked transistors, in DETFF2 there are 10 clocked
transistors while in our design there are only 8 clocked
transistors. We have reduced the number of clocked
transistors. Thus we have reduced the power dissipation. The
switching power, which is dominant power, is proportional to
the square of the supply voltage. We have reduced supply
voltage to 1V or even less voltages for the reduction of the
power.
4. PERFORMANCE COMPARISON
The performance of the proposed DETFF is evaluated by
comparing the average power dissipation, delay and PDP with
DETFF 1 and DETFF 2. Comparison of simulation results at
nominal condition for the three FFs is summarized in Table 2.
Table -2: Comparison of simulation results for the three FFs
at nominal condition
Specification DETFF 1 DETFF 2 Proposed
DETFF
Average Power
Dissipation (µW)
1.471 1.066 0.881
Clk-Q Delay (ps) 26.47 22.86 15.26
PDP (10-18J) 38.94 24.37 13.44
PDP
Improvement %
over 1
---- 37.42 65.48
No. of Transistors
(excluding clock
driver)
22 20 16
Since D to Q delay depends on when the data transition
occurs, here we measured clock to Q delay. The clock-to-Q
delay is the delay from the active clock input to the new value
of the output. We simulated all DETFFs with different clock
frequencies ranging from 400MHz to 2GHz and with different
supply voltages ranging from 0.8V to 1.2V. Each flip-flop is
optimized for power delay product. In general, a PDP-based
comparison is appropriate for low power portable systems in
which the battery life is the primary index of energy
efficiency.
Table 3 shows the PDP of all three flip flops at different
supply voltages, this indicates that the proposed design has an
average improvement of 65.28% and 45.98% in terms of PDP
as compared to DETFF1 and DETFF2 respectively. Table 4
shows the PDP of all three flip flops at different clock
frequencies, this indicates that the proposed design has an
average improvement of 70.82% and 44.25% in terms of PDP
as compared to DETFF1 and DETFF2 respectively. Chart 1
and chart 2 show the statistical variation of PDP for all the 3
flip-flops on varying supply voltage and clock frequency
respectively. This shows our design is suitable for low power
dissipation and high performance applications.
Table -3: PDP of 3 FFs with the variation of supply voltage
VD
D
(V)
DETF
F 1
(10-
18J)
DETFF
2
(10-
18J)
Proposed
DETFF
(10-18J)
Improv
ement
% Over
1
Improve
ment %
Over 2
0.8 37.84 20.81 11.06 70.77 46.85

_______________________________________________________________________________________
0.9 36.8 22.09 12.04 67.28 45.5
1.0 38.94 24.37 13.44 65.48 44.85
1.1 41.3 27.85 15.25 63.07 45.24
1.2 43.54 33.32 17.51 59.78 47.45
Table-4: PDP of 3 FFs with the variation of clock frequency
Clock
Freq.
(GHz
)
DETF
F1
(10-
18J)
DETF
F 2
(10-
18J)
Propose
d
DETFF
(10-18J)
Improve
ment %
Over 1
Improve
ment %
Over 2
0.4 38.94 24.37 13.44 65.48 44.85
0.5 43.84 26.92 15.52 64.6 43.46
1.0 67.47 33.89 19.2 71.54 43.35
1.5 95.09 41.54 22.94 75.87 44.78
2.0 114.69 48.55 26.8 76.63 44.8
Chart -1: Statistical variation of PDP for the 3 FFs on varying
supply voltage
0
20
40
60
80
100
120
140
0.4 0.5 1 1.5 2
PDP(10-18J)
CLOCK FREQ. (GHz)
DETFF 1
DETFF 2
Proposed DETFF
Chart -2: Statistical variation of PDP for the 3 FFs on varying
clock frequency
CONCLUSIONS
The proposed Double-edge-triggered D flip-flop is a low
power, low voltage, high speed and low transistor count flip-
flop designed in 65nm CMOS technology. This is static in
nature. The proposed DETFF design eliminates one feedback
loop by using common feedback inverter logic thereby
reducing the number of transistors and also by using pass
transistors the number of clocked transistors is reduced
compared to the existing DET flip flops. This improves the
power efficiency of the proposed DETFF. The three DET flip-
flops are simulated with different supply voltages ranging
from 0.8V to 1.2V and with different clock frequencies
ranging from 400MHz to 2GHz. At nominal condition, the
PDP of the proposed DETFF is improved by 65.48% and
44.85% over earlier designs DETFF1 and DETFF2
respectively. Simulation results show that the proposed design
has lowest power dissipation and least delay thereby it has
lowest PDP than existing designs. Therefore the proposed
design is very well suited for low power and high speed
applications operating at low voltages.
REFERENCES
[1] M. Pedram, Q. Wu, and X. Wu, “A New Design of
Double Edge Triggered Flip-Flops,” Proceedings of the
Asia and South Pacific Design Automation Conference
(ASP-DAC), pp. 417–421, 1998.
[2] Imran Ahmed Khan, Danish Shaikh and Mirza Tariq
Beg,“ 2 GHz Low Power Double Edge Triggered flip-
flop in 65nm CMOS Technology” IEEE Conference, pp
978-1-4673-1318-6/12, 2012.
[3] Xiaowen Wang, and William H. Robinson, “A Low-
Power Double Edge-Triggered Flip-Flop with
Transmission Gates and Clock Gating” IEEE
Conference , pp 205-208, 2010.
[4] Hossein Karimiyan Alidash, Sayed Masoud Sayedi and
Hossein Saidi, “Low-Power State-Retention Dual Edge-
Triggered Pulsed Latch,” Proceedings of ICEE 2010,
May 11-13, IEEE 2010.
[5] Peiyi Zhao, Jason McNeely, Pradeep Golconda and
Jianping Hu, ”Low Power Design of Double-Edge
Triggered Flip-Flop by Reducing the Number of
Clocked Transistors,” IEEE Conference, 2008.
[6] Yu Chien-Cheng ,“Design of Low-Power Double Edge-
Triggered Flip- Flop Circuit", Second IEEE Conference
on Industrial Electronics and Applications, pp 2054-
2057, 2007.
[7] Ahmed Sayed and Hussain Al-Asaad,“A New Low
Power High Performance Flip-Flop” IEEE Conference,
pp 723-726, 2006.
[8] Nikola Nedovic, Mark Aleksic and Vojin G.
Oklobdzija, “Comparative Analysis of Double-Edge
Versus Single-Edge Triggered Clocked Storage
Elements” IEEE Conference, pp V-105 to V-108, 2002.

_______________________________________________________________________________________
[9] Chandrakasan, W.Bowhill, and F. Fox, “Design of
High-Performance Microprocessor Circuits”, 1st ed.
Piscataway, NJ: IEEE, 2001.
[10] G. E. Tellez, A. Farrahi, and M. Sarafzadeh, “Activity-
Driven Clock Design for Low Power Circuits,”
Proceedings of the IEEWACM International
Conference on Computer-Aided Design (ICCAD), pp.
62- 65, 1995.
BIOGRAPHIES
A. Varalakshmi received her Bachelor Degree in
Electronics and Communication Engineering
from Jawaharlal Nehru Technological University
Hyderabad in the year 2004, Master’s Degree in
VLSI System Design from Jawaharlal Nehru
Technological University Hyderabad in the year 2008.
Presently working as an Assistant Professor in ECE
Department In SITAMS at Chittoor. Her interested areas of
research are Nano Electronics, VLSI Design, Low Power
VLSI Design.
Y. Yaswanth Kumar received his Bachelor
Degree in Electronics and Communication
Engineering from Jawaharlal Nehru
Technological University Ananthapur in the year
2011. Currently he is doing his Master’s Degree
in VLSI System Design in Jawaharlal Nehru Technological
University Ananthapur. His interested areas are Low Power
VLSI Design, Digital Circuits Design and VLSI Technology.

Low power and high performance detff using common

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Low power and high performance detff using common (20)

More from eSAT Publishing House (20)

Recently uploaded (20)

Low power and high performance detff using common