Design of Memory Cell for Low Power Applications

N.geetha Rani Int. Journal of Engineering Research and Applications www.ijera.com
ISSN: 2248-9622, Vol. 5, Issue 8, (Part - 1) August 2015, pp.83-86
www.ijera.com 83 | P a g e
Design of Memory Cell for Low Power Applications
N.geetha Rani*
, DR.P.Chandrasekhar Reddy**
*
Associate Professor, ECE Department, Ravindra College of Engineering for Women, Kurnool, A.P
**
Professor, ECE Department, JNTUH College of Engineering, Hyderabad, A.P
ABSTRACT
Aggressive CMOS scaling results in lower threshold voltage and thin oxide thickness for transistors
manufactured in nano regime. As a result, reducing the sub-threshold and tunneling gate leakage currents has
become crucial in the design of ICs. This paper presents a new method to reduce the total leakage power
dissipation of static random access memories (SRAMs) while maintaining their performance.
Keywords - Leakage currents, SRAM, power consumption
I. INTRODUCTION
With diminishing process feature sizes and
operating voltages, the control of leakage currents in
modern VLSI designs is becoming a significant
challenge. Leakage currents increase exponentially
with decreasing threshold voltages. The power
consumed by a design in the standby mode of
operation is due to leakage currents in its devices.
With the prevalence of portable electronics, it is
crucial to keep the leakage currents of a design small
in order to ensure a long battery life in the standby
mode of operation. The leakage current for a PMOS
or NMOS device corresponds to the Idsof the device
when the device is in the cutoff or sub-threshold
region of operation. The expression for this current
[2] is
Ids=W/L Ioe(Vgs-VT-Voff/ηVt)
(1-e(-Vds/Vt)
) (1)
Here Io and Voff (typically Voff = -.08V) are
constants, while Vt is the thermal voltage (26 mV at
300 K) and η is the sub-threshold swing parameter.
We note that Ids increases exponentially with a
decrease in VT . This is why a reduction in supply
voltage (which is accompanied by a reduction in
threshold voltage) results in exponential increase in
leakage.
Another observation that can be made from (1) is
that Ids is significantly larger when Vds>>ηVt. For
typical devices, this is satisfied when Vds~VDD. The
reason for this is not only that the last term of (1) is
close to unity, but also that with a large value of
VDS.VT would be lowered due to drain induced
barrier lowering (DIBL) VT decreases approximately
linearly with increasing Vds [3], [2]. Therefore,
leakage reduction techniques should ensure that the
supply voltage is not applied across a single device,
as far as possible.
In recent times, leakage power reduction has
received much attention in academic research as well
as industrial applications. Several means of reducing
leakage power have been proposed. In [8], the
authors propose a dynamic threshold MOSFET
design for low leakage applications. In this scheme,
the device gate is connected to the bulk, resulting in
high-speed switching and low leakage currents
through body effect control. The drawback of this
approach is that it is only applicable in situations
where is lower than the diode turn-on voltage. Also,
the increased capacitance of the gate signal slows the
device down, and as a result, the authors propose the
use of this technique for partially depleted silicon on
insulator (SOI) designs. Another methodology for
controlling leakage is the variable threshold (often
called VTCMOS) approach.
In such an approach, the device threshold
voltages are controlled dynamically by modifying the
device bulk voltage. This method offers the
advantage of decreasing the leakage in standby mode
while not increasing the delay in the active mode..
However, complex control circuitry is required to
generate and control the bulk voltages. Another
drawback is that the bulk terminals have to be
electrically isolated from the source terminals of the
devices and this may require some major changes in
cell layout. In another approach called the super cut-
off CMOS (SCCMOS) approach, the gate of a PMOS
device which gates the VDD supply is overdriven,
thereby reducing the leakage dramatically. This again
requires the design of complex circuitry to generate
and control the special over-driven voltage values.
II. MULT-THRESHOLD CMOS
The concept of Multi- Threshold CMOS
(MTCMOS) which has emerged as a very popular
technique for standby mode leakage power reduction.
In this technique, a high-threshold voltage
transistor is inserted in series with the power supply
and the existing design and ground as shown in
Figure. The working of this circuit is as follows.
During Active mode of operation, the high threshold
Vt transistors are turned on, thereby facilitating
normal operation of the circuit as there exists a direct
RESEARCH ARTICLE OPEN ACCESS

path from the output to ground and Vdd. During
Standby mode, these transistors are turned off
creating a virtual power supply and ground rail and
cutting off the circuit from supply.
Fig. 2.1 MTCMOS Structure
Since the high Vt transistors operating in standby
mode forces the circuit to go to”sleep”, they are also
known as sleep transistors. Thus, in this technique
both standard – and high - threshold voltages are
fabricated on a single chip. From a visual
perspective, these high Vt sleep transistors act as a
current gate to the designed circuit. For this reason,
this technique is also referred to as Power Gating.
Effect of Introducing Sleep Transistors in Active
Mode
Consider the Pull Up Network (PUN) of a static
CMOS circuit. A PMOS sleep transistor is inserted
between Vdd source and the PUN. During the normal
mode of circuit operation, the sleep transistors can be
modeled as a resistor R as shown in Figure.
Assuming that the current flowing into the transistor
is I, this resistance will cause a voltage drop across it,
say Vsleep. Therefore, the gate driving capability
reduces to Vdd −Vsleep from Vdd. This reduction in
driving capability causes degradation in circuit
performance.
To overcome this problem, it is essential to
lower the resistance R of the transistor as much as
possible. This in turn implies increasing the size
(width) of the the transistor, since the resistance of
the transistor is inversely proportional to its width.
This, however, comes at an expense of increased area
and dynamic power dissipation. Conversely, a small
size transistor would degrade the circuit speed. A
solution to this problem would be to reduce the
threshold voltage Vt but as seen from Equation 1, the
sub-threshold current and hence the leakage power
would increase exponentially. Hence, there is a clear
trade-off between area, power and delay metrics of a
circuit for low leakage designs.
Figure 2.2 Sleep Transistor modeled as Resistor
III. PROPOSED METHOD
This paper deals with low-leakage ASIC design
using specialized standard cells. Based on the
discussion of the previous section, we know that Ids
would be significantly larger When Vds>>ηVt. This
is because VT drops due to DIBL when Vds is Large.
This causes the first term of (1) to increase
exponentially, while the parenthesized term of (1) is
close to 1. Our approach to leakage reduction
attempts to ensure that the supply voltage is applied
across more than one turned-off device and one of
those devices is a high-VT device. This is achieved by
selectively introducing a high- VT PMOS or NMOS
supply gating device. By this design choice, we
obtain standard cells with both low and predictable
standby leakage currents. Our goal is to design
standard cells with predictably low leakage currents.
To achieve this purpose, we design two variants of
each standard cell. The two variants of each standard
cell are designated “H” and “L.” If the inputs of a cell
during the standby mode of operation are such that
the output has a high value, we minimize the leakage
in the pull-down network. So a footer device (a high-
VT NMOS with its gate connected to standby) is
used. We call such a cell the “H” variant of the
standard cell. Similarly, if the inputs of a cell during
the standby mode of operation are such that the
output has a low value, we minimize the leakage in
the pull-up network by adding a header device (a
high- VT PMOS with its gate connected standby), and
call such a cell the “L” variant of the standard cell.
This exercise, when carried out for a NAND3 gate,
yields Circuits shown in Fig. 1.Note that the
MTCMOS circuit is also shown here. Although the
PMOS and NMOS supply gating devices
(equivalently called header and footer devices
(devices shown shaded in Fig. 1) are shown in the
circuit for the MTCMOS design, such devices are in
practice shared by all the standard cells of a larger
circuit block.

Fig. 3.1 Transistor Level Description
Fig. 3.2. Layout flop-plan of HL-gates
We sized the header and footer devices so that
the worst-case output delay penalty over all gate
input transitions was no larger than 15% as compared
to the regular standard cell using low transistors
IV. SIMULATION RESULTS
Fig .4.1 High-Vt SRAM cell
Fig .4.2 D.C Response of High-Vt SRAM cell
Fig. 4.3 Low-Vt SRAM Cell
Fig. 4.4 D.C Response of Low-Vt SRAM cell

eFig. 4.5 MTCMOS SRAM Cell
Fig.4.6 D.C Response of SRAM Cell
V. CONCLUSION
In this paper, we have explored low-leakage
standard cell based ASIC design methodologies for
SRAM cell. We implemented low leakage SRAM
cell for different cases such as high-Vt, low-Vt and
MTCMOS methodology. By out technique we can
reduce the leakage currents to nearly 50-60% thereby
operating the cell at low power.
REFERENCES
[1] ITRS, San Jose, CA, “The international
technology roadmap for
semiconductors,”(2002).
[2] The Device Group, Dept. EECS, Univ.
California, Berkeley, CA, “BSIM3
Homepage,” (2001). [Online]. Available
[3] J. Rabaey, Digital Integrated Circuits: A
Design Perspective, ser. Prentice- Hall
Electronics and VLSI Series. Englewood
Cliffs, NJ: Prentice-Hall, 1996.
[4] J. T. Kao and A. P. Chandrakasan, “Dual-
threshold voltage tec hniques for low-power
digital circuits,” IEEE J. Solid-State
Circuits, vol. 35, no. 7, pp. 1009–1018, Jul.
2000.
[5] S. Mutoh, T. Douseki, Y. Matsuya, T. Aoki,
S. Shigematsu, and J. Yamada,“1-v power
supply high-speed digital circuit technology
with multithreshold-voltage CMOS,” IEEE
J. Solid-State Circuits, vol. 30, no. 8, pp.
847–854, Aug. 1995.
[6] T. Kuroda, T. Fujita, S. Mita, T. Nagamatsu,
S. Yoshioka, K. Suzuki, F. Sano, M.
Norishima, M. Murota, M. Kako, M. K. M.
Kakumu and T. Sakurai, “A 0.9-V, 150-
MHz, 10-mW, 4 mm 2, 2-D discrete cosine
transform core processor with variable
threshold-voltage (VT) scheme,” IEEE J.
Solid-State Circuits, vol. 31, no. 11, pp.
1770–1779, Nov. 1996.
[7] H. Kawaguchi, K. Nose, and T. Sakurai, “A
super cut-off CMOS (SCCMOS) scheme for
0.5-v supply voltage with picoampere stand-
by current,” IEEE J. Solid-State Circuits,
vol. 35, no. 10, pp. 1498–1501, Oct. 2000.
[8] F. Assaderaghi, D. Sinitsky, S. A. Parke, J.
Bokor, P. K. Ko, and C. Hu, “Dynamic
threshold-voltage MOSFET (DTMOS) for
ultra-lowvoltage VLSI,” IEEE Trans.
Electron Devices, vol. 44, no. 3, pp. 414–
422, Mar.1997.
[9] K. Kumagai, H. Iwaki, H. Yoshida, H.
Suzuki, T. Yamada, and S.Kurosawa, “A
novel powering-down scheme for low Vt
CMOS circuits,” in Dig. Tech. Papers,
Symp. VLSI Circuits, 1998, pp. 44–45.
[10] I. Hyunsik, T. Inukai, H. Gomyo, T.
Hiramoto, and T. Sakurai, “VTCMOS
characteristics and its optimum conditions
predicted by a compact analytical model,” in
Proc. Int. Symp. Low Power Electron. Des.,
2001, pp. 123–128.

Design of Memory Cell for Low Power Applications

More Related Content

What's hot (19)

Viewers also liked (18)

Similar to Design of Memory Cell for Low Power Applications (20)

Recently uploaded (20)

Design of Memory Cell for Low Power Applications