Paper 4.pdf

IEEE International Conference on Power, Control, Signals and Instrumentation Engineering (ICPCSI-2017)

Design Methodologies and Circuit Optimization Techniques for Low Power CMOS VLSI Design Dr.B.T.Geetha1 ,B.Padmavathi2 Associate Professor1,Assitant Professor2 Department of ECE Jeppiaar Maamallan Engineering College Chennai,India [email protected]

V.Perumal Associate Professor3. Department of CSE Saveetha Engineering College Chennai,India.

Abstract— Low power is the real test for late hardware businesses. Control scattering is an essential thought as far as execution and area for VLSI Chip outline. Control administration procedures are for the most part used to configuration low power circuits and frameworks. Insights demonstrate that 40% or considerably higher rate of the aggregate power utilization is because of the leakage of transistors. This rate will increment with innovation scaling unless streamlining methods are acquainted with bring leakage inside points of confinement. This paper concentrates on circuit improvement and plan mechanization strategies to fulfill this target. It additionally portrays many issues with respect to circuit outline at building, legitimate and gadget levels and exhibits different procedures to beat the previously mentioned issues. The initial segment of the paper gives a diagram of primary wellsprings of leakage current in CMOS transistor. The second part of the paper depicts various circuit streamlining procedures for controlling the standby leakage current. Some leakage current lessening procedures like rest approach; stack approach, and lector strategy are talked about for planning CMOS entryways which fundamentally chops down the leakage streams. The benefits of lector method are it doesn't require any extra control and observing hardware, in this way constraining the range increment and furthermore the power dissipation in dynamic state when contrasted with different systems and it doesn't influence the dynamic power which is the significant confinement with the other leakage diminishment strategies, is likewise examined in this paper.

circuits. In the zone of small scale controlled battery worked compact applications, for example, mobile phones, the objective is to keep the battery lifetime and weight sensible and bundling cost low. Scaling of CMOS gadgets has empowered the semiconductor business to take care of its demand for higher execution and higher coordination densities. However as the component measure gets to be distinctly littler, in light of short direct lengths it brings about expanded sub-threshold leakage current through a transistor when it is off condition. Another purpose behind expanded sub-threshold leakage current is that, transistors can't be turned off totally. Henceforth leakage control dissipiation has turned into a vital bit of the aggregate power utilization for silicon innovations. The significant three plan parameters are power, speed and area. In CMOS VLSI circuits, control dissipation is basically because of the three essential elements: dynamic, static and short out. A few enhancement strategies have been proposed for leakage current lessening. One essential thing in CMOS VLSI circuit configuration is to lessen the power dissipation while keeping up the superior of the circuit. The edge voltage must be scaled to keep up the execution of the circuit. In VLSI circuit outline the power utilization has turned into an essential issue. In this manner the real test in versatile framework configuration is to diminish the power utilization of coordinated circuits through outline improvements. Various strategies have been proposed by scientists to take care of the power utilization issue. In any case, there are no influences to meet trade-off between power, postponement and region .Appropriate systems must be picked that fulfils the application and item needs. At last, for better execution, no battery worked frameworks, for example, workstations, set-beat PCs and interactive media computerized flag processors, the general objective of force minimization is to lessen framework cost (cooling, bundling and vitality utilization) while guaranteeing long haul gadget unwavering quality. These diverse necessities impacts how control improvement is managed and how much the planner will bargain in cost or execution to get bring down power dissipation.

Keywords- VLSI, Power consumption, Dynamic power, Clock gating, lector method etc I. INTRODUCTION In the previous decades, the real test for the VLSI architect was region, execution, cost and power utilization. As of late, in any case, this has started to change and, progressively control utilization is being given practically identical weight to region and speed contemplations. The thoughts for lessening power utilization vary from application to application and circuits to

978-1-5386-0814-2/17/$31.00 ©2017 IEEE

1759

IEEE International Conference on Power, Control, Signals and Instrumentation Engineering (ICPCSI-2017) A.

II. SOURCES OF POWER DISSIPATION When we perceived power utilization as a plan limitation, Power per MHz is normally utilized as a recurrentation of a segment. Electric current is not steady amid operation and pinnacle power is a critical concern. The gadget will breakdown because of electro migration and voltage drops regardless of the possibility that the normal power utilization is low. The condition for the normal power utilization is given as Pavg = Pdynamic + Pshort + Pleakage + Pstatic So the aggregate normal power utilization relies on upon Dynamic power utilization, Short-circuit control utilization, Leakage control utilization and static power utilization. The leakage current is controlled by the manufacture innovation, which comprises of turn around inclination current in the parasitic diodes shaped amongst source and deplete dissipiations and the mass district in a MOS transistor and in addition the sub threshold current that emerges from the reversal charge that exists at the door voltages beneath the edge voltage. The short out current which is because of the DC way between the supply rails amid yield moves and the charging and releasing of capacitive burdens amid rationale changes. The short out and leakage streams in CMOS circuits can be made little with appropriate circuit and gadget outline procedures. The fundamental wellspring of force scattering is the charging and releasing of the junction capacitances. Exchanging movement is a measure for the quantity of entryways and their yields that change their bit an incentive amid a clock cycle. To flip between rationale zero and rationale one can released and charged the junction capacitor. The electric current that streams amid this procedure causes dynamic power dissipation Pdynamic. The dynamic power is rely on the capacitive yield stack Cout and the supply voltage Vdd and recurrence of clock flag. Pdynamic = K Cout Vdd2 f K is the average number of positive transitions during one clock cycle and f the clock frequency. By reducing the power supply will have a larger effect on saving power, taking into consideration that typically Pdynamic is responsible for 80% of Pavg. III. SOURCES OF STATIC LEAKAGE POWER DISSIPATION 1. Reverse-biased junction leakage current (Irev) 2. Gate induced drain leakage (Igidl) 3. Sub-threshold (weak inversion) leakage (Isub)

Junction leakage

At the point when a transistor is off, the junction leakage happens from the source or deplete to the substrate through the turned around one-sided diodes. This leakage current has two principle parts: one is minority bearer dissipation close to the edge of the exhaustion locale; the other is because of electronopening pair era in the consumption district of the turnaround one-sided junction. In the event of an inverter with low information voltage, the NMOS is OFF, the PMOS is ON, and the yield voltage is high. Thus the deplete to-substrate voltage of the OFF NMOS transistor is equivalent to supply voltage. This causes leakage current from the deplete to the substrate through the turned around one-sided diode. B.

Gate-induced drain leakage

The door instigated deplete leakage current, is the deplete tosubstrate leakage created by high field impact in the deplete junction of MOS transistors. At the point when the deplete of a NMOS is one-sided at the supply voltage (VDD) and the door is one-sided at either zero or negative voltage, an exhaustion locale is framed under the entryway and deplete cover district. This causes noteworthy band bowing in the deplete permitting electron gap combine era because of torrential slide augmentation. A profound exhaustion condition is made as the openings are quickly cleared out to the substrate. In the mean time electrons are gathered by the deplete, bringing about Gate Induced Drain Leakage current. C.

Sub-threshold leakage

Sub-threshold leakage is the current that streams between the source and deplete of a MOSFET. At the point when a transistor is working in the powerless reversal district, that is, for entryway to-source voltages beneath the limit voltage. This is brought about because of the diffusion current of the minority carriers in the channel for a MOS gadget. The subthreshold leakage current is much bigger than the other leakage current segments. This is mostly as a result of the moderately low VT in cutting edge CMOS gadgets. IV. LEAKAGE CONTROL IN ACTIVE MODE A.

Multiple threshold cells

To manage the leakage issue, multi limit CMOS circuit, which has both high and low edge transistors in a solitary chip can be utilized. Sub limit leakage current is stifled by the high edge transistors while low edge transistors are utilized to accomplish the superior. However utilization is constrained to utilizing low VT transistors, with addition rates on the request of 20% or less since the standby power is much bigger for low VT transistors contrasted with the high VT transistors. The dynamic power

1760

IEEE International Conference on Power, Control, Signals and Instrumentation Engineering (ICPCSI-2017)

part or the outline size is not unfavourably influenced since the TOX and entryway lengths are same for high and low VT transistors. The disadvantage of this strategy is that variety because of doping is uncorrelated between the high and low limit transistors and additional cover steps acquire a procedure cost. B.

Long channel devices

There is VT move off because of short channel impact. Consequently dynamic leakage of CMOS entryways can be diminished by expanding their transistor channel lengths. However the entryway capacitance increments for longer transistor lengths used to accomplish high edge transistors. This negatively affects the execution and element control dissipation. Long channel inclusion has comparative or lower prepare cost contrasted and various limit voltages taken as the size increment instead of the veil cost. The disadvantage of this system is that the dynamic power dissipation of the up-sized entryway builds relative to the compelling channel length increment. Consequently movement variable of the influenced entryways must be low to spare the circuit control scattering. V. STANDBY MODE LEAKAGE REDUCTION TECHNIQUE

power supply and ground. Amid standby mode, rest transistors are killed, along these lines making the virtual way between the power supply and the ground. It is important to have rest transistor with high limit esteem else it will have high leakage current. This innovation is actualized with a few sorts of transistors with various VT values. Transistors with a low VT are utilized to actualize the rationale, while high VT gadgets are utilized as rest transistors.

B.

Sub-threshold power reduction techniques

Sub-limit leakage current is turning into the essential wellspring of force dissipiation in CMOS beneath 90nm. Subsequently power is supplanting different components like execution and silicon zone in planning CMOS circuits. 1.

Conventional CMOS technique.

In this method reciprocal CMOS circuit is utilized that has a NMOS pull down system to interface with "0" (GND) and PMOS pull up system to associate with the yield to "1" (VDD). It is a fundamental approach that is utilized for the most part in all methods. Fig 2 demonstrates the piece outline.

Most electronic frameworks invest significant energy in a standby state. Thus battery life stays sit out of gear for long stretch aside from when being used and depletes off a large portion of the power, which thus decreases the battery life. Hence battery life can be spared by closing it down from the supply when not being used. A.

Multi-threshold CMOS technique

In this procedure, high edge voltage transistor is associated in arrangement with the power supply and existing PMOS and NMOS arrangement circuit to make a virtual ground and a virtual power supply as appeared in the fig. 1

Fig 2. Conventional CMOS circuit structure

2.

. Stack technique:

Another strategy for leakage control diminishment is stack approach, that depends on the way that characteristic stacking of MOSFET help to lessen the leakage current. At the point when two transistors associated in arrangement are killed together, instigated turn around predisposition between the two transistors causes in the decrease of sub-limit leakage current. However isolated transistors increment defers essentially restricting the execution. Fig.3 demonstrates the circuit schematic for the same.

Fig.1 Power gating circuit

In dynamic express, the rest transistor is ON, encouraging typical operation as there is an immediate way between the

1761

IEEE International Conference on Power, Control, Signals and Instrumentation Engineering (ICPCSI-2017)

VI. SYSTEM DESIGN Framework level low power plan systems ought to be most encouraging for decreasing vitality utilization. Two systems are meant in this segment. A.

Fig.3 Stack technique circuit structure

3.

Modified lector technique.

The essential thought behind LECTOR procedure for leakage current lessening depends on the successful stacking of transistors in the way from supply voltage to ground. This depends on the idea that "a state with more than one transistor OFF in a way from supply voltage to ground is less cracked contrasted with a state with just a single transistor OFF in any supply to ground way. Fig.4 demonstrates the circuit schematic for the LECTOR method. Between hubs N1 and N2 two LCTs (leakage control transistor) are currented. The entryway terminal of each LCT is controlled by the wellspring of the other. Henceforth called self controlled stacked transistors. As LCTs are self controlled, no outside circuit is required. In this manner the restriction with the rest transistor system has been overcome. The leakage current is diminished in light of the fact that the currentation of LCTs builds the resistance of the way from VDD to ground. In this strategy, two LCTs were included each CMOS entryway, a PMOS added to the draw up system and a NMOS added to the draw down system and the door terminal of one LCT is controlled by the wellspring of the other. Accordingly one of the LCTs is constantly close to its cut-off area of operation for any offered contribution to the CMOS entryway. This gives extra resistance in the way from supply voltage to ground consequently diminishing the subthreshold leakage current, thusly the static power.

HW/SW Partitioning

The MicroPP Plus will be an installed framework with two processor centres, a controller and a DSP. Administrations can be actualized either in programming running on these centers, or in committed equipment. In our plan stream this will be chosen amid the progression of hardware software dividing. This procedure has awesome impact on framework control. This is shown in the accompanying case: Specific equipment is for the most part more productive. 330mW are expended to play out an expansion utilizing a SPARClite processor centre in a commendable innovation (0.32mm/1.8V/16.8 MHz). A custom viper in a similar innovation expends just 2mW or more extra correspondence overhead. B.

Integration of Chip Components

Actualizing frameworks utilizing current day innovation brings about third or a greater amount of aggregate power being expended at the chip's information/yield (I/O) ports. The bigger capacitances of chip's limits contrasted with inner doors and higher voltages are the explanation behind this perception. Common qualities for inner capacitances dwell around 10's of femtofarads, where I/O pins achieve measurements of 10's of microfarads. These days supply voltages for chip cores have a tendency to be lower then 2.0V. In mechanical frameworks not all segments of an outline may be best in class and require higher voltages or specialized limitations require them. Still, these distinctive parts need to impart over their I/O. This makes double voltage frameworks (bring down voltage for the centers – higher voltage for I/O) very normal. VII. CONCLUSION The leakage current turns out to be more basic in compact frameworks with profound sub-micron and nanometre advancements, where battery life is of essential concern. Dealing with leakage power is an incredible test in nanometre scale innovations since sub-limit leakage control gets to be distinctly good to element control utilization. The upside of Lector system is that as it doesn't require any extra control and checking circuits it doesn't influence the dynamic power. A low voltage/low limit innovation and circuit configuration approach, focusing on supply voltages around 1 Volt and working with decreased edges.

Fig.4 Lector CMOS gate

1762

IEEE International Conference on Power, Control, Signals and Instrumentation Engineering (ICPCSI-2017)

•

•

Low power interconnect, utilizing propelled innovation, lessened swing or diminished action approaches. Dynamic power administration systems, changing supply voltage and execution speed as per movement estimations. This can be accomplished by parcelling the plan into sub-circuits whose vitality levels can be freely controlled and by shutting down sub-circuits which are not being used

References [1].

Anjana R, Ajay kumar somkuwar, "Analysis of sub threshold leakage reduction techniques in deep sub micron regime for CMOS VLSI circuits"

[2].

Farzan Fallah and Massoud Pedram, " Standby and Active Leakage Current Control and Minimization in CMOS VLSI Circuits".

[3].

Gary K. Yeap, “Practical Low Power Digital VLSI Design”, Kluwer Academic, Publishers, 1998.

[4].

Kanika Kaur and Arti Noor, “Strategies & methodologies for low power vlsi designs: a review” International Journal of Advances in Engineering & Technology, May 2011. ISSN: 2231-1963.

[5].

K.Srilaxmi, Y. Syamala and A. Suvir Vikram, "Static power optimization using dual sub- threshold supply voltages in digital CMOS VLSI circuits", International Journal of VLSI design & Communication Systems (VLSICS), Vol. 4, No. 5, October 2013.

[6].

K.R.N.karthik, M.Nagesh Babu, V.NarasimhaNayak, S.Rajeswari and Dr. FazalNoorbasha, "Cadence Design of Leakage Power Reduction Circuit in CMOS VLSI Design ", International Journal of Engineering Research and Applications, Vol. 3, Issue 2, March - April 2013, pp. 738 741.

[7].

Luca Benini, Giovanni De Micheli, Enrico Macii, "Designing LowPower Circuits : Practical Recipes

[8].

Massoud Pedram Design Technologies for Low Power VLSI. Department of EE-Systems University of Southern California Los Angeles, Encyclopedia of Computer Science and Technology, 1995.

[9].

W.C.Athas, L. J. Svensson, J.G.Koller, N.Thartzanis and E. Chou. " Low-Power Digital Systems Based on Adiabatic-Switching Principles. " IEEE Transactions on VLSI Systems, 2(4)398407:, December 1994.

[10]. Sonam Rathore, "Low Leakage Power SRAM Design Using Lector Technique In Various CMOS Technology", International Journal of Advanced Research in Computer Science and Software Engineering, Volume 4, Issue 6, June 2014, pp. 279 - 284.

1763