Int. Journal of Electrical & Electronics Engg.
Vol. 2, Spl. Issue 1 (2015)
e-ISSN: 1694-2310 | p-ISSN: 1694-2426
Energy Efficient Design of Multiplexer Using Adiabatic logic 1
Richa Singh, 2Prateek Raj Gautam, 3Anjali Sharma
1,2
Dept.of Electronics& Communication, Allen House Institute of Technology, Rooma, Kanpur,India Dept.of Electronics & Communication, AP Goyal University, Shimla, India
1
singhricha51@gmail.com, 2prateekrajgautam@gmail.com, 3Anjali.iitt@gmail.com
Abstract—the increasing prominence of portable systems and the need to limit the power consumption in very high density VLSI chips have led to rapid and innovative developments in low power design during the recent years. The CMOS technology provides circuits with very low static power dissipation, during the switching operation currents are generated, due to the discharge of load capacitances that cause power dissipation increasing with the clock frequency. The adiabatic technique prevents such losses, the charge does not flow from the supply voltage to the load capacitance and then to ground, but it flows back to a trapezoidal or sinusoidal supply voltage and can be reused.In this paper a low 2:1 multiplexer is designed using positive feedback adiabatic logic. The design is simulated at .12¾m technology using Microwind 3.1. Simulated results shows that proposed design saves 38% energy as compare to conventional CMOS design.
traversed by charge that flows on to the load capacitance [2-3]. By using smaller voltage steps or increments dissipation can be reduced [3]. The minimum power consumption during the charge transfer phase is termed as adiabatic switching.
Keywords- VLSI, PFAL, Adiabatic, BSIM, Multiplexer.
INTRODUCTION The need for low power design is becoming a major issue in high performance digital systems, such as microprocessors, digital signal processors, and other application. The common traits of high performance chips are the high integration density and high clock frequency. The power dissipation of the chip, and thus, the temperature, increases with the increasing clock frequency. Since the dissipated heat must be removed effectively to keep the chip temperature at an acceptable level. There are many ways t achieve low power in digital circuits, it involves reduction of the switching events, decrease the node capacitance, reduce the voltage swing or apply a combination of these methods. Yet in all these methods energy drawn from the power supply is used only once before being dissipated[1]. . In CMOS logic design half of the power is dissipation in PMOS network and stored energy is dissipated during discharging process of output load capacitor during the switching events. Most of the power consumption reduction techniques are based upon scaling of the supply voltage, reducing capacitance and switching activity. Yet in all these cases, energy drawn from the power supply is used only once before being dissipated. Thus to increase the energy efficiency of logic circuits, a technique is required that can reuse the energy stored on load capacitor. It has been found that there is a fundamental relation between computation and power dissipation. That is if somehow computation could be implemented without any loss of information, then the energy required by it could be potentially reduced to zero. This can be done by performing all computation in reversible manner. Also energy dissipation depends upon average voltage drop NITTTR, Chandigarh
EDIT -2015
Fig.1 Conventional CMOS logic circuit with pull-up (F) and pull-down (/F) networks
ADIABATIC PRINCIPLE Most of the power saving techniques involved scaling of the power supply, which results, substantial increase in subthreshold leakage current also it causes uncertainty in the process variation. Therefore some other technique is required which is independent of voltage scaling. It has been found that there is fundamental connection between computation and power dissipation. That is if somehow computation could be implemented without any loss of information, then energy required by it could be potentially reduced to zero. This can be achieved by performing all the computation in a reversible manner. Thus minimum power consumption during charge transfer phase is known as adiabatic switching [4]. In Fig. 2output load capacitance is charged by a constant current source instead of a constant voltage source used in conventional CMOS structures. This circuit is same as the equivalent model used in charging process in conventional CMOS. On resistance of pull up PMOS network is represented by R and C0 is the output capacitance. It is noted that constant charging current corresponds to a linear voltage ramp. Energy dissipated through adiabatic logic is given as [5-6]
104
Int. Journal of Electrical & Electronics Engg.
Ed
RC .C .V C2 T T
Vol. 2, Spl. Issue 1 (2015)
(1)
Fig.2 Equivalent model during charging phase in adiabatic circuits [4]
It can be seen from equation from equation 1 that If charging period T is larger than 2RC then dissipated energy can be made smaller than Conventional CMOS circuit. Thus dissipated energy can be made arbitrarily small by increasing the charging period. Hence by using constant current source energy can be transferred from supply to load capacitor with any dissipation and the energy stored on the load capacitance after charging process send back to the supply voltage by simply reversing the direction of current source. Thus recycling is very attractive feature in adiabatic logic. The constant current supply must be capable of retrieving the charge back to the power supply as shown in fig. 2. Adiabatic circuits does not use standard power supply instead of this it uses pulsating power supply which is also called as pulsed power supply[7-8].
e-ISSN: 1694-2310 | p-ISSN: 1694-2426
advantages of CPL logic are good output driving capability due to outputinverters, fast differential stage due to cross coupled PMOS structure and small input loads. The main disadvantage of CPL logic is largenumber ofnodes and high overhead due to dual rail signal. Schematic design of CPL MUX is shown in fig. 5 In the energy economized pass-transistor logic (EEPL), the sources of the PMOS pull-up transistors of a CPL gate are connected to the complementary output signal instead of Fig. 5 The main advantage is smaller delay and smaller power dissipation as compare to CPL. Becauseof regenerative positive feedback which provides shorter delay than CPL logic. It has same structure as CPL MUX employing two PMOS and four NMOS instead of a positive feedback [12-13]. It is shown in Fig. 6.
Fig.3 Schematic design of DCVSL 2:1 Multiplexer [31].
PREVIOUS WORK
A logic style is the way how a logic function is derived from a set of transistors. It affects the speed, size, and power consumption and wiring complexity of a circuit. All these characteristics may vary considerably from one logic style to another and thus make the proper choice of logic style crucial for circuit performance. Cascade voltage logic switch (CVSL) is developed by IBM. Later it is known as differential cascade switch logic (DCVSL) shown in fig. 3. The designing of DCVSL logic style requires both its true andcomplementary signal to be routed. It is made of two ntype switching networks and two p-type switching networks connected in a cross coupled manner to VDD. The MDCVSL stands for modified differential cascade voltage switch logic. Delay has been improved by adding two NMOS in the previous design. It is shown in fig. 4 this circuit also provides self checking feature that is if circuit is operating correctly, the values at the output may assume 0-1 or 1-0 means the combination such as 0-0 or 1-1 will never occur[9]. Complementary pass transistor logic (CPL) is based upon pass transistors networks. The CPL circuit requires complementary inputs and generates complementary outputs to pass on the next CPL that is in this logic for every signal its complement is generated. Elimination of PMOS transistors reduces the parasitic capacitances associated with each node in the circuit Gates are static, because the output is connected to either VDD or GND. Design is modular; same cell can produce various gates by simply permuting the input signals. CPL requires fewer transistors[10-11]. The threshold voltages of NMOS must be reduced to zero through threshold adjustment implants. It performs very fast operation as compare to CMOS. The 105
Fig. 4 Schematic design of MDCVSL 2:1 Multiplexer [31].
Fig. 5 Schematic design of CPL 2:1 Multiplexer [20].
Fig.6 Schematic design of EEPL 2:1 Multiplexer [20] PROPOSED WORK
This design 2:1 MUX is based upon a pair of cross coupled inverters. In this latch is made from two PMOS and two NMOS that avoids the degradation of the logic level at the output node. These NMOS devices areconnected between output and ground. A sinusoidal supply is applied. This logic familyalso generates both positive and negative outputs. The functional blocks are in parallel with the PMOSFETs of the adiabatic amplifier and form a transmission gate. The two n-trees realize the logic functions. This logic family also generates both positive NITTTR, Chandigarh
EDIT-2015
Int. Journal of Electrical & Electronics Engg.
Vol. 2, Spl. Issue 1 (2015)
e-ISSN: 1694-2310 | p-ISSN: 1694-2426
and negative outputs. It is known as positive feedback adiabatic logic [14-15]. An adiabatic 2:1 multiplexer based on PFAL is designed on DSCH 3.1. This circuit implements the function F A S BS shown in fig. When the select (S) line is low, the output node follows thesignal A and when the select (S) line is high, theoutput nodefollows the signal B, respectively. Schematic is shown in fig. 7 Fig.10 Layout simulation of 2:1 adiabatic multiplexer
Layout simulation of X-OR/X-NOR circuit is shown in fig. 11The waveforms verify the correct logic of the circuit. Range of the voltage used for analog signal is 0-1.2V.
Fig. 7 Adiabatic design of 2:1 multiplexer.
An adiabatic XOR gate based on positive feedback logic is designed and simulated. Timing waveforms and its schematic is shown in fig. 8. The minimum sized XORgate is implemented at 0.12µm technology. In the given circuit complementary output is obtained.
Fig 8 Design of a XOR gate using positive feedback adiabatic logic.
Fig. 11 Layout simulation of 2 input X-OR/X-NOR gate. RESULT & COMPARISON In this paper circuits based on various logic styles are compared with the adiabatic circuits. These logic styles include CMOS & complementary CMOS family that is DCVSL, MDCVSL, CPL, EEPL. Adiabatic circuits are designed using DSCH and simulation is performed on Microwind 3.1 tool with 120nm technology.Table 1 shows maximum darin current and energy for different logic style multiplexers. It has been observed from the table that proposed multiplexer is very energy efficient as compare to other multiplexers. Also darin current is minimum for proposed multiplexer which indicates low power consumption. Table.1Comparison table of drain current & energy for different muliplexer,
LAYOUT & SIMULATION
Physical layout of a positive feedback adiabatic multiplexer is designed using Microwind 3.1 and simulation is performed using BSIM4 model. Schematic of an inverter is designed using DSCH 3.1 tool. Layout of the circuit is achieved after compiling the verilog file, in the Microwind. A verilog file is a kind of netlist consisting all thecomponents and connections used in designing of a circuit [16-17]. Layout of inverter is shown in fig. 9
Different Logic Max Drain Energy (fJ) Multiplexers Current (mA) MDCVSL 0.672 179.3 DCVSL 0.604 98.2 CPL 0.511 11.74 EEPL 0.489 9.4 CMOS 0.462 5.9 PROPOSED 0.161 3.6 Drain current is a stong function of power consumption, means power dissipation largely depends upon darin current. Fig. 12 shows variation of drain current with supply volage at 270C temperature. PFAL is more power efficient than DCVSL multiplexer.
Fig. 9 Layout Representation of 2:1 adiabatic multiplexer.
Analog simulation is performed on the layout of multiplexer design. Fig. 10 shows time domain simulation of Multiplexer. Logic ‘0’ corresponds to a zero voltage and logic ‘1’ corresponds to 1.2V. A sinusoidal signal is applied as power clock supply with amplitude 0.8V. Simple clocks are applied as inputs and select lines NITTTR, Chandigarh
EDIT -2015
106
Int. Journal of Electrical & Electronics Engg.
Vol. 2, Spl. Issue 1 (2015)
Fig. 12 Power consumption vs supply voltage for CMOS and adiabatic compressors. 6)
CONLUSION Multiplexer through various logic style has been designed and simulated using DSCH & Microwind 3.1. These logic styles includes DVCSL, MDCVSL, CPL and EEPL multiplexer design. It is observed that proposed Multiplexer shows better performance in terms of power consumption. It is recorded that 49.56% and 68.32% improvement is obtained in terms power consumption as compare toCMOS It has been observed that proposed multiplexer saves 38.9% energy as compare to CMOS multiplexer. All results are verified at different supply voltage and temperature. Proposed Multiplexer shows good performance with supply voltage vs temperature & supply voltage vs drain current variations as compare to EEPL, CPL, DCVSL, MDCVSL multiplexer.energy as compare to CMOS multiplexer. All results are verified at different supply voltage and temperature. Proposed Multiplexer shows good performance with supply voltage vs temperature &supply voltage vs drain current variations as compare to EEPL, CPL, DCVSL, MDCVSL multiplexer REFERENCES 1) ....Sung-Mo Kang, Yusuf Leblebici, CMOS DigitalIntegrated Circuits: Analysis and Design TATA McGRAW-HILL, pp. 197-519. 2) ....N. Weste and K. Eshraghian, Principles of CMOS VLSI Design: A System Perspective reading pearson education, Addision-Wesley pp. 145-148. 3) H. J. M. Veendrick, “Short-circuit Dissipation of Static CMOS Circuitry and its Impact on the Design of Buffer Circuits,” IEEE Journal of Solid State Circuits, pp. 468473thinfilmsandexchangeanisotropy,” inMagnetism,vol.III,G.T.RadoandH.Suhl,Eds.NewYork:Academic,1 963,, August 1984. 4) I.S.JacobsandC.P.Bean, “Fineparticles,pp.271–350. A. P. Chandrakasan and R. W. Brodersen, Low-power CMOS digital design, Kluwer Academic, Norwell, Ma, pp. 45-89, 1995. 5) Richa Singh, Anjali Sharma, Rohit Singh, “Power Effcient Design of multiplexer based Compressor Using Adiabatic Logic,”
107
7)
8) A. 9)
10)
11)
12)
13)
14)
15)
16)
e-ISSN: 1694-2310 | p-ISSN: 1694-2426
International Journal of Computer Applications, pp. 45-50, Vol. 81, November 2013.. Richa Singh, Rajesh Mehra, “Power Efficient Design of Multiplexer Using Adiabatic Logic,” International Journal of Advances in Engineering & Technology, pp. 246-254, Vol. 6, March 2013. M. Young, TheTechnical Writer's Handbook. Mill Valley, CA: University Science, 1989. Richa Singh, Rajesh Mehra, “Power Efficient Design of Multiplexer Using Adiabatic Logic,” International Journal of Advances in Engineering & Technology, pp. 246-254, Vol. 6, March 2013. M. Young, TheTechnical Writer's Handbook. Mill Valley, CA: University Science, 1989. P. Chandrakasan and R. W. Brodersen, Low-power CMOS digital design, Kluwer Academic, Norwell, Ma, pp. 45-89, 1995. Varga, L.Kovacs, F. Hosszu, G. “An Improved PassGate Adiabatic Logic,” IEEEConference on Application Specific Integrated chip, pp. 208-211, 2001 Yingtao Jiang, Al-Sheraidah, Yuke Wang, Edwin Sha and Jin-Gyun Chung, “A Novel Multiplexer-Based Low-Power Full Adder,” IEEE Transactions on Circuit and Systems-2nd:Express Brief, Vol. 51, pp. 345-348, July 2004. Suhwan Kim, Ziesler, Conrad H. Ziesler and Marios C. Papaefthymiou, “Charge-Recovery Computing on Silicon,” IEEE Transactions on Computers, Vol. 54, pp. 651-659, June 2005. Hsu-Wei Huang, Cheng-Yeh Wang, Jing Yang Jou, “An Efficient Heterogeneous Tree Multiplexer Synthesis Technique,” IEEE Transactions on Computer-Aided design of Integrated Circuits and System, Vol. 24, pp. 1622-1629, October 2005. Muhammad Arsalanand Maitham Shams, “Charge-Recovery Power Clock Generators for Adiabatic Logic Circuits,” IEEE Conference on Embedded System designs, pp. 171-174, 2005. SambhuN.Pradhan, Gopal Paul, Ajit Pal, Bhargab B. Bhattacharya, “Power Aware BDD-Based Logic Synthesis Using Adiabatic Multiplexer,” IEEE Conference on Electrical and Computer Engineering, pp. 149-152, December 2006. XuJian, Wang Peng-jun, Zeng Xiao-yang, “Research of Adiabatic Multiplier Based on CTGAL,” IEEE Conference on Application Specific Integrated Chip, pp. 138-141, 2007. Massimo Alioto and Gaetano Palumbo, “Design of Large Fan-In CMOS Multiplexer Accounting for Interconnects,” IEEE Transactions on Circuits And Systems-2nd:Express Briefs, Vol. 54, pp. 484-488, June 2007.
NITTTR, Chandigarh
EDIT-2015