A Novel High Speed and Energy Efficient 10 Transistor Full Adder Design
A Novel High Speed and Energy Efficient 10 Transistor Full Adder Design
A Novel High Speed and Energy Efficient 10 Transistor Full Adder Design
1. INTRODUCTION
With the advance of VLSI technology, many computing intensive applications such as ultimedia processing, digital communication can now be realized in hardware to either speed up the operation or reduce the power/energy consumption. The essence of the digital computing lies in the full adder design. The design criteria of a full adder are usually multi-fold. Transistor count is, of course, a primary concern which largely affects the design complexity of many function units such as multiplier and algorithmic logic unit (ALU). Two other important yet often conflicting design criteria are power consumption and speed. A better metric would be the power delay product or energy consumption per operation to indicate the optimal design tradeoffs. Related to the power consumption is the lowest supply voltage in which the design can still operate properly. Numerous full adder designs [1][8] in the categories of static CMOS, dynamic circuit, transmission gate, or pass transistor logic have been presented in the literature. The full adder design in static CMOS with complementary pull-up pMOS and pull-down nMOS networks is the most conventional one but it requires as many as 28 transistors. Dynamic circuits can significantly reduce the transistor count but the incurred power consumption, including that of the clock tree, is usually high. Building logic in transmission gate is another alternative to reduce the circuit complexity. In [1], a transmission gate plus inverter based full adder design using 20 transistors was presented.The circuit can operate with full output voltage swing. In [2], the designs were further reduced to only 16 transistors while maintaining the full output voltage swing operation. To pursue even lower transistor count full adder designs, pass transistor logic (PTL) can be used in lieu of transmission gate. In [3], PTL based XOR/XNOR circuits were devised using only 4 transistors. Despite the saving in transistor count, the output voltage level is degraded at certain input combinations. This XOR circuit was adopted in [4] and the full adder design consists of only 14 transistors. In this design, an inverter is employed to generate the complementary signal of A xor B . The design also features full swing operations. The design was further improved in [5], where a 16-transistor full adder design was proposed. At the cost of two additional
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
transistors, this design can eliminate the inverter from the critical path and thus avoid the possible short circuit power consumption for low power operation. As the designs with fewer transistor count and lower power consumption are pursued, it becomes more and more difficult and even obsolete to keep full voltage swing operation. Note that in pass transistor logic, the output voltage swing may be degraded due to the threshold loss problem. That is, the output high (or low) voltage is deviated from the Vdd (or ground) by a multiple of threshold voltage VT . The reduction in voltage swing, on one hand, is beneficial to power consumption. On the other hand, this may lead to slow switching in the case of cascaded operation such as ripple carry adder. At low Vdd operation, the degraded output may even cause circuit malfunction. Therefore, for designs using reduced voltage swing, special attention must be paid to balance the power consumption and the speed. In [6], a PTL based new static energyrecovery full (SERF) adder with as few as ten transistors was presented. In spite of its claimed superiority in energy consumption, the design is relatively slower than peer designs and cannot be cascaded at low Vdd operation due to multiple-threshold loss problem. In [7], improved 10transistor full adder designs were derived based on systematic exploration of the combinations of various XOR/XNOR, Sum and Cout modules. Out of 41 circuit combinations, three designs with better power consumption were selected.Again, these designs suffer from the severe threshold loss problem and cannot operate properly in cascade under low supply voltage. The lowest possible power supply is limited to 2 VTn+VTp, where VTn and VTp are the threshold voltages of nMOS and pMOS respectively. In [8], another 10-transistor full adder design consisting of two pass transistor based XORs and a 2-to-1 multiplexer was presented. Its Cout (Vdd-VTn~VTp) voltage swing is degraded by a total of 3VTs. As usual, the cascaded operation in low supply voltage becomes problematic. In this paper, we will propose a novel 10transistor full adder design with alleviated threshold loss problem. This leads to faster ripple carry additions while maintaining the performance edge in energy consumption per operation. The design can also sustain lower V dd operation than peer designs.The different fulladders under consideration were compared for parameters such as power consumption and delay.The simulation results, based on .25-m process and were done by using DSCH and PSPICE. The performance of the proposed design becomes more significant as the word length of the adder increases .
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
From (3) and (4), we can easily identify two basic modules needed in implementing the functions, i.e., XOR and 2-to-1 multiplexer. As illustrated in Fig. 2.1, an XOR/XNOR function can be achieved with only 4 transistors in PTL [3]. Circuits in (a) and (b) are inverter based structures while circuits in (c) and (d) are PTL based structures. Assume both inputs have full voltage swing, the output voltages subject to different input combinations are compiled in Table I. In inverter based circuits, both output high and low voltages are degraded by the magnitude of a threshold voltage under certain input combinations. In PTL based circuits, the situation is improved and only output high or output low voltage is degraded. If the degraded output is used to control the gate of the subsequent stage in pass transistor logic, further voltage degradation may occur. However, degraded outputs do not necessarily imply adverse effect on the circuit performance .It depends on whether they cause further voltage drop in the following stage or not. As for the 2-to-1 multiplexer, possible circuit designs are illustrated in Fig. 3.2. In (a), two transmission gates are employed and complementary select control signals are required. Despite its merit of full output voltage swing, the circuit complexity is considered too high. In (b)(d), pass transistors are used in lieu of the transmission gate to reduce the circuit complexity. The price to pay is the degraded output voltage swing. Assume all inputs have full voltage swings, the output voltage levels of these circuits are summarized in Table I.
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
(a)
XOR |VTp | Vdd- VTn Vdd Gnd
(b)
XNOR Vdd Gnd |VTp| Vdd- VTn
(c)
XOR |VTp | Vdd Vdd Gnd
(d)
XNOR Vdd Gnd Gnd Vdd- VTn
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
Sum =( A xor Cin) and not(Cout) or (A xnor Cin) and B Cout=( A xor Cin) and B or (A xnor Cin) and A
(5) (6)
The logic block diagram of the proposed design is shown in Fig. 3.1. The design rationales are as follows: First, try to avoid the usage of degraded output in the following stage as gate control signals. This is the common problem existing in most 10-transistor full adder designs. It will lead to multiple threshold voltage losses and may hinder the cascaded circuit from correct operation. Second, try to eliminate unbuffered carry signal propagation in a pass transistor chain. According to Elmore formula, the propagation delay is a quadratic function of the number of cascaded pass transistors. Even for moderate number of cascade length, the delay is still intolerable.
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
As shown in Fig. 3.2, the XNOR circuit adopted in the proposed design is realized by a 2to-1 multiplexer followed by an inverter.The role of the inverter is 3-fold. Firstly, it is used as a level restoring circuit to combat the output threshold voltage loss. The level restored output is then fed to MUX 2/3 to generate Sum and Cout signals. The threshold voltage loss of Sum and Cout will be confined to only one |VT |away from the power supplies. Secondly, the inverter (INV 2) serves as a buffer along the carry chain to speed up the carry propagation. Thirdly, the inverter (INV 2) provides complementary signals needed in the following stage. The availability of complementary signals also helps simplify the XNOR design, where only one signal is needed in the selection control. The MOS circuit schematic design of a CLRCL full adder is depicted in Fig. 3.3 . We adopt the circuit in Fig.3. 2 (b) to implement the 2-to-1 multiplexer in that it is the one without using complementary select controls. The entire full adder circuit requires only ten transistors (5 pMOS and 5 nMOS)the one with the least transistor count . In the following section, we will conduct various analyses and simulations to demonstrate the performance superiority of the proposed adder over other designs.
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
Table II COMPARISION OF 2 TO 1 MULTIPLEXER CIRCUITS 2-to-1 multiplexer circits Transmission gate Pass transistor Double NMOS Double PMOS 2 2 2 No Yes Yes Vdd VTn Vdd VTn Vdd |VTp | Gnd |VTp| 4 Yes Vdd Gnd Transistor count Complementary control select Output high(min) Output low(max)
Fig. 3.3 MOS circuit schematic design of the CLRCL full adder
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
10
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
11
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
12
can be calculated by multiplexing a and cin controlled by (a exnor b). Let us consider that there is a capacitor at the output node of the first XNOR module. To illustrate static energy recovery let us consider an example where initially a=b=0 and then a changes to 1. When a and b both equals to zero the capacitor is charged by VDD. In the next stage when b reaches a high voltage level keeping a fixed at a low voltage level, the capacitor discharges through a. Some charge is retained in a. Hence when a reaches a high voltage level we do not have to charge it fully. So the energy consumption is low here. It is the adder so reported to be the circuit consuming power in 10t topologies. It should be noted that the new SERF adder has no direct path to the ground. The elimination of a path to the ground reduces power consumption, removing the Psc variable (product of Isc and voltage) from the total power equation. The charge stored at the load capacitance is reapplied to the control gates. The combination of not having a direct path to ground and the reapplication of the load charge to the control gate makes the energy-recovering full adder an energy efficient design. To the best of our knowledge this new design has the lowest transistor count for the complete realization. The circuit produces full-swing at the output nodes. But it fails to provide so for the internal nodes. As the power consumption by the circuit reduces the circuit becomes slower. Also it cannot be cascaded at low power supply due to multiple threshold problem.
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
13
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
14
5. PERFORMANCE ANALYSIS
In this paper, several comparable full adder designs are included for performance comparision.Since the design goals are low circuit complexity and high-speed operation subject to competitive energy consumption, we focus mainly on low-gate-count and static power transitor based full adder designs.Besides the proposed CLRCL full adder design, eight more designs are employed. It includes 28T static CMOS, TG-CMOS, 16T, 14T, SERF, 9A and 9B. For the analysis, first conduct static circuit analyses to determine the voltage swings of different adder designs. Comparing only 10 transistor designs, where the threshold voltage loss problem comes hand in hand with the low-gate-count approach.Among the three inputs of a full adder,inputs A and B are assumed to be perfect and have full voltage swing.Input Cin ,however is drawn from the Cout of another full adder .As a result Cin might be degraded due to the threshold voltage loss.Theoretical analysis reveals that among the 10 T designs,only proposed CLRCL design encounters only one threshold loss along the unbuffered Cout signal, while others encounters two threshold voltage losses. The simulations are based on 0.25m process and uses 3.3-V power supply.Typical employed. transistor sizes,i.e.,(W/L)p=2m/0.25m and (W/L)n=1m/0.25m are
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
15
Since many 10T full adder designs fail to function properly at low Vdd operations at low operations and use Vdd=3.3V in our worst case delay simulations. The rise time and the fall time are set to be 0.01 ns. The delay (in nanoseconds) is the period between the applying of input pattern and the availability of at MSB. The simulations are conducted for different adder sizes, ranging from 2-bit to8bit. The results are summarized in graph Since the proposed CLRCL design has complementary carry signals, two delay numbers are given. The maximum working frequency ( in megahertz) is the reciprocal of the larger delay. Among the eight designs under comparison, CLRCL, 9A, 9B and SERF are 10T designs. The remaining 4 designs, i.e., 14T, 16T, TFA, TG-CMOS and 28T, are higher gate count designs with full voltage swing
operations. They are considered to have better speed performance, when compared with 10T designs, at the cost of increased circuit complexity.it is clear that the proposed CLRCL design has the minimum delays among all 10T designs. And the delay gap between our design and the other designs becomes even wider as the size of ripple adder grows.
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
16
900 800 700 600 500 400 300 200 100 0 1 BIT 2 BIT 4 BIT 8 BIT WORD LENGTH
DELAY (CARRY)
16 T
Fig. 6.1: Delays of different ripple adder designs Next considering the power consumption of various adders at different bit length.From the Table .III ,it is clearly understandable that power consumption is low for the proposed CLRCL full adder among the 10-T designs.And for the high gate count designs,the power consumption increases with increase in wordlength.
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
17
TABLE .III
POWER CONSUMPTION FOR DIFFERENT ADDER DESIGNS
2 Bit Power(Watts) 1.81E-10 1.72E-10 2.03E-11 1.74E-11 3.20E-10 3.20E-10 3.50E-10 3.50E-10
4 Bit Power(Watts) 3.73E-10 3.80E-10 4.00E-11 3.37E-11 6.40E-10 6.40E-10 6.99E-10 6.99E-10
8 Bit Power(Watts) 7.57E-10 6.89E-10 8.13E-11 6.63E-11 1.28E-9 1.28E-9 1.40E-9 1.40E-9
On the basis of power consumption for different adder designs for different wordlength,can plot its relationship in graphically and is shown in fig. 5.2. Graphical analysis clearly show that the proposed CLRCL full adder has lower power consumption. This design adopts inverter buffered XOR/XNOR designs to alleviate the threshold voltage loss problem and to enhance the driving capability for cascaded for cascaded operations.It successfully embeds the buffering circuit in the full adder design,so the transistor count is minimized.
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
18
7. FUTURESCOPE
Instead of ripple adder designs ,can use carry look ahead logic for performance improvement. Instead of 4T xnor circuit, can use 3T xnor circuit for circuit complexity reduction.
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
19
8. CONCLUSION
In conclusion in this paper, presented a novel 10T full adder design with alleviated threshold loss problem.This design adopts inverter buffered XOR/XNOR designs.This enhances the driving capability for cascaded operations.Also it leads to faster ripple carry additions and also can sustain lower Vdd operation than peer designs.All the results are obtained in 250 nm CMOS process technology with a 3.3V of supply voltage.This paper compares various full adder configurations both the low gate count and high gate count full adder designs.Eight more full adder designs performances were evaluated and conclude that the proposed CLRCL design can successfully reduce the dc power dissipation caused by input signal degradation and thus performs better than the other 10T designs. If a specific input pattern is applied, the standby power dissipation can be further minimized to leakage components only.
Dept. of ECE
A NOVEL HIGH SPEED AND ENERGY EFFICIENT 10 TRANSISTOR FULL ADDER DESIGN
20
9. REFERENCES
[1] N. Weste and K. Eshraghian, Principles of CMOS VLSI Design, A System Perspective. Reading, MA: Addison-Wesley, 1993. [2] Jin-Fa Lin,Yin-Tsung Hwang,Ming-Hwa Sheu,Cheng-Che Ho, A Novel High-Speed and Energy Efficient 10-Transistor Full Adder Design,IEEE Trans.Circuits Syst.I regular papers,vol.54,no.5,may 2007. [3] N. Zhuang and H. Wu, A new design of the CMOS full adder, IEEEJ. Solid-State Circuits, vol. 27, no. 5, pp. 840844, May 1992. [4] [3] J. Wang, S. Fang, and W. Feng, New efficient designs for XOR and XNOR functions on the transistor level, IEEE J. Solid-State Circuits,vol. 29, no. 7, pp. 780786, Jul. 1994. [5] E. Abu-Shama and M. Bayoumi, A new cell for low power adders, in Proc. Int. Midwest Symp. Circuits Syst., 1995, pp. 10141017. [6] A. M. Shams and M. Bayoumi, A novel high-performance CMOS 1-bit full adder cell, IEEE Trans. Circuits Syst. II, Analog Digit. SignalProcess., vol. 47, no. 5, pp. 478481, May 2000. [7] R. Shalem, E. John, and L. K. John, A novel low-power energy recovery full adder cell, in Proc. Great Lakes Symp. VLSI, Feb. 1999, pp. 380383. [8] H. T. Bui, Y. Wang, and Y. Jiang, Design and analysis of low-power 10-transistor full adders using XOR-XNOR gates, IEEE Trans. CircuitsSyst. II, Analog Digit. Signal Process., vol. 49, no. 1, pp. 2530, Jan.2002.
Dept. of ECE