Inter national J our nal of Electrical and Computer Engineering (IJECE) V ol. 7, No. 2, April 2017, pp. 759 766 ISSN: 2088-8708 759       I ns t it u t e  o f  A d v a nce d  Eng ine e r i ng  a nd  S cie nce   w     w     w       i                       l       c       m     A New Instrumentation Amplifier Ar chitectur e Based on Differ ential Differ ence Amplifier f or Biological Signal Pr ocessing Zainul Abidin 1 , K oichi T anno 2 , Shota Mago 3 , and Hir oki T amura 4 1 Department of Materials and Informatics, Uni v ersity of Miyazaki, Japan 2,3 Department of Electrical and System Engineering, Uni v ersity of Miyazaki, Japan 4 Department of En vironmental Robotics, Uni v ersity of Miyazaki, Japan Article Inf o Article history: Recei v ed No v 9, 2016 Re vised Feb 24, 2017 Accepted Mar 9, 2017 K eyw ord: biological signal instrumentation amplifier dif ferential dif ference amplifier common-mode g ain resistor mismatches ABSTRA CT In this paper , a ne w Instrum entation Amplifier (IA) architecture for biological signal pro- cessing is proposed. First stage of the proposed IA archi tecture consists of fully balance dif ferential dif ference amplifier and three resistors. Its second stage w as designed by using dif ferential dif ference amplifier and tw o resistors. The second stage has smaller number of resistors than that of con v entional one. The IA architectures are simulated and compared by using 1P 2M 0 : 6 - m CMOS process. From HSPICE simulation result, lo wer common- mode v oltage can be achie v ed by the proposed IA archi tecture. A v erage common-mode g ain ( A c ) of the proposed IA architecture is 31 : 26 dB lo wer than that of con v entional one under 3 % resistor mismatches condition. Therefore, the A c of the proposed IA architecture is more insensiti v e to resistor mismatches and suitable for biological signal processing. Copyright c 2017 Institute of Advanced Engineering and Science . All rights r eserved. Corresponding A uthor: K oichi T anno Department of Electrical and Systems Engineering Institute of Education and Research for Engineering, Uni v ersity of Miyazaki 1-1, Gakuen Kibanadai Nishi, Miyazaki, 889-2192, Japan tanno@cc.miyazaki-u.ac.jp 1. INTR ODUCTION Biological signals processing by using wearable de vice is useful for health care system. Sensing of biological signals is a v ery challenging research. The bi ological signals, such as Electroencephalogram (EEG), Electrooculogram (EOG), Electrocardiogram (ECG), Electromyogram (EMG), and Axon Action Potential (AAP) ha v e amplitudes in the order of V to mV and frequencies span from DC to a fe w kHz, as sho wn in Fig.1 [1]-[3]. In order to detect and process the v ery weak and lo w frequenc y signals, design of analog front-end has to meet strict performance parameters. IA is often used in sensor interf ace. Three operational amplifiers (op-amps) based IA archite cture is often emplo yed to achie v e high signal-to-noise ratio in first block of sensor interf ace. In the case of biological signals sensing, the IA architecture needs lo w A c and it can be achie v ed by satisfying well-matched condition of resistors netw ork [4]-[10]. Ho we v er , in actual f abricat ed chips, res istors are often not well-matched and it deteriorates the A c . In reference [3], an IA architecture based on Fully Balanced Dif ferential Dif ference Amplifier (FBDD A) which its A c is lo w and insensiti v e t o resistor mismatches w as presented. In this paper , it is called as con v entional IA architecture. W ith same number of resistors (7 resistors i ncluding g ain-setting resistor), tw o op-amps in first stage of the 3 op-amps based IA architecture were replaced by using FBDD A. Lar ge common-mode input v oltage and limitation of circuit design of the FBDD A (in the first stage) produce some amount of common-mode v oltage. Under the resistor mismatches condition, common-mode v oltage reduction by the second stage is necessary . This paper focuses on reduction of remaining common-mode v oltage of first stage by ne w design of second stage. Furthermore, a ne w IA architecture based on Dif ferential Dif ference Amplifier (DD A) with smaller number of resistors, lo wer A c J ournal Homepage: http://iaesjournal.com/online/inde x.php/IJECE       I ns t it u t e  o f  A d v a nce d  Eng ine e r i ng  a nd  S cie nce   w     w     w       i                       l       c       m     DOI:  10.11591/ijece.v7i2.pp759-766 Evaluation Warning : The document was created with Spire.PDF for Python.
760 ISSN: 2088-8708 and more insensiti v e to resistor mismatches is presented. Figure 1. V oltage and frequenc y ranges of some biological signals Figure 2. Three op-amps based IA architecture 2. PR OBLEM OF CONVENTION AL INSTR UMENT A TION AMPLIFIER ARCHITECTURE In man y te xt books and literatures, under the condition of well-matched resistors netw ork ( R 2 = R 3 , R 4 = R 5 , and R 6 = R 7 ), deri v ation of output v oltage of the 3 op-amp based IA architecture sho wn in Fig.2 ( V out ), can be determined by V out = R 7 R 5 2 R 3 R 1 + 1 ( V in 2 V in 1 ) (1) Defining V in 1 and V in 2 as v cm v dm and v cm + v dm , respecti v ely ( v cm and v dm are common-mode v oltage and dif ferential input, respecti v ely). Representing resistor mismatch of R i as R i +1 (1 + i +1 ) j i 2 f 2 ; 4 ; 6 g , where i +1 is mismatch rate of R i +1 , the V out becomes [3] V out = R 7 R 5 2 R 3 R 1 (1 + + 3 ) + 1 + v dm + R 7 R 5 ( 1) v cm (2) where is coef ficient defined as follo ws = R 5 + R 7 R 5 (1+ 5 ) 1+ 7 + R 7 (3) IJECE V ol. 7, No. 2, April 2017: 759 766 Evaluation Warning : The document was created with Spire.PDF for Python.
IJECE ISSN: 2088-8708 761 Figure 3. IA architectures: (a) Con v entional (b) Proposed Eq. 2 indicates that the V out contains v cm . The is caused by resistor mismatches of second stage. Therefore, g ain of second stage is often set to 0 dB to a v oid deterioration of the A c . In this w ay , the A c of the 3 op-amps based IA architecture is sensiti v e to the resistor mismatches [3]. In order to o v ercome this problem, we proposed IA architecture sho wn in Fig. 3a [3]. The number of resistors in the con v entional IA architecture is as same as that in the Fig. 2. First stage of the con v entional IA architecture which is modified from [11] w as designed by using FBDD A, 2 ne g ati v e feedback resistors, and g ain-setting resistor . The FBDD A consists of 2 stages fully dif ferential g ain stage and Common-Mode Feed Back (CMFB) circuit as sho wn in Fig. 4. Since an ideal amplifier responds only to dif ferential v oltage, A c is zero in ideal case. Therefore, in ideal condition, the output v oltages of FBDD A ( V out 1 c; 2 c ) can be determined by [11] V out 1 c; 2 c = A f ( V in 2 V in 3 ) ( V in 1 V in 4 ) g (4) where A is the amplification of FBDD A. Using (4) and the defined V in 1 and V in 2 , under the same manner of resistor mismatch condition, output v oltages of the first stage ( V out 1 c; 2 c ) become as follo w . V out 1 c = R 3 R 1 (2 + 3 ) + 1 v dm (5) V out 2 c = R 3 R 1 (2 + 3 ) + 1 v dm (6) While, its s econd stage is op-amp based subtractor . Final output ( V outc ) of the con v entional IA architecture can be deri v ed as follo ws. V outc = R 7 R 5 R 3 R 1 (2 + 3 ) + 1 (1 + ) v dm (7) From the abo v e deri v ation, v cm can be theoretically rejected since passing first stage and lo wer A c can be achie v ed e v en though there are resistor mismatches. Therefore, we can set the g ain of the second stage lar ger than 0 dB (of fset v oltage must be considered). In actual condition, nonideality must be considered. Limitations in practical circuit design and de vice mis- matches produce some amount of v cm , especially FBDD A [12]. T ransistor mismatch often occurs due to channel width and length ( W =L ). Furthermore, mismatch of W =L v alue will af fect to mismatch of transconductance ( g m ) as mentioned in this deri v ation result [13, 14]. g m = @ I ds @ V g s (8) = s 2 C ox W L j I ds j (1 + V ds ) (9) = s 2 C ox W L j I ds j (10) A Ne w Instrumentation Amplifier Ar c hitectur e Based on Dif fer ential Dif fer ence Amplifier ... (Zainul Abidin) Evaluation Warning : The document was created with Spire.PDF for Python.
762 ISSN: 2088-8708 Figure 4. Circuit schematic of FBDD A Figure 5. Circuit schematic of op-amp Figure 6. Circuit schematic of DD A Since the FBDD A main component is cross-coupled amplifier (see Fig. 4), tra nsistor mismatch of cross-coupled amplifier is analyzed. In order to analyze the ef fect of transconductance mismatch of transistors M 17 , M 18 , M 19 , and M 20 (see Fig. 4), the cross-coupled amplifier can be simplified by replacing transistors M 7 and M 8 with R ss and transistors M 33 and M 34 with R d . Supplying the inputs with v cm , the common-mode g ain of the cross-coupled amplifier ( A ccr ) can be deri v ed as follo ws. A ccr = V out 2 c;p V out 1 c;p v cm = f ( g m 17 + g m 18 g m 19 g m 20 ) v cm ( g m 17 g m 18 ) V p ( g m 19 g m 20 ) V q g R d v cm = ( ( g m 17 + g m 18 g m 19 g m 20 ) g 2 m 17 + g 2 m 18 R ss ( g m 17 + g m 18 ) R ss + 1 g 2 m 19 + g 2 m 20 R ss ( g m 19 + g m 20 ) R ss + 1 ) R d (11) Since A c of FBDD A ( A cF B D D A ) is not infinite from (11), some amount of v cm may appear in V out 1 c; 2 c . Because IJECE V ol. 7, No. 2, April 2017: 759 766 Evaluation Warning : The document was created with Spire.PDF for Python.
IJECE ISSN: 2088-8708 763 of g ain or resistor mismatches of second stage, the v cm may appear and be amplified in V outc . Finally , this condition deteriorates the A c of the IA architecture. Ne xt, we discuss about ne w design of second stage in the proposed IA architecture. 3. PR OPOSED INSTR UMENT A TION AMPLIFIER ARCHITECTURE Proposed IA architecture is sho wn in Fig. 3b . The proposed IA architecture consists of 2 stages. First stage is as same as that of the con v entional one. It has the same output v oltages ( V out 1 p; 2 p ) as (5) and (6), respecti v ely . Second stage consists of Dif ferential Dif ference Amplifier (DD A) and tw o resistors which are independent each other . The DD A is 4 inputs single output amplifier . In ideal condition, the relationship can be defined as follo ws. V outp = A f ( V out 1 p V out 2 p ) ( V f V g ) g (12) Implementing DD A for second stage may reduce the remaining v cm of first stage (in V out 1 p; 2 p of Fig. 3b) because V outp sho wn in (12) has component of subtraction ( V out 1 p V out 2 p ) which is independent of resistor mismatch. Furthermore, re g arding resistors used in the second stage, number of resistors of the proposed IA architecture is smaller than that of the con v entional one. Using (12) and referring Eqs.(5) and (6), the output of the second stage ( V outp ) under the resistor mismatch condition can be deri v ed as follo ws. V outp = 2 1 + R f R g R 3 R 1 (2 + 3 ) + 1 v dm (13) From Eq. (13), e v en though A cF B D D A is finite, g ain of second stage is more than 0 dB, and resistor mismatch of second stage occurs, the v cm can be drastically reduced compared with con v entional one. As mentioned in Chapter 2, in IA sho wn in Fig. 2, resistor mismatches of second stage cause high A c (see Eq. (2)). Therefore, implementing 2 independent resistors ( R f and R g ), the second stage of the proposed IA architecture can be set to higher g ain to get hi gher dif ferential g ain with smaller ef fect to the A c . Furthermore, ne w design of second stage mak es the proposed IA architecture has lo wer A c and more insensiti v e to resistor mismatches than the con v entional one. 4. SIMULA TION RESUL T In this chapter , the IA architectures were e v aluated using 1P 2M 0 : 6 - m CMOS process. In order to compare the performance of the IA architectures, the transistor le v el circuit of 3 kinds amplifier were realized. The op-amp in second stage of con v entional IA architecture w as realized by widely used op-amp circuit sho wn in Fig. 5 [13, 15]. The DD A w as realized by circuit schematic sho wn in Fig. 6. It consists of 2 stages fully dif ferential g ain stage and phase compensation circuits ( R c 4 , C c 4 ) with single output. Fig. 4 sho ws the emplo yed FBDD A, which is modified from the reference [11]. The FBDD A is de v eloped from DD A by adding CMFB circuit and phase compensation circuits ( R c 2 and C c 2 ) because of its dif ferential output. In the CMFB circuit, V c is set to 0 V . T able 1. Simulation Condition Items V alue CMOS process 1P 2M 0 : 6 - m CMOS V dd [V] 2 : 5 V ss [V] 2 : 5 V c [V] 0 R bias [k ] 295 M 1 ; 3 [ m/ m] 1 : 3 / 2 , M = 2 M 4 14 [ m/ m] 1 : 3 / 2 , M = 4 M 15 28 [ m/ m] 16 : 1 / 3 , M = 2 M 29 40 [ m/ m] 3 : 3 / 2 , M = 2 R c 1 4 [k ] 9 C c 1 4 [pF] 0 : 5 Note: M means the number of parallel connection A Ne w Instrumentation Amplifier Ar c hitectur e Based on Dif fer ential Dif fer ence Amplifier ... (Zainul Abidin) Evaluation Warning : The document was created with Spire.PDF for Python.
764 ISSN: 2088-8708 Figure 7. FFT results of V outc and V outp Figure 8. Histogram of A c (dB) based on Monte Carlo analysis T able 2. Summary of the Simulation Results P arameters Con v entional IA Proposed IA A C analysis Dif ferential g ain [dB] 54 : 15 54 : 15 3 dB g ain bandwidth [kHz] 301 : 34 398 : 46 Po wer cons. [ W] 843 : 79 992 : 52 Monte Carlo analysis A v e. Ac [dB] 85 : 53 116 : 79 Noise perf ormance (PV) Input ref. noise [ V/ p H z ] 89 : 68 89 : 68 Output ref. noise [mV/ p H z ] 41 : 41 45 : 78 Note: PV = Peak V alue In order to e v aluate the A c , the V in 1 and V in 2 were supplied by v cm which is repre sented by sine w a v e signal with amplitude of 50 mV and frequenc y of 60 Hz. The resistors netw ork of both IA architectures (see Fig. 3) w as designed as R 1 = 10 k , R 2 = R 3 = R 6 = R 7 = 250 k , R 4 = R 5 = 25 k , R f = 9 k , and R g = 1 k . W e set same g ain for both stages and the ideal total dif ferential g ain of both IA architectures is 54 : 15 dB. The IA architectures were s imulated using HSPICE. The detailed simulation condition is sho wn in T able 1. Representing w orst case of resistor mismatches conditi on ( 3 %), HSPICE simulation w as done under the mismatch rates 3 = 7 = 3% and 5 = 3% . The resistors R 2 and R 6 become 257 : 5 k and R 4 becomes 24 : 25 k . Fig. 7 sho ws the FFT simulation result of V outc and V outp . At frequenc y 60 Hz, the v cm of the con v entional and proposed IA architectures reach 98 : 66 dBV and 137 : 34 dBV , respecti v ely . The proposed IA architecture has 38 : 68 dBV lo wer v cm than the con v entional one. Monte Carlo simulation w as done by 500 tim es to get data of A c with de viation of resistor mi smatch 3 % w as randomly gi v en to all resistors of both IA architectures. Fig. 8 sho ws histogram of A c . A v erage A c of the proposed and con v entional IA architectures are 116 : 79 dB and 85 : 53 dB, respecti v ely . The a v erage A c of the proposed IA architecture is lo wer than that of t he con v entional one. Lastly , the simulated performance of the IA architectures are listed in T able 2. 5. CONCLUSION In this paper , a ne w IA architecture based on DD A has been presented. Resistor mismatches ef fect to common-mode g ain of IA architectures has been identified and compared in theoretic al analysis and HSPICE sim- ulation. W ith same dif ferential g ain and smaller number of resistors, ne w design of second stage mak es the proposed IA architecture has lo wer common-mode g ain. Its ability to achie v e lo wer common-mode g ain under resistor mis- matches condition mak es it more suitable as a part of inte grated circuit for biological signal processing. This design w as submitted for f a b r ication. Actual chip e v aluation and de v elopment of lo w common-mode DD A and its transistor mismatch ef fect are considered as future w ork. IJECE V ol. 7, No. 2, April 2017: 759 766 Evaluation Warning : The document was created with Spire.PDF for Python.
IJECE ISSN: 2088-8708 765 A CKNO WLEDGEMENT This w ork is supported by VLSI Design and Education Center (VDEC), the Uni v ersity of T ok yo in collabo- ration with Synopsys, Inc. and Cadence Design Systems, Inc. REFERENCES [1] Xiaodan Zou, Xiao yuan Xu, Libin Y ao, and Y ong Lian, ”A 1-V 450-nW fully inte grated programmable biomed- ical sensor interf ace chip, IEEE J. Solid-State Circuits , v ol. 44, no. 4, pp. 1067-1077, Apr . 2009. [2] Chih-Jen Y en, W en-Y a w Chung and Mely Chen Chi. ”Micro-Po wer Lo w Of fset Instrumentation Amplifier IC Design F or Bio-Medical System Applications”. IEEE T ransactions On Circuits And Systems-I: Re gular P apers , v ol. 51, no. 4, pp. 691-699, Apr . 2004. [3] Z. Abidin, K. T anno, S. Mago, H. T amura, ”Lo w Common-Mode Gain Instrumentation Amplifier Architecture Insensiti v e to Resistor Mismatches”, IAES International Journal of Electrical and Computer Engineering , v ol. 6, no. 6, Dec. 2016. [4] IN A114, Precision Instrumentation Amplifier , T e xas Instruments Inc., Dallas, March 1998 [Online]. A v ailable: http://www .ti.com/lit/ds/symlink/ina114.pdf [5] R. P all as-Aren y and J. W ebster , ”Composite Instrumentation Amplifier for Biopotentials”, Annals of Biomedical Engineering , v ol. 18, no. 3, pp. 251-262, 1990. [6] A. A. Silv erio, W .-Y . Chung, and V . F . Tsai, ”A Lo w Po wer High CMRR CMOS Instrumentat ion Amplifier for Bio-impedance Spectroscop y”, 2014 IEEE International Symposium on Bioelectronics and Bioinformatics (ISBB) , pp. 1-4, 2014. [7] R. P all as-Aren y and J. G. W ebster , ”Common Mode Rejection Ratio in Dif ferential Amplifiers”, IEEE T ransac- tions on Instrumentation and Measurement , v ol. 40, no. 4, pp. 669-676, 1991. [8] K. K oli and K. A. Halonen, ”CMRR Enhancem ent T echniques for Current-Mode Instrumentation Amplifiers”, IEEE T ransactions on Circuits and Systems I: Fundamental Theory and Applications , v ol. 47, no. 5, pp. 622-632, 2000. [9] J. Szyno wski, ”CMRR Analysis of Instrumentation Amplifiers”, Electronics Letters , v ol. 14, no. 19, pp. 547-549, 1983. [10] Hw ang- Cherng Cho w and Jia -Y u W ang, ”High CMRR instrumentation amplifier for biomedical application, Proc. IEEE International Symposium on Signal Processing and Its Applications , pp. 1-4, Febr . 2007. [11] A. Hussain and I. Mohammed, A CMOS fully balanced dif ferential dif ference a mplifier and its applications, IEEE T rans . on Circuits and Systems-II: Analog and Digital Signal Processing , V ol. 48, No. 6, pp. 614-620, June 2001. [12] J.P . Martinez Brito, S. Bampi, ”A DC of fset and CMRR analysis in a CMOS 0.35 m op- erational transconductance amplifier using Pelgroms area/accurac y tradeof f, Microelectron. J (2008), doi:10.1016/j.mejo.2008.02.029 [13] P .E. Allen and D.R. Holber g, CMOS Analog Circuit Design , Second Edition. Oxford Uni v ersity Press, Ne w Y ork, 2002. [14] R. Dehghani, ”Design of CMOS operational amplifiers, Artech House (2013). [15] N. Mukahar and S. H. Ruslan, ”A 93.36 dB, 161 MHz CMOS Operational T ransconductance Amplifier (O T A) for a 16 Bit Pipeline Analog-to-Digital Con v erter (ADC) , IAES International Journal of Electrical and Computer Engineering , v ol. 2, no. 1, pp. 106-111, Feb . 2012. BIOGRAPHIES OF A UTHORS Zainul Abidin w as born in 1986. He recei v ed the B. Eng. from Uni v ersity of Bra wijaya and M. Eng. from Uni v ersity of Miyazaki in 2008 and 2011, respecti v ely , and is currently w orking for Uni v ersity of Bra wijaya and to w ard the PhD de gree in Department of Materials and Informatics at Uni v ersity of Miyazaki. He has been in v olv ed with design of analog inte grated circuit since Master De gree. His current research interest includes analog circuit for biological signal processing. He is af filiated with IEEE and IEICE as student member . A Ne w Instrumentation Amplifier Ar c hitectur e Based on Dif fer ential Dif fer ence Amplifier ... (Zainul Abidin) Evaluation Warning : The document was created with Spire.PDF for Python.
766 ISSN: 2088-8708 K oichi T anno w as born in Miyazaki, Japan, on April 22, 1967. He recei v ed B. E. and M. E. de- grees from the F aculty of Engineering, Uni v ersity of Miyazaki, Miyazaki, Japan, in 1990 and 1992, respecti v ely , and Dr . Eng. de gree from Graduate School of Science a nd T echnology , K umamoto Uni v ersity , K umamoto, Japan, in 1999. From 1992 to 1993, he joined the Microelectronics Prod- ucts De v elopment Laboratory , Hitachi, Ltd., Y ok ohama, Japan. He w as eng aged in research on lo w-v oltage a nd lo w-po wer equa lizer for re ad channel LS I of ha rd disk dri v es. In 1994, he joined Uni v ersity of Miyazaki, where he is currently a Professor in the Department of Electrical and Sys- tems Engineering. His main research interests are in analog inte grated circuit design and multiple- v alued logic circuit design. Dr . T anno is a member of IEEE and the Ex ecuti v e Subcommittee of the IEEE Computer Society T echnical Committee on Multiple-V alued Logic. Shota Mago w as born in 1992. He recei v ed the B.Eng from Uni v ersity of Miyazaki in 2015, and is currently studying to get mast er de gree of Electrical and Electronic Engineering at Uni v ersity of Miyazaki. His current research is Analog CMOS Inte grated Circuits. Hir oki T amura recei v ed the B.E and M.E de gree from Miyazaki Uni v ersity in 1998 and 2000, respecti v ely . From 2000 to 2001, He w as an Engineer in Asahi Kasei Corporation, Japan. In 2001, He joined T o yama Uni v ersity , T o yama, Japan, where He w as a T echnical Of ficial in Department of Intellectual Information Systems. In 2006, He joined Miyazaki Uni v ersity , Miyazaki, Japan, where He w as an Assistant Professor in Departm ent of Electrical and El ectronic Engineering. In 2012, He is currently a Professor in the Department of En vironment al Robotics. His main research interests are Neural Netw orks and Optimization Problems. In recent years, He has the interest in Biomedical Signal Processing using Soft Computing. IJECE V ol. 7, No. 2, April 2017: 759 766 Evaluation Warning : The document was created with Spire.PDF for Python.