Galley Proof 13/10/2017; 9:01 File: thc–1-thc171401.tex; BOKCTP/xjm p.

Technology and Health Care -1 (2017) 1–7 1

DOI 10.3233/THC-171401
IOS Press

1 Biosignal integrated circuit with

2 simultaneous acquisition of ECG and PPG
3 for wearable healthcare applications

Hyungseup Kima , Yunjong Parka , Youngwoon Koa , Yeongjin Muna , Sangmin Leeb and

on
4

5 Hyoungho Koa,∗
a Department

si
6 of Electronics Engineering, Chungnam National University, Daejeon, Korea
7
b Department of Biomedical Engineering, Kyung Hee University, Yongin, Korea

er
fv
8 Abstract.
9 BACKGROUND: Wearable healthcare systems require measurements from electrocardiograms (ECGs) and photoplethysmo-
oo

10 grams (PPGs), and the blood pressure of the user. The pulse transit time (PTT) can be calculated by measuring the ECG and
11 PPG simultaneously. Continuous-time blood pressure without using an air cuff can be estimated by using the PTT.
12 OBJECTIVE: This paper presents a biosignal acquisition integrated circuit (IC) that can simultaneously measure the ECG and
pr

13 PPG for wearable healthcare applications.

14 METHODS: Included in this biosignal acquisition circuit are a voltage mode instrumentation amplifier (IA) for ECG acqui-
sition and a current mode transimpedance amplifier for PPG acquisition. The analog outputs from the ECG and PPG channels
ed

15
16 are muxed and converted to digital signals using 12-bit successive approximation register (SAR) analog-to-digital converter
17 (ADC).
ct

18 RESULTS: The proposed IC is fabricated by using a standard 0.18 µm CMOS process with an active area of 14.44 mm2 . The
19 total current consumption for the multichannel IC is 327 µA with a 3.3 V supply. The measured input referred noise of ECG
rre

20 readout channel is 1.3 µVRMS√with a bandwidth of 0.5 Hz to 100 Hz. And the measured input referred current noise of the PPG
21 readout channel is 0.122 nA/ Hz with a bandwidth of 0.5 Hz to 100 Hz.
22 CONCLUSIONS: The proposed IC, which is implemented using various circuit techniques, can measure ECG and PPG signals
co

23 simultaneously to calculate the PTT for wearable healthcare applications.

Keywords: Biosignal acquisition, simultaneous biosignal acquisition, electrocardiogram (ECG), photoplethysmogram (PPG),
un

24
25 pulse transit time (PTT), chopper stabilization

26 1. Introduction

27 Recently, healthcare, mobile, and wearable applications for the Internet of Things (IoT) have been
28 gaining attention. Wearable healthcare applications require measuring biosignals such as electroen-
29 cephalogram (EEG), electrooculogram (EOG), electromyogram (EMG), electrocardiogram (ECG), pho-
30 toplethysmogram (PPG), and body temperature. In addition, by measuring the ECG and PPG, the pulse
31 transit time (PTT) can be calculated. The continuous-time blood pressure without using an air cuff can

∗
Corresponding author: Hyoungho Ko, Department of Electronics Engineering, Chungnam National University, 99 Daehak-
ro, Yuseong-gu, Daejeon 34134, Korea. Tel.: +82 42 821 5664; Fax: +82 42 823 5436; E-mail: hhko@cnu.ac.kr.

0928-7329/17/$35.00
c 2017 – IOS Press and the authors. All rights reserved
 Galley Proof 13/10/2017; 9:01 File: thc–1-thc171401.tex; BOKCTP/xjm p. 2

2 H. Kim et al. / Biosignal IC with simultaneous acquisition of ECG and PPG

32 be estimated by calculating the PTT [1–3]. The previous researches reported that the estimated blood
33 pressure using PTT and the types of the blood pressures, including the systolic blood pressure (SBP),
34 the diastolic blood pressure (DBP) and the mean arterial pressure (MAP), have strong relationships with
35 the correlations greater than 87% [2]. One of the important parts of an ECG monitoring IC is the fre-
36 quency response of the amplifier of the AFE. The capacitive coupled instrumentation amplifier (CCIA)
37 scheme is adopted for the instrumentation amplifier of the ECG monitoring IC. However, a mismatch
38 between the input capacitor and feedback capacitor rejects the common mode rejection ratio (CMRR)
39 and degrades the performance of the biosignal acquisition IC [4–7]. To overcome this problem, a current
40 balanced instrumentation amplifier (CBIA) scheme has been reported [5–7]. Because the CBIA scheme
41 is stable without a feedback loop between the input and output, no passive elements are needed. The mis-
42 match between passive elements such as resistors and capacitors causes a degradation of the CMRR. In
43 addition, the CBIA scheme is suitable for low power applications because of its low power consumption
44 and small active area. The PPG monitoring system is composed of photodiodes, light emitting diodes

on
45 and a transimpedance amplification stage to optically detect the blood volume changes from the skin.
46 In the PPG monitoring system, the large DC offset current exists, and the DC offset currents limit the

si
47 dynamic range of the PPG ICs, thus, the DC offsets of the PPG signals should be removed. Several re-

er
48 searches have been reported for reducing the DC offsets of the PPG monitoring ICs [8–10]. The previous
49 DC rejection schemes adopt the correlated double sampling (CDS) [9], or on-chip large time-constant
fv
50 circuits including pseudo resistor [8,10].
oo

51 2. Proposed biosignal acquisition IC for wearable healthcare applications

pr

52 A block diagram of the proposed biosignal acquisition IC for wearable healthcare applications is
ed

53 shown in Fig. 1a. The ECG signals are amplified by the IA after filtered by the HPF and modulated by the
54 chopper. The HPF is designed with a pseudo resistor to gain a low cutoff frequency. The implemented
ct

55 CHPF of HPF is 482 pF and RHPF of HPF is controllable through 2 GΩ to 58 GΩ by 4-bit control
register. Therefore, the cutoff frequency of the HPF is controllable approximately 0.0596 Hz to 0.1 Hz
rre

56

57 to reject the DC voltage. The amplified output signal is amplified again by the programmable gain
58 amplifier (PGA) using the differential difference amplifier (DDA). On the other hand, the PPG readout
co

59 channel has the inputs TIA_INP, TIA_INN and Sink_LED. The Sink_LED input is for the LED driver.
60 The LED driver can control the current of the photodiode of the pulse oximeter to monitor peripheral
un

61 oxygen saturation (SpO2 ). The TIA amplifies the input signals by a current-to-voltage conversion. The
62 PPG signal is amplified again by the PGA as the ECG readout channel. Each amplified ECG and PPG
63 signal is buffered out after passing through the LPF. The channel selection mux allows for selection of
64 a desired channel by using a digital control pin. The buffered analog signals that are selected by the
65 channel selection mux are converted into 12-bit digital code by the 12-bit SAR ADC. The proposed IC
66 can monitor sensed signals and acquire the data as digital code for the back end DSP.

67 2.1. Instrumentation amplifier (IA) for ECG readout channel

68 The ECG IA of the proposed IC is shown in Fig. 1b. The IA input stage is a transconductance amplifier
69 stage, and the IA output stage is a transimpeance amplifier stage. The currents of MP1 and MP3 are fixed
70 by each current source of current IB1 and IB2 . The input transistor MP1 is a source follower so that the
71 input dependent current flow at R1 is determined as
 Galley Proof 13/10/2017; 9:01 File: thc–1-thc171401.tex; BOKCTP/xjm p. 3

H. Kim et al. / Biosignal IC with simultaneous acquisition of ECG and PPG 3

on
si
er
fv
oo
pr
ed
ct
rre
co

Fig. 1. Top level overview of the proposed biosignal acquisition IC.

INP − INN
un

IR1 = (1)
R1
72 The current mirrored to MP5 of the IA output stage can be expressed as (where K is the W/L ratio of
73 the current between MP2 and MP5)
OUTP − OUTN
= KIR1 (2)
R2
74 Therefore, the transfer function of CBIA is given as
VOUT OUTP − OUTN R2
= =K (3)
VIN INP − INN R1

75 2.2. Transimpedance amplifier (TIA) for PPG readout channel

76 A schematic of the TIA of the proposed PPG readout channel is shown in Fig. 1c. The current mode
 Galley Proof 13/10/2017; 9:01 File: thc–1-thc171401.tex; BOKCTP/xjm p. 4

4 H. Kim et al. / Biosignal IC with simultaneous acquisition of ECG and PPG

77 TIA amplifies the input current signal into an output voltage signal by a current-to-voltage conversion.
78 The photocurrent IP D of the photodiode flows through the inputs INP and INN of the TIA. The pho-
79 tocurrent is modulated by the chopper of the input stage. The TIA input stage is biased by the two fixed
80 bias currents IB . The two currents IB + IP D and IB − IP D , which flow through MP1, are mirrored by
81 MP3 of the TIA output stage. The currents mirrored from MP1 to MP3 are amplified by the W/L ratio
82 of K between MP1 and MP3. The mirrored currents are converted into output voltage signals after being
83 demodulated by the current flow of resistor R. The currents mirrored by the TIA output stage to MP3
84 can be expressed as (where K is the ratio of the current between MP1 and MP3)
OUTN − OUTP
= KIP D (4)
R
85 Therefore, the output voltage of TIA is given as
VOUT = OUTP − OUTN = −KRIP D (5)

on
86 2.3. Programmable gain amplifier (PGA) and DC servo loop

si
The schematic of the PGA integrated in the proposed IC is shown in Fig. 1d. The PGA stage is a two

er
87

88 stage resistive amplifier implemented using a differential difference amplifier (DDA). The gain of the
fv
89 PGA is controllable by adjusting the 4-bit programmable resistor RF . The programmable resistor can
90 adjust the resistance from 500 kΩ to 7.5 MΩ, and the resistance of Rref has a fixed value of 1 MΩ. the
oo

91 output voltage of the PGA can be written as

 
RF
pr

Vout = 1 + (Vinp − Vinn ) + Vref (6)

Rref
The DC servo loop of the proposed IC is shown in Fig. 1e. The DC servo loop removes the DC offsets
ed

92

93 (VOS ) by negative feedback into the input stage after modulating the integrated output signals OUTP
94 and OUTN of the IA. The output voltage affected by the DC servo loop can be expressed as
ct

 Z 
R2 1
rre

VOUT (t) = OUTP − OUTN = K (VINP − VINN ) + VOS − VOUT (t)dt (7)
R1 RDSL CDSL
co

95 3. Prototype IC implementation and measurements

un

96 3.1. Measurement environment

97 The measurement environment for the proposed IC is shown in Fig. 2, in which information about
98 the measurement equipment can be seen. The fabricated chip is implemented on the PCB board by the
99 chip on board (COB) process. A wet electrode and clip electrode are used for ECG signal acquisition.
100 SpO2 is used for PPG acquisition. The output data is stored in an Arduino Due, and data is acquired by
101 a laptop computer. In addition, the direct analog output signal of the proposed IC can be measured. A
102 dynamic signal analyzer is used to measure the input referred noise and the frequency response of the
103 proposed IC.

104 3.2. Measurement results

105 The measurement results are shown in Fig. 3. The total current consumption of the proposed IC is
 Galley Proof 13/10/2017; 9:01 File: thc–1-thc171401.tex; BOKCTP/xjm p. 5

H. Kim et al. / Biosignal IC with simultaneous acquisition of ECG and PPG 5

Fig. 2. Measurement environment for proposed biosignal acquisition IC.

on
si
er
fv
oo
pr
ed
ct
rre
co
un

Fig. 3. Measurement results for proposed biosignal acquisition IC.

106 327 µA with a 3.3 V power supply. The measured input referred noise of the ECG readout channel is
107 1.3 µVRMS with a bandwidth of 0.5 Hz to 100 Hz. This is shown in Fig. 7a. The frequency response
108 of the ECG readout channel is shown in Fig. 7b. The measured DC gain of the transfer function of the
109 ECG readout channel is 70.4 dB. The input referred current noise of the PPG readout channel is shown√ in
110 Fig. 7c. The measured result of input referred current noise of the PPG readout channel is 0.122 nA/ Hz
111 with a bandwidth of 0.5 Hz to 100 Hz and a TIA gain of 14.25 MΩ.
 Galley Proof 13/10/2017; 9:01 File: thc–1-thc171401.tex; BOKCTP/xjm p. 6

6 H. Kim et al. / Biosignal IC with simultaneous acquisition of ECG and PPG

Table 1
Performance comparison
[6] [8] [11] [12] [13] This work
Process 0.5 µm 0.18 µm 0.18 µm 0.13 µm 0.18 µm 0.18 µm
Supply voltage 3V 1.2 V 3.3 V 1.2 V 1V 3.3 V
Total current consumption 20 µA 160 µA 3.8 µA 22 µA – 327 µA
(without LED (with ECG + PPG
power) readout + LED
power)
Simultaneous biosignal NO NO NO NO NO YES
acquisition
DC servo loop YES YES YES YES YES YES
ECG readout channel input 0.574 µVRMS 1.3 µVRMS 0.94 µVRMS – 1.3 µVRMS 1.3 µVRMS
referred noise (100 Hz BW)
ECG readout channel transfer 60 dB 50 dB 71.9 dB – 60 dB 70.4 dB
DC gain (100 Hz BW) √ √

on
PPG readout channel input – – – 0.026 nA/ Hz – 0.122 nA/ Hz
referred current noise
Die area 1.95 mm2 40.00 mm2 10.50 mm2 19.84 mm2 6.25 mm2 14.44 mm2

si
er
112 The simultaneous acquisition of the ECG and PPG signals is shown in Fig. 7d. All ECG and PPG
113 signals are clearly visible. In addition, the different waveforms of P and QRS are clearly distinguishable
fv
114 by the high performance of the proposed IC. The proposed biosignal acquisition IC can measure the ECG
115 and PPG signals at the same time. This simultaneous acquisition makes it able to acquire continuous-
oo

116 time blood pressure by calculating the PTT.

pr

117 4. Conclusions
ed

118 This paper presented a biosignal acquisition integrated circuit for wearable healthcare applications.
ct

119 The feature of the proposed IC is that ECG and PPG signals can be acquired simultaneously. The si-
120 multaneous measurement of the ECG and the PPG signals can be used to calculate the PTT. The PTT
rre

121 can be used to estimate continuous-time blood pressure. The proposed biosignal acquisition IC has two
122 channels: an ECG readout channel for ECG signal acquisition, and a PPG readout channel for PPG
co

123 signal acquisition. A performance comparison with previous works is listed in Table 1. The fully inte-
124 grated biosignal acquisition IC has functions and performance that are suitable to various applications
un

125 of wearable healthcare systems.

126 Acknowledgments

127 This research was supported by the Nano·Material Technology Development Program through the
128 National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT and Future
129 Planning (NRF-2015M3A7B7045527). This research was also partly supported by a grant from Kyung
130 Hee University in 2017 (KHU-20171202) and supported by IC Design Education Center (IDEC).

131 Conflict of interest

132 None to report.

 Galley Proof 13/10/2017; 9:01 File: thc–1-thc171401.tex; BOKCTP/xjm p. 7

H. Kim et al. / Biosignal IC with simultaneous acquisition of ECG and PPG 7

133 References

134 [1] Yoon Y, Cho J, Yoon G. Non-constrained blood pressure monitoring using ECG and PPG for personal healthcare. J.
135 Med. Syst., 2009, 33(4): 261–266.
136 [2] He X, Goubran R, Liu X. Evaluation of the correlation between blood pressure and pulse transit time. Proceedings of
137 IEEE International Symposium on Medical Measurements and Applications, 2013: 1–4.
138 [3] He X, Goubran R, Liu X. Secondary peak detection of PPG signal for continuous cuffless arterial blood pressure mea-
139 surement. IEEE Trans. Instrum. Meas., 2014, 63(6): 1431–1439.
140 [4] Qian C, Parramon J, Sancjez-Sinencio E. A micropower low-noise neural recording front-end circuit for epileptic seizure
141 detection. IEEE J. Solid-State Circuits, 2011, 46(6): 1392–1405.
√
142 [5] Yazicioglu R, Merken P, Puers R, Hoof C. A 60µW 60nV/ Hz readout front-end for portable biopotential acquisition
143 system. IEEE J. Solid-State Circuits, 2007, 42(5): 1100–1110.
144 [6] Yazicioglu R, Merken P, Puers R, Hoof C. A 200µW eight-channel EEG acquisition ASIC for ambulatory EEG system.
145 IEEE J. Solid-State Circuits, 2008, 43(12): 3025–3038.
146 [7] Helleputte N, Kim S, Kim H, Kim J, Hoof C, Yazicioglu R. A 160 µA biopotential acquisition IC with fully integrated
147 IA and motion artifact suppression. IEEE Trans. Biomed. Circuits Syst., 2012, 6(6): 552–561.

on
148 [8] Harrison R, Charles C. A low-power low-noise CMOS amplifier for neural recording applications. IEEE J. Solid-State
149 Circuits, 2003, 38(6): 958–965.
150 [9] Rankow A, Rodriguez-Villegas E, Lee M. A novel correlated double sampling poly-si circuit for readout systems in large

si
151 area x-ray sensors. Proceedings of IEEE International Symposium on Circuits and Systems, 2005: 728–731.
152 [10] Normandin J, Sawan M, Faubert J. A new integrated front-end for a noninvasive brain imaging system based on near-

er
153 infrared spectroreflectometry. IEEE Trans. Circuits Syst. I, Reg. Papers, 2005, 52(12): 2663–2671.
154 [11] Song H, Park Y, Kim H, Ko H. Fully integrated biopotential acquisition analog front-end IC. Sensors, 2015, 15(10):
fv
155 25139–25156.
156 [12] Kim J, Kim J, Ko H. Low-power photoplethysmogram acquisition integrated circuit with robust light interference com-
oo
157 pensation. Sensors, 2015, 16(1): 1–11.
158 [13] Verma N, Shoeb A, Bohorquez J, Dawson J, Guttag J, Chandrakasan. A micro-power EEG acquisition SoC with in-
159 tegrated frature extraction processor for a chronic seizure detection system. IEEE J. Solid-State Circuits, 2010, 45(4):
pr

160 804–816.
ed
ct
rre
co
un