ECG Circuit: Analyzation and Application

Wenjun Ge

1,† a

, Junyang Sui

2,*,† b

and Pinzhang Wu

3,† c

Electrial and Mechanical Engineering, Jilin Institute of Chemical Technology, Jilin, Jilin, 132000, China

Bell Honors School, Nanjing University of Posts and Telecommunications, Nanjing, Jiangsu, 210000, China

Youpei College, Yancheng Institute of Technology, Yancheng, Jiangsu, 224007, China

Corresponding author

†

These authors contributed equally

Keywords: Electrocardiogram (ECG) Circuit, ECG Monitoring, Wearable ECG Sensors, Heart Rate Variability,

Wearable Devices.

Abstract: Cardiovascular disease has become the most common cause of death worldwide. According to the 2013

Global Burden of Disease Study, cardiovascular diseases are estimated to cause 17.3 million deaths

worldwide. As the primary technique for monitoring the heart's activity, the ECG plays an irreplaceable role

in the management of heart disease. Currently, there are two main types of devices used to measure ECGs.

One is the larger devices used in hospitals, such as medical ECG machines, and the other is smaller devices

for home use, mainly wearable devices such as the Samsung Galaxy Watch and Apple Watch. Compared to

traditional devices used in hospitals, wearable devices offer the advantages of small size, low energy

consumption and portability. In the future, lower noise, greater noise immunity, lower energy consumption

and higher ECG accuracy are necessary for developing wearable devices in the field. However, there are still

many difficulties at present. For example, as smart wearables need to minimize the device's size, the power

supply circuitry is also somewhat limited, and the device's battery life becomes a major issue. This review

introduces the application of ECG monitoring devices on wearable devices, introduces different wearable

ECG devices, and analyzes the performance of wearable ECG devices, points out the gap between current

wearable ECG and large medical ECG monitoring devices, analyses the causes of wearable ECG noise

generation, and proposes to reduce wearable ECG from electrodes, circuitry and other aspects monitoring and

improve the monitoring accuracy. Finally, the review summarizes the gap between wearable ECG and

medical-grade ECG monitoring devices, predicts the future challenges of wearable ECG, and expresses an

outlook on the future of wearable ECG devices.

1 INTRODUCTION

Cardiovascular diseases (CVD) have become the

most common cause of death worldwide. According

to the 2013 Global Burden of Disease study

estimation, CVD caused 17.3 million deaths globally.

CVD replaces 31.5% of all deaths and 45% of all non-

communicable disease deaths, more than twice that

caused by cancer, as well as more than all

communicable, maternal, neonatal, and nutritional

disorders combined (Townsend, Wilson, Bhatnagar,

Wickramasinghe, Rayner and Nichols ,2016)

Electrocardiogram (ECG), as the main technique for

https://orcid.org/0000-0002-5944-3215

https://orcid.org/0000-0001-7064-3659

monitoring cardiac activity, plays an irreplaceable

role in treating cardiac diseases. ECG diagnosis is one

of the most reliable methods for treating arrhythmia,

and it has great application value (R., B., and K. D.

2005). Currently, the devices that measure the ECG

are divided into two main categories. The one is the

large devices used in the hospital such as medical

ECG monitor, and the another is the home devices

that measure the ECG mainly with wearable devices

such as Galaxy Watch and Apple Watch.

The devices that measure the ECG are constantly

evolving with the development of technology.

Compared with conventional devices applied in

hospitals, the wearable takes advantage of small-size,

https://orcid.org/0000-0002-5511-0560

Ge, W., Sui, J. and Wu, P.

ECG Circuit: Analyzation and Application.

DOI: 10.5220/0011367500003444

In Proceedings of the 2nd Conference on Artiﬁcial Intelligence and Healthcare (CAIH 2021), pages 283-290

ISBN: 978-989-758-594-4

283

low energy consumption, portable, etc. In the future,

Lower noise, stronger anti-noise ability, less energy

consumption and higher accuracy of an

electrocardiogram are necessary for wearable device

development. However, there are still many

difficulties remaining. For example, because smart

wearable devices need to minimize the device's size

and the power supply circuit is also limited to a

certain extent, the device's battery life has also

become an important issue.

This paper summarizes the circuit used in

wearable devices which measure the ECG, including

the comparison with the devices applied in hospital,

as well as analysis of the typical ECG amplify circuit,

describing different types of ECG electrode, listing

some parameters about op-amp design, and the

current application of wearable devices which can

measure the ECG. Finally, the existing problems in

ECG research are discussed, and the future

development direction has prospected to bring some

references for related research.

2 THE APPLICATION OF ECG

CIRCUITS IN WEARABLE

DEVICES

The traditional medical ECG monitoring equipment

applied in hospitals (see Fig.1) usually uses the

common technique of 12-leads ECG to perform heart

analysis. It records heart electrical activity through

electrodes on the body surface and represents it into

a grid.

Figure 1: Medical ECG monitoring equipment and dry

electrodes used in traditional ECG monitoring.

Indeed, this technique offers different views of

heart electrical activity, allowing cardiologists to gain

a full, complete view of the patient heart, so some

anomalies of the ECG record symbolize pathologies,

can be easily detected, and lots of deaths prevented.

But this approach has two main drawbacks: one is

that patients and physicians should be in the same

place together with the electrocardiograph, which

means it cannot be detected anytime and anywhere.

As a result, this approach can’t predict and escape

some heart-related complications, such as cardiac

arrest, irregular heartbeat, congestive heart failure,

coronary artery disease, etc. Another is the traditional

disposable dry electrode which is often used for ECG

monitoring in hospitals at present (see Fig.1)

(

Beniczky, Conradsen, Henning 2005). The

conductive glue and other substances contained in the

dry electrode will penetrate the skin of patients,

which will cause skin allergies and other adverse

reactions in some patients. At the same time, using

the dry electrode for a long time will cause the

electrode strip to fall off and poor contact due to the

drying of conductive adhesive, so the measured ECG

signal will not be accurate (

Zhang, Bai, Zhou

1997

In the current year, wearable devices have

suddenly appeared on people's horizons. Compared

with large and heavy equipment in the hospital,

wearable devices take advantage of low energy

consumption, portability, and higher accuracy of

electrocardiogram, so people start to pay attention to

applying the wearable devices to monitor ECG. The

ECG circuits are the key portions of wearable

devices.

2.1 The Measurement Principle of

ECG Circuits in Wearable Devices

The biological electric change of the heart itself

passes the conductive organization and humoral fluid

around the heart, reflecting on coming up to the body

surface. It makes each part of the body also produce

regular electric change activity in each cardiac cycle.

ECG is a technology that uses an ECG circuit to

record the heart's electrical activity produced by each

cardiac cycle of the graph from the body surface.

2.2 The ECG Circuits in Wearable

Devices

2.2.1 A Wearable ECG Acquisition System

with Circuit Board-based Shirt

Fig.2 is a wearable ECG acquisition system based on

a planar-fashionable circuit board (P-FCB)-based

shirt. The system removes cumbersome wires from

the traditional Holter monitor system for

convenience. Dry electrode screen printing directly

on the fabric allows long-term monitoring without

skin irritation. The ECG monitor shirt uses a monitor

chip with a set of electrodes around the body.

Electrodes and interconnects are implemented using

P-FCB to enhance wearability and reduce production

costs (

Yoo, Yan, and Lee 2009).

CAIH 2021 - Conference on Artiﬁcial Intelligence and Healthcare

284

Figure 2: (a) Clinical ECG monitoring system. (b)

Conventional Holter monitor system. (c) Proposed

ECG monitoring shirt with P-FCB electrodes.

Fig.3 shows the structure of the ECG monitoring

chip. It consists of an amplifier (1A) with a

programmable gain amplifier, a 10 b SAR ADC, a

compression accelerator, an AES-128 encryption

accelerator, an internal memory, an MCU, I/O

interface, and a pair of P-FCB electrodes. The

electrode is data compressed using a secondary

compression accelerator (

Randazzo, Ferretti, and

Pasero 2019

). Then encrypted and stored in internal

memory. To improve safety, the AES-128 accelerator

is used (

Kim and Yoo 2008). If the internal memory

is full, external memory can be used through the

external memory interface. Internal memory is

accessed through the fast-mode 𝐼



𝐶 interface.

Figure 3: ECG monitoring chip architecture.

2.2.2 ECG WATCH

The ECG WATCH (Fig.4) is developed for standard

techniques such as 12-lead electrocardiography

(ECG) or dynamic electrocardiogram systems that

are insufficient to fully resolve sporadic ECG

abnormalities such as atrial fibrillation. It is an

inexpensive, wearable, easy-to-use health device that

can monitor the heart activity of patients with

cardiovascular disease anytime and anywhere

without the need to go to the hospital or cardiologists.

The recording takes 10 seconds. It also embedded an

algorithm to detect possible atrial fibrillation

episodes.

Figure 4: The ECG WATCH.

Fig.5 shows the ECG WATCH PCB. A TI

MSP430 series low-power microcontroller collects

the ECG signal with a 10b 200kbps SAR ADC. ECG

signals are collected at 1kbps, which are sufficient to

obtain good temporal resolution. The external voltage

reference provides an accurate DC reference voltage

for the ADC. In order to identify the risk of atrial

fibrillation, a 10-second ECG acquisition is sufficient

for the implementation of the algorithm. On the other

hand, the flash memory on the microcontroller has

enough space to store a few seconds of collected data

on the board at 1kbps, eliminating the need for

external storage modules and thus reducing the size

of the circuit (Kim, Kim, and Yoo 2008).

Figure 5: ECG WATCH PCB.

2.2.3 Wearable Mobile Ear-based ECG

Monitoring Device

This work presents the design and evaluation of a

wearable mobile ear-based ECG monitoring system

based on a highly conductive material graphene

electrode. Smartphones and headphones are

ECG Circuit: Analyzation and Application

285

becoming more common across generations. Here, a

novel design aims to advance the development of an

ear-based graphene sensor that, via a mobile

connection, generates high-quality, long-term, real-

time ECG measurements in a system more familiar to

the end user.

A typical ECG monitoring system consists of

electrodes and a front-end data acquisition circuit. A

three-electrode system is used to collect ECG signals.

In a three-electrode system, two active electrodes are

used for differential input to the ECG amplifier. The

third electrode is connected to eliminate common-

mode interference and improve signal quality. In the

proposed work, two electrodes are placed near the

ear-the one is attached behind the ear, the other is

placed on the neck, and a third electrode is connected

to the arm on the ground to obtain ECG signals, as

shown in Fig.6. a. The electrodes are tested in a

developed wearable device (see Fig.6b) that sends the

raw ECG signal from an ear-based electrode to a

measurement circuit mounted on the arm. The ECG

data is digitized, amplified, filtered in the

measurement circuit, and then transmitted to the

smartphone through Bluetooth for continuous

monitoring (

Celik, Balachandran, and

Manivannan 2019

Figure 6: (a) Electrode placement and block diagram of the ECG monitoring system; (b) Prototype design of ear-based ECG

monitoring system.

3 METHODS TO OPTIMIZE THE

PERFORMANCE OF ECG

SENSOR

In the ECG monitoring of wearable devices, the

accuracy of ECG signal detection for the wearer has

become an important index to measure the

performance of wearable ECG monitoring devices,

which is one of the main reasons why wearable ECG

monitoring devices cannot replace medical-grade

ECG monitoring devices at present. For well-known

reasons, the interference of ECG signals collected by

wearable devices is significantly greater than that of

medical-grade ECG monitoring devices.

The basic structure of major wearable ECG

designs is shown in Fig.7. To detect the ECG signal

from the human body, the system needs a sensor that

consists of 2 parts (electrode and amplifier). The

sensor can get the ECG signal and convert it to an

electronic signal that the signal processor can

process.

Figure 7: The basic structure of major wearable ECG designs.

CAIH 2021 - Conference on Artiﬁcial Intelligence and Healthcare

286

Every step of these processes causes problems

(consuming power, generating noise etc.),

influencing the system's performance. Therefore, it is

important to find the methods to optimize the

performance of the ECG sensor.

3.1 Electrodes of Wearable ECG

Equipment

Fig.8 shows various bio electrodes: wet electrode, dry

electrode, non-contact electrode, which are widely

used in ECG equipment.

Figure 8: Some ECG electrodes.

Wet electrodes are a kind of electrode of common

ECG equipment used by hospitals because it uses gel

to keep good contact of skin to make sure the signal

is good. However, most wearable ECG sensors do not

use this kind of electrode. The reason is simple:

Wearable devices should be comfortable and

portable. But the gel of wet electrodes is

inconvenient. People feel bad when putting gel on

them, and it is hard to replace gel after old gel

becomes useless.

The dry electrode is a replacement for the wet

electrode. They do not need gel to keep the

connection between electrode and skin at the expense

of the performance. Fig.9 shows the noise of a wet

electrode and a dry electrode. We can find that dry

electrodes generate more noise than wet electrodes,

which means the result of dry electrodes is more

imprecise.

Figure 9: The noise of a wet electrode (blue) and a dry electrode (red).

Meanwhile, using dry electrodes is

uncomfortable, too. People need pressure on the

electrode and skin to keep the electrode on the skin,

which causes a strong pressing feeling. Using a non-

contact electrode can solve this problem. However, it

brings some problems: Firstly, non-contact electrodes

generate more noise than wet and dry electrodes,

making a huge challenge in the amplifier design.

Secondly, it is hard to keep the distance between

electrodes and skin, especially using wearable

devices like smartphones or T-short sensors. As a

result, the distance of the electrode will change

rapidly and causes more noise. Fig.10 shows the

effect of sensor separation distance on input-referred

noise of an ECG device. When the distance between

the electrode and skin increases, the input-referred

noise increases.

Figure 10: The effect of sensor separation distance on

input-referred noise of an ECG device (Brain Support.

2020).

In conclusion, every electrode has some

advantages and disadvantages. And a better user-

friendliness electrode has lesser accuracy. So, it is

important to select an appropriate electrode by

considering all working situations of the design.

ECG Circuit: Analyzation and Application

287

3.2 Typical ECG Amplify Circuit

Table 1 is the comparison of some parameters of the

ECG signal. The voltage range of the ECG signal is

0.5-4 mV. It is too small to be processed. As a result,

we need an amplifier, whose gain is recommended

larger than 40 dB, to amplify the signal. And the

signal frequency range is 0.01-250 Hz. So, a filter is

recommended to decrease another useless signal.

Table 1: Some parameters of ECG signal.

Parameter or Measuring

Technique

Principal

Measurement

Range of

Paramete

Signal

Frequency

Range

Standard

Sensor or

Method

Electrocardiography

(

ECG

)

0.5-4 mV 0.01-250 Hz

Skin

electrodes

Wearable ECG sensors always use typical

instrumentation amplifiers (INAs) and common-

mode feedback circuits, which connect to the human

body, to solve the biasing problem.

Fig.11 is a typical INA circuit for ECG devices.

There are 3 parts to it: The First part is the full

differential input buffer, which consists of A1 and

A2. The second part is the common-mode feedback

circuit. The voltage of the human body is unknown,

so using it to set the biasing voltage to the body

voltage. In the third part, a differential amplifier

converts the differential input into a single end

output.

Figure 11: Typical INA circuit for ECG devices.

Instrumentation amplifiers amplify small input

signals accurately. It is appropriate to build an ECG

sensor. But, because of the small input signal,

reducing noise is very important. However, building

the noise model of the whole circuit is very complex.

But using an integrated INA noise model is simpler.

Fig.12 is the noise model of integrated INAs. In this

figure, Eno is the noise of the output stage, and en is

the RMS sum of the noise of the input and output

stages. G is the gain of INA. in is the noise current.

Figure 12: Integrated INA noise model.

Sometimes, manufacturers may give eni, which is

the noise of the input stage. In this situation, we can

use (1) to calculate en:

𝑒



𝑒

+(𝑒

/G)

(1)

Then we can find the result of output noise of the

circuit is Eno:

Eno=G∙

√

𝑁𝐸𝐵

∙



𝑒

()



+2𝑖



(𝑅



/2)

(2)

In equation (2) (Sullivan, T. J., Deiss, S. R.,

Cauwenberghs 2008), NEB is the noise equivalent

bandwidth. According to (2), we can reduce the

signal bandwidth through filters or use lower resistor

values to reduce noise. But it also brings some

problems: It is impossible to reduce the bandwidth to

a very small value. Because the frequency of the ECG

signal is 0.01-250 Hz, we cannot make the bandwidth

lower than 250 Hz. Secondly, using lower resistors

cost higher power consumption. As a result, users

should recharge their devices frequently. In

conclusion, every method to reduce noise leads to

other problems. So, selecting values according to the

product requirements is the best choice.

3.3 Pretreatment of ECG Signal

In the preprocessing of ECG signal, the baseline drift

caused by power line interference and interference of

body breathing movement can be removed first.

CAIH 2021 - Conference on Artiﬁcial Intelligence and Healthcare

288

The interference amplitude of the power line is

usually with volt magnitude, which is far larger than

that of the ECG signal with millivolt magnitude

(

Renesas 2020). Therefore, some methods, such as

adaptive zero-phase shift notch filter based on least

mean square (LMS) algorithm, are used to remove

power line interference and avoid phase distortion of

signals (

Gu, Hu, Zhang, Ding, Yan 2020).

And the baseline drift caused by breathing and

body movement is also an inevitable interference in

ECG signal acquisition by wearable devices. To

eliminate the baseline drift, we can estimate or extract

the baseline component, remove the component

caused by the drift by subtraction (Kuo, Morgan

1995

), or use a high-pass filter (Blanco-Velasco,

Weng, Barner 2008

4 CONCLUSION

After years of development, the gap between

wearable ECG devices and medical-grade ECGs is

getting smaller and smaller. Today, wearable ECG

devices can be found everywhere, such as mobile

phones, smartwatches, and headphones, which can

measure your heart rate, pressure, and blood oxygen

saturation. We can use this wearable ECG data to

analyze your physical health. to prevent disease. But

it must be acknowledged that although wearable

ECGs have developed rapidly over the years, they

still cannot replace large medical ECG machines. In

some cases, wearable ECG monitors are not as

accurate as medical-grade devices. For example,

noise and voltage. The increasingly miniaturized

wearable ECG devices are also becoming more and

more problematic in terms of battery life. Of course,

we have a lot to look forward to in the future of

wearable ECG devices.

In the future, wearable ECG devices will become

even smaller and more accurate. It may also rely on

the body's energy to provide a long-life span, with

functionality not limited to ECG monitoring but even

whole-body health monitoring. Although there is still

a long way to go, we expect this day to come.

REFERENCES

Beniczky S, Conradsen I, Henning O, et al. Automated real-

time detection of tonic-

Blanco-Velasco M, Weng B, Barner KE. ECG signal

denoising and baseline wander correction based on the

empirical mode decomposition. Comput Biol Med

2008;38(1):1–13.

Brain Support. (2020) Wet, dry, active and passive

electrodes. What are they, and what to choose?

https://www.brainlatam.com/blog/wet-dry-active-and-

passive-electrodes.-what-are-they-and-what-to-

choose-413

clonic seizures using a wearable EMG device. [J].

Neurology, 2018, 90(5)

Gu, X._ Hu, J._ Zhang, L._ Ding, J._ Yan, F. - An Improved

Method with High Anti-interference Ability for R Peak

Detection in Wearable Devices (2020)

H. Kim, Y. Kim, and H. Yoo, "A low-cost quadratic level

ECG compression algorithm and its hardware

optimization for body sensor network system," in Proc.

IEEE Eng. Med. Biol. Conf. (EMBC), Aug. 2008, pp.

5490–5493.

H. Kim, Y. Kim, and H.-J. Yoo, "A 1.12 mW continuous

healthcare monitor chip integrated on a planar-

fashionable circuit board," in Proc. IEEE Int. Solid-

State Circuits Conf. (ISSCC) Dig. Tech. Papers, Feb.

2008, pp. 150–151.

Hyeonjeong Lee, Jaewon Lee, and Miyoung Shin. "Using

Wearable ECG/PPG Sensors for Driver Drowsiness

Detection Based on Distinguishable Pattern of

Recurrence Plots". Electronics 2019, 8, 192;

doi:10.3390

Jearld Yoo, Long Yan, and Seulki Lee. "A Wearable ECG

Acquisition System with Compact Planar-Fashionable

Circuit Board-Based Shirt". IEEE. Novembei, 2009.

Pp:1089-7771

Köhler BU, Hennig C, Orglmeister R. The principles of

software QRS detection. IEEE Eng Med Biol

2002;21(1):42–57

Kuo SM, Morgan D. Active noise control systems:

algorithms and DSP implementations. John Wiley &

Sons, Inc; 1995.

N. Townsend, L. Wilson, P. Bhatnagar, K.

Wickramasinghe, M. Rayner and M. Nichols,

"Cardiovascular disease in Europe: epidemiological

update 2016," European Heart Journal, pp. 1–14,

August 2016.

Numan Celik, Wamadeva Balachandran, and Nadarajah

Manivannan. "Wearable Mobile Ear-based ECG

Monitoring Device Using Graphene-Coated Sensors".

IEEE. July. 2017, PP. 978-1-5090-1012.

R., B. and K. D. Telemetric ECG evaluation using

einthoven-leads. in Computers in Cardiology, 2005.

2005.

Renesas. (2020) Noise Calculations of Instrumentation

Amplifier Circuits.

https://www2.renesas.cn/us/en/document/apn/r13an00

11-noise-calculations-instrumentation-amplifier-

circuits-rev100

Sullivan, T. J., Deiss, S. R., Cauwenberghs, G.. (2008). A

Low-Noise, Non-Contact EEG/ECG Sensor.

Biomedical Circuits & Systems Conference. 154-157.

Vincenzo Randazzo, Jacopo Ferretti, and Eros Pasero.

“ECG WATCH: a real time wireless wearable ECG”.

IEEE. Mar, 2019. Pp:978-1-5386-8428.

ECG Circuit: Analyzation and Application

289

Zhang Y, Bai J, Zhou X, et al. First trial of home ECG

and blood pressure telemonitoring system in Macau[J].

Telemedicine Journal the Official Journal of the

American Telemedicine Association, 1997, 3(1):67.

CAIH 2021 - Conference on Artiﬁcial Intelligence and Healthcare

290