Content uploaded by Varad Choudhari
Author content
All content in this area was uploaded by Varad Choudhari on May 06, 2019
Content may be subject to copyright.
2018 International Conference on Communication, Information & Computing Technology (ICCICT), Feb. 2-3, Mumbai, India.
A Portable and Low-cost 12-Lead ECG device for
sustainable remote healthcare
Varad Choudhari, Vaishnav Dandge, Natasha Choudhary
Electronics Department
Sardar Patel Institute of Technology (Mumbai)
varadchoudhari@icloud.com, vaishnav099@gmail.com
natasha.choudhary1912@gmail.com
Rajendra G. Sutar
Dy. HOD, Electronics Department
Sardar Patel Institute of Technology
Mumbai, India
rajendra sutar@spit.ac.in
Abstract—The diagnosis of cardiac patients is commonly done
using Electrocardiogram (ECG). But in the frequently used 3-
Lead ECG device, all the cardiac diseases are not covered in the
diagnosis. This paper illustrates an approach towards developing
a 12-Lead ECG system. The existing 12-Lead ECG systems are
costlier and require a large setup size. Also, these machines
are not that user-friendly to be operated in rural areas. The
proposed system is compact which makes it portable and shows
effective results in ambulatory systems. This manifesto presents a
customized approach towards designing an ECG amplifier, and
Digital data converting and storing unit, enabling the domain
of ‘Internet of Things’ which makes it a very effective mode of
Telemedicine. The received digital data is directly uploaded to an
online database. To minimize the cost, reduction in hardware has
been suggested without compromising on the quality of signal
captured. Use of basic embedded and open source tools made
the system simple and easy in terms of handling. This study has
fostered a successful working model which helps in acquiring
12 ECG outputs altogether for a thorough diagnosis of cardiac
patients. Thus, making good healthcare facilities available to
remote and underprivileged locations of the country.
Index Terms—12 Lead ECG, Portable ECG system, Remote
healthcare
I. INTRODUCTION
Along with the population, the risk of chronic heart diseases
is also on the rise, thereby increasing the cost of healthcare.
Latest statistics suggests that there are roughly 30 million
heart patients in India out of which 1 million surgeries are
carried out every year. Of the 30 million heart patients, 14
million reside in urban areas and 16 million in rural areas[1].
There comes the shortcoming that, due to unavailability of
intelligent medical apparatus in the rural areas the number of
cardiac patients in the rural areas is on the rise. Telemedicine
is the use of telecommunication and information technology
to diagnose and provide clinical health care at a distance. It
helps in eliminating the distance barriers and can enhance
access to medical services which could not be consistently
available in distant rural communities. It is also used to provide
critical healthcare assistance during an emergency. Developing
a portable device will lead to the efficient implementation of
telemedicine over a wide range of diseases.
This paper discusses an optimised and reliable way to design
a 12-lead ECG device and transfer the electrocardiograph from
a remote location to professional doctors anywhere around the
world. The motivation of 12-Lead ECG system came from the
purpose of making the ECG system more accurate and to cover
a wide range of cardiac diseases. A 12-Lead configuration
provides a more detailed view of the heart’s three areas -
anterior, lateral and inferior[2]; thereby increasing detection
and diagnosis of the patient’s heart condition.
The design complications increase while expanding the
system from a 3-lead to a 12-lead architecture. As seen in
the paper[3], for the collection of data from a single lead, the
analog front end consists of a right leg drive, pre-amplifier,
filters, amplifier and power-line noise attenuation circuit. Also,
this design displays only single channel output at a time and
a switch between three channels. In order to reduce the cost,
the number of operational amplifiers has been reduced to six
from eight per channel. That is an overall 25% reduction in
a number of Op-amp, even after reducing the op-amps, the
results can be obtained accurately.
Author of[4] has mentioned that for obtaining an ECG
signal, the circuit design must meet standard specifications
while designing hardware blocks. For example, the gain of the
instrumentation amplifier, cut-off frequencies of filter block
and sampling frequency of the ADC. By referring to this
document it was understood that using an instrumentation
amplifier with a gain value of 1000 can causes problems of
noise amplification and amplifier saturation. As a solution to
this, the overall gain can be achieved by a cascade combi-
nation of instrumentation amplifiers followed by a secondary
amplification section, making gain value considerably large.
In case of bandpass filter design, with second-order filtration,
the response becomes much ideal as the slope of magnitude
response changes from -20db/decade to -40db/decade[5].
The author of paper[6] has developed an android based
application to display the ECG graph on a smartphone. The
application framework provides access to Bluetooth function-
ality through Android Bluetooth APIs. However, the use of
Bluetooth to transmit the data limits the maximum allowable
range between the patient and the doctor. From study presented
in[6] it is understood that use of Bluetooth becomes range
limiting and redundant. Hence a terminal device with accurate
data transmitting capabilities and easy user interface becomes
better choice when it comes to bio-medical tool installation in
rural parts of the country.
978-1-5386-2051-9/18/$31.00 ©2018 IEEE
2018 International Conference on Communication, Information & Computing Technology (ICCICT), Feb. 2-3, Mumbai, India.
Fig. 1. Schematic block diagram of entire 12-Lead ECG system describing complete process flow
From the literature, it can be observed that the aim of
several studies is to get higher accuracy, with reduction of
enlarged complexity and cost. The problem of expense and
reliability, towards the design of an ECG monitoring system
dedicated to the remote area, are yet to be resolved. To
resolve such issues, a system with features like portability,
cost-effectiveness, reliability and easy accessibility for under-
privileged location is presented in this paper. Fig.1 illustrates
the schematic block diagram of the proposed system. When
the patient is connected to a system via electrodes, hardware
part containing analog circuit wherein the signal obtained from
electrodes is amplified, filtered and processed. The controller
receives the analog signal converts into a digital signal with
help of ADC and transmits it serially to RTU(Remote Terminal
Unit) with Internet access. RTU uses open source software to
monitor and transmit data to an online database.
II. MATE RI AL S
The experimental set-up was carried out with certain tools
and materials. Starting with the use of Simulation Tools, to
understand the circuit parameters like gain, frequency response
also to verify variation in output with change in components
and change in their values. The experiment was carried out
on more than 10 patient from the age group of 20-50, all
patients had the normal heart condition and their heartbeat
varying between 70-120 bpm. The system was connected to
the patient with ECG cable. The interface between patients
body was done by electrodes, for understanding purpose
document[7] was referred. Industry standard and commonly
used Disposable Pre-gelled Ag/AgCl electrodes were used.
PCB Designed and layout tools were used in the experiment
to improvise circuit from breadboards to fixed Printed Circuit
Board. For data acquiring purpose, Arduino Uno was used
containing ATMega328 single-chip microcontroller by Atmel.
Open source software tools like MySQL, phpMyAdmin, and
Telemetry Viewer are also part of this system.
III. METHODOLOGY
A distributed approach was used to build hardware part of
the system, entire circuit was divided into small segments
for ease of experiment. Dedicated to each parameter every
segment represented an individual circuit block on which
simulations were performed. Simultaneously gain or frequency
response values of individual breadboard mounted circuit
block were examined and compared. After multiple patient’s
ECG analysis, this breadboard mounted circuitry was replaced
by equivalent PCBs in a customised manner. This helped in
making the system portable as card interface technique was
used, explained in the interface section.
A. Hardware Implementation
1) Driven Right Leg Circuit: The circuit shown in Fig.2
is a customised ECG amplifier built in our project, which
is based on the circuit in[3] but with reduced hardware and
suitable gain and frequency values. The driven right leg (DRL)
circuit was implemented for patient safety and noise reduction
purpose as referred in the study of document[8]. The circuit
shown in [9] was modified to lower the effect of common
mode signals. From Fig.2 amplifier U1, U2, and U3 represent
the DRL circuit block. Resistors of 10MΩare coupled at the
output of buffers, this minimises the common mode signal.
Each Right leg buffer has a 1.8KΩresistor in series with
output and also at the input side of the buffer to balance DRL
circuit. This helps in limiting current from unwanted external
sources. Thus DRL circuit for all 10-Leads i.e. three limb
leads, three augmented leads and six chest leads were mounted
on a separate PCB. This is connected with rest of the circuit
with external wire connects through a rotary switch.
2) Instrumentation amplifier: The circuit flow from instru-
mentation amplifier till the ECG output was printed on a
dedicated circuit board called ECG card in system. This pro-
vision was made to make DRL input switching easier between
two simultaneously working channels on two different ECG
card. It is commonly known as biomedical amplifier. Opams
U4, U5 and U6 from Fig.2 used in circuit represents the
instrumentation amplifier circuit block. TL084 were used for
instrumentation amplifier design. Gain of this circuit can be
calculated with eq.1, resistors mention in the equation are with
reference to Fig.2. Because of it’s high CMRR, better noise
over margin and high input impedance as seen in[10], this
amplifiers are widely used for bipotential amplification.
Gain =1 + 2R6
R7R11
R9
(1)
considering the amplifier saturation condition the desired gain
value was reached by cascading amplifiers and hence this
amplifier was provided a gain of 25. The circuit effectively
amplifies 0.5-3mV biopotential signals from patient’s body
with a gain of 25.
2018 International Conference on Communication, Information & Computing Technology (ICCICT), Feb. 2-3, Mumbai, India.
Fig. 2. Circuit diagram consisting of 3 DRL amplifiers and one complete ECG card
3) Bandpass Filter Design: ECG is a biomedical signal
produced by electrical activity of the heart, although there can
exist such multiple electrical signals while body is connected
to system. These unwanted signals can result in noise and
distort the actual ECG signal and hence signal characteristics
were taken in consideration during design. In study of [4]
author has mentioned bandwidth of ECG signal, can be
seen in Table I. The op-amp U7 in Fig.2 act as buffering
element to separate Low-pass and High-pass filters. The cutoff
frequencies are set to 0.5 Hz on highpass side and 40 Hz on
lowpass side. The filter design for both cut-offs was done using
fc=1
2πRC (2)
equation-3, where for highpass filter R=R13 ,C=C1and
for lowpass filter design R=R17,C=C3. Both filters are
2nd order filter showing better result than single order filter.
To reach the system gain this filter was given gain of 83. The
Gain = 1 + Rf
Ri
(3)
gain was set by using equation-3, where U7 is working as
non-inverting amplifier with Rf=R15,Ri=R1 6 referring
to Fig 2. However, the practically achieved gain was less than
the designed gain and comes out to be 37.
4) Notch filter: During the testing period, considerable
ECG signal was achieved up till previous circuit block but
it did carry huge flickering noise. This noise was mainly
because of powerline interference. The 50Hz mains supply
produces a very low amplitude of signal and ECG signal
shows noisy output due to the presence of its first two
harmonics. Hence to attenuates a specific frequency from the
spectrum i.e. 50Hz AC supply noise frequency notch filter
has to be used. As shown in Fig.2 circuit before Op-amp
U8 and after bandpass filter signifies Twin-T notch filter. In
design procedure theoretically notch frequency of 48.22 was
calculated by using equation-2 with R=R19 and C=C7.
Simulation output for the same circuit can be seen in Fig.3,
shows effective attenuation for 50.5 Hz frequency. While
Fig. 3. Magnitude spectrum of Twin-T notch filter
mounting the circuit practically 20nF capacitor was taken as a
parallel combination of two 10nF capacitors. Circuit showed
better attenuation for frequency variation between 48-52 Hz
and can prove effective on problem discussed by author of[11]
in recording considerations section.
5) DC Shift: Till previous circuit blocks ECG signal could
be examined on a DSO but now the task was to convert the
data into digital format, interface to the microcontroller was
2018 International Conference on Communication, Information & Computing Technology (ICCICT), Feb. 2-3, Mumbai, India.
Fig. 4. Output of two channels observed simultaneously on DSO
needed. The Arduino board used, supports ADC for voltage
values between 0-5V but the ECG signal obtained had some
negative potential in some part of its wave. This is because of
dual power supply used for powering the Op-amps, providing
freedom for positive as well as negative voltage swing. In case
of such series connections of amplifiers negative shift is added
in every stage and hence DC shift was provided to elevate
the signal with desired voltage level similar to study done in
[12]. This shift was provided at one of the inputs to adder
circuit as referred in Fig.2, where U9 is Op-amp working as
an adder. This DC-shift voltage was produced by potential
divider combination with fixed resistor connected to Vcc and
variable potentiometer to ground.
TABLE I
COMPARISON BETWEEN REQUIRED AND OBSERVED VALUES
Parameters Simulated/Ideal Practical
Voltage gain >1000 1850
Bandwidth 0.5-40Hz 0.5-35Hz
Notch suppressed frequency 50Hz 49Hz
CMRR >80 86
After following these circuit design blocks, quality ECG
signal was observed on DSO can be seen in Fig.4. Consid-
erable efforts were taken to reach the parameters required
or simulated, Table I lists all parameters and their ideal and
practically achieved values.
B. Interface
1) Card interface: The entire circuit is developed in card
format as explained in DRL circuit section. These cards were
connected with each other in a manner that all 12-Lead ECG
signals can be acquired on only two channels. Therefore two
ECG cards and a DRL card were connected via two, 2 poles 6
way rotary switch refer Fig.2. This switch can have six differ-
ent combinations and hence two of such switches complete
requirement for 12-Leads. Three limb-leads with a simple
subtraction equation like RA−LA, and three augmented
leads with slightly tricky equation like RA−1
2(LA+LL)are
calculated by 1st ECG card. The remaining six chest leads
following equation V(1−6)−RLon the other card.
Fig. 5. Programming logic flow
2) Controller Interface: Controller plays an important part
in this system by converting the analog signal into digital
format and sending it to the computer. Arduino Uno was used
in the application as its a cheap and very powerful embedded
tool. 10 bit ADC of Arduino UNO kit plays an important role
in capturing the signal. This 10 bit ADC produces 0-1023
for the equivalent of 0-5V. The program flow of controller
shown in Fig.5 express a simple but effective program. ADC
polling mechanism is used for continuous data collection
and transmitted over the serial port in the same loop. The
importance of storing the ADC data into a text buffer before
serial transmission is that, when data is stored from the serial
port it becomes easier to achieve it in a CSV (comma separated
values) format on the computer. This buffer is transmitted to
the computer with a standard baud rate of 9600 bps, which can
record 57samples/sec with precise resolution in the ECG plot.
The gain set in this system is quite greater than required or
preferred but optimised to boundaries for ADC capture, which
increase the noise margin. A very small amplitude flickering
noise sitting on wave does not show up in digital signal
captured. During testing period maximum signal amplitude
observed of various patients for all twelve leads was <3V.
C. Software
Telemetry systems possess a feature which is essential
for every remote application, i.e data transmission over long
distance. In order to make this system useful for remote health-
care, software features were necessary. Hence a systematic
user flow for online database upload was developed using few
open source software applications:
1) Data Monitoring Tool: The visualization of digital data
transmitted by the controller on a serial port of was done
by using a computer-based java application called Telemetry
Viewer [13]. This application reads data on serial port with the
same speed as the baud rate and represents it in a graphical
2018 International Conference on Communication, Information & Computing Technology (ICCICT), Feb. 2-3, Mumbai, India.
format. The controller had programmed accordingly so that
data transmitted should be interfaced properly without any
mismatch. This software was successful to retrieve ECG plot
Fig. 6. ECG Output on Telemetry Viewer data monitoring software.
without any data loss and almost same as that of the analog
plot, Fig.6 is a screenshot of this application while the patient
is connected to the system. This tool is not developed for
storing that data received and hence it was necessary to search
a tool can be capable for this task. The data obtained on
monitoring application screen was verified and the interpreta-
tion was made that hardware and controller settings can easily
retrieve ECG wave from the database.
2) Storing and Upload of Data: Data logging for the
system was done using open source software called PLX-DAQ
[14]. It connects the serial port of PC with excel sheet which
uses Macros feature to collect the data by automating common,
repetitive tasks. Once excel sheet containing time and 12-
Leads digital data is saved in .csv format, it was imported
into MySQL table. This workbench was used to create an
offline database. This database was designed with thirteen
columns one for time input and 12 for ECG signals once the
table is updated, it can be exported as .sql in the workspace.
The most important part of this project was to make the
ECG data available online and hence a website was hosted
using free web hosting services [15]. The hosted website i.e.
https://spitetrxecgdb.000webhostapp.com was fed some basic
information about the project. This website was developed
to work as a server so that it’s database can be accessed
through an Android application. The data was uploaded to
a server by importing the SQL file to an online database.
At the back-end of the website, the database was managed
through a PHP script. We used phpMyAdmin which is an
open source administration tool for MySQL. This standard-
ized process was followed every time we uploaded the new
patient’s data, a new file in the database was created for every
Fig. 7. Data received on online database
patient. Fig.7 is chart view of Lead II column of the online
database through phpMyAdmin. Our aim was to give access
of their electrocardiograph to every patient, by giving them
an individual account. However, it was much of coding part
which was considered for future scope of this experiment.
IV. RES ULTS AND DISCUSSION
Purpose of the design, to acquire 12-lead ECG using two
channels is fulfilled. It’s a very uniquely designed portable
device, with a compact size and easy functionality. The total
costing of the system is less than INR 4,500. During the
testing of the prototype, the substantially high gain of 65dB
became a key factor over the noise. The designed filters with
reduced hardware were able to provide very good attenuation
of unwanted noise frequencies. Table II shows amplitude
measurements for two limb leads of normal heart condition
comparing it to study in [16], our design shows equivalent
readings. The amplitude reading mentioned in this table repre-
sents practically observed electrical voltage produced by heart
in millivolt, these values are calculated by using equation
ActualOutput(V)/T otalGain. Whereas, general amplitude
values are taken from explanation done in [17].
TABLE II
AMP LIT UD E OF DI FFER EN T SEG ME NTS O F ECG WAVE
Intervals on the ECG Lead I(Ra-La) Lead II(Ll-Ra) General Amplitude
P wave 0.21-0.27mV 0.30-0.35mV 0.25mv
QRS complex 0.59-0.72mV 1.40-1.56mV 1.60mv
T wave 0.37-0.43mV 0.21-0.32mV 0.10-0.50mv
The arrhythmia detection using ECG requires proper iden-
tification of time intervals for every single segment in ECG
wave. During testing of our prototype we recorded readings for
5 volunteers, The observed readings are mentioned in Table III.
On comprising these readings with standard timing referred in
[18], are quite good between the range stated by the author.
However, this intervals are calculated from analog reading and
scaling in digital data is still upcoming work.
TABLE III
TIME PERIOD OF DIFFERENT SEGMENTS OF ECG WAVE
Intervals on the ECG Observed Standard
P wave 70-75ms 60-120ms
PR Segment 140-170ms 120-200ms
QT Segment 360-410ms 360-440ms
QRS complex 85-95ms 60-100ms
T wave 150-160ms 160ms
The communication with the controller is not done wire-
lessly instead it’s done using serial communication. At the
serial speed of 9600 bps sampling rate of 56 Samples/Sec
is achieved. It provides a clean ECG plot but, with even
higher rate recording can be much better. Fig8 is a graphical
representation of the data retrieved from online ECG database
from our website. This six different plot indicates a complete
cycle of ECG wave for three limb leads and three augmented
leads. The Y-axis scale is ADC numbers, starting from 0 till
2018 International Conference on Communication, Information & Computing Technology (ICCICT), Feb. 2-3, Mumbai, India.
Fig. 8. Six leads output on excel graph
1023 equivalents to 5V. Signals do not carry any flickering
artifacts even though these waveforms are at their highest
resolution indicates good system advantage over the noise.
Although the software method is a bit difficult to coop up with,
therefore we are looking forward to developing a dedicated
software. A software that can automate the steps followed for
database upload and also provide a better singular GUI to the
user.
ACKNOWLEDGMENT
We thank Dr. Pranjali Save (Medical Officer, PHC-Palghar
District) for sharing her experiences while practicing in rural
area, and also for examining our system working as well as
results achieved. She believes that with more upgrades in the
same methodology can create a revolutionary effect in the
transformation of remote healthcare.
V. CONCLUSION
It has been shown that it is possible to design a 12-
Lead ECG monitoring device at minimal cost and simplified
portable design. The circuit attains accurate ECG systems
for remote patient monitoring. When a patient is connected
to the system it collects the data and administrator have to
upload data on internet manually. This system can boost the
healthcare facilities in primary healthcare centers at rural part
across the nation. Eventually leading an awareness on various
types of arrhythmia and blockage detection at early stages.
Future work is required to make this data available in making
digital data available in MIT-BIH[19] format i.e. .txt and
compatible for arrhythmia detecting algorithms. Comparing
our analog recordings with ECG Homecare device by the
author of [4] results are quite similar in terms of clarity. They
have used AD8232 which is a dedicated ECG chip by analog
devices. This system uses advantage over cost-effectiveness in
comprising with study of[11] which involves high fabrication
cost.
REFERENCES
[1] “Heart disease in india is a growing concern.” Available on-
line http://food.ndtv.com/health/world-heart-day-2015-heart-disease-in-
india-is-a-growing-concern-ansari-1224160. Accessed: 2016-06-11.
[2] http://www.emtresource.com/resources/ecg/12-lead-ecg-placement/. Ac-
cessed: 2016-06-11.
[3] R. Sutar and A. Kothari, “Intelligent electrocardiogram pattern classifi-
cation and recognition using low-cost cardio-care system,” IET Science,
Measurement Technology, vol. 1, pp. 134–143, 2015.
[4] M. W. Gifari, H. Zakaria, and R. Mengko, “Design of ecg homecare:12-
lead ecg acquisition using single channel ecg device developed on
ad8232 analog front end,” in Proc. International Conference on Electri-
cal Engineering and Informatics, no. 5, 2015.
[5] R. Dozio and M. J. Burke, “Optimised design of the front-end analogue
high-pass filter for a diagnostic quality ecg monitoring system,” in Proc.
31st Annual International Conference of the IEEE EMBS Minneapolis,
2009.
[6] P. T. P., O. Manjarres, M. Quijano, and A. U. Blanco, “Remote
monitoring system of ecg and temperature signals using bluetooth,”
in Proc. 2012 International Symposium on Information Technology in
Medicine and Education (ITME), 2012.
[7] J. D. Bronzino, The Biomedical Engineering Handbook, ch. 48. CRC
Press LLC, 2 ed., 2000.
[8] B. W. Winter and J. G. Webster, “Driven-right-leg circuit design,” IEEE
Transactions on Biomedical Engineering, vol. 30, pp. 62–66, 1983.
[9] J. Gomez-Clapers, E. Serrano-Finetti, R. Casanella, and R. Pallas-Areny,
“Can driven-right-leg circuits increase interference in ecg amplifiers?,” in
Proc. 33rd Annual International Conference of the IEEE EMBS Boston,
2011.
[10] M. Yildiz, F. Guler, A. Turkmen, and D. Yllmaz, “Biopotential instru-
mentation set,” in Proc. 23rd Annual EMBS International Conference,
2001.
[11] S. Nag and M. Shojaei-Baghini, “Multi channel electrocardiogram using
custom analog chip,” in Proc. PCEA-IFToMM International conference
PICA, 2006.
[12] K. F. Akingbade, I. A. Alimi, and T. Oni, “Design of a data logger for
biomedical signals,” International Journal of Electronics and Electrical
Engineering, vol. 3, no. 2, pp. 149–152, 2015.
[13] “Telemetry-viewer, data monitoring tool.” Available online
https://github.com/farrellf/Telemetry-Viewer. Accessed: 2016-10-
07.
[14] “PLX-DAQ gui based data storing software.” Available online
https://www.parallax.com/downloads/plx-daq. Accessed: 2016-12-07.
[15] “Website hosting service.” Available online https://in.000webhost.com/.
Accessed: 2017-02-07.
[16] A. Bharadwaj and U. Kamath, “Accurate ecg signal processing,” in CA:
Cypress Semiconductor, 2011.
[17] E. A. P. Leslie Cromwell, Fred J. Weibell, Biomedical Instrumentation
and Measurement, ch. 6, pp. 105–107. Prentice-Hall of India, 2 ed.,
2002.
[18] L. W. Wilkin, ECG Interpretation Made Incredibly Easy, ch. 3, pp. 42–
56. Wolters Kluwer, 5 ed., 1997.
[19] “MIT-BIH open source arrhythmia database.” Available online
http://www.physionet.org/ physiobank/database/mitdb. Accessed: 2017-
03-16.
