108
Chapter 5
Design and Development of
ECG amplifier and testing
the developed sensor
109
5.1 Design and development of ECG amplifier with the
conventional sensor
5.1.1 Design considerations
Accuracy and precision are very important aspects to be
considered in handling diagnostics or other medical applications. Any
small fluctuation in the waveform generated could carry critical
diagnostic value and hence it is very essential to design any bio-
medical instrument with the highest precision possible. It becomes
very necessary that the designed ECG system must faithfully display
the actual ECG signal, such that any irregularity detected should be
due to an unhealthy cardiac cycle and not from the equipment that was
used. Therefore, there are many special considerations to be taken into
account when designing the ECG amplifier.
1. The input impedance of the pre-amplifier should be high:
The developed sensor uses TiO2 as a dielectric material. Since
the optical band gap of this material is high, around 3.61eV, the
material offers high resistance. The resistance offered by the sensor
was found to be around 3.4kΩ. The input impedance of the per-
amplifier should be high enough to match the sensor resistance. This
is important to achieve the entire sensor signal amplitude to the input
of the pre-amplifier. Op-Amp amplifiers are usually considered for
such requirements. The main advantage of am Op-Amp amplifier is
that it is an AC / DC amplifier. Since the frequency of the ECG signal
110
is very low, these amplifiers are ideally suited. LM324 is a low
power quad Op-Amp with a large voltage gain and internal frequency
compensation. They operate from a single power supply over a wide
range of voltages. They are available as a convenient 14 pin DIPs.
2. Gain of the ECG amplifier should be such that the
amplified ECG signal can be displayed:
The amplitude of the ECG signal is in the range of micro volts.
The amplifier should provide a proper gain to the ECG signal over the
low frequency range so that the signal is amplifier without distortion.
The signal needs to be displayed after being amplified and hence, a
good gain is required. The LM386 is a power amplifier designed for
use in low voltage applications. The gain is internally set to 20 to keep
external part count low, but the addition of an external resistor and
capacitor between pins 1 and 8 will increase the gain to any value up
to 200. The IC is available as an 8 pin DIP.
3. Pre-amplifier should be able to accept the low power
signals form the sensors:
The sensor output signal is of very low and very low amplitude.
The pre amplifier required should be able to accept and amplify this
signal. The LM324 is such an amplifier whose important application
includes transducer signal amplification.
111
4. The pre-amplifier should be able to amplify low
frequency signals with constant gain:
The LM324 has a constant voltage gain from 0Hz to a few kHz.
Since the ECG signal frequency is very low, LM324 is an ideal
amplifier which can be used as a pre amplifier which can directly be
fed by the sensors
5. The amplifier output should be distortion free:
Since the sensor signal amplitude is very low, it gets easily
affected by noise signals. The main interfering signal is the movement
artifact. The amplifier should be able to reject the common mode
signals and provide a high signal to noise ratio. The circuit should be
able to identify the sensor signal and amplify it.
5.1.2 Block diagram of the amplifier
Block diagram of the amplifier system
The ECG amplifier system block is as shown in figure 5.1. The
preamplifier was designed using IC LM 324 and the ECG amplifier
was designed using IC LM 386. Two output options are provided one
at the output of the pre-amplifier and the other at the output of the
ECG amplifier.
112
Figure 5.1: The ECG amplifier system block
Sensors: Since the proposed work is to develop a three lead
ECG system, three sensors, Sensor 1, Sensor 2 and the reference
sensor was used. These developed sensors were connected to the Op-
Amp amplifier inputs.
Pre-amplifier block: This block is constructed to amplify the
low amplitude, low frequency sensor output and to present the same
for further power level enhancement process.
To get the required characteristics of high input impedance and
high CMRR, an Instrumentation Amplifier which Op-Amps is usually
used as the first stage of a bio-potential amplifier. The architectures of
an Instrumentation Amplifier can be either the 3 Op-Amps system or
the current-feedback type [1, 2]
. The three Op-Amps system is easy to
construct while the current feed type has inherent higher CMRR. Both
113
types can be constructed using commercial available integrated-circuit
(IC) form.
ECG amplifier block: This block is constructed to enhance the
power levels of the pre-amplified signal. The output of the pre-
amplifier feeds the input of this block. This block uses a low voltage
audio power amplifier in the IC form. The output of this block is the
amplified ECG signal, ready to be transmitted.
5.1.3 Circuit Design
This section deals with the design and development of the
amplifier and the filter required to process the ECG signal for display
and further transmission. The inputs for this section are derived from
three sensors, developed using Titanium dioxide as the base material.
The circuit is designed to amplify the signals and at the same time
eliminate as much noise as possible. The high frequency noise and the
common mode noise both have to be addressed to in designing the
circuit.
The duration of the QRS complex, which gives the highest
frequency is between 0.07 to 0.1 seconds, which gives a frequency of
10 to 14.285Hz. Hence the cut off frequencies for the filter is chosen
to be 15Hz.
According to the sampling theorem, good sampling of the signal is
possible only if the bandwidth of the circuit is two times the highest
frequency to be handled. Hence, the cut off for the antialiasing filter is
chosen to be 30Hz.
114
The first stage of the circuit is the preamplifier stage which is a
very crucial stage which has to produce high gain and should provide
a good Common Mode Rejection. The input level shifter circuit shifts
the signal level required for the digitizing of the signal at a later stage
[3, 4]. The buffer amplifier ensures that the resistive loading is avoided
and also provides high input impedance.
The voltage difference between the two electrodes serves as the
signal input that is amplified through the Op-Amp circuits. These
signals are then differentially amplified and passed through a low pass
filter whose cut-off frequency is around 33Hz. The next stage is a
second order, Sallen-Key, low pass filter which is used for anti-
aliasing.
Pre-amplifier Designing:
To nullify any wandering DC effects, a simple RC high pass
filter with a very low cut off frequency has been implemented. The
cut-off frequency has been chosen to be 1.6Hz and the capacitor value
is chosen to be 10μF.
Since f = 1/(2πRC), R = 9.9522k Ω, which has been taken to be 10kΩ.
To reject the noise frequency at 50Hz from the power mains,
the acquired signal requires low pass filtering below 50Hz. The cut-
off frequency has been selected to 33Hz to reject the power line
frequency.
Let R = 100kΩ, f = 1/(2πRC) and f = 33Hz, hence C = 47nF = 47kpF
115
The two signals are added and passed through an anti-aliasing
Low Pass Filter which also provides a gain of 100.
Gain A = -R2/R1 and the required gain is 100
Chose R1 = 100kΩ, then R2 = 10MΩ
f = 1/(2πR2C1) and f = 33Hz
Then C1 = 483pF and hence chosen to be 470pF
(a)
(b)
Figure 5.2: (a) Low pass filter and (b) simulation result of the filter
116
This signal is passed through a Second order Sallen-key, unity
gain low pass filter to get a narrow bandwidth of about 15.31 Hz.
F = 1/√(2πR1C1R2C2); when R1 = R2 = R and C1 = C2 = C;
then f = 1/(2πRC)
Let R = 100kΩ, then C = 106.15nF. Hence C is chosen to be 104kpF
(a)
(b)
Figure 5.3: (a) Second order Sallen-key, unity gain low pass filter, (b)
simulation result
117
The output of this filter is fed to two stages of passive filters, as
shown in figure 5.3, for step wise elimination of noise. These two
stages of passive LPFs give a second order response with a roll-off
rate of -40 dB/ decade. The two step attenuation of the signal is
compensated by providing an equivalent gain by the non-inventing
amplifier.
F = 1/(2πRC); f = 15.31Hz,
Let R = 100kΩ, then C = 106.15nF. Hence C is chosen to be 104kpF
(a)
(b)
Figure 5.4: (a) Two stages of passive filters (b) simulation result
118
The final output is ready to be connected to the CRO, and is fed
to the ECG amplifier. The same output is fed to an LED through a
transistor switch to indicate the pulsing of the heart.
(a)
(b)
Figure 5.5: (a) Final stage of the amplifier system, (b) simulation result
119
ECG amplifier designing:
The LM386 is a power amplifier designed for use in low
voltage consumer applications. The gain is internally set to 20 to keep
external part count low, but the addition of an external resistor and
capacitor between pins 1 and 8 will increase the gain to any value up
to 200. The capacitor also maintains the DC bias levels in the gain
adjustment circuit. The unused input is grounded to keep the DC
offset voltages low [5]
.
Also, when using LM386 with higher gains, it is necessary to
bypass the unused inputs, preventing degradation of gain and possible
instabilities. The capacitor connected between pins 2 and 6, provides
power supply decoupling. The RC circuit between the output point
and ground acts as a high frequency load to provide stability. The
potentiometer at the input pin provides adjustable input level
attenuation.
Figure 5.6: ECG amplifier
120
5.1.4 Circuit diagram of the amplifier system
Figure 5.7: Circuit diagram of the ECG amplifier system
5.2 Testing the fabricated sensor with the developed amplifier
The circuit diagram was translated on to a printed circuit board.
The components were procured and soldered on to the board as shown
in Figure 5.7.
121
Figure 5.8: ECG amplifier circuit on a PCB
The developed sensors were connected to the circuit with the
help of crocodile clips as shown in Figure 5.8. The sensors were
placed on the subject with help of adhesive tape.
Figure 5.9: ECG sensors placed on the subject with the help of adhesive
tapes. Conventional sensors placed on the subject but not connected to the
system can also be seen
The output of the circuit was connected to a digital oscilloscope
and tested for signal acquisition as shown in Figure 5.9.
122
Figure 5.10: ECG signal display on the CRO from the developed sensors
and amplified by the developed amplifier system
The developed sensor was connected to the designed amplifier
circuit and connected to a digital oscilloscope. The ECG signals were
obtained on the CRO. The obtained signals were compared with the
signals from conventional, commercially available pre-gelled sensors.
The signals obtained from the developed sensors were on par with that
obtained from the commercial sensors. The signals were validated in
Shanthi Hospital and Research center, Bangalore, by Doctor Sanjay
Gururaj, Director, SHRC.
123
5.3 Skin Impedance measurements
Skin resistance measurements, Variation of resistance with
different skin conditions:
Skin resistance measurements without and with sensors were made as
shown in Figure 5.11
Figure 5.11: Skin resistance measurements
The skin resistance of the volunteer was found to be 5.26MΩ
without any sensor. When the conventional sensors were attached to
the skin of the same person, the resistance was found to be 2.28MΩ as
shown in Table 5.1. The reduction in the resistance can be attributed
to the gel used between the skin and the conventional sensor material.
When the developed dry sensor was attached to the skin of the same
124
person, a still lower resistance was examined, around 1.68MΩ,
indicating a better signal achievement from the developed sensor.
Skin Resistance in MΩ
Normal
Skin
Hairy
Skin
Moist
Skin
Without Sensors 5.26 8.2 8.85
With Conventional Sensors
2.28 5.4 9.2
With Developed Dry Sensors
1.68 0.794 7.54
Table 5.1: Skin resistance measurements
Skin impedance measurements were carried out using the
following experimental set up as shown in Figure 5.12.
Figure 5.12: Experimental set up to measure skin impedance
A small current of 1mA is given as input from the current
source. Three sensors were connected to the forearm as shown. The
input frequency is varied from 10Hz to 10 KHz and the corresponding
125
output voltages are noted down. The impedance is calculated for these
different frequency values.
Skin Impedance in MΩ
Frequency in kHz
Normal Skin
Hairy Skin
Moist Skin
0.01 5.26 8.2 8.85
0.1 2.68 3.98 4.12
1 0.5 0.5 0.5
10 0.5 0.5 0.5
100 0.5 0.5 0.5
Table5.2: Skin impedance measurements without sensors on the skin
Figure 5.13: Skin impedance measurements without sensors on the skin
Without any sensors attached to the skin, the impedance variations of
the skin are shown in Figure 5.14. It is observed that the impedance is
the maximum for moist skin and it decreases with frequency.
126
Skin impedance measurements were done for various skin types
with conventional Silver / Silver Chloride sensors placed on the skin.
Table 5.3 and Figure 5.14 give the details of the measurements.
Frequency in
kHz
Normal
Skin
Hairy
Skin
Moist
Skin
0.01 2.28 5.4 9.2
0.1 1.09 3.001 3.89
1 0.52 0.53 0.5
10 0.52 0.53 0.5
100 0.52 0.53 0.5
Table 5.3: Skin impedance measurements with conventional Ag/AgCl sensors
Figure 5.14: Skin impedance measurements with conventional Ag/AgCl sensors
on the skin
127
With conventional sensors, there is a clear reduction in the skin
impedance for normal skin, but the impedance increases with moist
skin.
Skin impedance measurements were done with the developed dry
sensors on the skin. The details are shown in Table 5.4 and Figure
5.15.
Frequency
in kHz
Normal
Skin
Hairy
Skin
Moist
Skin
0.01 1.68 0.79 7.54
0.1 0.85 0.36 3.21
1 0.5 0.5 0.5
10 0.5 0.5 0.5
100 0.5 0.5 0.5
Table 5.4 Skin impedance measurements with the developed dry sensors on the
skin.
Figure 5.15: Skin impedance measurements with the developed dry sensors
128
With the dry developed sensors, the skin impedance is
considerably reduced, but shows a large value for moist skin. But the
large moist skin impedance of the moist skin with the developed
sensor is lesser compared to the moist skin impedance with
conventional sensors and without sensors.
In conclusion, the developed sensors tend to provide low skin
impedance compared to the conventional sensors indicating better
performance and avoidance of skin preparation.
129
References:
[1] Chia-Nan Chien, Fu-Shan Jaw, “Miniature Ultra-Low-
Power Bio-potential Amplifier for Potable Applications”,
Biomedical Engineering Applications - Basis &
communications vol. 17 no. 2 April 2005, 108-110
[2] Baghini, M. S., Lal Rakesh, Sharma, D. K, “A Low-
Power and Compact Analog CMOS Processing Chip for
Portable ECG Recorders”, Proceedings of the Asian
Solid-State Circuits Conference, Hsinchu, Taiwan, 1-3
November 2005, 473-476
[3] D Rowlands, D. A. James, C Vanegas, “Design and
fabrication of an ECG amplifier on silicon using standard
CMOS process”. Sensors (Peterboroug (2003), Volume:
2, Publisher: IEEE, Pages: 1348-1352
[4] G.M. Patil et. al, “Embedded Microcontroller based
Digital Tele-monitoring system for ECG”, J. instrum.
Soc. India, 37(2), 134-149
[5] IC LM386 data sheets