Problem with a heartbeat monitor circuit

[aredhel]

Hello there. I am having a problem with my heartbeat monitor circuit. The project is to count the number of heartbeat per minute. I am planning to use an LCD to display the result, but before that I wanted to make sure that I am getting the right ECG signal. However, I can't seem to get the signal or maybe it was swamped with noise. The active filter implemented worked just fine, but whenever I connect my LL to the ground the signal went flat. Anyone knows what is wrong?

 

[Audioguru]

heart beat monitor schematic

Your Sallen and Key lowpass filter has a gain of 11 which is higher than 3.0 which will cause it to oscillate. It is a Butterworth filter if its gain is 1.6.

 

[aredhel]

heart beat monitor circuit

Your Sallen and Key lowpass filter has a gain of 11 which is higher than 3.0 which will cause it to oscillate. It is a Butterworth filter if its gain is 1.6.

Why must it be 1.6 to get a butterworth response? Does this mean it is better to amplify the signal after the filtering is done, for example amplify the signal through the active notch filter. I am not sure what is wrong, but my active notch filter is not working. I have checked my connections for more than five times and yet the filter refuses to work.

 

[ywzslm]

heart beat display on cro

depend on the filter i think

 

[Audioguru]

single supply ecg twin t

A Butterworth response for a second-order Sallen and Key filter requires that the opamp's gain must be 1.6 if the resistors and capacitors have equal values, or the feedback capacitor must have double the value of the other capacitor and the opamp must have a gain of only 1.

 

[aredhel]

heart beat monitor circuit(pdf)

I've fixed my design and attached is the new schematic. I am wondering about the instrumentation amplifier: AD620ANZ. That amplifier requires a dual power supply of around the range from ±2.3V to ±18V. In order to supply this from the dual power supply usually found in labs do I need to do some sort of bridging? That is apply the red to the positive supply pin and the black to the negative supply pin, while the remaining red and black socket I need to connect them together in order to create a ground?

In addition, the pin 5 of AD620ANZ ... does it need to be grounded because when I grounded it there is no signal at all, however in the sample circuit shown in the datasheet pin 5 is grounded.

 

[Audioguru]

schematic waveform output for heartbeat monitor

Medical electronic products are supposed to have perfect isolation from an AC powered power supply to avoid electrocuting a patient if something shorts.

Pin 5 of the instrumentation amplifier is supposed to be grounded. Yours doen't work probably because the inputs do not have a path to 0V. Usually the patient is connected to 0V.

Your instrumentation amp has a very low gain. Then the output opamp amplifies the very high noise level of the lousy old 741 opamps.
The instrumentation amplifier is low noise and should provide most of the circuit's gain.

You have U3 connected as a Sallen and Key highpass filter at about 5Hz. Why?

Most of your resistor values are way too low. U3 has a load of only 81 ohms (R14 plus R15) which is like a dead short. An ordinary opamp can drive 2k ohms but not less.

Even though U3 is not needed, its frequency determining resistors are not 1:1 like they should be when the opamp has a gain of 1.6.
Your 1k resistor value is so low that the capacitor values need to be very high. Where are you going to find a 33uF non-polarized accurate capacitor? An electrolytic has a very wide tolerance. A metalized plastic film capacitor is perfect but a 33uF one is huge and very expensive.
If the 1k resistors are changed to 100k then the 33uF capacitors can be small and accurate 0.33uF ones

The same applies to your lowpass filter. The capacitors should have equal values and the value of the resistors should be much higher.
The same for the twin-T filter.

 

[LvW]

project on heart beat monitor using arm7

The filter looks like an highpass filter, doesn´t it ?

 

[Audioguru]

heart beat monitor project

I corrected the highpass filter but why is it used?
Heartbeats are low frequencies and you want high frequency noise and low frequency mains hum reduced.

 

[FvM]

schematic output waveform for heartbeat monitor

In addition to odd filter dimensioning (the low pass is even unstable), the DC gain is so high, that a considerable part of usable output range is wasted for amplified offset voltage, at worst case the output could be saturated in bias point.

 

[aredhel]

filter heartbeats

The instrumentation amplifier has low gain because I read that a low gain is to avoid gain saturation. As for the 741 opamps ... well they are the cheapest and most importantly easiest ones for me to obtain. When I constructed the circuit with a high gain for the instrumentation amplifier, the signal was swamped with noise; even filtering doesn’t help.

As for the HPF and LPF … sorry, I didn’t realise that I typed in the wrong values. The HPF was supposed to have a cutoff at 0.05Hz and 100Hz for the LPF ... since the ECG signal ranges from 0.05 – 250Hz.

Ah, so an ordinary opamp can drive 2k ohms and above ... I did not know that. Thank you. I am wondering about one thing, why is a non-polarised capacitor much preferable to a polarised one?

I built my circuit based on:
https://www.bme.utexas.edu/ugrad/UGLab/221/Biopotentials.pdf.
**broken link removed**

Let me get everything sorted as I think I understood it. Large values of resistors are preferable due to the opamps and non-polarised capacitors with small value are better.

Q1. Would cutoff of 0.5Hz (HPF) and 100Hz (LPF) be enough?
Q2. If I was to make use of the AD620 gain ... how do I know the gain limit without saturating the output?
Q3. The instrumentation amplifier is able to drive below 2k ohms?

Attached are the new calculated values and the waveform obtained from the previous design. Pin 5 (not shown) is grounded because I could not get a signal without grounding it. Input is only from my Left and Right Wrist.

 

[LvW]

ad620 gain

Hi aredhel,

one comment to the highpass stage: What is the required cut-off frequency ?
In your text, you have mentioned 0.5 as well as 0.05 Hz. But in the referenced document it is 0.05 Hz.
However, your design is according to 0.5 Hz. Please check again.

In addition, be aware that it will be very problematic to design an active filter circuit at a cut-off of 50mHz (large component values, bad capacitors, finite opamp input resistance).

 

[Audioguru]

lm324 ecg circuit

In the filters if you use high value resistors then you can use low value small accurate metallized poly capacitors.

The circuit from the university is very old. It uses very low noise opamps that have a very low input offset voltage. The lousy old 741 opamps that you use have the noise that shows on your 'scope.

The university circuit has an overall very low gain so maybe their 'scope provides most of the gain.
The schematic shows only 2 inputs on the wrists but the discussion also mentions a "left ankle" wire which is probably the circuit's ground.

Here is a more modern ECG circuit that uses low noise opamps. It feedsback the average input common-mode DC voltage to the ankle to cancel DC offset.

 

[aredhel]

digital heart beat monitor

Hi aredhel,

one comment to the highpass stage: What is the required cut-off frequency ?
In your text, you have mentioned 0.5 as well as 0.05 Hz. But in the referenced document it is 0.05 Hz.
However, your design is according to 0.5 Hz. Please check again.

In addition, be aware that it will be very problematic to design an active filter circuit at a cut-off of 50mHz (large component values, bad capacitors, finite opamp input resistance).

Yes I wanted a cutoff at 0.05Hz, but as I built the circuit and used the frequency generator (I wanted to test whether the circuit is working as it should) could not supply such low frequency. In addition, it seems completely useless to have a HPF with a cutoff at 0.05Hz since my circuit is not even functioning properly, in fact its cutoff turned out to be 4Hz ... nearly killed off the ECG signal. I am giving 0.5Hz HPF a try and hope the cutoff is not too high that it kills my signal.

I have one question ... I tried out my Notch filter from the previous schematic (the one before the new one I just posted). I got an square-like waveform when I supplied a sinusoidal wave from the frequency generator. However for my LPF I ended up with a sinusoidal wave for an output. I tested these two circuits on separate breadboards. Anyone knows what may be the reason?

Audioguru are there any other alternatives to the old 741 that I'm using? The one provided is hard to find where I'm at. TL071 and TL081, are they good options?

 

[Audioguru]

it project topics on heart beat moniter

the lousy old 741 opamp is 40 years old. It and the lousy old LM324 and LM358 opamps were not designed for low noise wideband audio.

I don't know where you are and I don't know why you can't buy ordinary inexpensive modern low noise opamps.
The TL071 has half the noise of a 741.
The OPA134 has 1/5th the noise of a 741.
The LM4562 has 1/14th the noise of a 741.
There are opamps with even less noise.

Make a lowpass filter with a sharper cutoff to reduce the hiss noise from your lousy old 741 opamps. But then its low frequency popcorn noise will be a problem.

 

[LvW]

heart beat monitoring circuits

Yes I wanted a cutoff at 0.05Hz, but as I built the circuit and used the frequency generator (I wanted to test whether the circuit is working as it should) could not supply such low frequency. In addition, it seems completely useless to have a HPF with a cutoff at 0.05Hz since my circuit is not even functioning properly, in fact its cutoff turned out to be 4Hz ... nearly killed off the ECG signal. I am giving 0.5Hz HPF a try and hope the cutoff is not too high that it kills my signal.
I have one question ... I tried out my Notch filter from the previous schematic (the one before the new one I just posted). I got an square-like waveform when I supplied a sinusoidal wave from the frequency generator. However for my LPF I ended up with a sinusoidal wave for an output. I tested these two circuits on separate breadboards. Anyone knows what may be the reason?

Highpass: As 0.5 Hz perhaps are OK for you - did you make already some measurements ? All cut-offs around 1 Hz (and below !) are problematic.
Notch: There are better alternatives than double-T topologies. Double-T is hard to tune. As you got a squarewave at the output (which frequency ?) this indicates bad suppression (notch not exactly tuned) and overdrive of the opamp with a gain of "9".

 

[aredhel]

heartbeat count at 78

Highpass: As 0.5 Hz perhaps are OK for you - did you make already some measurements ? All cut-offs around 1 Hz (and below !) are problematic.
Notch: There are better alternatives than double-T topologies. Double-T is hard to tune. As you got a squarewave at the output (which frequency ?) this indicates bad suppression (notch not exactly tuned) and overdrive of the opamp with a gain of "9".

I only know of the Wein Bridge and the Twin-T ... been looking for other notch topologies. Could you recommend some? The square wave was well ... a whole range of frequencies. I tried from 4Hz (and below) up to more than 10kHz.

If I applied sinusoidal input from the frequency generator at Pins 2 and 3 of the instrumentation amplifier is the output supposed to be a flat line (since it is a differential amplifier)? I seemed to be getting a normal sinusoidal wave back at Pin 6.

 

[Audioguru]

ecg monitoring circuit through cro

The input common-mode rejection ratio of the AD620B instrumentation amp is a minimum of 100dB. So if both inputs are fed the same signal then the output is reduced by 100 thousand times or better.

The datasheet also shows an opamp feeding back the offset voltage to the ankle of the patient.

 

[aredhel]

heart monitor circuit ad623

The input common-mode rejection ratio of the AD620B instrumentation amp is a minimum of 100dB. So if both inputs are fed the same signal then the output is reduced by 100 thousand times or better.

The datasheet also shows an opamp feeding back the offset voltage to the ankle of the patient.

Yes, thats is the right leg drive to drive the common-mode voltage to a low value and to increase the patient's safety. However, I am not sure of what kind of op amp I could use. The one shown in the datasheet is not available from Farnell (online electronic shop) and the electronics shop that I go to. Will OPA134 or OP177GPZ do?

OP177GPZ: it can be an alternative to my 741?

 

[Audioguru]

straight line heart beat monitor

Why don't you compare the noise of the opamps on their datasheets?

Most 741 opamp manufacturers don't say the noise level of a lousy old 741 opamp. Philips rates the typical noise at 38 (very noisy).
The OP177 is rated at 10 (pretty quiet).
The OPA134 is rated at 8 (pretty quiet).