Heart beat measuring analog circuit

[logicnibble]

Hi!
I was trying to make a circuit to measure the heart rate using those electrodes you stick in the skin (can be a coin).
I've tried an instrumentation amplifier (LT1167) with gain set to 1000, so I was expecting a signal of 1V, but I got so much 50Hz mains noise, that the amplifier output saturates.
Several circuits on the internet use filtering after the amplifier, but I need to get rid of the noise before it enters the amplifier.
How can I reduce this mains noise?
Do you suggest using any particular circuit?

Thanks!
LN

[Zad]

Have you tried using a 3 terminal system? The use of a third (virtual ground) terminal allows the removal of the induced mains hum. Analog Devices have a huge amount of info on ECG systems http://healthcare.analog.com/en/patient-monitoring/ecg-consumer-grade/segment/health.html

[alm]

Note that care should be taken to isolate the body from any transients or leakage current. If you had to choose how to best electrocute someone without cutting open their body, ECG electrodes would be it. Resistance is much lower this way. Look at the requirements for medical grade power supplies for some inspiration.

[johnmx]

The differential mode of amplification should remove the 50Hz mains interference.
Advice, limit the amplifier’s bandwidth from 0.5Hz to 100Hz.

If you want more help you need to show us the schematic.

I made once an ECG, see the result in the attached picture.

[NiHaoMike]

Note that care should be taken to isolate the body from any transients or leakage current. If you had to choose how to best electrocute someone without cutting open their body, ECG electrodes would be it. Resistance is much lower this way. Look at the requirements for medical grade power supplies for some inspiration.

EKG signals are very low bandwidth so it would be trivial to build a battery-operated wireless sensor. In fact, that's going to be part of my senior design project. The DC/DC converter that steps up 10-22V from the pedal alternator to 170-190V for the inverter needs an EKG signal for MPPT operation. Wireless just happens to be the best and most convenient way to get the necessary isolation (for both safety and noise reasons) as well as not requiring the user to be physically connected to the device.

[logicnibble]

The differential mode of amplification should remove the 50Hz mains interference.
Advice, limit the amplifier’s bandwidth from 0.5Hz to 100Hz.

If you want more help you need to show us the schematic.

I made once an ECG, see the result in the attached picture.

Can you show your design schematic?
To limit the amplifier's bandwidth - should I just choose one with low bandwidth or do that by circuit design?

[scrat]

You have to set the transfer function by design,while the amplifier has to be in excess of Gain-BW.

There are ready-made instrumentation amplifiers, which are made for differential input, otherwise you can use an opamp in differential configuration, which allows to use both input and feedback for filtering, but a gain of 1000x would be difficult this way.

[logicnibble]

I'm using the LT1167 with +-5V supply. Reference pin is on GND.
The gain resistor RG was first set to 49R (G=1000).
Output saturates either on ~4V or ~-4V.
I've tried with 49k (G=2) and I got the same results.
Should it be the Common Mode Voltage that is beyond the +-5V and then the outputs saturates immediately?

[scrat]

Sorry, you're already using an instrumentation amplifier.

BW can be set on the input (capacitors in series with R-C parallel), and then you can do further filtering on the output.

If the input terminals come both from body and are treated as differential (i.e. symmetrical paths, and twisted cable if it is worth), you should be able to eliminate the common-mode mains signal.

[johnmx]

Check the following resources:

http://www.robots.ox.ac.uk/~neil/teaching/lectures/med_elec/downloads.html#lecture
http://www.picotech.com/applications/ecg.html
http://www.docstoc.com/docs/70647467/ECG-(electrocardiogram)-is-a-test-that-measures-the-electrical

[logicnibble]

Check the following resources:

http://www.robots.ox.ac.uk/~neil/teaching/lectures/med_elec/downloads.html#lecture
http://www.picotech.com/applications/ecg.html
http://www.docstoc.com/docs/70647467/ECG-(electrocardiogram)-is-a-test-that-measures-the-electrical

Thanks! These are good information sites regarding ECGs.

I've made some progress but I think I'm having problems with Common-Mode voltage at the instrumentation amplifier inputs:
the output is often at V+(+5V) or V-(-5V) and sometimes it goes from V+ to V- (or vice-versa) slowly, taking ~10s.
During this time, I have the ECG waveform after filtering it between 0.5Hz and 30Hz.
Can I just connect one input to GND to avoid this?

[scrat]

Can I just connect one input to GND to avoid this?

I'm not an expert, but as said above and IIRC in some of the articles linked, it seems that a third electrode connected to the body and to your gnd could solve the problem. Besides this, I'd twist the two input wires (which should be the same length) and avoid using a too much high input resistance.

[mjkuwp]

on fitness equipment there are contact assemblies for the LH and for the RH.  Each side has two paddles.  This makes 4 total contact pads but two of them are the same.  You need the RH, the LH and one of the others.  In other words, yes you need 3 connections.

this link looks to have some good information.  i go there via one of the previous links...I think.

http://www.analog.com/library/analogDialogue/archives/37-11/ecg.html