I'm not sure I would trust a homebrew instrumentation amplifier made from 741 ICs when it comes to monitor heart (or brain) signals. If something went wrong - it could go very, very wrong. There do exist proper medical-equipment grade amplifiers out there - but they aren't cheap. Then again, I value my life, so...?
Quote from: cr0sh on Sep 27, 2011, 09:48 pmI'm not sure I would trust a homebrew instrumentation amplifier made from 741 ICs when it comes to monitor heart (or brain) signals. If something went wrong - it could go very, very wrong. There do exist proper medical-equipment grade amplifiers out there - but they aren't cheap. Then again, I value my life, so...?I'd want to make the device battery operated (no high voltages anywhere) and optically isolated from anything it connects to such as a PC. One possibility would be to base it on the Arduino Uno design but split it in two, with opto isolators on the Tx and Rx signals between the mcu and the USB-to-serial chip. The mcu could be run from a 9v battery via the regulator. An Arduino PWM output pin and a couple of capacitors and diodes can be used to generate -3.7v (or -7.5v using a transistor as well), so the op-amps can be run from about +/- 4v or +/-8v.At the very least, I'd replace the 741 op-amps with something more modern with jfet or mosfet inputs.
Isolation is a good idea - but the issue isn't so much voltage - it's current. Even a 9 volt has enough current to send you into defib (though not via skin resistance). The problem is, in order to run the circuit, you need enough current to run everything; I think the way medical inst amps work is by isolating the pads so only very low currents are on the pad side (probably way less than 100mA). I don't know how you would do this "homebrew" to make things safer...
I'm quite aware that it is the maximum current flow that has to be kept within safe limts. However, we're not talking open-heart surgery here, we're talking about electrodes on the skin - so surely skin resistance will always be present in the circuit, and 9v should be safe? If there is any concern about the maximum current flow under fault conditions, then I would have thought that adding a 100k resistor in series with each input (thereby limiting the possible fault current to around 0.1mA) would provide sufficient protection. However, I don't claim any expertise in this area.
I just found this: http://www.analog.com/library/analogDialogue/archives/37-11/ecg.html. About half way down, it says that the medical standards require fault current to be less than 50uA. To achieve this they run the instrumentation amplifier from +/- 5v and use 220K series resistors on the inputs. They also run it from a battery and use optical isolation from the PC, just as I suggested in my original post.