0
\$\begingroup\$

I built an amplifier following this schematic:

enter image description here

And the gain is approximately 1000. When I use 10mV sine wave from a function generator, it works perfectly (it also work with other signals: triangle, pulse, square etc). However, when I get the input from electrodes to measure ECG (2 on each shoulder, 1 ground on the abdomen), it always saturates at op amp power voltage (10V). How do I fix this problem?

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Do the electrodes provide a DC path for the bias currents of A1 and A2? You could also have a noise problem where AC noise is saturating the output of the amplifier which is slow to recover. \$\endgroup\$
    – John D
    May 7, 2015 at 18:50
  • \$\begingroup\$ No. Two electrodes on two shoulders are connected to non inverting input of A1 and A2. The electrode on the abdomen is connected to the ground of the circuit. The rest (power pin, etc) is connected to a power supply (10V, -10V, and ground). \$\endgroup\$ May 7, 2015 at 19:01
  • \$\begingroup\$ What op-amps are they? \$\endgroup\$
    – Andy aka
    May 7, 2015 at 21:22

1 Answer 1

3
\$\begingroup\$

The NIOSH states "Under dry conditions, the resistance offered by the human body may be as high as 100,000 Ohms. Wet or broken skin may drop the body's resistance to 1,000 Ohms," adding that "high-voltage electrical energy quickly breaks down human skin, reducing the human body's resistance to 500 Ohms."

Let's say that body resistance is 100k and the op-amp chosen is a TL082. This op-amp has an input bias current of 50pA and through 100kohm will generate an offset voltage of 5uV. Multiply this by the gain of 1000 and no problems - the output voltage will be 5mV and not end-stopping. However, the TL082 has an input offset voltage of 5mV so this would offset the output by 5V when gain is 1000. But remember that two op-amps are present and their offset voltages could be additive.

What about the LM324? Bias current is 40nA and with 100kohm this generates an input offset of 4mV and not surprisingly produces a whopping 4V offset after amplification of 1000. It looks like the beginning of the problem if the LM324 is used but there's another issue - the LM324 has an input offset voltage of typically 2mV so, after a gain of 1000 this might look more like 2V added to the 4V due to input bias currents. Given also that there are two amps the offset voltage could be additive making a total of 8V offset.

An LM741 has a bias current of 80nA and this, given the same rationale as above produces an 8V offset on the output. Plus, the input offset voltage is maybe 2mV thus adding another 2x 2V. Grand total is 12V offset.

My favourite quad op-amp (OP4177) has a typical offset voltage of 25uV and a bias current of 0.5nA - this would, in the configuration above produce 50mV at the output due to bias currents and 2x 25mV due to offset voltages. 100mV total offset, ahhh that's better....

So, what op-amp are you using?

\$\endgroup\$
2
  • \$\begingroup\$ I am using LM741 for all three. I think that is the reason, gain is too high making the offset outside the range of power voltage(-10 to 10). How do I fix the problem? \$\endgroup\$ May 8, 2015 at 19:57
  • 1
    \$\begingroup\$ I recommend the OP4177 if you want a quad op-amp solution (as per my answer). Never ever consider using a 741 these days - I seriously began using op-amps about 1980 and I rejected them back then in favour of the now ubiquitous LM324. Think on, it's bias current and input offset voltage that are the likely killers of your project so if you can't afford the OP4177 look for something nearly as good. \$\endgroup\$
    – Andy aka
    May 8, 2015 at 20:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.