Where is the offset occurring? Generally you could counter that with an adjustment pot at the proper spot in the circuit.

 

Where is the offset occurring? Generally you could counter that with an adjustment pot at the proper spot in the circuit.

At the output of the AD8226. Signal had a dc offset slightly higher than Vref instead right at it.

Id rather not have to employ adjustment pots.

 

Okay. But I don't understand why capacitive coupling is causing AC pickup. It should have nothing to do with that.

 

Hello,

Do you have the AC I bias resistors?
from the same datasheet:

Bertus

 

Hello,

Do you have the AC I bias resistors?
from the same datasheet:

Bertus

Yup, using 1uF caps and 33kOhm resistors. cut off freq = ~5Hz

I'm using this setup to measure biopotentials so the noise is most def coming from the body. Usually the high CMRR eliminates this noise but I don't really understand why the HPF at the inputs degrades this.

 

Yup, using 1uF caps and 33kOhm resistors. cut off freq = ~5Hz

I'm using this setup to measure biopotentials so the noise is most def coming from the body. Usually the high CMRR eliminates this noise but I don't really understand why the HPF at the inputs degrades this.

At that cutoff frequency any mismatch between the RC time-constants can cause a differential output signal for common-mode 60Hz pickup. A simulation shows that a 10% difference in time-constants (which can easily occur due to the large tolerance of electrolytic caps) would give a 17mv difference signal for a 1V common-mode input signal. Try changing the capacitors to at least 100uF and/or increasing the resistor values.

 

At that cutoff frequency any mismatch between the RC time-constants can cause a differential output signal for common-mode 60Hz pickup. A simulation shows that a 10% difference in time-constants (which can easily occur due to the large tolerance of electrolytic caps) would give a 17mv difference signal for a 1V common-mode input signal. Try changing the capacitors to at least 100uF and/or increasing the resistor values.

Would choosing a different cutoff freq help? I chose 5Hz cause its the minimum frequency the true signal would produce. I could move it lower if it would help remove the 60Hz noise.

 

Would choosing a different cutoff freq help? I chose 5Hz cause its the minimum frequency the true signal would produce. I could move it lower if it would help remove the 60Hz noise.

That's what increasing the value of C and/or R, as I suggested, will do.

 

That's what increasing the value of C and/or R, as I suggested, will do.

Ah OK. I took it to mean up the cap or resistor but keep the same freq. (Eg 0.01uF and 3.3MOhm).

 

Ah OK. I took it to mean up the cap or resistor but keep the same freq. (Eg 0.01uF and 3.3MOhm).

The idea is to make the corner frequency much lower so variation in the filter gain at 60Hz is negligible, even if the capacitor/resistor values vary due to their tolerances. And anyway you should have the corner frequency much lower than your lowest frequency of interest, since you lose 3dB of the signal at the corner.

 

The idea is to make the corner frequency much lower so variation in the filter gain at 60Hz is negligible, even if the capacitor/resistor values vary due to their tolerances. And anyway you should have the corner frequency much lower than your lowest frequency of interest, since you lose 3dB of the signal at the corner.

What cutoff frequency would you recommend?

 

What cutoff frequency would you recommend?

It's not critical. Probably no higher than 1/10 of the lowest frequency of interest or 0.5Hz in this case. I previously suggested 100uF caps to minimize the 60Hz pickup, which would make it 0.05Hz.