We have run into issues during testing of our AFE4950 based design towards a AECG100 simulator (simulator for ECG and PPG devices).
If our device is connected to a computer controlling the AECG100 has any ground reference connected to the device under test common mode interference swamps out any and all signals.
Two samples at 250sps and 1000sps provided to demonstrate as ground link is disconnected. PRF is set to 1000 Hz but we decimate to 250sps to get a longer view in plot.
Recovery is practically instant so it does not look like a saturation issue. Amplitude is also well within ranges.
Device is run with 2 electrodes and 100pF connected to RLD_Out with RLD amplifier in unity gain mode.
Path lengths and capacitive coupling to ground may be slightly different on ECG_INP / ECG_INM after the protection resistors.
Protection resistors are 2x 100k in series per input.
We are wondering if this is a CMRR issue for the INA or if this is picked up by the ADC due to reference or other issues.
Connecting the AECG100 reference to DUT groundplane clears up the noise completely but since we run 2-lead we would like to identify if there is a improvement to be made in the circuit. Working under the assumption that we should have better CMRR.
Any guidance in debugging this issue would be appreciated. Also, any information regarding importance of matching of Rprot etc would be good. The example waveforms were captured after DUT had manually matched Rprot as 200.17k and 200.07k but problem is still present (although possibly reduced slightly).
Attached images show examples of noise and "noiseless" result at 250 and 1000sps. Disconnecting noise source (chassis of computer) from DUT or connecting DUT ground and AECG100 reference both result in same clearing up of signal.
