Customer used R=10ohm, with 0.1uF common mode cap for the input filter of INA280. But they found when the actual cap value is not well matched, such as with 3nF difference, the output of INA280 would show oscillation, and the amplitude would reduce.
Do you think it makes sense?
Hi Aki,
any imbalance in the common mode input low pass filter will ruin the awesome common mode rejection of INA280. The TINA-TI simulation demonstrates that an imbalance of only 3% results in a degradation of common mode rejection of nearly 40dB!!!!!:
I guess you have a common mode input signal looking like an "oscillation", which with a good balance of input filter (C1 = C2 = 100nF) is almost entirely suppressed at the output by the INA280, but becomes visible suddenly when there is an imbalance of input filter (C1 = 100n, C2 = 103nF) because of the degraded common mode rejection. This would look like an oscillation which seems to be caused by the imbalance of input filter. But in reality the "oscillation" is merely the consequence of a heavy degradation of common mode rejection.
When it comes to common mode input filtering. "less" is sometimes "more". You should not choose all too high common mode filter capacitances, because the higher the capacitances the higher the degradation of common mode rejection. In any case the caps should be state-of-the-art referring to manufacturing tolerances, temperature drift and long term drift. Best use very narrow tolerated ceramic NP0 (C0G) caps. Also, avoid the introduce of additional imbalance by uneven stray capacitances at the input. Avoid any unnecessary cabling between the shunt and the INA280.
Kai
