This morning I changed the matrix that sends DCPD ADC channels to the 524 kHz test channels that Jeff set up and described in 82686, hoping that this might be useful for following up on some of investigations that Kiet has done on violin mode contamination. 85026

Right now, I've set them up to send the 4 individual ADC channels for DCPD A to all 4 filters, so that A CH0 is A1, A CH1 is B1, A CH2 is A2, and A CH3 is B2. 

One striking thing in the first look I've taken at these spectra is that the rms is dominated by high frequency lines (and somewhat the 1kHz violin modes), 10215.3 Hz (f1) is the loudest followed by 10436.9 Hz (f2) and 8154.8 Hz (f3).  We have been wondering if a quadratic response of the ADC could explain the violin mode contamination, so I checked the 6 combinations of these frequencies, at this resolution I only see a peak at f2 + f3 = 18591 Hz.