I’ve also measured similar noise on some channels when recording to a portable USB drive with hot, cold, and ground inputs of all channels shorted together. It manifests as short pulses that occur at regular intervals of approximately 1 second and raises the noise spectrum at frequencies up to 2 kHz. The increase on the worst channels is 20 dB at 100 Hz and more at lower frequencies. The worst channels are those with the separate 2-channel A/D converters (7, 8, 15, 16, 23, 24) although the pulse-like noise is also apparent at lower levels on channels 5, 6, 13, 14, 21, 22. It does not appear on the remaining channels. Doing some microsurgery on the USB cable so I could power the same drive from a separate bench power supply resulted in a reduction in this noise of about 20 dB, so it is highly likely to be interference due to hard drive current pulses. Externally powering the drive gives acceptable results for our purposes but this problem could certainly catch people out.