After swapping the cables at the interface, the noise on the A seismometer was gone (see attached)--should have been now on the B channel (it has never been seen on the B channel.) Okay, one thing in common, the B cable has no shell so it has never been tightened onto the chassis and I had not tightened the A cable when connected to the B chassis. Watching a running spectra, I tightened the A cable into the B chassis--no change observed. Powered off the chassis and swapped the cables back A to A & B to B but did not tighten the screws--no noise seen once the signal settled down. Then tightened the screws slowly while watching the spectra--no change in the noise seen.
The second attachment below is after the cables were swapped back to their correct positions. Now looking closer and talking to Robert, we see that while the noise in the magnitude looks much better, it is still not as good in the coherence especially if you compare it to the Z channel. With all the swaps and tests, it may be the powering off of the interface during swaps or moves that causes the problem. This will be my next campaign; otherwise, move the HAM2 unit back into the BierGarten and swap the signals at the instruments.
Tested the shut down/restart theory. Over about an hour, I powered off the interface chassis and restarted it ~10 seconds later, five times over the course of this hour. No obvious change in the spectra observed. This power cycle is certainly quicker than the time to restart after switching cables much less to move the instrument from location to location, so, maybe, not a valid test. Maybe the 'change' requires being off longer. This morning when the extreme noise became just marginal, it was just a chassis cable swap but still probably took a couple minutes, not 10 seconds.
The Y channel signals for the HAM2 sensor have a bit worse coherence than the X channel, but the X channel too does standup to the solid coherence between ITMY & HAM5.