Just a quick look here that I'll follow up on in a day or 7.

On HAM3, it looks like the gain switch on H1 corner for the GS13 produces a step and possibly a DC shift.  That does not seem good.

I just quickly scanned back through the data and looked for a GS13 BIO switch event that had a trip.  I don't know whether it is guardian or Jim's python script doing the switching.

On the first plot there are all six HAM3 GS13INFs; it looks like H1 switches first (red).  The Step and Offset are fairly clear.  In the second plot you can see this step/offset with the BIO channels showing that the step/offset occurs during the switching process.  H2 also switches before the WD trip registers (3rd plot) and although it might be influenced by the trip that is commencing, it does not appear to have the step/offset seen on H1.  All other GS13s don't actually switch until after the WD has registered the trip state and their signals have already blown up.  Not to say that these other corners wouldn't look the same if the switching were reordered.  Mind the channel names--the colors are not consistent.

Bottom line, the HAM3 H1 exhibits a step/offset that is not good, IMHO.