Last weekend we ran the GG again throughout the weekend.   The conditions were:

1.  Full TPX/iTPC
2.  Pedestal run was taken with GG open, so the initial GG peak is in the data.   Data with GG set normal.
3.  (I turned on the old GG crate, which is the nearest electronics to the new GG crate)
4.  We did see several instance of the secondary GG peaks:
5.  We had oscilloscope leads on sector 6 of the itpc, taking scans every 10 seconds.

The extra data was primarily in sector 13 (which was the less frequently observed mode during last years cosmic run).  Here are the sector sizes for a bad event in this run and a good event in the same run:
  

Same for TPX:



There is about 20X more charge in sector 16 than in the other sectors.   Several other sectors seem to be affected as well (23 & 24 have the most significant effects).   These are close to sector 13 both in real space and also in terms of GG board number (23 & 24 are in Boards 6 and 7, sector 13 is in GG board 5).

The charge step for these events looks as follows:

For the TPX, the GG pulse looks nearly the same but is > 20 x larger.  It is slightly shifted in time, but this might simply be a slewing effect:

For the iTPC we see a similar factor of 20, but in this case there are double peaks:

Finally, the GG traces do consistently show a very small artifact during the periods when the problem occurs

Here, the small notches on the turn on / turn off seem to be consistently present when the pulses are present on sector 13.  Of course, these might be more significant if the right sector's wires were being watched.

The next step is to repeat the same thing, but with an additional scope trace on sector 13 (as well as sector 6 which was the worst failing sector previously)