After yesterdays 10am mtg the east/west GG driver cables were swapped.   

The signature of the TPC noise near the 50th timebin seems to have moved with the cables.   Before the move I saw the noise occur in both the east and the west TPC, but in the 24 hours since the change the noise has only showed up in the west.

I looked event by event at the adc values for the noise.   As we already knew, the noise affects almost all of the pads/padrows in the sector, showing up as a sheet near 50 timebins.   The structure is extremely stable from event to event and from run to run, but the structures are different in each sector:

The following plots show the summed ADC values for the entire inner sector at a specific timebin for a single event vs timebin (for the first 100 timebins).    The time is during an outbreak of noise spikes, and many events are superimposed on the same plot.

I show 12 plots, for each sector from run 20041001 (before the swap of cables) and run 20043007 (after the swap).   I put the sectors that correspond to the same GG Driver port on the same plots:

I show the plots for all 12 sectors, despite the fact that many show no real effect.   The reason I show them all is because its obvious for the sectors that have the most significant noise that the signature doesn't change a bit.   These would be sectors (5, 6 & 7) corresponding to sectors (13, 24, and 23) respectively.   The other sectors, however, show effects that are much less correlated.

I'm guessing that the ultimate issue is related to some kind of junk being passed around on the backplane of the GG driver crate, perhaps caused by a power supply or a bad board.  The biggest effects are at least somewhat stable, but the minor effects less so.