In short, the OMC DCPD ground problem we experienced in the past (e.g. 61549) seems to come from the OMC DCPD assy itself, either DCPD1 assy (D1300372) or DCPD2 (D1300371) (cannot tell which, but one of them is bad). The most likely thing is that the case GND (pin3 in the PD assy) is somehow connected to the MightyMouse (MM) connector shell. See attached.

The connection of PZT and QPD shield to the metal connector bracket on the OMC breadboard is according to Rich (https://alog.ligo-wa.caltech.edu/aLOG/uploads/47227_20190301152409_OMCGroundingatLHO1March2019.pdf). The new A+ design (that's what we're using now) is supposed to isolate OMC DCPD shield from the metal bracket, certainly D2100716 seems to be made as designed, but somehow that's bypassed on the breadboard (red connection in the attached).

Here's how we reached that conclusion:

1. Disconnect the in-chamber PZT cable from the feedthrough.
2. Disconnect the DB9-DB9 cable (that goes from the OMC DCPD preamp to the OMCS cage top cable bracket) at the OMCS cage.
3. At this point, DCPDs are connected to 2xMM-1xDB9 cable (D2100599) on the cable bracket on the OMC breadboard itself, but that cable is not connected to anything downstream.
4. At this point, there is a connection from pin6 of DB9 side of D2100599 to the pin1 of the in-vac PZT cable (shield ground).
   1. pin6 of DB9 side of D2100599 is the case ground for one of the PDs.
   2. No other DB9 pins are connected to the PZT cable shield ground.
   3. FYI, once connected to the preamp, pin6 as well as 3 and 9 of DB9 are grounded inside the preamp.
5. pin6 to PZT shield connection goes away if we disconnect one of the MM connectors (J2 on D2100599) on the cable bracket on the OMC breadboard. J2 is to the left viewed from +X door.
6. pin6 to PZT shield connection is only dependent on the J2 connection, it's not affected by the other MM connector (J3).
7. When we disconnect J2 MM connector from the OMC breadboard (thus no connection between DB9 pin6 and PZT shield ground) to test the D2100599 cable itself, there's no bad connection from any pin, that is,
   1. DB9 pin6 is only connected to one pin of J2 (we assume pin 3, case GND), not to the J2 shell.
   2. DB9 pin1 is only connected to another pin of J2 (we assume pin 1, cathode), not to the J2 shell.
   3. DB9 pin2 is only connected to yet another pin on J2 ( we assume pin 2, anode), not to the J2 shell.
   4. DB9 pin3 (shield) is not connected to any pin of J2 nor J1 shell nor J2 shell
   5. J2 pin4 (NC) is not connected to anything on DB9.
8. When we disconnect J3 too, there's no bad connection either.

As such, we eliminated the possibility that D2100599 cable is bad. This only means that the case of one of the DCPDs, the one that is connected to J2, is connected to the metal cable bracket on the OMC breadboard.