Update (Betsy, Keita)

Prep work.

SMA feedthroughs are an isolated type and nothing is connected outside of the chamber, and there's no continuity between the shells and the chamber ground, so that's not an issue. We left it connected inside the chamber to both the preamp and the feedthrough during our check.

I disconnected in-air DCPD cable as well as in-air PZT cable at the feedthrough and attached breakout boards. QPD cable was still connected at the feedthrough but was disconnected at the rack.

Before doing anything.

From outside of the chamber, I checked continuity between things and it was the same as yesterday. See 1st attachment top where green arrow with X in the middle means isolated and the red arrow means continuous. PZT pin 13 was short circuited to DCPD pin 13 as well as DCPD signal ground (pin 10), but not to the chamber ground.

DB9 disconnected from the preamp in chamber.

With preamp isolated from the OMC breadboard bracket, the connection between PZT pin 13 and DCPD pin 13 as well as DCPD signal ground are gone, but DCPD pin13 was still connected to DCPD signal ground. See the first attachment bottom.

At this point we measured the continuity between the preamp box and the DB25 shell and they're continuous. Note that in-vac DB25 has an all metal shell on the preamp and PEEK/metal shell on the feedthrough. 

We removed the DB25 from the preamp and found that the metal shell at the preamp end and the metal part of the shell on the feedthrough are continuous, so that's one thing that was wrong.

While that cable is still connected to the feedthrough I confirmed that pin13 is isolated from the signal ground in air unless it's connected to the preamp, so the short in the cable was excluded.

Installed  a temporary in-vac DB25 cable that has PEEK shell at both ends.

To our disappointment, this didn't change the short-circuit between the DCPD pin 13 and the signal ground. See the second attachment.

We measured the continuity between the preamp and the aluminum collar that is seemingly bundling the shielding on the PEEK DB25 connector, and they were isolated.

It seems to me that pin 1 of in-vac DB25 cable (which is connected to pin 13 in air) for DCPD is somehow connected to the signal ground inside preamp. We'll talk to others.

Attached is the pic confirming the PEEK spacer in the mounting of the new OMC PreAmp box - good.

Now picture of cabling disconnected from the PreAmp while ground loop checking.  Note, old style D1000225 25pin cable to feedthru is being reused from before from old PreAmps.

Few pix of an all-PEEK end 25pin cable to feedthru which we tried, thinking maybe it would solve the ground loop seen.  Pictures show both flavors of cable (old D1000225 and new D2000456).

Updated update (Jim, TJ, Betsy, Daniel, Rich, Dean, Keita)

DCPD in-air cable pin 13 (which becomes pin 1 in the chamer) is shorted to DCPD signal ground inside the preamp.

I disconnected DB25 and DB9 from the preamp in chamber and confirmed that the pin1 of DB25 connector on the preamp is continuous with the preamp body.

We cannot do anything about it, and Rich proposed to change the grounding scheme in air. We still need to decide how, though Rich suggested to modify DB25 in-air cable at the rack (e.g. remove pin 13).

This is not the end of the world as nothing is grounded to the chamber.

I connected the old in-chamber DB25 cable back on the preamp. This is a cable with metalized shell at the preamp end (D1000225-S1105035). That's a wrong cable according to Rich and Dean, and we already knew that it was a wrong one from my observation in the parent alog above, but using the correct cable with PEEK shells at both ends won't do anything good (because the wrong connection is already inside the preamp).

I also connected the DB9 back on the preamp.

Summary of grounding as of now.

In-air PZT cable and In-air DCPD cable disconnected from the feedthrough, and QPD cable connected to the feedthrough but disconnected from the chassis, this is the current state.

• DCPD signal GND <- continuous -> DCPD pin 13 <- continuous -> PZT pin 13
  • The reason why DCPD signal GND is connected to PZT pin 13 is because of the MightyMouse connector thing documented elsewhere. See Rich's figure from March 2019 in alog 47227.
• but none of these are grounded to the chamber.

Later I connected the in-air DCPD cable and in-air PZT cable to the feedthrough, but kept them disconnected from the chassis, and checked the continuity between the pins at the chassis-end of the cables.

• QPD pin 13 was connected to DCPD pin 13 and DCPD signal ground.
  • Again, see Rich's diagram.
• PZT pin13 was NOT connected to DCPD pin 13 and DCPD signal ground.
  • I disconnected the PZT cable from the feedthrough and confirmed that there's no continuity from pin13 at one end to the other end. Same for the shell. Apparently somebody did this in the past, but I don't see any label on the cable.

PZT pin 13 to DCPD signal ground as well as This is good enough for now.

Checking if things are still connected in chamber, burnt by a wrong in-vac SMA cable.

Note: I stole a function generator and an oscilloscope from SQZ rack. I returned them later but I didn't connect things back, sorry.

• I powered the preamp on and used a flashlight. Both DCPDs responded and I was able to see signals in H1:OMC-DCPD_A_OUT_DQ and B. Good.
• I checked the SMA connections for 3MHz and PDA didn't work at first. Turns out that that was a bad in-vac SMA cable. Replacing that with a new one, it started working. Good.
  • Used a function generator to inject signal from the DB9 on the DCPD whitening chassis back panel. (Pin 1 of DB9 was connected to CH1 of the function generator, pin6 to CH2, and these were set to 200kHz, 5Vpp with 180deg different phase, ground was connected to the chassis ground. Though I didn't have to do that, that's what I did.) I was hoping to be able to see something from SMA connectors at the feedthrough, and I did, but PDA was a factor of 10-ish smaller than PDB (1st attachment, blue-> PDB, yellow -> PDA). Not good.
  • We swapped the in-vac A and B cable at the feedthrough, we swapped A and B cable on the preamp, and the signal followed the cable. We concluded that something should be wrong with the cable for A.
  • We installed a new in-vac SMA cable and PDA started working (2nd attachment). Good.
• Weird coupling from injection DB9 to SMA. Not preventing us from closing the chamber, but weird.
  • Before swapping the bad cable, I was changing the amplitude and frequency of the injection, and it peaked at around 4MHz, which is OK, but at around that frequency the signal came from SMA (for PDB because that's the only thing I was looking at) regardless of the relay status. Both ON, both OFF, A=OFF/B=ON, A=ON/B=OFF, no matter what the signal from PDB was big.
  • At 200kHz, though, the signal goes on/off according to the relay. That was the case before the cable swap and that's still the case after the cable swap. There is apparently an unknown coupling that is highly frequency dependent. Rich thinks that this might be coupling via bias. I didn't check it at 4MHz after swapping the cable, but I expect that it's still there.
• Checked the BIAS MON output voltage, and it makes sense for both PDs.
  • I measured it at the DB9 of the back of the whitening chassis. When the preamp was ON, pin1 to pin 6 (PDA) and pin 2 to pin 7 (PDB) were both -0.987V.
  • According to Dean, the bias voltage is 5V. But there's a resistive divider of 1/11 (at DC) in the preamp, and then differential driver, so the overall gain is 2/11. And there's a negative gain of -1 in the whitening. Overall, 5*2/11*(-1)=-0.909V is expected, I'm getting -0.987V.

All in all, the most important thing is that things in chamber are connected as intended, which seems to be the case.

We're ready to put the door on.

Another update.

It turns out that pin 1 on the preamp box DB25 was intentionally tied to the signal ground (T2100403), but the schematic wasn't updated. One mystery solved.

Another small thing is that, in one of my previous alogs above (alog 61559) I wrote about the cable with PEEK shells at both ends: "We measured the continuity between the preamp and the aluminum collar that is seemingly bundling the shielding on the PEEK DB25 connector, and they were isolated." That bothered me because, if that collar thing is tied to the shield, and if the shield is connected to pin 1, and pin 1 to signal ground, and signal ground to the preamp case, why are they isolated? So before installing the old cable back I quickly made another continuity test, and everything made sense. See attached. The "collar thing" I refered to in my alog is the band that goes over the insulation mesh and doesn't necessarily touch the shield. But the metal tube that sticks out of the PEEK definitely touches the shield. While the PEEK DB25 connector was still on the preamp, there was continuity between the tube and the preamp body.

Here's my diagram to explain/guess the current status.

Things to check (see attached):

• Are we modifying the OMC DCPD in-air cable or the chassis?
• Was anything done to QPD chassis? The in-air cable for QPD seems to be unmodified.
• How PZT cable was modified? Apparently it was modified but I cannot tell how from the external continuity test.

Sigh.

Rich was confused by my original alog 61549 above (the very first step to measure continuity between pin 13 and everything on DCPD feedthrough) because I wrote a confusing half-information. Basically what I did was to measure in-chamber connection of signal GND and shield GND of DCPD through FT.

(What I actually did was to remove the DCPD cable from FT, put a breakout board on the the DCPD cable, stuck it back to FT, and then checked, among other things, the connection from pin13 to everything else , while all three cables were still disconnected from the rack. )