keita.kawabe@LIGO.ORG - posted 22:32, Wednesday 16 July 2014 - last comment - 08:48, Thursday 17 July 2014(12794)
TMSX IR QPD alignment: Success (Corey, Dan, Sheila, Keita)
Two things that were important:
- Better mode matching to the telescope so that the IR beam becomes smaller than the QPDs.
- "Fixed" a crappy binary output cable that goes to the QPDB whitening chassis.
We still need to fix the ground loop problem of TMSX BOSEM, and we still need to look at the IR beam position on the viewport path, but we're basically done with the alignment.
"Fixing" the crappy cable.
Though we were able to see our IR beam on both of the QPDs, at first I was never able to center QPDB while QPDA was fine. After much trial and error, we used a flash light on QPDB and found out that the output from segment 4 was at least 100 times bigger than the other segments.
We swapped the in-air IR QPD cable with the green QPD cable, and suddenly things look more reasonable.
I looked at the whitening chassis of the IR QPDB, and the DB25 cable for BIO looked crooked on the front panel. I seated it properly, but the two screws that attach the connector to the chassis (let's call these the connector screws) were loose, it was sort of difficult to tighten both at the same time. The connector screws themselves looked crooked relative to the connector shell.
Then I disconnected it again, had a look, and found that the shell was half broken (attached, this is as discovered, but after disconeccting from the chassis).
Two halves of the connector shell are put together by two screws that go into tapped holes inside the shell (let's call these the shell screws), but the thread of one of the tapped holes is stripped off. The other hole was fine, but the screw was loose, therefore the connector shell was open on one side, making the connector and the connector screws misaligned with each other.
Anyway, crooked connector meant that only the whitening gain of segment 4 was controlled by the Beckhoff binary IO module while other segments were fixed to some low gain (probably zero dB). We needed to use the highest whitening gain during the HIFO test as well as today, so there should have been a huge gain imbalance there. This is likely the reason why people couldn't "center" the IR beam on the QPD sled.
My dirty "fix" was to simply tighten the working shell connector until everything is held together, and then connect the thing back to the chassis. With nothing crooked, the connector screws went in without much fuss.
But this is not really a fix, I think the shell should be replaced.
I don't remember seeing a huge imbalance on QPDB segments when we initially checked out the electronics for HIFO-X. Maybe the connector was seated properly but the connector screws were not fully screwed in because it was difficult, and later the strain on the cable eventually made the connector crooked.
We checked particle counts in the middle of the work, we had
90 0. 3 micron.
30 0.5 micron