First, yesterday we had a WFS interface cable that would short circuit -100V bias to the ground. A new cable was put in (or maybe the old one was fixed), and today I restarted WFSY-ing.

I immediately noticed that WFS whitening binary IO readback doesn't agree with the output I intended to send, for all four BIO cables.

Upon looking at the whitening chassis, I've noticed that +15V power indicator LED was really dim for WFSA whitening (first picture), but when I removed the second BIO cable from the WFSA whitening chassis the LED got bright again (second picture). But that didn't solve the readback problem for that specific cable.

I and Daniel started disconnecting cables from the binary IO output chassis, and immediately the readback got sane again, and we confirmed that BIO chassis is working fine by connecting a BIO tester. As soon as we connected the cables (that connect the BIO chassis to the field rack whitening interfaces) back to the BIO chassis (the othe end of the cable is still disconnected) the readback got mad again, and the madness is somehow dependent on how we strained and wiggled the cables.

We opened the shell of the cables, and were disappointed to find that it seemed to be really easy for the exposed conductor of the flat cable to be pressed against the shell to cause short circuit (third picture).

By removing all 8 shells (four cables both ends), and then connecting the cable back to the BIO output chassis and the whitening chassis, we were able to eliminate the readback problems for most (but not all) of the requested states. This means that most of the short circuits were gone but there are still some that is not related to the short circuiting to the shell.

And the +15V LED of WFSA whitening isn't dim any more.

Anyway, for the moment, short-circuiting between high and low is a problem but not for high-high and low-low, so we decided to find a seemingly good state where the readback don't complain. For both of the WFSs, this is +33dB whitening gain, first two filters on, third filter off.

And then I connected H1:LSC-Y_EXTRA_AO_1_EXC output to the B input of the PDH CM board, locked Y arm on green, injected a sine wave to the CM board and started measuring the demod phase of WFSs. And I noticed that there are huge differences in signal levels for quadrants, and for some quadrants I couldn't make a decent job of minimizing the injected signal in one demod quadrature.

I got suspicious about whitening and measured the dark noise coming into the ADC (4th attachment, bottom), and sure enough, whitening gains seem to be all over the place, and it seems like two channels for WFSB have only one stage of whitening filter engaged. I then found that the readback wouldn't complain at 45dB too, so I repeated the same thing with 45dB whitening gain (top), and it got better for WFSA but not for WFSB. Maybe there's still some cable-related problem though I don't know how, or maybe this is the whitening chassis problem.

The next step would be to go to the floor and look at the whitening chassis. It would also make sense to see if the whitening gain for X arm is OK.