[Paul, Cheryl, Giacomo]
Yesterday (Friday) evening the reading from WFS_A and WFS_B didn't quite seem to make sense, despite them having been aligned a few hours before by Cheryl. Paul and I went back to the IO table to do some more systematic testing. Bottom line is that quadrant 3 of WFS_A seems defective. It has a large offset and possibly a reduced sensitivity (although the fact that it has some sensitivity might be partially good news).
Note that this is the same unit that showed the anomalous heating described in entry 5131. The tests performed in that occasion were too superficial to notice the defect.
Although it might be initially possible to still use this sensor for some ASC test (by compensating for the offset and sensitivity in software), we will eventually need to remove WFS_A from the table and test it. To our knowledge, no spare is readily available.
This is the set of measurements we performed (values are in adc counts). We first (column 2) measured the dark current (note that for some reason the single quadrants have a sign inversion wrt the sum), then centered the beam on the WFS witht he IMC locked (column 3), then intentionally unlocked the IMC (column 4), then realigned the beam again (column 5) and finally re-locked the IMC (column 6). Note that the change in reading between IMC locked/unlocked in not necessarily due to beam wondering, as the shape of the beam itself is very different between the unlocked (mostly gaussian) and locked (mostly junk) states.
| Dark | IMC Locked, centered | IMC Unlocked | IMC Unlocked, centerd | IMC locked | |
| WFSA_SUM | -1215 | 500 | 9700 | 9350 | 750 |
| WFSA_Q1 | -0.5 | -450 | -1200 | -2750 | -650 |
| WFSA_Q2 | -3.75 | -450 | -6600 | -3000 | -225 |
| WFSA_Q3 | 1220 | 875 | -1200 | -1000 | 925 |
| WFSA_Q4 | -0.3 | -400 | -600 | -2600 | -900 |
| WFSB_SUM | 21 | 2000 | 11250 | 11250 | 2250 |
| WFSB_Q1 | -10.75 | -500 | -2750 | -3350 | -550 |
| WFSB_Q2 | -5.5 | -500 | -5050 | -2300 | -300 |
| WFSB_Q3 | -3 | -475 | -2650 | -3050 | -650 |
| WFSB_Q4 | -1.75 | -475 | -725 | -2350 | -750 |
I measured the noise spectra of both DC and RF channels with no light on the sensors (IOT1 enclosure closed and lightpipe shutter closed). The defective channel doesn't show any noticeable excess noise (although some other wierdness is going on in the RF channels of WFS_B).
Also, we found the test document of the "defective" WFS_A unit (LIGO-S1203264-v2) and realized that it shows a dark current of quadrant 3 much higher (~45 mA) than all the others quadrants of both this and its twin (~1 mA). This is not quiet the factor ~1000 difference wrt the other quadrants that we observe now, but I guess a small offset somewhere could easily change that ratio.
The question at this point is if this offset has been deemed acceptable at some point, and we should just consider the unit as "normal"...