Displaying report 1-1 of 1.
Reports until 17:35, Tuesday 25 January 2022
H1 CDS (CDS, ISC, OpsInfo, SUS, SYS)
jeffrey.kissel@LIGO.ORG - posted 17:35, Tuesday 25 January 2022 - last comment - 09:49, Wednesday 26 January 2022(61444)
OMC DCPD Signal Chain Confusion: Nothing Broken, but Lost for Hours
J. Kissel for F. Clara, J. Driggers, K. Kawabe, himself, R. McCarthy, D. Sigg, B. Weaver

Continuing efforts in trying to get a power budget of the DCPDs, we lost the afternoon trying to investigate apparently non-functional DCPDs. Long story short, we all forgot that an early 2019 ECR E1900064 renders the OMC DCPD side of the Split Whitening Chassis completely non-functional when the whitening *gain* is set to anything besides 0 dB. 

The chassis and the entire signal chain has indeed now been confirmed functional when in the correct 0dB state, but we're all exhausted so we'll pick up the fight tomorrow.

%%%% Details %%%% (Aka long story long.)
OMC signal chain wiring diagram D1300502
OMC DCPD Split ISC Whitening Chassis D1300520

After lunch, and confirming that we can get a beam consistently in HAM6 (LHO:61430), and figuring out that the OMC cavity length PZT cable was unplugged at the chamber (LHO:61440), we still were not getting any cavity flashed on the OMC DCPDs when sweeping through the entire range of the PZT.

So we got lost in investigating why:
    - Sweeping through the OMC cavity PZT does not show any flashes. Stays "ADC noise 0.0" throughout.
    - Flashlight test on DCPDs reports "ADC Noise 0.0" nothing throughout.
    - Manually adjusting the OMC DCPD whitening gains (H1:OMC-DCPD_[A/B]_WHITEN_GAINSTEP) makes some electrical glitching, but does not change the signal level. In retrospect, here's where we lost. The "electrical glitching" was the signal chain coming alive at 0dB, but we didn't stay their long enough to figure it out.
    - Suspecting that the OMC ASC loop offsets were steering the beam input to the OMC somewhere bad, we confirming that OMC ASC Offsets -- in both H1:ASC-OMC_[A/B]_[PIT/YAW]_OFFSETs as well as H1:OMC-ASC_QPD_[A/B]_[PIT/YAW]_OFFSET -- are the same as when we last locked the OMC. 
    - Checked that other LSC channels on LSC IO Chassis ADC_O are functional (eg IMC_F and L monitors from the common mode board).
    - Richard / Fil / Daniel checked that AA chassis was on and alive by injecting with a known voltage source via DB9 breakout on the front in CER rack ISC C1 U9 Port 4.
    - Keita also confirms the same thing from the back of the Whitening chassis and confirmed the "to AA / ADC" chain was alive.
    - AS_C on the other card of the split ISC whitening chassis, and we see real signals on the AS_C QPD.
    - Power cycling the (split whitening) whitening chassis confirms that AS_C goes away. Causes a gltich on DCPD channels but turning chassis back on does not regain/recover any signal (i.e. we still get essentially ADC noise 0.0 off of the ADC).
    - Keita continues toward the chamber from the front of the whitening chassis (via DB25 breakout), and find that he cannot send signals through the chassis with a hand-held voltage source.
    - Daniel / Fil / Keita replace the chassis with a spare, and find the same non-functionality in retrospect, because the digital BIO request was still set to non-0 dB whitening gain
    - An hour or two more go by, but then, after pulling the BIO input from the chassis and confirming function with 0 dB request input, BOOM, signal chain comes alive.

Again -- all involved forgot that E1900064 renders the OMC DCPD side of the Split Whitening Chassis completely non-functional when the whitening *gain* is set to anything besides 0 dB. 

And so tomorrow, we'll make one (and hopefully more) further step(s) forward!
Happy Tuesday, er' body!
Comments related to this report
jeffrey.kissel@LIGO.ORG - 17:54, Tuesday 25 January 2022 (61445)ISC, Lockloss, OpsInfo, PSL, SUS, SYS
Attached is proof that the signal chains are alive, and I can sweep through some cavity flashes when the DCPD whitening chassis gains are set to 0dB.

First attachment shows each DCPD channel ["calibrated" into rough mA by the OMC-DCPD_[A/B] bank] during a sweep through the cavity's length PZT. Recall that -50 "V" requested from the slider means 0 [V] on the PZT.

There's no frequency dependent whitening on the DCPDs (when in O3-era observation, we usually had the 1st frequency-dependent stage ON, or all three stages ON).

There are some offsets in the DCPD banks, but are roundish numbers that are different from (probably O3 era) nominal so it's probably left from Keita setting the dark offsets by hand earlier today.

Power into the IMC is 2W; into PRM is 1.5 W.

OMC ASC QPD servo is ON.

Second attachment shows the state of all optics alignment sliders during this test. PRM and SRM are intentionally MISALIGNED.

Third attachment shows OMC Control screen, specifically to highlight how much the OMC ASC QPD loops are requesting to drive OM3 and OMC.

Test was at Jan 26 2022 01:43 UTC.
Images attached to this comment
jeffrey.kissel@LIGO.ORG - 09:05, Wednesday 26 January 2022 (61449)
See LHO aLOG 61447 for the greater context surrounding this aLOG.
jenne.driggers@LIGO.ORG - 09:49, Wednesday 26 January 2022 (61452)

To help us remember this tidbit for the future, I have added a reminder + warning light on the OMC DCPD screen (we had definitely been looking at that screen yesterday during the diagnosis, so hopefully it's obvious enough, without crowding the overview control screen).

First attachment: Screen showing the gentle reminder that will always be there

Second attachment: Screen showing the red warning light should the gain be changed to a non-zero value

Images attached to this comment
Displaying report 1-1 of 1.