stefan.ballmer@LIGO.ORG - posted 07:27, Thursday 27 February 2014 (10368)
ISI ITMX ST1 blend filters messed up
Couldn't turn on ISI ITMX...
Images attached to this report
Couldn't turn on ISI ITMX...
I added an X arm guardian today, and tested as much of it as I could without having an arm cavity to lock. This also involved some changes to the ALSEndStateMachine Library in Beckhoff (for error reporting), and changesto the guardian overview and ALS Main medm screens.
ITMY TF will be running overnight on opsws8. Started at around 20:45
Measurement completed and lasted ~10hours.
In prepartation for tomorrows guardian deployment on ISI HAM2/3, I checked the configurations of the filter modules in the HAM2/3 ISIs, to make sure they are configured as needed for the guardian. I found a couple of discrepencies that I've tried to sort out.
The BLEND configuration is not currently touched by the guardian code, so we're not working about their configuration for the moment.
The configuration of the DAMP filter modules seems to be as expected (FM1 loaded only with 'Damp').
The ISO modules were a different story. ISI_HAMX expects the following filter banks to be populated for the vaious isolation levels:
HAM2/3 has nothing in FM3/5/6/7. I therefore added unity gain filters into those banks, and gave them the appropriate names.
The Boost_* filters also didn't have the 5 second ramp times that we decided we needed during the HAM4 testing. I added those ramps as well.
I have saved the filter files, but I have not yet loaded them into HAM2/3. I will load them when I start my tests tomorrow morning.
There is a typo in the HIGH - HIGH uses FM 4,5,6, possibly 7, and boosts with FM8 (using the 1-10 numbering)
Below are a series of reference alignment values that allow hitting the various BAFFLE PDs: # TMS pointing straight shot # ITMX PD1 caput H1:SUS-ETMX_M0_OPTICALIGN_P_OFFSET 0 caput H1:SUS-ETMX_M0_OPTICALIGN_Y_OFFSET 0 caput H1:SUS-TMSX_M1_OPTICALIGN_P_OFFSET 203.3 caput H1:SUS-TMSX_M1_OPTICALIGN_Y_OFFSET -228.1 # TMS pointing straight shot # ITMX PD3 (PD4) caput H1:SUS-ETMX_M0_OPTICALIGN_P_OFFSET 0 caput H1:SUS-ETMX_M0_OPTICALIGN_Y_OFFSET 0 caput H1:SUS-TMSX_M1_OPTICALIGN_P_OFFSET 272.7 caput H1:SUS-TMSX_M1_OPTICALIGN_Y_OFFSET -289.7 # ITM pointing # ETMX PD1 caput H1:SUS-ETMX_M0_OPTICALIGN_P_OFFSET 0 caput H1:SUS-ETMX_M0_OPTICALIGN_Y_OFFSET 0 caput H1:SUS-ITMX_M0_OPTICALIGN_P_OFFSET 35.3 caput H1:SUS-ITMX_M0_OPTICALIGN_Y_OFFSET -69.2 caput H1:SUS-TMSX_M1_OPTICALIGN_P_OFFSET 238.0 caput H1:SUS-TMSX_M1_OPTICALIGN_Y_OFFSET -258.9 # ITM pointing # ETMX PD4 caput H1:SUS-ETMX_M0_OPTICALIGN_P_OFFSET 0 caput H1:SUS-ETMX_M0_OPTICALIGN_Y_OFFSET 0 caput H1:SUS-ITMX_M0_OPTICALIGN_P_OFFSET 67.4 caput H1:SUS-ITMX_M0_OPTICALIGN_Y_OFFSET -36.3 caput H1:SUS-TMSX_M1_OPTICALIGN_P_OFFSET 238.0 caput H1:SUS-TMSX_M1_OPTICALIGN_Y_OFFSET -258.9 # ETM pointing # ITMX PD1 caput H1:SUS-ETMX_M0_OPTICALIGN_P_OFFSET 231.4 caput H1:SUS-ETMX_M0_OPTICALIGN_Y_OFFSET 90.8 caput H1:SUS-ITMX_M0_OPTICALIGN_P_OFFSET 51.35 caput H1:SUS-ITMX_M0_OPTICALIGN_Y_OFFSET -52.75 caput H1:SUS-TMSX_M1_OPTICALIGN_P_OFFSET 238.0 caput H1:SUS-TMSX_M1_OPTICALIGN_Y_OFFSET -258.9 # ETM pointing # ITMX PD3 (PD4) caput H1:SUS-ETMX_M0_OPTICALIGN_P_OFFSET 255.4 caput H1:SUS-ETMX_M0_OPTICALIGN_Y_OFFSET 60.7 caput H1:SUS-ITMX_M0_OPTICALIGN_P_OFFSET 51.35 caput H1:SUS-ITMX_M0_OPTICALIGN_Y_OFFSET -52.75 caput H1:SUS-TMSX_M1_OPTICALIGN_P_OFFSET 238.0 caput H1:SUS-TMSX_M1_OPTICALIGN_Y_OFFSET -258.9
Fabrice and Dave.
We completed the build, install and restart of the BSC ISI systems using the older version of the WD code (minus HWWD and reset inputs).
h1isi[bs,itmx,itmy,etmx,etmy] were restarted. Fabrice verified the WD code is running correctly.
I verified the safe.snap restores went without any errors. I cleared the IPC errors caused by the restarts. No DAQ restart is required.
9:00 Kris to EX, back at noon
9:30 HughR, MitchR to HAM4, done at 10:30
10:00 Kiwamu PR2 Model changes, multiple DAQ restarts
10:00 ETMX ISI & ITMX ISI and HPI found to be down. ETMX came back up, ITMX HPI would not. ITMX HPI has been down since yesterday's boot fest. Looks like the HPI IPS targets got lost during a burt restore. This was not recovered until the afternoon. It required finding an older burt file, then redoing the safe.snap.
10:15 CoreyG to EX TMS lab. Travis to LVEA SUS test stand.
10:30 HughR, MitchR to EY for cabling. Done at 12
11:00 I went to HAM4 for a fruitless troubleshooting mission. Problems fix themselves.
11:15 DaveB restarting BS-ISI.
12:30 Cleaning crew to EY. Done 15:00
13:00 Ed and Evan to IOT2R. Transitioning to laser hazard.
14:00 CraigC, RIckS, ChrisS to H2 PSL.
14:00 JeffB, AndresR to HAMs 2,4&5 hunting for install hardware. Done 15:00
Replaced the incorrect feed thrus with the needed 3-Pin power style. Then completed the cable attachment, cable strain relief and finally the easier half of the Dial Indicators. Almost that is--I was short three 5" bolts to strain relieve some SUS cables on the South side. I must have cut myself a little to close wrt inventory. But generally, all the really needed dirty work is done.
The Apollo crew got Test Stand, extra Spiral stair case, 3P spreader, and a few other things from the VEA.
The cleaning crew jumped right on as soon as I was out of the way--a few dust alarms with that work.
Crud--I forgot the CPS sync cables...Will be routed on the South & West side--away from the open door.
Fabrice's sensor correction for ETMX stage 2 was just turned back on (in the last few minutes). I did this by turning on the output of H1:ISI-ETMX_ST2_SENSCOR_X_MATCH and setting the gain H1:ISI-ETMX_ST2_SENSCOR_X_MATCH_GAIN=1.15. This helped the pitch motion.
In the attached screenshot you can see that it clearly helped to reduce the pitch motion.
The 17 V power to the ALS fiber distribution box was loose, I'm not sure if this happened while we were searching for the cause of the IMC wandering peak yesterday. At least the TwinCAT PLL autolocker code gave us a usefull error message, and the problem was easy to track down. They are tightened now.
Deleted 4 channels, inserted 1,405 for a current total of 123,768.
Mitchell and I installed the TMDs on the HAM ISI Blade Springs similar to as Hugo and I did on WHAM3. Mitchell has the serial numbers. There were also a couple feedthrus that needed to have the in-vac cable attachment screws secured. Mitchell did this.
Serial numbers for the Blade Spring Tuned Mass Dampers (D0900703) are 025, 026, 027.
WP4464. Brian, Fabrice and Dave.
Following the restarts of all the recompiled front ends yesterday Fabrice noticed that the BSC ISI watchdogs were being reset a few seconds after tripping.
We tracked the problem down to a modified common C-code file which was installed on the HAM ISI systems but not on the BSC ISI.
The C-code is isi/common/src/ISIWD_GPS.c. Its number of inputs was increased to accomodate the hardware watchdog system. The HAM ISI common model (isihammaster.mdl) was updated to the new code but the BSC common model (isi2stagemaster.mdl) was not.
The solution to get the BSC ISI systems operational today is to make a copy of the older version of ISIWD_GPS.c called ISIWD_GPS_OLD.c and change the isi2stagemaster.mdl to use this file (with the smaller number of inputs) in both stages. Brian and Fabrice will now work on isi2stagemaster to use the new code. For now LHO's isi2stagemaster.mdl will remain a locally modified not submitted to SVN, and the ISIWD_GPS_OLD.c will remain a local file not added to the repository.
We built,installed and restarted h1isibs and Fabrice tested the watchdog, it is functioning correctly. We will restart the rest of the BSC ISI this afternoon when the opportunity arises.
As mentioned previously, we have been testing the new ISI_HAMX guardian infrastructure on HAM4, in preparation for deployment on HAMs 2 and 3.
At this point things are looking good to go. We've interated with the Stanford SEI team to iron out all the configuration and behavioral issues. We made one last change to the DEISOLATION state to smooth out the isolation gain ramp down (userapps svn r7300), and it seems to be working fine. We've now gone through many cycles of changing to all the full isolation levels, and everything looks good. Therefore I think we're ready to move on to HAM2 and HAM3:
We will seize this target of opportunity tomorrow morning while the SUS and ISC teams are running suspension actuator diagnolization measurements on ITMX. They have confirmed that theeir measurements will likely run through the early afternoon, so we will start early to make sure we have plenty of time to test the code. The basic testing procedure will be:
If the behavior looks good, we will leave the ISI_HAM2 and ISI_HAM3 guardians running.
Yuta, Kiwamu,
We continued working on the alignment automation this morning. Today's emphasis was on the dither alignment of PRY. It is now successfully coded in the LSC guardian.
The next step is to make the PRX and PRY dithering systems sequential so that we can run them by pressing only a single button.
Dither alignment:
Currently we use Stefan's IAL to dither ITMY. The POP18 signal is temporarily routed to the ADC and demodulated in IAL. Since the middle masses didn't give us a big excitation in the angle, we decided to use the top mass to excite the angular motion at around 2 Hz for pitch and yaw. The demodulated signal is also fed back to the top mass. This worked pretty well.
The guardian was modified to include the dither alignment. We checked the functionality a couple of times by going through the initial state to the dither state. It is functional although we are not 100% confident with our guardian code.
I zero-ed out the optical levers today, HAM3 looked especially bad but this was because the QPD cable was unplugged from the whitening chassis. Please don't do this.
Daniel, Keita, Stefan For yesterday's IMC mystery we wanted to go back and look at the HAM2 optical lever to assess HAM2 rotation. Turns out it didn't move, but HAM2 is doing something funny over a 2day or so period.
Written by Yuta
This morning, I could dither PR2 and PRM from ASC-ADS, but I can't now.
For example, even if I put on H1:ASC-ADS_PRM_PIT_OSC, the signal doesn't show up at H1:SUS-PRM_M3_ISCINF_P_IN1.
Since it did this morning(6AMish~10AMish), I suspect this is related to the boot festival today.
If I lookup the simulink diagram, h1asc.mdl has H1:ASC_SUS_PRM_PIT_DITHER, but h1prm doesn't.
[PRMI dither alignent work this morning]
We want the alignment automation to speed up the PRMI commissioning.
I checked the guardian script for PRX dither alignment Evan and I made yesterday. It now works for PR2 dither alignment. PRX_ALIGN state closes PR2 (and PRM) dither alignment loop, wait for a while, and offloads the feedback offset (e.g. H1:SUS-PR2_M1_LOCK_P_OUT) to the alignment slider values (e.g. H1:SUS-PR2_M1_OPTICALIGN_P_OFFSET).
For PRM, I had to lower the excitation frequency for pitch/yaw from 14/11.5 Hz to 5.4/7.2 Hz, and excite not only M3 but also M2 to get the error signal. I haven't checked fully yet, but the guardian script should also work for PRM dither alignment.
Next work is to do BS(or ITMY) alignment using PRY.
Just to make it clear:
The excitation was not active because of a change in the ASC model rather than the SUS models. I was in the middle of splitting the ASC paths into the low frequency feedback and dither paths. Currently only PR2 is capable of having the split dither signal.
This is from last week.
We briefly measured the distance between optics and such, and used nanoscan to measure beam width/profile at various places though they were very very ugly.
See attached scribbling for layout (all dimensions in inches). Mirror names are arbitrary.
Measurement points are indicated by alphabet from A to I. Measurement points with prime (G'-I') means that the first lens in WFSB path (LWFSB1) was removed during the measurement.
The rest of the attachements are the nanoscan result. WFSB1.jpg-WFSB4.jpg=measurement point A-D. WFSA1.jpg and WFSA2.jpg=measurement point E and F. WFSC1.jpg-WFSC3.jpg=measurement point G', H' and I'.
The lenses as installed are:
LWFSA1 +250mm
LWFSA2 -75mm
LWFSB1 +250mm
LWFSB2 -100mm
This is an adaptation of the mode-matching solution by Bram Slagmolen used for the OAT WFS in D1100607-v10, and shown in the attached plots. This is not the simplest possible solution, but it has the very desirable quality of having great flexibility with WFS placement while keeping 90 (+/-10) degree Guoy phase separation.
It is important to note that the nominal mode-matching solution used to generate this layout did not account for the beam quality issues. In theory, the input beam should have been small enough that the beam expansion properties of the layout would have been necessary to get a usable beam size on the WFS. If this ends up being an issue, the mode-matching can be modified to reflect reality.
WFSa1.jpg in the original entry was bogus (it was a copy of WFSb1). Attached is the correct one.
