kyle.ryan@LIGO.ORG - posted 15:57, Thursday 09 January 2014 (9183)
Soft-cycled GV7 for craning of HAM4 doors + misc.
Started in 'screen' on h0epics2.
This morning, Gerardo and I prepped the ITM03 test mass (side S3) and it's associated ear (s/n 95) for ear bonding. He then silicate bonded the ear to S3. After the usual 2 hours of intermittant inspection, he remeasured and confirmed that the position was within spec. Tomorrow morning we plan to do the S4 side ear bonding.
Rolf, Jeff, Ben, Dave
Rolf added the susHWWD part to the h1susitmy model, which permits the front end to remotely reset the hardware watchdog and set the operational parameters (trip levels and time-to-trip). Communication is acheived via a single binary output channel from SUS ITMY.
We tested the binary I/O communication lines in both directions. The readback was tested by removing one or both of the 37pin connectors on the HWWD (from the Satellite Amp monitor port) which simulates a loss of LED current. The readback was not initially received by the model, but started working after the cable was reseated. We tested the control function by remotely resetting the watchdog and changing the time-to-trip from the defaut of 20mins to 2mins.
New MEDM screens have been built. An alarm handler was also built.
The running h1susitmy model has been built against the trunk (as of 9th January) to permit the new HWWD part to be used. I have subsequently reset the rtscore/release pointer back to RCG2.8.2
After lunch Jeff will put the ITMY SUS and SEI through a series of tests to verify the watchdog functions correctly.
Alexa, Kiwamu, Koji, Stefan Today we started with centering the TMS pointing on the ITM using the baffle diodes (with ITMX misaligned) : Baffle PD1: H1:AOS-ITMX_BAFFLEPD_1_POWER maximized: H1:SUS-TMSX_M1_OPTICALIGN_P_OFFSET 105,5 H1:SUS-TMSX_M1_OPTICALIGN_Y_OFFSET -208.5 Baffle PD4: H1:AOS-ITMX_BAFFLEPD_3_POWER maximized: H1:SUS-TMSX_M1_OPTICALIGN_P_OFFSET 174.0 H1:SUS-TMSX_M1_OPTICALIGN_Y_OFFSET -270.0 Center position: 1:SUS-TMSX_M1_OPTICALIGN_P_OFFSET 139.75 H1:SUS-TMSX_M1_OPTICALIGN_Y_OFFSET -239.25 Next we established the visibility into HAM1, i.e. the range of PR3 alignments that result in the x-arm green beam reaching HAM1. The orthogonal slider was left at the initial position. The numbers below are PR3 alignment numbers (i.e. H1:SUS-PR3_M1_OPTICALIGN_P_OFFSET and H1:SUS-PR3_M1_OPTICALIGN_Y_OFFSET): initial position: pit -263.8, yaw -272 disappear up pit -189 disappear down pit -330 (spot visible in ham 2, glow starting at -320) disappear north yaw -188 disappear south yaw -320 on swiss cheese baffle down pit -353 on swiss cheese baffle up pit -176 on swiss cheese baffle south yaw -366 (glow on cam from tower yaw -320) on swiss cheese baffle north yaw -178 center visible: pit -254.5 yaw -254 (using -320 instead of -330 for pit down limit) center swiss cheese baffle: pit -264.5 yaw -272
Nominal running of GV20 motor for 4 minutes and 16 seconds did not result in effective isolation of X-end (see attached) Additional running of motor was required to get full O-ring contact as follows: Ran motor 4 minutes 16 seconds - stop - accelerate to 500 rpm - stop - accelerate to 300 rpm
This entry pertains to activity on 1/7/2014
JimW FabriceM HughR
Jim identified a problem that I've been harping on for months (Input matrix errors.) While for me it was improper calibration into the cartesian basis, it was also affecting the ISI & HEPI commissioning. Once we rebuilt the matrices corrrectly, the ISI commissioning we did before the holiday allowed us to bring the ITMx ISI on line with 100mHz blends (on the translational dofs) and 750s elsewhere. This greatly lowered the ITMx motion hopefully making IFO commissioning easier; especially with the notch Fabrice added to the X&Y dofs to help reduce Optic Pitching. ITMX HEPI still remains uncommissioned but that should fall away soon.
The ETMx was brought on too the help the ALS crew with similar blends and this too helped quiet the cavity. However, just as in the past, the watchdogs tripped on the T240 after a couple hours. Still looking at that one.
Here is 30 minutes second trend of the T240 glitch that tripped the ISI. It is like most all the others, pretty much just out of the blue. BTW, I'm pretty sure the Isolation was just level1 as well.
New safe.snaps were generated this morning for ITMX ISI. I also checked the same matrices at ETMX and they at least agree with what is currently supposed to be there.
Kiwamu, Sheila
Today we saw that the power on the als refl PD drops when we misalign the ITM. (ITM misalinged, 13000 counts, ITM alinged, fringing up to 18000 counts). This would suggest that the ETM has a low reflectivity for green.
We went out to the end station and measured 34mW going into the chamber, and by misaligning the ETM measured 13mW returning. According the Keita the TMC efficiency is 90% each way, if this is correct for the polarization we are injecting the ETM reflectivity is 47%. (or if we assume the etm transmitts 24% of the green, it could mean that the TMS efficiency one way is 70%) We also measured 10mW rejected by the Faraday and 1.4mW in the hartman path.
The fringes that we saw in the reflected PD DC output and the signal out of the demod seem consistent with a verry low cavity finesse.
It may be worth checking the polarization of the light leaving the table. Keita noted that there was more green light in the IR QPD path this time than in end Y: alog 8705 However, the polarization should only change the efficiency of the TMS, it would not explain the signal on the refl PD.
According to the coating report that Betsy sent the ETM transmission at 532 is 24% https://dcc.ligo.org/DocDB/0059/C1103233/002/Coating%20Characterization%20Report_ETM08.pdf
Alexa and stefan measured the polarization at the bottom of the periscope to be spol: 8558
Looking at the data from last night more carefully, with the ITM misaligned we get 12675 counts on refl B LF, the top of the fringe is around 17500 counts and the bottom of the fringe is 15300 counts. The attached plot shows the ratio of the top of the fringe to the prompt reflection, and the bottom of the fringe to the prompt reflection, with predictions assuming the ITM R=99% and no other losses. Both the top and the bottom of the fringe are consistent with the ETM R=73%, not so different from the 76% on the nebula page. The other ETMs on the nebula page (including livingston) are similar. It seems like at least several of our ETMs are out of spec (spec was 3%-15%) , the cavity fringing is not terribly far off from what we would expect given the measurements from LMA. Also, estimating this using the fringes means that our measurement is sensitive to mode mismatch and misalingments, and we don't know how our mode matching or alingment are right now.
If the ITM has a reflectivity of 99% (we don't have a measurement of this, but that was intended) and the ETM 76%, we would get a cavity finesse of 22, and the reflected power on resonance would be about 93% of the reflected power on resonance. The cavity pole would be at 840 Hz. This will probably mean that the performance of our PDH lock is worse, but this may not be a serious problem since the noise of the PDH lock wasn't limiting us in HIFO Y.
The efficency of the TMS table still seems to be worse than expected. Assuming the ETM has R=76% , our measurement of the reflection off the ETM indicates that we loose 50% of our power in the TMS, or 70% of the power on each pass through the TMS. At end Y (alog 3077) Keita measured 61% of the injected green light returning to the table, so this was at worse 78% efficiency each pass through the TMS.
Summary:
the ETM is out of spec, but this was reported by LMA in the coating documentation, and pretty much explains the low finesse cavity we have.
TMS efficiency is worse at end X than at end Y for reasons we do not know yet.
https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=3077
Look at the above alog entry from OAT and its followup below that.
I was wrong about silver mirror reflectivity (I looked at Thorlabs, Newport and CVI catalog and they're more like 97% or less rather than 98%), and this 1% difference makes a huge impact because we have four mirrors double path so it's 8 reflections.
Anyway, all included, TMS itself is supposed to have 67% double path efficiency.
During OAT the double path efficiency was supposed to be 66% including the 99% ETM but was measured to be 61%, but note a large uncertainty regarding the protected silver mirror coating. If we put this discrepancy in the silver coating reflectivity, it is 96% per silver coating.
One thing to note is that we changed the coating vendor of some (but not all) of the TMS optics, we now use Newport protected silver mirrors for F1 instead of Edmund for all but H1 TMSY.
J. Kissel, R. Bork, B. Abbott, D. Barker, F. Clara, A. Sevigny Satisfied with a few days of debugging and testing a single chassis of the new Hardware Watchdog System on the H1 DAQ Test Stand, we have installed a similar chassis in the SUS BSC123 rack, and hooked it up to the H1 SUS ITMY M0 sensors. We have deliberately *not* connected any of the trip signals (either to the SEI or SUS actuators) while we characterize the RMS trigger system. In addition to the hardware, we've installed temporary configurations of the h1iopsusb123 and h1susitmy models, in which - h1iopsusb123: Unfortunately, neither analog or digital readbacks of the RMS calculated by the board are exposed in any way. Instead, we've installed a mock RMS system to replicate the RMS logic that is performed inside the hardware watchdog itself, using the raw ADC values as inputs. Though the changes have been made to the model, it has not yet been compiled, installed or restarted, since the green team were still actively searching for beams and needed ITMX up all afternoon. Hopefully we can install tomorrow morning. - h1susitmy: the status bits of the hardware watchdog *are* digitized via a binary input chassis, and fed into spare channels of the already present Contec BIO card. Similarly, the remote control reset bit are spit out of the same card, through a binary output chassis, to the SUS hardware watchdog. As such, for the status bits, we needed to add a few new connections inside the BIO_DECODE block, which are turned into a bitword and fed into a new EPICs readback channel, H1:SUS-ITMY_HWWD_STATUS. Once we get the readback installed, we'll run the ITMY chamber in various configurations of isolation and excitation, to determine a good value for the RMS threshold. The test cases we plan to use are: - Ambient (no actuation from any layer) - SUS Undamped, ISI Damped, HEPI Position Loops - SUS Undamped, ISI Damped, HEPI Position Loops (With large alignment offset) - SUS Damped, ISI Damped, HEPI Floating - SUS Damped, ISI Damped, HEPI Driven (Normal level TFs) - SUS Driven (LOUD White Noise TF), ISI Damped, HEPI Floating - SUS Damped (With large alignment offset), ISI Damped, HEPI Floating - SUS Undamped, ISI Driven (LOUD White Noise TF), HEPI Floating - SUS Undamped, ISI Floating, HEPI Driven (LOUD White Noist TF) These should give us a good feel for the most quiet and most loud situations for the M0 OSEMs. Based on these results, we'll make an assessment of where we want to set the threshold.
J. Kissel, J. Warner After ensuring all seismic isolation and suspension systems were safely ramped down, I re-compiled, re-installed the h1iopsusb123 front end model. After which, I killed all processes running on that front end, h1susitmx, h1susbs, and h1susitmy, restarted the h1iopsusb123 process, and restarted all the suspension processes. I then finally restored damping on SUS ITMX, ITMY, and BS, and restored the alignments of ITMX and BS to what they had been set before I got started. Both temporary configurations of the h1susitmy.mdl and h1siopsusb123.mdl have been committed to the userapps repo.
Alexa, Sheila, Daniel, Keita, Kiwamu, Stefan
We tried to lock the x-arm today, but we stumbled over a couple of mysteries.
- For the initial alignment we first pointed TMS onto ITM baffle PD1, and found TMS PIT=107, YAW=-207. This was -2urad, -4urad less in PIT, YAW than the previous alignment measurements (alog 9126), so we subtracted these values from the subsequent ones. We repeated this for the ITM using the ETM baffle PD. The ITM was PIT=57, YAW=-78.
- With both optics aligned we get about 18000 counts on REFL_B_LF. Surprisingly though, misalignment of ITMX (yes ITM) dropped this to 12000 counts
- We do get a PDH signal of about 5V out of the Imon demod. However, with the cavity just slightly misaligned, we still get a sinusoidal PDH signal of about 2V. We are not sure why, but are afraid it could be some scattered light.
- The PDH loop does lock temporarily, but the arm motion was slightly too large -- the sus pro-team is working on this.
- Next we misaligned PR3 in yaw by -150urad (from -272urad to -422urad). This put the straight through beam nicely on the Swiss cheese baffle.
- We also noticed a ghost beam to the right of the straight shot. It moves with ITMX and CPX, which pretty much nails down what it is.
- There is also a 2nd copy of the straight though and ghost beam visible to the left of the POP hole in the Swiss cheese baffle.
Here are Keita's rough alignment numbers, as well as two screen shots of the Swiss cheese baffle.
The fiber pulling machine upper clamp was binding up when we advance the trolly to the end of its stroke after the fiber is pulled. This advanced travel gives you the clearance to allow you to get the fiber assembly out after the fiber is pulled. This issue has been happening more and more recently and causes damage and gross mis-alignments of the translation stages and goniometer mounts. The mounting hardware fit to the fiber clamps was tightly constrained adding to the problem. There is a bit too much wiggle to the mounting stages that were allowing further issues. Bottom line, I loosened up some of the fit tolerances and coaligned the top and bottom mounts followed up by a beam realignment. I will revisit this in the AM and pull a test fiber. Hopefully we will not have any further issues. I hope to install more robust stages in the near future.
Attached is a picture of the top bracket of what would be a fiber holder stuck in the top stage of the fiber puller, when in fact it should have released (hence Doug's hand prepared to catch it during the photo shoot!). This sticky mating is what Doug fixed. Currently Travis and Giles are pulling fibers.
Now that the alignment of the cavities is roughly good, I aligned the ITMX, PR3, HAM2, HAM3, and ETMX optical levers to their respective optics/HAMs. It's relevant to note that the connector of the controller for the translation stages broke, we were able to re-solder the wires but the connection is very fragile. I'll just have to be careful when moving it around but it'd be best to have a system where we're not constantly unplugging and plugging in delicate wires every time we want to re-align the pointing.
