Plots show the last two trips. All other plots look normal.
Last week we added the OPO and some of the optics to the VIP.
[Sheila, Terry, TVo, Jenne]
Beam go through hole, all the way to HAM6!
Attached are 3 screenshots, with titles indicating what they mean. The important one is the last, BeamOn_AS_C.png, which is the slider locations where we see some beam on AS_C. That means that we're making it through the output Faraday!
(The 2nd screenshot is where we think we're roughly centered through the OFI).
In Friday, Travis finished cleaning up the ETMY PUM horns by filing them to the proper length and then cleaning up the debris in the flats. I then thoroughly cleaned the lower main structure using a tiger vacuum and wipes. Gerardo aided us in weighing the new ETM test mass (ETM16) and installing it in the structure using the ergo arm.
Jason also worked on setting up IAS equipment to look at the suspension in the afternoon. His equipment is parked in the far side of the weld cleanroom away from the man-door entrance into the VEA and he has caution tape around some of it. Please limit activities around this equipment and watch your step near it if you need to get to the other side. We anticipate the equipment to be removed after next week.
We are in good shape to start the fiber welding process on this suspension next week.
Jenne, Sheila, TVo
We spent most of Friday using cameras to get the pointing of SR3 onto SR2 and back down to SRM correctly but it was much more difficult without a sensor to tell if you're actually getting the center of the baffle aperture. We think we were able to eventually get the right beam down to the SRM but it's not obvious how well centered you are on any of the optics (except SR3 since we have the input chain well aligned). We were hoping to get the optics close enough so that an excitation on all three optics would show up all the way back at LSC-POP but unfortunately this didn't work.
The next hopeful resolution is to get into HAM6 and try to align through the Faraday with an IR card.
Re-finding that place where we are hitting SRM and getting some beam back toward SR2, I realized that we hadn't put in any screenshots of the alignment for a while. So, here's an alignment that gets us back toward SR2, although I can adjust the SR mirrors quite a bit and still get this.
In the photo, the beam returning from SRM is at the bottom of the top TV. It's hitting something much lower than the baffle aperture (bottom of cage, or side of ISI or something?). On the lower TV, the smaller circle on the left is where we think the SRM is, and the larger circle on the right is the SR3 baffle aperture. The vertical lines are where we think the cages are for the suspensions, and the dotted line at the bottom is the eye-baffle between HAMs 4 and 5. The top TV has markings for the SR2 baffle, but it's hard to capture. The SRM return beam is directly below the center of the circle part of the SR2 baffle aperture.
J. Kissel, M. Pirello After finishing out the OFI, I moved on to H1SUSZM2. It doesn't yet work. Symptoms: - The OSEM sensor readbacks (the ADC inputs on the MEDM screen) are showing little-to-no signal (only ~1000 cts), however transfer functions still have plenty of coherence. - Although ZM2 is driven with an ISC configuration HAM-A driver, which supposedly has all binary control jumpered such that the low-pass filter is engaged, the pitch and yaw transfer functions show excess response at high frequency, which can be "calibrated out" by using exactly the same poles and zeros as the low pass filter (zpk(zeros,poles,gain,normalization) = zpk([0.9;211.883],[10;21],1,"n")). So it's as though the filter *has not* been jumpered. However, because there's plenty of coherence with the TFs, the Test/Coil enable jumper must be good. - In a quiescent state (damping loops OFF, but no excitation), the BOSEMs ASD shows noise elevated way above the noise floor. - Earlier in the day, when the external OFI's signal chain was unknowingly hooked up to the ZM2 in-vacuum chain, we saw lots of signal (i.e. the typical ~15000 cts), and transfer functions showed a normal frequency response (except for the poor basis conversion; see MAGENTA comments below). Attached are - Transfer Functions. BLACK is an old OM1 result, RED is as it stands now, MAGENTA is driven and readout with the OFI chain. Regarding MAGENTA -- Unfortunately, when driving a 4-OSEM tip tilt through the 3 OSEM OFI chain with the wrong coordinate transformation matrices means that all three "long" "pitch" and "yaw" drives were some wonky combination of only UL, LL, and UR. That's why you see a poorly diagonalized result, and you see all three DOF's modes in each TF. - Some MEDM screen shots relevant to the discussion - The DTT and Foton comparison for the "calibration" of the Pitch and Yaw TFs. Note, the TFs are *not* shown with the calibration, but trust me, when it's applied it makes the high frequency response of the RED match BLACK - The OSEMs ASD compared against the target / spec BOSEM noise. Because of the incorrect frequency response, I suspect the coil driver either hasn't been jumpered or it's just a bad board. Note, also, as is standard on the HAM-A driver, US field sat-amp systems, the sensor signals are plugged directly from the sat-amp into the AA chassis towards the ADC / IO chassis. That means that the sensor signals, which are polluted with noise are *not* being polluted by the coil driver. So maybe both are problematic? We know the in-vac suspension and electronics are OK, because they tested well with the OFI signal chain. We'll investigate further on Monday.
After discussing with Jeff, I pulled the ZM2 coil driver S1700079 and found that the jumpers were not installed correctly. I installed jumpers on the W2 links and P3 (Run/Acquire Controls and Indicators) matching this unit to the ones we worked on prior.
It looks like the ZM1, VOPO 1-4, VOPO 5-6 also need the same treatment as their indicator lights are not correct on the front of the chassis.
Sheila, Nutsinee
Here I attached the SHG conversion efficiency measurement as a function of 1064 input power (red dots) plotted against theoretical prediction (the equation came from Polzik and Kimble 1991 paper on frequency doubling) . ~75% conversion efficiency at 400mW input, 48% conversion efficiency at 100mW input. For the fit I tweaked a couple parameters, intracavity losses (L) and the single-pass nonlinear conversion efficiency (Enl).
Here's everything I wrote down.
| P in (1064, mW) | P out (532, mW) | P trans (1064, mW) | Common mode board gain (dB) | responsivity (A/W) |
| 100 | 48.6 | 2 | 24 | |
| 151 | 88 | 2.6 | 23 | |
| 102 | 49.5 | 0.313 | ||
| 206 | 135.5 | 3.1 | 22 | |
| 250 | 175 | 3.45 | 22 | |
| 300 | 220.2 | 3.8 | 21 | 0.297 |
| 350 | 267 | 4.1 | 21 | |
| 400 | 306.5 | 22 | ||
| 401 | 315.5 | 4.5 | 22 | 0.293 |
The common mode board common path gain was adjusted to keep the UGF at 3kHz.
The responsivity was adjusted to match the photodiode readback to the power meter.
P in and P out was measured with a 1W Thorlabs power meter. P trans was read out of the photodiode.
The optimized crystal temperature for green output was the same for all input power, at 35.9 C
More SHG characterizations:
[Kyle, Chandra]
We leak checked the gate of this valve after suspecting a leak. Kyle cycled it five times on leak detector and four out of the five times it leaked (e-6 mbar-L/s range leak). Then we sprayed the gate with IPA to clean it and re-tested with three cycles. No more leaks! It could be that particulate fell on the gate gasket during rough down when it was installed on OMC NEG housing. We have since received six more of this valve model and decided to leak check all at once right out of packaging. All six are leak tight to a level of 5e-8 mbar-L/s background. I cycled each valve once while on leak detector.
https://services.ligo-la.caltech.edu/FRS/show_bug.cgi?id=9593
J. Kissel, M. Pirello, R. McCarthy
Continuing along with the new HAM5/6 electronics adventure, we were able to make the H1 SUS OFI signal chain completely functional today. The electronics are final and as designed. As such, I was able to confirm that
(a) The suspension remains free after pump-down (yes!)
(b) The damping loop design from yesterday works well (yay!)
(c) The AOSEM noise floor is below spec / requirements (yeah!)
Attached are screenshots of the (newly taken) undamped transfer functions, (newly taken) damped transfer functions, and (newly taken) amplitude spectral densities.
%% Details %%
Undamped Data Files:
/ligo/svncommon/SusSVN/sus/trunk/OFIS/H1/OFI/SAGM1/Data/
2018-01-19_2151_H1SUSOFI_M1_WhiteNoise_L_0p01to50Hz.xml
2018-01-19_2151_H1SUSOFI_M1_WhiteNoise_T_0p01to50Hz.xml
2018-01-19_2151_H1SUSOFI_M1_WhiteNoise_Y_0p01to50Hz.xml
Damped Data Files:
/ligo/svncommon/SusSVN/sus/trunk/OFIS/H1/OFI/SAGM1/Data/
2018-01-19_2312_H1SUSOFI_M1_WhiteNoise_L_0p01to50Hz.xml
2018-01-19_2312_H1SUSOFI_M1_WhiteNoise_T_0p01to50Hz.xml
2018-01-19_2312_H1SUSOFI_M1_WhiteNoise_Y_0p01to50Hz.xml
How we made it work:
(1) The OFI's coil driver is *not* an ISC configuration HAM-A driver chassis (No chassis drawing exists, only front and back panel drawings; see complaints about that in LHO aLOG 8772), it's a regular ol' HAM-A driver (D1100687)-- but both use the HAM-A board D1100117. That means we'd forgotten to jumper the "Test/Coil Enable" switch. That means out OFI's DAC drive wasn't making it into the coil driver circuit. Richard realized this and had Marc pull the chassis and jumper the appropriate pins on the board. While at it, we made the executive decision to jumper the Lowpass Filter ON as well, such that it behaves identically to an ISC configured HAM-A driver.
(2) The drawing for cable arrangement on the vacuum flanges of HAM5 (D1002876-v4) is not up-to-date. As such, I discovered (after driving ZM2 with the OFI signal chain!!) that ZM2 and OFI cables were swapped at the vacuum flange, because whomever installed them couldn't possibly have known. We now have the cables correct, and
- ZM2 (on cable SUS-SQ-21) is plugged into
- Conflat D3 (the furthest +X, on the beam splitter side of HAM5),
- Flange 2 (lower left, if you're looking at the chamber from the air side with the beam splitter to your back)
- Connector 2 (the bottom DB25)
- OFI (on cable SUS-SQ-30) is plugged into
- Conflat D3
- Flange 3 (lower right, if you're looking at the chamber from the air side with the beam splitter to your back)
(3) I had not yet installed the proper COILOUTF driver electronics frequency response compensation (primarily because there's an innocuous bug in the coil driver state machine code, 'cause it was built for a 4 OSEM system, and we're using it here on a 3 OSEM system).
Having fixed those three problems -- we're good to go!
The process segfaults at startup. The reason appears to be an enum that went out of bounds. It was requesting a description string that did not exist. This appears to be an issue only when the connection is first made and a determination is made if the enum should be treated as a string or a numeric value. See FRS https://services.ligo-la.caltech.edu/FRS/show_bug.cgi?id=9749
I branched the advLigoRTS code and implemented a fix. It is being tested on the h1sysecatc1plc4sdf. The h1sysecatc1plc4sdf is running.
Attached are trends over an hour where the BS HEPI actuators were driven individually to look for limits. The plot is likely only useful for me but it does contain the local sensors and the OUTPUT drive and the Z location_mon. I've zoomed into the X-axis of each corner during that corner's drive, hence the screwed up DV x axis. Anyway, it shows some clipping and limits but again mostly for me. I drove sine of various amplitudes.
In the end the IPS should have good signal to +-1mm but the signal is better the closer it is to zero. The mechanical limit of the Actuator is design to be 1.375mm and the sensor gap is 1.625mm. So best to have the IPS operating at zero but you can't just zero it out where you sit. If the platform has sagged for some reason you can get close to the mechnical limits and if you muck about too much with the sensor position, the sensor can be put in jeopardy.
The take away is the following:
V1--Appears to be hitting a mechanical limit at +22K (~0.8mm) and is reaching the IPS limit at -32k counts. Suggests the actuator could be lowered and/or have the IPS rezero'd, but really, the range is fine.
V2--With the sensor centering I did yesterday, V2 is moving freely for +- 1.1mm, plenty of range.
V3--This sensor is showing some sign of clipping at 0.4mm but I think these signs are just showing how the V4 limits impact the other (mainly corner 1 & 3.) Should repeat after V4 is fixed.
V4--Despite the closeness of the mechanical stop on this corner (see last photo on alog 40171) the corner is still moving about 0.8mm. This test isn't good for performance evaluation and I'd still like to do the Actuator reset to make sure we are not too close.
Closing WP 7297
Cold cathode pressure gauge (PT-120) read back is noisy starting ~ 11 UTC (3am local) on Jan. 18th. Not sure why.
J. Kissel I've designed and installed some primitive damping loops in the H1 SUS OFI's DAMP bank. I mocked up a plant model in Foton using the in-air measured data from LHO aLOG 39033, copied over some basic filters from the OMs, and had foton model the open loop gain of the loops. Without any requirements or specifications, I've arbitrarily chosen a loop gain of ~5 around the resonances as a start. Once we get all electronics figured out, I'll characterize the new loops via measurement and adjust the gain / design accordingly. Design details: H1:SUS-OFI_M1_DAMP_L: FM1 (0:100,100) -- a complex pair of poles at 100 Hz with a Q of 0.87 between them FM5 (cheby5) -- a 2nd order Chebyshev low pass, with a 5 Hz corner frequency and 2 dB of pass-band ripple FM6 (gain) -- a x200, gain only "filter" FM9 (PendL) -- [DO NOT USE] the plant model, to be used for offline design assistance only H1:SUS-OFI_M1_DAMP_T: (Same as L, but the FM6 gain "filter" has a value of x4000) H1:SUS-OFI_M1_DAMP_T: (Same as L, but the FM6 gain "filter" has a value of x50)
Turned the RF driver and the power supply on for the CO2Y laser, also connected the IR sensor, and was getting green lights for the control board. Using the PWM controller I tried to turn the laser on but wasn't getting any results on the power monitor. After double checking the power supply was set to 28V and 28A limit, and the RF driver was set to 40.68MHz, Fil and I found that the wrong RF cable was plugged in. After plugging the right one in the laser powered up in PWM mode and I was able to verify the rotation stage is working as well. Tomorrow I will check that the alignment hasn't changed significantly and that the periscope picomotors are operating correctly.
CO2Y is now running in remote operation mode. The beam is dumped before the periscope though. I quickly checked the beam path and everything looks clear, although I did remove the alignment laser as it could possibly be clipping the IR camera path. When the laser proves to be running stably with the new chiller the beam dump will be removed.
CO2Y ran for just under 3 days before faulting out to a flow rate trip. A quick inspection of the chiller shows nominal water level and no faults indicated, so flow rate sensor is the main suspect. Also note the odd drop in the power supply and power output. Not sure why that happened.
J. Kissel, R. McCarthy, M. Pirello While trying to finally get around to finishing out the MEDM infrastructure for the new SQZ suspensions and the re-arranged OMs, we discovered that the RT and SD OSEMs are dead on SRM. They apparently died on Jan 10 ~19:30 UTC (about 11:30a PT). Remember, we know the problem is highly likely out of vacuum, since I was able to take successful transfer functions on Jan 3rd (LHO aLOG 40002). Much more likely -- Gerardo's Vacuum work mentioned in LHO aLOG 40090, or Fil and Liz's cable pulling mentioned in LHO aLOG 40085) After tracing and confirming integrity and correctness of all HAM5/6 SUS cables from the AA chassis to the remote satellite amplifier, I noticed that two of the SRM (T3 LF RT SD) sat amp fault lights were on (see attached picture). Power cycling the respective SRM T3LFRTSD coil driver chassis didn't find the problem. Talking with Richard, he suggested it was the last cable run from the sat amp to the chamber, since it's a tight squeeze getting the cable to seat properly at the feed through. As suspected, he went out to diagnose and was able to wiggle the connection at the feed through enough to intermittently restore life to RT and SD, but didn't have the right tools to really make the connection secure, so he (or Fil) will finish out the fix tomorrow morning.
This morning I was able to remove the strain relief cover and get to the connector. The aligo feedthrough are more finicky than iligo so clear access is needed to insure the 25 pin cables are seated properly. All OSEM signal are present on the MEDM
J. Kissel, R. McCarthy
Ran a (damping loops closed) set of top mass transfer functions for a health check after Richard has properly seated the cable. Functionality appears to be completely restored. Nice work!
I haven't exported and properly processed the data yet, but the data files are committed to the SVN here:
/ligo/svncommon/SusSVN/sus/trunk/HSTS/H1/SRM/SAGM1/Data/
2018-01-19_1618_H1SUSSRM_M1_WhiteNoise_L_0p01to50Hz.xml
2018-01-19_1618_H1SUSSRM_M1_WhiteNoise_P_0p01to50Hz.xml
2018-01-19_1618_H1SUSSRM_M1_WhiteNoise_R_0p01to50Hz.xml
2018-01-19_1618_H1SUSSRM_M1_WhiteNoise_T_0p01to50Hz.xml
2018-01-19_1618_H1SUSSRM_M1_WhiteNoise_V_0p01to50Hz.xml
2018-01-19_1618_H1SUSSRM_M1_WhiteNoise_Y_0p01to50Hz.xml
Corresponds to FRS Ticket 9744.
Daniel, Nutsinee
Here's a quick note of what we measured out there with the RF power coming in through (and monitored via) temporary cables. These RF power drive AOM1 and AOM2. There seems to be too much loss at Pmon2 compared to Pmon1 (almost 11 dBm loss versus 5dBm loss). The cable length going from CPL to Pmon are the same. The top RF input gets its power all the way from the CER while the bottom RF input gets its power from the SQZ rack. So the difference in the incoming RF power makes sense. Not sure if the factor of two difference coming straight out of the Cpl makes sense though.
Terry, Nutsinee
After realizing that the way the cables were hooked up didn't match the wiring diagram we decided to make it right once and for all. The output from the top RF amp used to drive AOM2, now it's driving AOM1 and the 2dB attenuator was taken off. The RF power measured on the table through the helix cable was 34 dB (2.5W), still within the max drive power allowed (2.9W). AOM1 (IntraAction ATM-200) diffraction efficiency is now 82%.
The bottom RF amp is now driving AOM2 (AA Opto Electronic MT200). The 6dB attenuator is still there. The power measured on the table was 33.2 dB (2.1W). The maximum power allows is 2.2W. The efficiency of AOM2 is still not great currently. The best we've had was 70%.