First some background; there are two ways of installing an ADC card in the front end IO Chassis: either as a single PCIe card or as a PCIX card sitting on a PCIe adapter card. Recent ADC errors when running RCG2.9 on the IOP model for SUS EY have been resolved by Keith and Rolf as being caused by ADC/adapter cards being used in certain slots. The hope was that the installed H1 system was not using any adapted ADCs, but when I scanned all the front end IOCs I found the following front ends do indeed have cards on adapters:
h1susey (2 ADC)
h1seiex (3 ADC)
h1oaf0 (1 DAC-16bit)
During next Tuesday's maintenance, we will replace these with PCIe cards to ensure all front ends are hardware identical. We will also check the 3rd IFO IO Chassis are hardware identical to H1 and L1.
no restarts reported
We tried locking PRMI with no success. Some optics seem loud. More details will be posted later.
Dan, Kiwamu
We tried PRX, PRY, MICH and PRMI. PRMI did not lock, but the other configuration locked. Also I am worried that either ITMs or BS may be moving a lot.
(MICH)
(PRX and PRY)
(PRMI)
(ITMX mysterious oscillation in pitch)
At the beginning, we saw a quite steady oscillation in pitch of ITMX at 0.5 Hz with an amplitude of 2 urad according to the uncalibrated oplev. We suspected a hidden excitation somewhere and therefore cleared all the excitation using the diag command. This did not solve the oscillation. Then I increased the pitch damping gain by a factor of 3 to suppress the oscillation and this worked well. Setting the gain back to the nominal value (i.e. -1), I saw no ring up in pitch and the oscillation still stayed small. So I left the damping gain as nominal. Strange. Since ITMY oplev is off from the QPD, we did not check the angular motion in ITMY.
(I feel that the mirror motion is bigger than it used to be. Maybe ?)
Since the last PRMI commissioning in this winter, we have kept a couple of limiters in LSC and SUS for preventing the servos from unnecessarily kicking the optics during lock acquisition. The limiter values and their associated filters had been carefully adjusted so that they don't limit the in-lock DC values but they limit some impulsive signals. However, the limiters were limiting the in-lock feedback signal at DC when PRX(Y) or MICH was locked. I had to increase some of the limiter values by more than a factor of 5 or take them out last night.
I have not done a quantitative spectral analysis yet, but I am feeling that the motion of the suspensions are greater than before.
As for the limiter issue, we need only two limiters to be enabled, which is BS_ISCINF and PR2_M3_LOCK (see this old entry 10559). In addition, I had a wrong limiter on at LSC_PRCL last night.
Anyway, the one limiting the DC value yesterday was BS_ISCINF. This used to be 5 x105 and it was 2 x 105 in eariler yesterday for some reason. I then increased this limiter to 2 x 106 in order to avid the signal being cut at this limiter.
32 channels added
Will continue tomorrow
With Dave's help, and Hugh struggling along with me, I've set a tf to start on ITMX ISI. Should start at 12:30, some time after Kiwamu leaves tonight. It should also: pause the SEI guardian, ctrldown the ISI, undamp the ISI, set the GS-13's to low gain and bump up the watchdog threshold on the GS-13's before starting the excitation. All any commissioners need do is make sure the ISI is in its normal operating state before they leave, and not interfere with the measurement while it's running. Running from opsws0
08:40 Peter King - to H2 enclosure
08:52 Jeff/Andres - craning over X-Arm
09:01 Kiwamu - doing alignment work @ ISCT1
09:20 Fil - doing camera work in LVEA
09:49 Hugh - to Ey
10:00 Jason - starting work @SR3 op/lev
10:10 Betsy/Travis - working in West bayl
10:19 Sudarshian - to EX tiltmeter
11:11 Fil - to Ex to look at Gig-E cabling
11:23 temp loss of read access to Conlog
12:00 HFD - investigatingtrouble alarms at Ex/My
12:32 Cyrus - restarting camera servers
12:34 R McCarthy - Ey HEPI investigation
12:37 King Soft taking samples
13:16 Restarting Ey HEPI controller - Richard?
13:26 D Barker/J Batch - to Ex
13:37 Karen - @ My
13:43 Cyrus - into CER for camera/video switch
13:47 J Bartlett - to W bay rummaging for parts
13:56 R McCarthy - back from Ex
14:13 Cyrus - back from CER
14:18 P King - back to H2 enclosure
15:11 Sudarshian/Christine - to beam tube between corner and Mid y
15:35 Sudarshian/Christine - back from beam tube
When Kiwamu locked PR-X (CH16), POP_A (CH1-7) saw some decent beam, about right in PIT and off in YAW.
POP_B (CH9-15) also saw something, it's not clear if it was the real beam or some ghost, the power is about a factor of 10 smaller than POP_A, and mostly in seg1 and 2.
Anyway, the important thing is that the POP_A already see the beam and POP_B is close.
Don't use pico to center these, though, as POP sled might be a good initial alignment reference.
We should first align the IFO to the POP sled and see if it makes things better. POP sled was aligned perfectly in the last days of HIFO-X/Y. Though a beam splitter was inserted after the vent, the insertion procedure was written such that the sled centering is preserved.
The above entry is simply wrong.
When I centered these QPDs before vent, with 45dB whitening gain and with a straight-shot beam, DC sum of QPDA was something like 56 counts with a bogus calibration factor of 0.0056 that is in QPD segment filters (screen shot from before the vent, May 30).
Since then we installed 90:10 to attenuate the light by a factor of 10, but increased the PSL power by a factor of 5, so when the beam hits the QPD I expect something like 28 counts, but we got 0.4 at most, it's a factor of 70 smaller than expected.
S. Karki The zeroing of Applied Geomechanics Tiltmeter is completed at ENDX and should be up and running. It was so off-scale, using a bubble-level and voltmeter did the trick to bring it down to reasonable level and after that I used the DTT to level it more.
Jeff, Krishna, Robert here is the calibrated (against seismometer) amplitude spectral density.
That doesnt look quite right. Lets debug via email. -Robert
As planned (see alog from yesterday), I restored the POP path on ISCT1. The POP beam is now detectable with both POPAIR_A and POPAIR_B. Also, I did some clean up on the table.
Here are summary of the work:
Summary
- OM1/2/3 showed huge angle to angle off-diagonal coupling of the order of 100% during the commissioning last week.
- Yesterday I figured out that the reason was incorrectly and sneakily set non-identity DRIVEALIGN matrix.
- Once DRIVEALIGN was fixed, I could measure the Pitch-to-Yaw (P2Y) and Yaw-to-Pitch (Y2P) couplings.
They were about 30% and 10%, respectively.
- A small arithmetic calculation revealed that these P2Y and Y2P can't be elliminated by the four coil gains.
Namely we need different coil balance for elliminating P2Y and Y2P independently.
- This has been done by adjusting the offdiagonal elements of DRIVEALIGN. As a result the P2Y and Y2P have been reduced
smaller than ~3%, even though the decoupling matrix (DRIVEALIGN) so far is a frequency independent one.
- One question: Why do we have the decoupling matrix only in the ISC signal chain on the every suspension?
This way, we can't enjoy the benefit of diagonalization for the other actuation signals such as alignment offsets,
damping servo outputs, test inputs. Is there a philosophie behind the current design?
Details
1. DRIVEALIGN
It turned out that the all elements of the DRIVEALIGN matrix (3x3) were the unity. Alas! This meant that any input (L, P, or Y) result in
the same actuation. This was overlooked because the matrix button still indicated only the diagonal elements were active (green)
even though that was not true. The warning in the DRIVEALIGN screen tell (in a very dark texts), the indicator criteria is different
between the diagonal and off-diagonal elements. It saids FM1 needs to be turned on to turn on the green indicator for the off-diagonal
elements. This is confusing as the frequency independent element in the off-diagonal element does not trigger the indicator. I strongly
feel that we should use the same criteria for the off-diagonals as the one for the diagonal elements.
For now, FM1 of the P2Y and Y2P elements have dummy k=1 filters so that we can activate the indicators.
2. P2Y/Y2P coupling measurements
Once the DRIVEALIGN was modified to be an identity matrix, more reasonable angle-to-angle coupling started to show up in the measurements.
The measurements have been done with no damping in order to avoid confusion caused by the coupling coming from the damping loop.
140826_TT_P2A and 140826_TT_Y2A show the measured transfer functions from the Pitch or Yaw excitations to the Pitch and Yaw local readings.
As I did not have the beam, the OMC QPD signals were not used. The coupling ratios (P2Y/P2P or Y2P/Y2Y) were picked up at 0.5Hz and 5Hz.
P2Y is ranging from 0.18 to 0.39 while Y2P is from ~0 to 0.13. They are summarized in the following table
| | P2Y@0.5Hz | P2Y@5Hz | | | Y2P@0.5Hz | Y2P@5Hz | |
OM1 | | | 0.394 | 0.326 | | | 0.088 | 0.100 |
OM2 | | | 0.236 | 0.176 | | | -0.015 | 0.014 |
OM3 | | | 0.369 | 0.313 | | | 0.109 | 0.126 |
The OM2 Y2P coupling is basically such small that the measurement crossed the zero due to noise and had unstable sign.
For the decoupling calculation, the average of the LF (0.5Hz) and HF (5Hz) numbers were used.
3. Coil balance is incapable for P2Y/Y2P decoupling
Suppose we have perfect coil balance. The pitch and yaw actuations (P and Y below) are distributed to each coil equally (the middle matrix).
The actuation forces are recombined at the suspension (the left most matrix below). This reproduces pure pitch and yaw motions (the right most P&Y)
Now we have the actuator response modified by alpha, beta, gamma, delta.
As you can see, the decoupling terms come in symmetrically. We can't decouple P2Y and Y2P independently by the coil balance.
Note that this is obviously not true for A2L, L2A couplings.
4. Decoupling by DRIVEALIGN
For the above reason the DRIVEALIGN matrix was used for the decoupling. This was easy to implement as we just plug in
the inverse matices of the 2x2 coupling matrix. The below is the matrix plugged in.
OM1 | Lin | Pin | Yin |
Lout | 1 | 0 | 0 |
Pout | 0 | 1.0352 | -0.0976 |
Yout | 0 | -0.373 | 1.0352 |
OM2 | Lin | Pin | Yin |
Lout | 1 | 0 | 0 |
Pout | 0 | 1 | 0 |
Yout | 0 | -0.206 | 1 |
OM3 | Lin | Pin | Yin |
Lout | 1 | 0 | 0 |
Pout | 0 | 1.0419 | -0.123 |
Yout | 0 | -0.356 | 1.0419 |
5. Resulting improvement of the coupling
After the new DRIVEALIGN matrices were plugged in,the same measurements wererun. The result for Pitch and Yaw excitations are found
in 140826_TT_P2A_after.png and 140826_TT_Y2A_after.png. The coupling is reduced to ~0.03 or better. This is enough for the further
work on the alignment servo.
If the further decoupling we should install a frequency dependent decoupling.
6. Decoupling topology
But a question arose:
In this topology, only the ISC signals enjoy the benefit of the decoupling.
The alignment offsets (i.e. the alignment sliders), test inputs, and damping actuations do not go through the decoupling matrix.?
Is this what we want? Don't we want to have a DRIVEALIGN FM matrix after the summing node before the euler-to-osem matrix?
Of course, we can convolve the decoupling into the euler-to-osem matrix, in principle, but I don't want to mess up that matrix.
This has been down since 12 Aug. The Input Line Lugs were tightened during the power outage. The VFD while powered, does not respond in the usual manner. It displays STOP and does not respond to the FWD button. It should not be looking like this anyway as the the FWD lugs have been jumpered and it should just run as soon as it's given command voltage. At McCarthy's suggestion, I toggled the main power which then required restarting the 'FE' but this changed nothing...
no restarts reported.
The OM1 and OM2 coil drivers were found to have bad Vmon channels. Chassis S1200614 was found to have -14VDC present on the output of IC5 (OP27) on Ch2. Similarly, S1200612 exhibited the same symptoms on Chs 2&3 (-13V and +14V respectively). In all cases, IC5 was replaced and this DC issue was reolved and the chassis were placed back into service. IC5 is the Vmon-N circuit. All Vmon circuits were examined (P & N). If there are any further issues with this circuit please let me or Filiberto know. Original post: https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=13585
I continued working on the PRMI alignment (see alog 13592) today and finally obtained good alignment. This should be good enough for proceeding with the next steps.
With the new alignment, the PRMI flashes without loosing the POP beam at ISCT1. In fact, one can now see PRMI fringing in the POP camera.
We are almost ready for some locking activity.
(Next steps)
(alignment)
From the experience yesterday, I learned that a combination of PR2 and PR3 did not improve the POP/PRMI alignment situation for some reason. In fact, it made the forward propagating beam hit the west side of the PR3 suspension cage. Therefore, I wanted to move the input pointing in order to introduce a translation in the cavity axis in the downstream after PR2 while keeping the POP beam angle (almost) the same. To do so, I touched IM4 in yaw. This then required re-alignment of PRM in order to maintain the beam at the REFL port. Also, every time I changed the alignment of IM4, I compensated it by steering PR2 such that I keep the POP beam at ISCT1. I repeated this process multiple times. This was successful. I could eventually obtain an alignment setting where the PRMI flashes with the POP still visible at ISCT1. The attached is the new alignment settings.
After the PRMI alignment, I re-adjusted the PR3 alignment such that the beam reaches HAM6 and ISCT6. I did not check the beam centering on the SR mirrors.
(fun locking)
Since I was running out my energy, I decided not to try the PRMI locking tonight. Instead, I locked PRX for fun. I just used REFL_A_RF45 for now. The ASAIR_LF signal was used to check if it is locked on the carrier or rf sidebands. The attached is time series of some DC signals at the AS port when locking PRX. It is locked on the carrier.
Since PR2 and PR3 are close in the Gouy phases, this combination is not very good for the prc alignment. IM4, PRM and PR2 (which is almost same as PR3 and BS in G phase) have some separations each other. A combination of these three will work.
Kyle, John, Keita, Dan, Koji, Filiberto, Gerardo, Kiwamu
The ISCT6 enclosure is now in place by HAM6 with two light pipes installed. ASAIR and OMC_R are successfully extracted to the table.
We craned the ISCT6 table from the squeezer bay over to the north side of HAM6. We carefully adjusted the position of the table by rolling it. The table legs were then secured on the floor. With help from Gerardo and Filiberto (a big thank you for Gerardo and Filiberto who machined the holes for us even though the working hour was over !), we made a 10 inch beam penetration hole on two back panels of the enclosure . These are for the ASAIR and OMC_R beams respectively. Note that we could not see a bright enough OMC_T beam this time and therefore we did not try to make a hole on the last back panel. This will be don once we reliably confirm the OMC_T beam position. We put the flanges and light pipes for the two beams and now the beams are successfully coming out to the table.
Also, in the process of trying to find the OMC_T beam, we turned on the PZT high voltage. We need to make sure that the high voltage is off when we pump down the chamber tomorrow.
We made a rough alignment of the table and connected ASAIR_A and ASAIR_B to the electronics rack. The alignment will change once the pumpdown starts tomorrow, but for now we have checked that the RFPDs are alive.
We're ready to make a beam scan measurement of the AS and OMC_R beams on the table to characterize the mode matching, we'll do this tomorrow. (Tonight Kiwamu was adjusting the alignment of the PRC, I thought it best to let him finish.)
The high voltage to the PZT has been turned off.
Borja
I just came back from a quick visit to End-X (UTC 2014-08-27 00:25:00).
I took some pictures (see attached) at the feedthrough wiring for the ESD at BSC9 (ETMX chamber). I compare it with the ones I took a while back at ETMY (see attached).
There is an obvious difference, the red insulating tape at ETMY wiring which does not show at ETMX. The wire which has the insulating tape carries VBIAS to the ESD at ETMY. The reason for the insulation is that the coax cable's shield is cut off because the corresponding coax cable inside the chamber has the known issue of having the shield connected to the center lead. This problem is non existen at ETMX.
This is another big difference between both masses which may be related to the more stable results on the charge measurements observed at ETMX.
We need to fix the shielding in any event. The shield now acts as a large capacitance coupling of the electrode to the chamber and will inevitably cause cross talk and noise in later operations.
Removed the Ring Heater Assembly from the ITMy quad structure going into 3IFO storage. Before starting I noticed the lower ring heater assembly had a crack in the glass former. Not sure when the break happened, but this one also displayed a lot of movement inside the ring heater shield. The pictures show how the former has pushed pretty far out of position. The ring heater cables were also removed. This includes the "antlers" (D1001755) but not the plate (D1002420) which the earthquake stop is attached to. The assemblies removed were: ASSY-D1001895-V5 #001 (lower ring heater, broken) ASSY-D1001517-V7 #612 (cable assembly) ASSY-D1001838-V6 #005 (upper ring heater)
That should read ITMx (not y) and ASSY-D1001838-V6 #002 (upper ring heater) not #005. typos.
Note, when the 40kg optics are installed and removed from the QUAD lower structures that these ring heaters are mounted to, there is some torquing of the structure (and therefore the RH). Possibly this adds to these failure modes.
One more fix. The lower ring heater should be: ASSY-D100195-108 Made a mistake with a V5 and V6 part.