So the week has been spent trying to get the chamberside testing area up and running. Satellite amplifiers were put in place, field cables run and the MC2 suspension (which currently sits on the table in the testing area) was all connected up. The models/medm screens (which were ported over from LLO I believe) for the MC2 suspension has all the matrices, filters, etc filled in as well as I have used the values from log entry 3125, to put in the correct offset and gains. I made a burtbackup using Joe B's, makeSafeBackup script and this has been put into the SVN
The BOSEMs were centered (in and out around the magnet, not up and down and side to side, this can be done after the optic is in) so that the speedo indicators now all stand straight up at a half open light count value. The aOSEMs are still pulled back, and can probably do so until after the optic has been put in
I did however notice that the open light readings of the MC3-LR aOSEM (S/N 188) was quite different to log report 3125. After first checking that this S/N OSEM was indeed in this location I remeasured the open light voltage of this particular OSEM using tdsavg 10 H1:SUS-MC2_M3_OSEMINF_LR_IN1. This gave an open light value of 19455.2 counts. I modified the offset and gain in the medm screen appropriately (I can't remember if I made a new burt backup of this or not. I am pretty sure I did, but cant be certain. Be aware of this if you need to Burt restore this model).
A dtt template of the channels used for a pitch TF measurement was brought over from LLO, and the channel names renamed appropriately. However when a TF was attempted to run the error message came back "Unable to select test points". A long period of time has been spent on the phone Tuesday and Thursday with Joe B at LLO and Dave Barker here at LHO trying to resolve this (and other software problems have run into especially in terms of permissions.........thanks very much to the two of them for their patience) and the usual suspects of looking that had correct ADC channels connected , INI files were as they should be, etc and various other things tried (including a DAQ restart) we were unable to fix the problem.I have copied and pasted below a couple of extracts of emails sent between the three of us giving what I believe is the latest on the unable to select test points problem:
The "Unable to select test points" message seems to be related to the excitation channel. Providing no excitation channels allows data to be taken with diaggui. Similarly opening up awggui and trying to even select the H1:SUS-MC2_M1_TEST_P_EXC channel brings up an error message "Cannot connect to channel".
and then the latest from David Barker
Looking at the network switch configuration (which does the UDP broadcast forwarding to the front ends), the environment variables and the GDS parameter files; all look correct for an H1 system. But when I run diag it always finds the H2 awgtpman systems and not H1 so I'm wondering if it is hard coded somewhere.
So the problem still exists but is being looked into. I believe (but dont hold me to this), that once this problem is fixed I should be good to go to start taking TF's on the suspensions chamberside
This was before we replaced the ALS Faraday today.
I tilted the beam by the PIT and YAW offset sliders of the TMS table. Then I used Keita's calibration of the sliders from counts to radians (see aLOG entry 3244) to estimate the actual beam tilt.
I swept the beam over the ITMY PD2 baffle PD. The beam was centered on this PD when the offsets were:
PIT -79500, YAW +49041
Initial values were:
PIT -60691, YAW +56641
Sampling a Gaussian beam in X or Y with a PD of finite radius r0, produces and intensity profile defined by the function attached below: xc is the beam center location, and w the beam radius.
I measured the beam radius by fitting the data from the PD scan with this function. Assuming that the ITMY baffle PDs are Perkin Elmer YAG-444AH (although a document by Mike Smith says that they're YAG-444-4AH QPD), then r0=16.3/2mm and the resulting radii are:
The beam is astigmatic and not well matched to the cavity mode.
We are going to improve the mode matching soon.
We replaced the green Faraday due to clipping problem.
We've already aligned the injection path (i.e. the beam goes to the TMS and the QPDs are centered) but we need to work on the PDH and WFS path tomorrow.
Before/After picture
First picture is the beam with the original Faraday. You can easily see some really ugly lines. According to Alberto and Dan, this gets uglier when the beam is centered in the Faraday, so it was aligned closer to the Faraday edge. Combined with large TM motion, this caused a serious clipping problem.
The second picture is with the new Faraday taken at the same position. No apparent ugliness here. If you look closely you can tell that there's some funny stuff still there at the top (it would become more apparent in the far field), but this is much better than before. We expect that the clipping problem will be gone, though the large TM motion is still a problem on its own.
Turns out that segment 1 and 2 of the legacy WFS DC interface for WFS1 are not working. When we injected some offset using a break out board, these segments didn't show anything in the output. It's not the WFS head, it's the interface.
Pulled the board to investigate in the lab.
Two op-amps in the interface board were broken. WFS1-Anode 1-U2 (OP27) and WFS1-Anode 2-U3 (AD8672) were removed and replaced, resolving the issues observed yesterday. The board is now functional and can be returned to the end station. According to Richard McCarthy, the board should be replaced after we're done with the one-arm test.
Attached are plots of dust counts > .5 microns in particles per cubic foot. Also included are plots of the modes for a particle counter in the optics labs, LVEA and end Y. The plot of the mode of the dust monitor in the optics lab (H0:PEM-LAB_DST1_MODE) shows approximately when the communication was lost and restored for the three dust monitors in optics lab, vacuum prep, and bake oven room.
The following is an update to the status of the H2 SUS ETMy L2 stage (PUM) actuation. The L2 stage has not been completely vetted with clean transfer function measurements and good coherence seen in all of the OSEMs. The drive amplitudes needed for these measurements are approaching the DAC limit of ~131,000cts for individual OSEMs coils. Note that this value is much higher than the actuation watchdog limit set in the SUS QUAD User model, which was initially set at 15,000cts. The limits were increased to allow these large excitations to pass. The first plot is a whitenoise transfer function measurement taken on 07/02/2012 in the Longitudinal DoF on ETMy L2. There is great coherence above 3.5Hz for one of the OSEMs (the "LR" OSEM), however, the drive is sent equally to each OSEM. The remaining OSEMs have some coherence, but is much lower than expected for a healthy OSEM response. The DC coherence is essentially zero for all OSEM responses. The WhiteNoise amplitude for this measurement was ~300,000cts in the "*_TEST_L_EXC" channel, corresponding to roughly a 75,000ct amplitude per OSEM coil. The reason why one OSEM has such a high coherence is still an ongoing investigation. Possible measurements would be to measure the impedance of the Coil Driver box or the Satellite amplifier for each of the OSEM drive signals. For the transfer function plot itself, the resonances are somewhat visible but not sharp enough to be distinguished. The second plot is the same measurement taken back on 06/04/2012 but with a lower excitation amplitude of 90,000cts in the "*_TEST_L_EXC" channel. There is great coherence above 6Hz for one of the OSEMs (the "LR" OSEM). The remaining OSEMs have some coherence, but is still not quite as good enough for a clean transfer function measurement. For the DC coherence, each of the four OSEMs do respond in equal magnitude with the "UR" OSEM having a slightly lower DC response. The same can be seen in the transfer function at DC, where the "UR" OSEM has a slightly lower DC response. Resonances are still not clearly visible as well.
Activities: July 5, 2012 - Dave B. - restarted models: "I have just restarted the H1 SUS models for PR2 and SR2 on h1sush34 to complete this system. I pressed the DIAG RESET button to clear the latched IPC error, so everything should be green on the medm screens." Dave - later, a frame builder restart on H1 - Dust monitors went invalid, Patrick restarted, alarm handler did not update, I had to reopen the alarm handler and truly false alarm states cleared, however a couple false false alarms remained (see note below) - Matt H. in LVEA working on triple and in CR working on making transfer functions, but instead uncovering issues with files and channels for H1. Dave B. is looking into it. - EY visited by numerous people: Keita, Alberto, Mike L. - OAT related - EX visited by numerous people: - Terry S., Terisha, others(?), for cleaning - Apollo for building a clean room - Kyle, changed out some GN2 pressure gauges --------- Alarms that really aren't…. - LVEA dust monitors DST7 and DST12 - false alarm when invalid
Attached are plots of dust counts > .5 microns in particles per cubic foot.
Though Beckhoff model was generated for the WFS DC signals, it wasn't reading anything at all.
We eventually found that the Legacy PD concentrator that routes WFS DC and PDH DC signal to Beckhoff chassis (https://dcc.ligo.org/cgi-bin/private/DocDB/ShowDocument?docid=61861) had a manufacturing error: DB37 male and female connectors are mixed up.
DB37 female should be used for the output end that is connected to Beckhoff input, but DB37 male was used instead.
Likewise, DB37 male should be used for the input end that is connected to the output of the legacy LSC diode interface, but DB37 female was used instead.
Due to this, apparently the Beckhoff cable and the legacy LSC diode cable were swapped as it's impossible to connect things up correctly.
We used two 37pin gender changers (one male-female, one female-male) to "fix" this problem. After this, it seems that the EPICS screen for WFS DC is displaying something sensible.
[Joe B, Matt H]
1) Make sure you have the matlab simulink .mdl model ready in userapps.
2) ssh into the build machine "ssh controls@h1build"
3) cd /opt/rtcds/lho/h1/rtbuild-2.5/
4) make model name (i.e. make l1susmc2)
5) make install-model (i.e. make install-l1susmc2)
6) If possible, create a safe.snap file before hand in /opt/rtcds/lho/h1/target/MODELNAME/MODELNAMEepics/burt/
6a)Copy a safe.snap from LLO (say l1susmc2_safe.snap, renamed properly) and run
sed -i 's/L1:/H1:/g' safe.snap
6b)Otherwise use the burtgooey command to make a backup *after* the model is running, using the autoBurt.req file in /opt/rtcds/lho/h1/target/MODELNAME/MODELNAMEepics/ directory, create a safe.snap.
6c)Another way is to use makeSafeBackup script in /opt/rtcds/userapps/trunk/cds/common/scripts/ and hand it susbsystem (i.e. sus) and model name (i.e. h2susmc2)
./makeSafeBackup sus l1susmc2
Only works as controls, so be logged into h1build or a front end, and only after the model is up and running, and set to the right values.
7) ssh into the front end the code will be running on (h1sush34)
8) lsmod (tells you whats currently running)
9) If you need to load a new IOP model, kill any currently running models
10)Run the "startMODELNAME" script (i.e. starth1iopsush34 or starth1susmc2). IOPs always need to be running before user models.
11) Quick filter loading has two options:
11a) Open your empty foton filter file in foton, and uncheck Read Only (under the file menu), and hit save. This preps the file to receive filters.
Then run the prepare.m script from /ligo/svncommon/SusSVN/sus/trunk/Common/MatlabTools/ in matlab.
prepare('H1','SUS','MC2','HSTS')
11b) Copy the foton filter file from LLO and rename it to H1. I.e. copy L1SUSMC2.txt from LLO and rename it H1SUSMC2.txt and put it in /opt/rtcds/lho/h1/chans/
12) Hit the "Coeff Load" button on the GDS_TP screen associated with your model.
in steps 4,5; of course the l1 should be replaced with h1
4) make model name (i.e. make h1susmc2)
5) make install-model (i.e. make install-h1susmc2)
I installed a new h2peml0 model for the corner station and restarted the H2 DAQ
The HAM6 ISI has been under purge for several weeks now and we've seen a continuous drop in dew point. These latest data are a continuation from previous logs.
You can see on the 24th the numbers spike. This was caused by me allowing (unintentionally) the Nitrogen dewar to run out over the weekend. Luckily after refilling and restarting the test on the 25th you can see the dew point dropped quickly enough. Most likely this was air intermixing at the probe and less mixing inside the shipping container. I've decided to continue to run the purge at 10 liters/min until the numbers stagnate for a long period of time.
Last Friday, I replaced Chassis Power Regulator board D060279 with D1000217 on the following AA/AI Chassis. S1102693 S1102694 S1202505 S1202506 S1102693 S1000250 This was to address overheating/over stressing of components on old power board. The units are located in the SEI and SUS Test Stands in the staging building. Filiberto Clara
I removed the final Horiz & Vert Actuators from BSC4's SW corner. All of these Actuators are on pallets out near BSC3, for installation on H1 HEPI locations. The crane is parked in its normal parking spot.
It turns out that most of the power in the HWS path is just dumped. We installed a PDA55 to monitor this otherwise wasted beam so that we can measure the reflected beam power without being clipped.
According to this PD our contrast (1-min(locked)/unlocked) is 65%. I'm taking min(locked) because, when locked, the REFL power fluctuates much because of the alignment fluctuation (but not clipping).
We can refine alignment further, but doing it manually is no fun. We're working on WFS. Also, though there's no fast ADC monitoring the refl DC, we might be able to auto-dither connecting REFL DC to AUX channels.