I sat down in the control room to find that ETMX and PRM's watchdogs had been tripped. I don't know how long they've been crazy, but there was a big something that showed up in the seismometers around 16:30UTC, or ~11:30 this morning. I don't find any significant earthquakes on the USGS site for that time though, so it might be more local, i.e. work next door or trucks or whatever.
I take back the suggestion that it was that seismic event. Clearly the PRM and the ETMX were kicked at different times, neither of which is the same as the seismic action. Mystery. You can see they have been ringing down for a while though, which is neat.
All of the SUS used to have only 1 filter module for SIDE. They now have 3 filter modules for SIDE just like the other DOFs.
Today I moved the filters around so that the sensor filters are in SDSEN, the servo filters are in SUSSIDE, and the dewhitening for the coil is in SDCOIL.
I noticed along the way that the bounce/roll mode notches for all of the suspensions are still set for the frequencies of the previous suspensions. Suresh has 'volunteered' to find the new frequencies and make the new bandstop filters by looking up the seminal work on this by Dan Busby / Sam Waldman.
The f2p measurements are done on ETMX and ITMX with the real time lockin systems.
The f2p measurements are done on ETMX and ITMX with the real time lockin systems.
I don't explain what is the f2p measurement in this entry, but people who are interested in it can find some details on an old elog entry here or somewhere on DCC.
So far the resultant filters looked reasonable compared with the previous SRM f2p filters.
- backgrounds -
Some times ago I found that the coils on ETMX had not been nicely balanced, and it made a POS to angle coupling when I tried green locking (see here).
In addition to that, accuracy of A2L kind of measurement including the dithering techniques depend on how well the coils are balanced. Therefore we need to balance the coils basically at all the suspended optics.
There used to be a script for this particular purpose, called f2praio.sh. This script does measure the imbalances and then balance the coils.
However this time I used the realtime lockin system to measure the imbalances instead of using the old f2p script.
One of the reasons using the real time system is that, some of the ezca and tds commands for the old script don't work for some reasons.
Therefore we decided to move on to the real time system like Yuta did for the A2L measurement a couple of months ago.
The f2p measurement finally gives us parameters to generate a proper set of filters for POS and also the coil gains. We apply those filters and the gains in order to eliminate the POS to angle coupling and to balance the coils.
- results -
The followers are the resultant filters and coil gains.
The plots below show new f2p filters according to the measurement.
ITMX (assuming pendulum POS has f0 = 1 Hz and Q = 1)
ULPOS fz = 1.009612 Qz = 1.009612
URPOS fz = 1.125965 Qz = 1.125965
LLPOS fz = 0.873725 Qz = 0.873725
LRPOS fz = 0.974418 Qz = 0.974418
ETMX (assuming pendulum POS has f0 = 1 Hz and Q = 1)
ULPOS fz = 1.055445 Qz = 1.055445
URPOS fz = 1.052735 Qz = 1.052735
LLPOS fz = 0.944023 Qz = 0.944023
LRPOS fz = 0.941600 Qz = 0.941600
C1:SUS-ETMX_LLCOIL_GAIN = 1.07233
The precision of the coil gains looked something like 1% because every time I ran the measurement script, the measured imbalances fluctuated at this level.
The precision of the filter gain at DC (0.01 Hz) could be worse, because the integration cycles for the measurement are fewer than that of the coil gains done at high frequency (8.5 Hz).
Of course we can make the precisions by increasing the integration cycles and the excitation amplitudes, if we want to.
The plot below shows how the f2p filters work.
At -2 min I turned on the f2p filters.
Cheater cables for SRM sus tied up. They were dangling aimlessly on the floor.
As a part of the DRMI preparation,
I leave all the suspensions free from the watchdogs for 5 hours from now.
Please DO NOT touch them.
I will check the spectra and the mechanical resonant frequencies on Monday.
Also I will renew all the input matrices of the local dampings based on these free swinging spectra.
There is a useful script for this particular job : shutting down all the suspensions and bringing it back to operation after 5 hrs.
It is called opticshudown, which resides in /cvs/cds/rtcds/caltech/c1/scripts/SUS/.
Also I added this script on the list in the wiki where all the scripts will be listed.
If you find any other useful scripts, please add them on the wiki.
I had a quick look at PRM optical lever.
The He-Ne beam is still successfully coming out from the chamber and I could guide it to the QPD by using steering mirrors.
But the beam size looks too big for the QPD. We should slide the lens which is standing before the injection to get a moderately smaller beam size at the QPD.
- activation of PRM oplev
The returning spot diameter on the qpd ~10 mm. In order to reduce the spot size I moved the f 1145 mm lens toward the PRM ~ 25 cm. The spot size was reduced to ~8 mm, 3200 counts.
I'll try to find an other lens tomorrow.
Here are the free-swinging spectra for the BS, ETMX, ETMY, ITMX, ITMY, MC1, MC2, MC3, and PRM chambers. Kiwamu left the suspensions free for 5 hours this weekend, starting at Sat Apr 30 00:15:26 2011.
This is GPS time 988 182 941. Quick tip: you can do local to GPS time conversions using lalapps_tconvert, which is a lot like tconvert but with special powers. It is installed on pianosa.
$ lalapps_tconvert Sat Apr 30 00:15:26 2011
I generated these figures with the attached Python script, measure.py.
Notice that the C1:SUS-ITMX_SENSOR_UL and C1:SUS-MC3_SENSOR_UL spectra fall as 1/f. Jenne suggested that this might indicate that there is a loose electrical connection.
Also, notice that C1:SUS-ETMY_SENSOR_LR, C1:SUS-ITMY_SENSOR_LL, and C1:SUS-PRM_SENSOR_SIDE are a lot noisier above 10 Hz.
Jenne went through all the suspension racks and pushed all the connectors.
After pushing them, we had a quick look at those spectra and found no funny noise spectrum except for C1:PRM-SENSOR_UL.
We then checked connection around the SCSI cables and eventually found the connection between ADC_card_0 and a SCSI was loose.
We put short standoffs on the ADC card so that the screws from the SCSI can nicely reach to the ADC card. Now everything looks fine.
SUS diagnostic is quite useful !
Atm 1, PRM oplev inward path with 2 lens solution: 14 cm gap between F 1145 and F 1545 mm lenses.
Atm 2, The PRM beam size 3 mm and the beam quality is still bad. The BS path only needed alignment.
[Leo w/ a little help from Kiwamu]
Leo summarized the mechanical resonances of all the suspensions, based on the free-swinging spectra taken on Sat Apr 30.
Since Leo doesn't have the wiki account I helped him putting the information on the wiki.
Good work, Leo !
I am tuning the notch filters for the bounce modes in the suspensions, starting with the ITMs and ETMs. I'll do the MCs, the PRMs, and the SRMs next.
I noticed that the filter for ITMX (in the file C1SUS.txt, the module ITMX_SUSPOS, the selection BounceRoll) that the filter was composed of two bandstops (and a constant gain). It looked like this:
Valera said that one of these was for the roll mode and the other for the bounce mode. However, looking at the spectra that Kiwamu and I made this week, I don't perceive a resonance between 11.4 and 12.2 Hz. So, we're taking a guess that this was for a mode that has moved due to new pendulum designs. For many of the suspensions, in the free swinging test we noticed a line around 23 Hz; we thought we might as well re-use one of these elliptical filters to avoid exciting this line. Of course, if this line does *not* result from excitation of an uncontrolled degree of freedom, this will not help and could be detrimental. When we talk to Valera again, we can review this decision and at that point we might decide just to take out that bandstop.
ITMX is done. I'll continue tomorrow. I've attached closed-loop spectra for before the tuning (itmx-before.pdf) and after (itmx-after.pdf).
(Update: the following day, I took closed loop spectra with (itmx-withbounceroll.pdf) and without (itmx-nobounceroll.pdf) the bandstops. It looks like the bandstops made the bounce mode slightly worse, but the roll mode slightly better.)
I tuned the ITMY bandstops -- 'before' and 'after' spectra attached. Note that the after the tuning, the bounce mode at ~16 Hz is about twice as quiet!
However, notice that in the 'before' plot the roll mode at about 23.5 Hz did not show up at all, whereas it is quite prominent in the 'after' plot. I was concerned that this line could have been a result of placing the bandstop there, so I made another plot with the BounceRoll filter turned off. Sure enough, the 23.5 Hz line is still there. So I'm not crazy: the roll mode did start acting up at some time between my 'before' and 'after' plot, but not as a result of the tuning.
The SRM qpd cable was removed from the BS-table. It's path was changed from 1x4 to ITMY-table following the inner cable tray.
New f2p filters were installed on ETMY.
The statistical error of the coil gain settings are now about 0.8% at high frequency (i.e. above the resonant freq of the pendulum mode)
What I did :
- measured and corrected the coil imbalances on ETMY using a script called F2P_LOCKIN.py
- made the new f2p filters based on the measurements and installed them.
Next step :
- do the same adjustment for all the suspensions including PRM, SRM, BS, ITMs and ETMs
(Notes on F2P_LOCKIN.py)
F2P_LOCKIN.py is a script that I've made in python. This is basically the same as the old script, f2pratio, but uses the realtime LOCKINs instead of ezcademods.
The script automatically measures the coil imbalances on an optics of interest by driving the local LOCKIN oscillators.
In the first step the script automatically balances the coil gains at high frequency (8.5Hz).
In the next step it gives some coefficients, which basically represent the coil imbalances at low frequency (0.01Hz)
Then with those coefficients one will be able to design the f2p filters.
It is not well polished yet, so I will spend some more times to make it user-friendly and readable.
Example usage : F2P_LCKIN.py -o ETMX
It currently resides in /cvs/cds/rtcds/caltech/c1/scriptss/SUS/
(new f2p filters)
The plot below shows the new f2p filters. Note that they are already installed.
I found that a He-Ne laser which has been used for ETMX_OPLEV was NOT giving the light.
Since I didn't find the switch key for it I have no idea if the laser is simply off or dead.
The dead laser was replaced by new JDSU 1103P of 2.6mW. The return beam is big ~5 mm diameter of 0.3 mW, 1400 counts
Laser diode oplev SRM is working. Qpd matrix values were reset like others.
In 0.44mW, returning 0.1mW, -500 counts.
New f2p filters were installed on ITMY this morning. The statistical error of the coil gain setttigs are about 0.6 % at high frequency.
NEXT : PRM, SRM, BS, ITMX and ETMX.
Pendulum mode = 0.988 Hz, Q = 1
C1:SUS-ITMY_ULCOIL_GAIN = 1.02482
C1:SUS-ITMY_URCOIL_GAIN = -1.06831
C1:SUS-ITMY_LLCOIL_GAIN = -0.996671
C1:SUS-ITMY_LRCOIL_GAIN = 0.91079
UL: fz = 1.014824 Hz, Q = 1.027150
UR: fz = 0.975038 Hz, Q = 0.98688
LL: fz = 1.000229 Hz, Q = 1.012378
LR: fz = 0.972688 Hz, Q = 0.946116
The coil imbalances are now about 0.8% at high frequency (i.e. above the resonant freq of the pendulum mode)
The dead laser was replaced by new JDSU 1104P of 2.6mW. The return beam is big ~5 mm diameter of 0.3 mW, 1400 counts
Whenever replacing any Oplev laser, please also put into the ELOG when it was installed so that we have an electronic record of the laser lifetime.
I tuned the bounce and roll mode bandstops for ETMY, although it was difficult for me to tell if there was improvement with the bandstops on relative to the bandstops off because it seemed like the bounce and roll modes were being excited intermittently. I'll take spectra with the filters both on and off during an evening next week.
The f2p adjustment for all the suspensions are done (except for MC1,2,3)
Last night I found that the sign of the oplev control of PITCH on ETMY was wrong. I flipped it to the correct sign.
We've been locking the Y arm by feeding a signal back to ITMY because pushing ETMY somehow made the lock unstable in the angular motion.
After the correction of the oplev contol sign, I was able to keep the lock robustly by pushing ETMY.
Today Steve was working around the 1X5 rack to strain relief the cable jungles and the jungle is now getting less jungle.
During the work he disconnected and reconnected some cables.
So for a doublecheck I checked all the suspensions to see if the suspensions are still healthy or not.
Aha, then I found a mistake.
See the pictures below. It's a very subtle difference. This wrong connection prevented MC1 and MC3 from damping.
The UL signal of the shadow sensor on ETMY went to zero this evening.
This was due to a loose connection on the cross connection board on the 1Y4 rack.
In order to make them tighten, a combination of stand-offs and screws were installed on the connectors. They won't be loose any more.
ETMY's watch dogs were found tripped. They were restored.
A new library part was made for the single suspension controller (it was originally made from the c1scx controller), using the following procedure:
Once the new sus_single_control library part was made and the library was committed to the cds_user_apps repo, I replaced all sus controller subsystems with this new part, in:
All models were rebuild, installed, and tested, and everything seems to be working fine.
Some weeks ago, Joe, Jamie, and I reworked the ETMY controls.
Today we found that the model rebuilds and BURT restores have conspired to put the SUS damping into a bad state.
1) The FM1 files in the XXSEN modules should switch the analog shadow sensor whitening. I found today that, at least on ETMY and ETMX, they do nothing. This needs to be fixed before we can use the suspensions.
2) I found all of the 3:30 and cts2um buttons OFF AGAIN. There's something certainly wrong with the way the models are being built or BURTed. All of our suspension tuning work is being lost as a consequence. We (Joe and Jamie) need to learn to use CONLOG and check that the system is not in a nonsense state after rebuilds. Just because the monitors have lights and the MEDM values are fluctuating doesn't mean that "ITS WORKING". As a rule, when someone says "it seems to work", that basically means that they have no idea if anything is working.
3) We need a way to test that the CDS system is working...
We measured the OSEM PD whitening transfer function of the ETMX OSEM UL whitening stage (D000210) by comparing the input signal to the whitening amplifier (single pin LEMO monitor) to the output signal - both were piped into the DAQ. The transfer function was close to constant 0 dB/180 deg independent of the whitening switch selection (FM1 filter engaged/disengaged) up to ~20 Hz where we run out of coherence. All other ETMX and ETMY spectra at the input of the digital whitening compensation don't change when the whitening is switched on/off so by induction we conclude that all the ETMX/ETMY OSEM PD hardware whitening filters are not on.
Joe discovered today that ETMY in fact has no binary output module at all, so there is actually no digital control of the whitening filters at ETMY.
We suspect that the ETMY binary output module was maybe harvested to put in the LSC rack, but we're not sure.
We found a spare binary output adapter pcb, which I will try to assemble into a module to install in ETMY.
This does not explain what's going on with ETMX, though. ETMX has a binary output module, that appears to be properly hooked up. I'll try to debug what's going on there as well.
In the mean time, I've removed the ETMX binary output module to use as a reference for putting together another identical module for ETMY.
I checked the state of the whitening filters for the ETMY shadow sensors.
Result : They've been OFF (i.e. flat response).
(measurement and setup)
I measured the transfer functions of the whitening board (D000210) by looking at the signal before and after the whitening stage.
The whitening board handles five signals; UL, UR, LR, LL and SD, and there are five single-pin lemo outputs for each signal on the front panel.
A good thing on those lemo monitors is that their signals are monitored before the whitening stages.
Rana suggested me to use these signals for the denominator of the transfer functions and consider the sensor signals as excitation signals.
So I plugged those signals into extra ADC channels via an AA-board and measured the transfer functions.
In the measurement the coherence above 4 Hz was quite small while the suspension was freely swinging.
Therefore I had to excite the ETMY suspension by putting random noise in a frequency band from 5 Hz to 35 Hz to obtain better coherence.
The response is flat over frequency range from ~ 0.2 Hz to ~40 Hz, see the plot below.
According to the spectrum of each signal the measurements above 10 Hz are just disturbed by the ADC noise.
If the whitening filters are ON, a pole and zero are expected to appear at 30 Hz and 3 Hz respectively according to the schematic, but no such features.
- I was investigating the SUS whitening issue.
- I could not find any suspension which can handle the input whitening switch correctly.
- I went to 1X5 rack and found that both of the two binary output boxes were turned off.
As far as I know they are pulling up the lines which are switched by the open collector outputs.
- I tried to turn on the switch. Immediately I noticed the power lamps did not work. So I need an isolated setup to investigate the situation.
- The cables are labelled. I will ask steve to remove the boxes from the rack.
- I went to 1X5 rack and found that both of the two binary output boxes were turned off.
As far as I know they are pulling up the lines which are switched by the open collector outputs.
I shut down damping to the Vertex optics and removed Binary IO Adapter chassy BO0 and BO1
About a week ago I discussed the BO0's power indicator lights with Kiwamu. They were not on or they were blinking on-off.
I put screws into ps connectors in the back, but it did not helped.
- We found the reason why some of the LEDs had no light. It was because the LEDs were blown as they were directly connected to the power supply.
The LEDs are presumably designed to be connected to a 5V supply (with internal current-limiting resistor of ~500Ohm). The too much current
with the 15V (~30mA) made the LED blown, or the life-time of them shorter.
- Jamie removed all of the BO modules and I put 800Ohm additional resister such that the resultant current is to be 12mA.
The LEDs were tested and are fine now.
- The four BO boxes for C1SUS were restored on the rack. I personally got confused what should be connected where
even though I had labeled for BO0 and BO1. I just have connected CH1-16 for BO0. The power supplies have been connected only to BO0 and BO1.
- I tested the whitening of PRM UL sensor by exciting PRM UL sensor. The transfer function told us that the pendulum response can be seen
up to 10-15Hz. When the whitening is on, I could see the change of the transfer function in that freq band. This is good.
So the main reason why I could not see theis was that the power supply for the BOs were not turned on.
- I suppose Jamie/Joe will restore all of the BO boxes on the racks tomorrow. I am going to make a test script for checking the PD whitenings.
ITMY sus damping restored.
I have installed a new binary output module in ETMY, where there was none previously. It is installed, powered (with working LEDs), hooked up (to the binary output card and the cross connect), but it hasn't been fully tested yet.
I also re-installed the binary output module in ETMX, with newly modified power-indicator LEDs.
Both modules are fully installed, but they have not yet been fully tested to confirm that they are indeed switching the whitening and de-whitening filters.
ETMX sus damping restored
Here is my work plan for this week:
1) Help Steve clean small table for experiment
2) Remove aluminum base from TT suspension
3) Mount shaker onto table base
4) Mount horizontal slider onto table base
5) Connect TT suspension, shaker, and horizontal slider
1) Begin building circuit for displacement photosensors
2) Calibrate photosensor using linear regions of power versus distance curves
3) Circuit box for photosensors?
The small optical bench (next to the MC-2 Chamber and the tool box tower) has been cleared of the misc. object previously on it, cleaned, and leveled (after much calibration X___X).
PLEASE, PLEASE, PLEASE do NOT MOVE OR HIT THE TABLE! It was incredibly painful to level.
This is how leveling the table made me feel...
VERY SAD...so do not move please!
The shaker has already been moved to the table and the amplifier for my shaking experiment is located behind the table (not on the table, as to prevent scratching).
I have made my transfer function model and posted it to the suspension wiki. Here is the link to my model!
Bode Plot Model
Please let me know if there need to be any adjustments, but I have posted the bode plots, a model image, and an explanation of why I think it's right! ^ ___^ V
I am currently working on the photo sensor circuit for the displacement detector. So far, I have gotten the infared LED to light up! ^ ___^ V
I am now trying to get a plot of forward voltage versus current for the LED. HOPEFULLY it will match the curve provided in the LED datasheet.
I'm using the bread board circuit box and when I'm not working at the bench, I have signs posted. PLEASE DO NOT REMOVE THE CONNECTIONS! It is
fine to move the bread board circuit box, but please do not disturb the connections > ____<
Here is a photo of the workspace
NOTE: The potentiometers on the bread board circuit box (the one I have been using with the signal generator, DC power, LED displays, and pulse switches) is BROKEN!
The potential across terminals 1 and 2 (also 2&3) fluctuates wildly and there dial does not affect the potential for the second potentiometer (the one with terminals 4, 5, and 6).
This has been confirmed by Koji and Jaimie. PS I didn't break it! >____<
NEVERTHELESS, using individual resistors and the 500 ohm trim resistor, I have managed to get the current versus forward voltage plot for the Hamamatsu L9337 Infared LED
We have fixed the counts-to-micron (cts2um) calibration for the suspension sensor filters. Each suspension sensor filter bank (e.g. ULSEN) has a "cts2um" calibration filter. These have now been set with the following flat gains:
40 V 10^3 um um
-------- * -------- = .36 --
2^16 cts 1.7 V ct
The INMTRX was also fixed with proper element values:
This was done for all core optic suspensions (BS, PRM, SRM, ITMX, ITMY, ETMX, ETMY).
I recorded a burt snapshot of these settings: /opt/rtcds/caltech/c1/burt/autoburt/snapshots/2011/Jun/23/21:40
I have updated the TT suspension wiki to include a new page on my transfer function model. In this new page, an introduction and analysis of my transfer function (including a comparison of the transfer functions for a flexibly- and rigidly-supported damper) are included. This page contains linear and logarithmic bode plots. Here is a link to the transfer function page.
I have also updated my photosensor page on the TT suspension wiki so that the experimental data points in my current versus voltage plot are plotted against the curve provided by the Hamamtsu data sheet. I have also included an introduction and analysis for my mini-experiment with the forward voltage and forward current of the LED. Here is link to the photsosensor page.
The PRM sus damping restored. C1:SUS-PRM_SDPD_VAR is still 20-30mV and going up. Side gain turned on. This pulled it down to 5-8 mV
Why is the side osem sensing voltage 4.4V ? It can not be higher than ~2.4V.......something is rotten in the state of Denmark?
Edit by KI:
It's because Valera increased the transimpedance gain of the PRM SIDE OSEM to match the signal level to the new ADC range (#3913 ).
Today Ishwita, Sonali, and I completed basic laser safety training with Peter King. I completed the Laser Safety Quiz and have turned in my certificate sheet.
I just need to turn in a signed copy of the Lab Safety Checklist to SFP (which I can now have signed by Koji after completing the course).
Steve and I have removed the TT mirror from the clean box. It is now on the small optical table in the lab that I have been working on. Thanks to Steve, all of the mechanical components for the horizontal displacement measurement experiment are compiled and on the small optical table. Here is a photo of the small optical table with the gathered components.
The plan is to attach the slider and the shaker directly to the black mounting plate. On the slider, we we then place the smaller black mounting plate (with the lip). The lip will attach to the shaker. We know exactly where to drill and everything is lined up. The shaker will be placed on the smaller black mounting plate (with the lip). The assembly will begin on Monday.
Here is a photo of the planned set-up for the shaker and the horizontal slider + mounting base.
I used scripts/SUS/freeswing-all.csh to give the optics a kick and then turn off their watchdogs and collect the free swinging data. Final script end time = 993173551. Start taking data ~ 993173751
I had to fix up the script a little: it had amateur stuff in there, such as undefined variables.
It still doesn't work that well. On the new Ubuntu workstations, pianosa, it fails by just not setting some of the EPICS variables using the EZCA stuff.
On Allegra, it failed on ~1 out of 10 commands by returning "epicsThreadOnce0sd epicsMutexLock failed" ???
On Pianosa, it sometimes says, instead, "epicsThreadOnceOsd: pthread_mutex_lock returned Invalid argument.". Ah...now I understand?
epicsThreadOnceOsd: pthread_mutex_lock returned Invalid argument.
So finally, I had to run the script on op340m to get it to actually run all of its commands. That's right; I used a 15 year old Solaris 9 Blade 150 because none of our fancy new Linux machines could do the job reliably.
Fixing our EZCA situation is a pretty high priority; if the locking scripts fail to run ~1 command every hour its going to completely derail the lock acquisition attempts.
If you want to use the IFO tonight, just run the script again on op340m again when you're done.
Chris Wipf tells me that the EPICS Mutex Jumbo Mumbo can be overcome by upgrading our EPICS. We should get one of Jamie's assistants to get this going on one of the Ubuntu workstations.