The weird jumps at the beginning of each TRX peak are due to the triggered switching between the Thorlabs trans PD and the QPD trans PD. Clearly we need to work on their relative normalizations. There are also little jumps after each peak as the triggering sends the signal back to the Thorlabs PD.
I was unhappy with the discontinuities between the Thorlabs and QPD versions of our transmitted light powers. I realized that in the olden days, we just used the Thorlabs PD, and we set the no-light offset in the LSC version of the TR[x,y] filter banks. However, now that we have brought the QPDs back, we are setting the dark offsets in the end suspension models, so that the signal chosen by the trigger already has its offset taken care of before we send it to the LSC model.
Anyhow, having the offsets script try to put a value in the C1:LSC-TR[x,y]_OFFSET was giving us an extra offset and then when we did the normalizations, the numbers came out all wrong. So. I have removed the C1:LSC-TR[x,y] filter banks from the offset list, since they were made redundant.
I have redone the normalizations for both arms (after running the ASS scripts). I checked by watching the _OUT16 versions of the Thorlabs and QPD diodes before the triggering happens, and as I put offsets into the LSC servos to change the transmitted power, the diodes both change in the same way. So, we'll have to see if this holds true for more than just values 0-1 the next time we lock, but hopefully it won't need changing for a while.
The IFO is being uncooperative tonight, and I have an early morning meeting, so I'm calling it a night.
Koji's filter module changes have been propagated from the Xarm to the Yarm, to CARM and to DARM. (Actually, Q overwrote the changes to Xarm on Sunday accidentally, so first he reverted those for us, and then we propagated the changes).
Today, with careful measuring, we find that for X and Y arms individually locked with the ALS, we want the gains to be +17 for the Yarm, and -17 for the Xarm (with the beatnote up-is-up convention). This puts the UGFs at 150 Hz.
We then switched over to CARM and DARM locking. We guessed that the gains should be a factor of 2 lower since we're pushing on both ETMs for DARM, and the MC2 actuator is roughly the same strength as the sum of the ETMs. In the end, after measuring the CARM and DARM loops, we find that the gains should be +7.5 for CARM, and +8.0 for DARM to set the UGFs at 150 Hz. The servo is a little bit delicate, so having too low of gain is not okay.
For some reason, we seem to be utilizing more actuator range with the new setup, so the limiters in the filter banks have been set to 11,000 (previously were 8,000), and the ALS watch script (ALSdown.py) threshold has been increased to 10,000 (previously 7,000).
When finding the IR resonances with the new scheme, we are having trouble holding lock throughout the scan. I have set the tramp for the coarse part of the scan to be 0.05 seconds (previously 0.01 seconds), which is an increase of a factor of 5 in the ramp time. This helps, but may still not be enough, since we don't always hold lock until both IR resonances are found.
Probably the most annoying thing from tonight is the fact that ETMY keeps drifting off, particularly in yaw, when locked. I don't have an explanation of why this is happening, but you can watch it happen sometimes, and the lock will be lost shortly thereafter. Definitely when we lose lock and the ETM gets kicked, it is far enough away in yaw alignment that I have to completely redo the Yarm alignment. This happens whether or not the ETMY oplevs are on.
To summarize, 3 scripts have been modified:
(1) ALSdown - threshold increased (Modification from last week - turns off the slow temp servos for the end lasers, clears histories)
(2) ALSfindIRresonance - increase ramp time
(3) Lock_ALS_CARMandDARM - final gain values set to 7.5 for CARM and 8 for DARM, no filters come on until gains all the way up, turns on new set of Koji filters. (Modification from last week - turns on the slow temperature servos for the end lasers)
New ALS servo performance
Comparison between the old (orange) and new (red). The new error signal (red) is suppressed like a white noise as expected.
Comparison between the out-of-loop evaluation (black) and the in-loop signal (red). Below 50Hz the out-of-loop is limited by the sensor-noise like something.
This out-of-loop stability was measured with the ALS stayed at the top of the resonance and calibrated the POX11 error signal.
New ALS servo with the LSC PDH signal. When the PDH signal is used for the control, FM4 is additionally used.
In this condition, the error signal was measured and calibrated into frequncy noise (Hz/sqrtHz).
By comparing the POX (with the new servo) and POY (with the old servo) signals, one can see that the new servo has better supression below 30Hz with almost no cost at ~100Hz.
The Y arm locks stably for IR PDH now.
The reason for ETMY getting kicked during lock acquisition was finally found to be related to the limiter value set in the Y arm servo. We reduced the limiter value unintentionally and found that the lock acquisition stayed smooth. The limiter value was stepped in 1000s from 7000 and eventually found that the ETMY suspension was kicked when we try to acquire lock with the limiter value was set at 11000.
The limiter for X arm at 11000 is not causing any trouble at the moment.
In the process, we did a bunch of things through the evening to troubleshoot IR locking of the Y arm.
Earlier today running the IFO configure script did not restore the arms to lock and both the ETMs needed to be aligned to lock the arms. The arms stayed locked for 15 minutes and the Y arm lost lock eventually leaving the ETMY in a misaligned state.
The state of the Y arm was similar to what Jenne has explained in ELOG where the ETMY was kicked during lock acquisition and would move to a misaligned state.
To trouble shoot, there were several things that were done. A few of them might not have any direct correlation to the locking issue but could just be a coincidence.
1. The trigger time for the filters in the arm filter modules were set such that they switch on after the SUS violin filters. Arm FM trigger time = 3 s (previously set at 0.1s) and SUS violin trigger time = 1s. This reduced the number of lock loss events.
2. There was some drop in transmission when the bounce filter of Y arm (FM6) turned ON. This was fixed by changing its ramp time (initially set at 1s). The filter has been modified to turn on immediately upon arm lock acquisition before the other triggered filters in the filter module turn on.
3. The QPD and SUS signal cables running to the rack were checked to be intact. Koji found some of them to be loose. But this had no evident correlation with the arm locking problem.
4.The oplev and PD alignment was checked at the Y end. The high gain trans PD for Y arm was checked for good alignment by looking at TRY. It was found that the EXIT light at the Y end is injecting some noise to the transmission PD.
5. The ETMY was given offsets in PIT, YAW and POS and the OSEM sensor values were checked to see if the suspension is behaving well. It was behaving well.
I. The Y arm stayed stable through last night and I have saved the arm lock settings to IFOconfigure.
II. ALS X arm noise measurements
I looked at the before and after noise of ALSX.
Phase tracker gain = 85
Xarm servo gain = -17
The rms in loop noise has dropped from 3KHz to 500 Hz.
Attachment 1: Phase tracker OLTF
Attachment 2: Free running noise and in loop noise
Attachment 3: Out of loop noise (measured with arms locked using PDH for IR)
Attachment 4: ALS arm servo OLTF
xml data files can be found in /users/manasa/data/140430/
I found the YARM LSC feedback going to MC2 and the MC2 violin mode (at 644.69 Hz) rung up. The existing notch was just a second order Twin-T style notch (so not a good idea) and also not turned on, since it was in the FM4 spot of LSC-MC2 and the vio triggers are ganged between all mirrors and don't touch FM4.
I copied the PRM vio bandstop into FM2 of this bank, deleted the old notch, and tuned the bandstop frequencies a little to get the violin peak into one of the zeros of the elliptic bandstop. Attached are the Y-arm / MCF spectrum with the mode rung up as well as the new filter's TF compared with the old notch.
P.S. I installed http://en.wikipedia.org/wiki/Midnight_Commander on pianosa.
I saw big misalignment on the GTRX camera, I went to the PSL table and aligned the beat beams.
I disconnected the RF out of the X beat PD and connect an oscilloscope.
The beat amplitude was 15mVpp at the beginning and is 60mVpp right now.
I checked the alignment on this RF PD and the DC PD as well as the spot on the CCD.
The RF cable was connected again.
Jenne and I ran the ALS and scanned the arm cavity. We had the impression that the noise level of the ALS improved,
but I don't have correctly calibrated measurement. Let's do it tomorrow in the day time.
The Yarm beat alignment look awful. We should align this too.
[Rana, EricQ, Jenne]
We locked the Yarm by using the CM board this evening.
POY is going from its demod board to the CM board, and then the slow output of that is going to the POY channel of the whitening, and then on to the ADC. So, with no AO path engaged, this is basically like regular Yarm locking.
First of all, Den and Koji back in December were concerned that they were seeing some EOM saturation in the fast path, but we don't think that's an issue. We looked at the FSS PCDRIVE while we increased the AO gain. In fact, it looks like the offset is coming from the MC board's IN2 slider. Even with no input on that slider, increasing its value puts an offset into the MC. To fix this, I am going to put a 6.8uF cap in series with R30 in the MC board, which is part of the crossbar switch where the IN1 and IN2 get summed. This should AC-couple the output of the IN2 slider before the summing node.
We aren't sure which sign to use for the AO path of the CM board...Eric is doing some modelling to see if he can figure it out. He's going to try to see which spectra (below) his model matches.
For the spectra, we have a reference trace with no AO path, a trace with "Plus" polarity on the CM board which started to show a peak when the value of the MC IN2 slider was at about -6 dB, and a trace with "Minus" polarity, which started to show a peak when the value of the MC IN2 slider was at about -16 dB.
We took loop measurements for each of the Plus and Minus cases. Something that seems a little weird is how shallow of a slope we have in both cases near our UGF.
POY is going from its demod board to the CM board, and then the slow output of that is going to the POY channel of the whitening, and then on to the ADC. So, with no AO path engaged, this is basically like regular Yarm locking.
Just to be clear, the normal POY signals are not currently present, so the restore POY script will not result in the arm locking. POY11_I is turned off, POY11_Q is the output of the CM board, which can be used to lock the arm, as we did tonight.
The POY digital demos angle went -56 -> 90, to get all of POY11_Q_IN1 to POY11_I_ERR
To fix this, I am going to put a 6.8uF cap in series with R30 in the MC board, which is part of the crossbar switch where the IN1 and IN2 get summed. This should AC-couple the output of the IN2 slider before the summing node.
MC board is out, so don't be surprised that the MC isn't locking.
MC board is back in place, MC is locked.
If I disable all of the AO path bits of the CM servo (disable switch, and also gain slider to -32dB), and then move the MC IN2 slider around, the MC does not get an offset anymore (as seen by reduced transmission and increased reflected power), so I think the DC coupling is working. I do lose lock of the MC if the slider goes above ~22 dB in this situation, but I don't see any effect before then, whereas we were able to see a steady increase in the reflected power as we moved the slider around last night. So, it seems like things are good with the DC coupling of the IN2 slider.
Here are some photos from before I modified the board (front, back, and zoom of the area I was working in):
And here is my modification, putting the 6.8uF cap in series with (a new) 2k thin film resistor, in the spot for R30:
The board is https://dcc.ligo.org/DocDB/0004/D040180/001/D040180-C.pdf
[Edit, 20140721: It looks like this is actually D040180 rev B, not rev C. —Evan]
(Edited this post; Forgot to account for the FMs other than 4 and 5... it now agrees better!)
I did some quick MATLAB simulation of the relevant loops to try and understand what was going on. I put the digital UGF around 200Hz, and then brought in the AO path with both signs.
In these plots, blue is digital only, green is AO+digital with the crossover happening at the UGF, and red is the AO gain set to five times of what it was in the green curve.
Based on the phase curves in the loop measurements, I would be inclined to say the pink -AO case corresponds to the opposite sign plot, and the +AO case to the same sign plot.
This correspondence also holds for the appearance of the peaks in the noise curves, the Opposite sign case has a dip in loop gain at ~50Hz (pink curve, -AO), same sign around ~30Hz (brown curve, +AO).
However, both of these look like they become unstable at some point in the transition! This agrees with our experience last night...
I'll fiddle around and try to come up with some compensating digital filter that will make the Opposite sign scenario work.
The MATLAB code I used to make these plots is attached.
cycleT = 60e-6;
% AI, AA shapes from http://nodus.ligo.caltech.edu:8080/40m/8555
[z,p,k] = ellip(4,4,60,2*pi*7570,'s');
AI = zpk(z,p,k*10^(4/20)) * zpk(,-2*pi*13e3,2*pi*13e3);
AI.OutputDelay = 1*cycleT;
[z,p,k] = ellip(8,0.001,80,2*pi*7570,'s');
I think that's about halfway there. Since this needs to be a precise comparison, we cannot use so many approximations.
We've got to include the digital AA and AI filters as well as the true, measured, time delay in the system. Also the measured/fitted TF of the CM board with the 79:1.6k filter engaged. We want an overall phase accuracy between Jenne's measured TF from last night and this model (i.e. on the same plot with the residual plotted).
Is there a cavity pole in the model? Should be at ~1.6 kHz.
After some more matlab loopology (see Qlog), we turned on the AO path successfully. The key was to turn on the 300:80 filter in the MCL path so that it could cross stably with the AO. Then we ramp up the AO gain via the newly AC coupled AO path into the MC servo board.
The POY11 signal looks nice and smooth. For the final smoothness after the overall common gain is ramped up, I turned on a FM7 pole at 300 Hz so that the MC path would keep falling like 1/f^2 and not interfere with the AO path around 1 kHz.
There's not enough gain yet to be able to turn on the Boost. PCDRIVE is ~3 V. Earlier tonight we were seeing the EOM saturation effect maybe, but we re-allocated the gain more to the front end and its all fine now. I think we can get another ~10-15 dB of gain by using the POY whitening gain slider + the CM AO slider. Then we can get the Boost on and take some TFs with the SR785 (as long GPIB allows).
CM REFL1 = +31 dB, AO = +16 dB, MC IN2 = +16 dB. SUS-MC2_LSC = FM6, FM& ON
** Everything has been pretty stable tonight except some occasional MC/EOM locking oscillations. This means that its been easy to keep trying some different CM steps since the Y-Arm relocks using MCL within seconds.
I touched up the alignment of the Ygreen on the PSL table.
Please check the X&Y ALS out-of-loop stability. Use fine resolution (BW0.01). Calibrate the POX/POY in Hz.
I have taken out of loop spectra for both arms, by looking at POX/POY while the arms were controlled with ALS.
To do this, I put the POY11 Q signal directly to the whitening board, which meant that I removed the cable coming from the common mode board. (Now that we're doing CM stuff again, I have put it back, so POY is still in the slightly weird "going through the CM slow path" situation).
For the locking, both arms had FMs 1, 2, 3, 5, 6 engaged. Yarm had a gain of +17, Xarm had a gain of -17.
Y beatnote was 98.6MHz with a peak height of -22 dBm. X beatnote was 45.0MHz with a peak height -11 dBm.
I drove ITMY at 503.1 Hz with 100 counts. I drove ITMX at 521.1 Hz with 25 cnt.
Koji helped me match up the peak heights between the FINE_PHASE_OUT_HZ calibrated signals and the PDH signals.
The out of loop noise is definitely below 1kHz rms now, which is better than it was! Hooray!
Today, we got a ~2kHz bandwidth lock of the YARM with the AO path. We weren't able to turn any boosts on, due to POY noise.
Rana and Koji have written scripts (/scripts/PRFPMI/cm_step and cm_down) that work very reliably.
Here is an OLTF. (Violin filter was off, the crap around 600Hz goes away with them on)
My MATLAB modeling was useful is predicting the features of the loop shape, and the dependence on AO gain/crossover. Still, I need to check it out, because there is nonzero discrepancy between reality and my model (this may be hiding in the non flat MC AO response, i.e. the bump at ~35kHz. Alternatively, the crossover frequency is a free parameter...)
In any case, we have confidence that the CM board is mostly working predictably. We presume that our current obstacle is the very noisy nature of POY, and thus it's not worth spending more time in this configuration.
RXA: we should also use AS45 instead of POY11. It has better SNR and I think our whole problem is too little light on POY.
Today I finally implemented a feature in the whitening triggering that I should have a long time ago: an override button.
Now, on each RFPD's phase rotation screen, there is a button to either allow triggering for that PD (both quads) or to be in manual mode.
If you are allowing triggering, they will behave as they have for the last ~year.....if any degree of freedom is using either quadrant of that PD, and that degree of freedom is triggered, then engage analog whitening and digital de-whitening.
If you chose manual mode, then you can engage or disengage the whitening as you please. (The analog whitening and digital de-whitening are still tied together).
This evening, we were able to lock the Yarm through the common mode board, using AS55 as our error signal. Our final UGF is about 5kHz, with 60 degrees of phase margin.
Before dinner, Rana switched the input of the CM board's REFL1 input to be AS55I rather than POY11Q, in the hopes that it would have better SNR. Demod phase of AS55 was measured to be 14 deg for optimum Yarm->I-phase and has been set to 0 degrees. Since the POY demod phase had been 90 degrees, which puts in a minus sign, and now we're using 0 deg which doesn't have a minus sign, we're using the plus (instead of minus) polarity of the CM board.
We re-allocated gains to help lower the overall noise by moving 15dB from the CM board AO gain slider to the MC IN2 gain slider, so we weren't attenuating signals.
We see, by taking loop measurements even before engaging the AO path (so, just the digital loop portion) that we've gained something like 20 degrees of phase margin! We think that about 5 degrees is some LSC loop re-shaping of the boost filter. We weren't sure why there was a hump of extra gain in the boost filter, so we've created a new (FM8) boost filter which is just a usual resonant gain: resgain(16.5,7,50)
The cm_down and cm_step scripts in ..../scripts/PRFPMI/ were modified to reflect the settings below, and their current states are included in the tarball attached.
Also, throughout our endeavors this evening, the PC fast rms has stayed nice and low, so we don't suspect any EOM saturation issues.
Now our Yarm digital servo has a gain of -0.0013, with FMs 2, 4, 5, 7, 8 engaged (2, 7, 8 are triggered).
Our final CM board settings are:
REFL1 gain = +22dB
offset = -2.898V
Boost = enable
Super Boost = 0
option = disable
1.6k:79 coupled cavity compensator = enabled
polarity = plus
AO gain = 15dB
limiter = enable
MC board: IN1 gain = 18dB, IN2 gain = 0dB.
Here is a measurement of the Common Mode MCL/AO crossover. The purple/orange trace here is after/before the boost was engaged.
We also have a measurement of the total loop gain, measured with the SR785. The parameter file, as well as the python script to get the data, are in the tarball attached. Noteably, the excitation amplitude was 500mV, whereas Q and Rana yesterday were using 5 or 8 mV. We aren't sure why the big change was necessary to get a reasonable measurement out. This measurement is with the boost enabled.
Finally, here is a measurement of the MC error point spectra, with the CM boost on, after we reallocated the gains. There's a giant bump at several tens of kHz. We need to actually go out with the fast analyzer and tune up the MC loop.
Yes, we still need to do these things, day team. Please tune up the MC loop first, before anything else.
I took a look at the MC OLTF and AO path TFs with the fast agilent analyzer.
I played with the relative gain of the EOM and PZT, but couldn't really change the MC OLTF shape much without making the PC Drive RMS angry.
However, it turns out we have plenty of phase headroom to up the MC UGF from ~100kHz to ~180, with about 40 degrees of phase margin and ~7dB of gain margin. As I write this, PC drive RMS is around 1.1, and FSS Fast at 5.6, so I think the extra gain is fine for now.
This pushes up and smoothens out the gain peaking in the AO path; see this figure:
(why does ELOG hate my python plots?! argggg)
Rana's rule of thumb was "We need at least +3dB MC loop gain at our CM servo UGF," so it looks like high tens of kHz bandwidth may be doable from the AO standpoint.
RXA: No, no, no, no, no, noooo. Rana said we need a gain of 3-10 at the CM UGF, not +3 dB.
This is an effort to get rid of our ground loops by isolating the electronic components from the optical table.
Aluminum mounting base plates of Thorlabs BA2 and Newport B-2 were replaced by plates or post made out of delrin material.
This is an insulator. DELRIN base plates were installed 6 places. The oplev-qpd has Nylon base plate.
The NPRO and HE/NE lasers are not isolated from the table. S8 and S9
I'm not sure about the doubling oven S10
The optical table is grounded at G11 through ~1 Mohms to the ETMY chamber.
Alignment touch up needed at all D-marked component!
To get ready for the PRFPMI CM trials tonight, I put AS55's cables back to their nominal state, and now have REFL11 I going to IN1 of the CM board. The OUT1 of the CM board goes to the REFL11I whitening channel.
REFLDC was not disconnected in the last few days, so it is still set up for IN2 of the CM board, with an external offset adjust.
Also, we found the PRM OL off and turned it back on. The ETMY was swinging a lot after lock loss, so we set its SUSPOS damping gain to match the ETMX and it stopped swinging so much.
Next up: more of the same, make this sequence more stable, turn on CARM OSC and watch the LOCKI outputs while we slowly ramp between signals.
Also, what should be the sign of the CARM offset ???
D: component delrin isolated
N: component nylon isolated ( or Delrin )
S: component shell is shorting to optical table (except oven)
G: optical table ground
I failed to maximize TRY the pds.
I've checked that the 2pin lemo connector that was run some time ago from the LSC rack to the MC board does indeed transmit signals. To try and evaluate its suitability I did the following:
No real difference was seen between the two cases. The signal peak was the same height, width. 60Hz and harmonics were of the same amplitude. Here are the spectra out to 200k, they are very similar.
Mode cleaner was locked during this whole thing. This may interfere with the measurement, but is similar to the use case for the AO path. If ground loop / spurious noise issues keep occurring, it will be worthwhile to examine the noise of the CM and MC servo paths, inputs and outputs more carefully.
Last night after checking cabling and turning on ISS, we tried several times to handoff to REFL_DC but it didn't work at all.
I worked at the ETMY-ISCT this morning and late afternoon. I will continue the 60 Hz noise hunt tomorrow.
As part of checking the common mode board before we get too carried away with using it, I looked at the time series of the AO servo output when I turned on various boosts, or changed gain values. As it turns out, basically anything that I did caused glitches. Oooops.
I plugged a function generator to the IN1 port of the CM board, with a freq of 400Hz, and a voltage of 10mVpp (which is the smallest value that it would allow). I plugged the BNC version of the servo output into a 300MHz 'scope.
First I looked at "boost" and "super boost", and then I looked at various steps of the AO gain slider. For all of the button presses that gave me glitches, I saved .png's of the 'scope screen (on a floppy, so I'll have to fetch the data tomorrow...).
Both enabling, and disabling the "Boost" button gave me glitches.
For "Super Boost", I saw glitches for all of the steps, 0->1, 1->2, 2->3.
For the AO path, I only started at 0dB, and only captured screenshots of glitches when I increased the gain, since presumably that's when we'll care the most during acquisition. I found that going down in gain caused glitches at every step! For increasing the gain, steps from an odd number of dBs to an even number consistently caused glitches. Steps from an even number to an odd number occasionally caused glitches, but they weren't very common. For the steps that did cause glitches, some were worse than others (7dB to 8dB, 15 dB to 16 dB, and 23 dB to 24 dB seemed the worst.)
After my work, I put all of the cables back, so that we should be ready to utilize the CM board for locking this evening.
For posterity, here are the notes that I took while I was working - I'll make them more coherent when I fold them in with my images tomorrow. The "first .png, next, etc." are because the 'scope numbers them in order as a default.
1st png = boost enable, then disable
2nd png = super boost, start at 0, then 1, then 2, then 3
3rd png = AO gain from 1 to 0
4th is AO gain from 0 to 1 (happens less often than 1->0, which is every time I get a glitch)
Next is AO gain 1->2, got 2 glitches!
3->2 glitch often, 2->3 much less often
next is 2->3
next png is 3->4, 2 glitches with weird dip
next png is 5->6
6->7 is rare
next png is 7->8, which is nasty!!
8->9 is rare
10->11 is rare
png 11-> 12, 3 glitches
png 13->14, 2 glitches
png 15->16, kind of nasty
png 17->18, 2 glitches
png 19->20, 3 glitches
png 21->22, 2 glitches
png 23->24, kind of nasty
png 25->26, 2 glitches
png 27->28, 3 glitches, at least
png 29->30, 2 glitches
Somehow, the images got put into a whole new entry, even though I thought I was editing this one. Anyhow, please see elog 9938.
Still no success tonight
Note: I thought I was editing elog 9935, but somehow this became a whole new entry. Either way, all the info is in here.
The screenshot of the Boost enable / disable I'll have to re-take. Apparently I instead caught a screenshot of the list of files on the floppy...ooops.
This is a shot of enabling the Super Boosts. At the beginning, it's at "0", so no superboosts (also, regular boost was off). Then, I switch to "1", and the trace gets a little fuzzy. Then I switch to "2", and it gets very fuzzy. Then I switch to "3", and a lot of the fuzz goes away. There's a glitch at each transition.
The following screenshots are all of various steps of the AO gain slider. For all of these, both the "boost" and "super boosts" were off. Each screenshot is a single gain step, even if there are several glitches captured.
First, 0dB to 1dB:
Next, 1dB to 2dB:
2dB to 3dB:
3dB to 4dB:
While increasing the gain, I didn't find any more steps from an even to an odd number where I got a glitch. They would glitch when I undid that step (decreased the gain), but over ~5 trials for each increase, I didn't ever catch a glitch. The odd to even steps still had glitches while increasing the gain though.
5dB to 6dB:
7dB to 8dB:
9dB to 10dB:
11dB to 12dB:
13dB to 14dB:
15dB to 16dB:
17dB to 18dB:
19dB to 20dB:
21dB to 22dB:
23dB to 24dB:
25dB to 26dB:
27dB to 28dB:
29dB to 30dB:
[Jenne, Rana, EricQ]
We tried a few times to engage the AO path while holding CARM on sqrtTrans and DARM on ALS, but failed every time. Since we cannot stably hold lock at arm powers of 1, even though we were able to do so early last week, we think that we have a problem (obviously). One noticeable thing is that while held with ALS, the Xarm transmission fluctuates almost full power.
As we were seeing late last week, the Xarm IR transmission while held with ALS was fluctuating wildly, whether we were locked on individual arms or on CARM/DARM.
Tonight we took some out of loop spectra, with different HEPA settings, to see how that affected things. It looks like HEPA at the nominal ~20% is okay, but anything higher than that starts to affect the Xarm beatnote sensing, while it mostly leaves the Yarm beatnote sensing okay. Perhaps this is something that isn't tightened enough in the Xarm path, or something on a skinny floppy mount that needs to be more secure.
I am still a little confused though, why we don't see large power fluctuations in *both* arms while using ALS to control CARM/DARM. Why are we not seeing this Xarm noise being fed back into the Yarm, through either the ETM via DARM, or common stuff via CARM actuating on MC2.
Note that the change at high frequency is because I switched from using non-DQ channels to DQ channels, so that's not anything to pay attention to. The noise reduction we see is below about 20Hz.
Rana pointed out that our POPDC level was very small. We don't have screens for them, but the DC PDs have the same analog whitening as the RF PD signals do. I changed C1:LSC-POPDC_WhiteGain.VAL from 0 to 11. Now our POPDC while locked on sidebands is about 8,000 counts.
We also swapped the cables between the SR785 and the CM board around. Now channel 2 of the 785 goes to TP2A, channel 1 goes to TP2B, and the source goes to EXCB. This is so that we can break the AO loop with the disable switch just after the slow/fast split, and look at the transfer function before we close the loop.
Tonight was mostly cleaning up some scripts, including the re-writing the restore and align scripts for the optics.
The new script is in the same folder as the old one (/opt/rtcds/caltech/c1/medm/MISC/ifoalign/NewAlignSoft.py), but is not yet called from the align screen. However, I'm using it in the carm up and down scripts, and it works nicely for the PRM. I need to check that the offset value is okay for all the other optics (i.e. are they getting misaligned enough?), but then I'll have the new script called from the screen. The new script, per Rana's suggestion, does not touch the bias sliders. Rather, it puts an offset in the pitch filter banks in the coil driver output matrix-of-filter-banks. Then the misalign routine turns the offset on, and the restore routine turns the offset off. This way we have a nice ramp time, but don't have to do the weird calculation of number of steps to take as is done in the current script. Also, the "save" functionality will be obsolete, since we're never touching the bias sliders except for actual alignment needs.
I'm not sure what changed, other than the HEPA being on lower, but the Xarm ALS was much better behaved tonight. I was able to hang out around arm powers of ~1 for as long as I wanted.
I didn't try to hand over to digital REFLDC, but I was trying a few times to engage the AO path. With the CM board set to Plus, I hear hooting when MC IN2 is about 4dB. With the CM board set to Minus, I didn't hear hooting, but I lost lock when I went from 13dB to 14dB.
Also, I put the cables for the SR785 back to the "A" set of test points and excitation, so that I could take a closed loop transfer function. However, I don't know where the latest working scripts to make a remote measurement are, so maybe we can take some loop measurements tomorrow.
The carm_cm_up script is good (for tonight) up to the prompt "Press enter to indicate that it is okay to turn on MC2 LSC FM8". There are "read"s every step of the way, so it goes nice and slow, but it'll do everything for you except any last tweaks of the PRM alignment after the PRMI is locked.
I tried many times this evening to engage the AO path, with limited success.
Q's new scripts worked really well, and so I have some transfer functions! To take these measurements, in ...../scripts/general/netgpib, I am running ./TFSR785 TFSR785_CARMloop_May2014.yml, where the file name is the name of my parameter file. The data, and the saved pdfs, are in /users/jenne/PRFPMI/CARM_loop_measurements/2014May14/ . For these measurements, the SR785 is hooked up to the "A" set of excitation and test points on the CM board.
All of these traces were taken while the IFO was PRFPMI, with PRCL and MICH on REFL33, DARM on ALS diff, and CARM on InvSqrtTrans. carm_cm_up.sh is up to date, through the echo "REFL_I should now be zero" (~line 111 in the script). All you need to do is set the beatnotes, and then run the script. Follow instructions in the prompt (such as "press enter to confirm PRMI is locked").
Here are my notes for the various times:
23:01:44 - MC IN2 = 0dB, CARM gain = 5.0
23:13:45 - MC IN2 = 10dB, CARM gain = 5
23:26:10 - MC IN2 = 6dB, CARM gain = 10 (after Q suggested increasing overall gain, rather than just AO path)
00:13:07 - MC IN2 = 6dB?, CARM = 6ish? don't remember exactly.
00:45:00ish, Realigned IFO using IR with arms.
01:03:17 - MC In2 = 0dB, CARM gain = 5
01:07:42 - MC IN2 = 8dB, CARM gain = 6.295 (AO went up to 6dB, then +1dB steps to both simultaneously using ezcastep C1:LSC-CARM_GAIN 1dB C1:IOO-MC_AO_GAIN 1)
ezcastep C1:LSC-CARM_GAIN 1dB C1:IOO-MC_AO_GAIN 1
01:08:57 - MC IN2 = 10dB, CARM gain = 7.92447
01:10:08 - MC IN2 = 12dB, CARM gain = 9.97631
lockloss when trying to add 1 more dB to both.
01:41:36 - MC IN2 = 12dB, CARM gain = 9.97
lockloss when just MC IN2 up by 1dB, left CARM gain alone.
The 60Hz noise in TRY is back. Since I thought I remembered someone suggesting that it was leakage light from the exit sign, Koji went in and wrapped the end table in foil, however the lines are still present.
In addition to a transparent legend, we need the corresponding CM crossover measurements from DTT to compare with the Q-Mist-Loop model results. The xover tells us when the AO gain is high enough so that they can be ramped up together.
Also, I wonder how much power fluctuation we get from the large ALS DIFF noise and if that demands we get the TR signals normalized by POPDC.
It's taken a lot of trial and error, but I've found a path through MATLAB loops that seems like it may be stable at all points.
CAVEAT: This doesn't give any indication as to why we weren't able to turn up the AO gain more last night, as far as I can tell, so it's not all good news.
However, it's still ok to at least have a plan that works in simulation...
Based on the location of the optical resonance peak in the CARM plant, we estimate our CARM offset to be 200pm. I haven't simulated TFs there exactly, but do have 100pm and 300pm TFs. This procedure works in MATLAB starting at either, though 100pm is a little nicer than 300. MATLAB data and code is attached in a zip.
The steps below correspond to the attached figures: Bode plots and step response of the Loop at each step.
0. [Not Plotted] DCTrans sensing, MCL actuation on CARM. FMs1,2,3,5,8; UGF = 120. (DARM not considered at all)
#4 Seems like the most sticky part. While both sides of this look stable as far as I can tell. I feel that flipping from the red phase curve to the teal might not actually be ok, since they are on either side of the bad phase of 0 degrees. It isn't immediately evident to me how to easily model the transitions between steps, rather than just the stability of of each step in the steady state.
The qpd was removed from the east end table and threaded adaptor ring epoxied on it's shell.
This tube will cut down the amount of emergency exit light getting into the qpd.
Den and I spent some time with the interferometer last night with hopes of bringing in the AO path, but were stymied by the (re)occurrence of the anomalously high low frequency motion of the Xarm, as seen by fluctuations of TRX from .9 to .2 while "held" on resonance.
Jenne reported that they weren't seeing it earlier in the evening, and then it started again when I showed up. Holding the arms on IR, we could see a fair amount of excess low frequency noise in the BEATX_FINE_PHASE_OUT_HZ channel, as compared to BEATY, bringing its RMS to 5 times that of the Y arm. From the shape of the excess noise (broad slope from DC to tens of Hz), Rana suspected air currents and/or scattering effects being the culprit.
Den poked around a bit on the PSL table, which didn't really change much. He then went down to the X end table to inspect the table, and while he was there, I noticed the noise go down to being in line with the Yarm. I joined him at the end, and we found the beat phase noise in the frequency region of concern to be hugely sensitive to tapping on the enclosure, air current, etc. There is also a ton of green light everywhere, and multiple spots of green light around the green refl PD.
At that point however, the quiescent noise was acceptable (TRX fluctuations of <.2), so we went back to the control room to try to lock. Unfortunately, after a few attempts, the noise was back. At this point, we went home. The layout of the end table likely needs some attention to try and minimize our susceptibility to excess scatter effects.
Turn off the AC and flow bench please.
Electronic components should be ISOLATED as they are installed on the optical tables.
This is essential to avoid ground loops, 60 Hz and harmonic peaks in the spectrum. We have just got some made.
Please only use it for this reason. Earlier black delrin base plates were used up in not needed places.
The anodized Aluminum base plates with magenets certainly will conduct.
We noticed that PRMI RIN is high when the cavity is locked on RELF33 I&Q signals. We compared the level of power fluctuations when PRMI was locked using REFL11, REFL55 and REFL 33. Attached plot "prmi_rin" shows the spectra.
Then we excited PRM and measured length to RIN coupling when PRMI was locked on REFL33 I&Q. DC offset of PRCL is only 3%. But MICH offset seems to be ~nm. When we gave offset of -15 cnts to the servo, power fluctuations improved by a factor of few.
Then we looked at DRMI. It seems that SRC macroscopic length is off but we still could lock it stably. To account for macroscopic length detuning we had to rotate REFL55 phase from 25 degrees to 50 degrees. Power at AS port increased by factor of ~2 compared to PRMI configuration. SPOP18 is decreased only by 30%. Attached plot "drmi_power" shows POP18, POP90, POPDC and ASDC in PRMI and DRMI configurations.
Plot "src_ol" shows srcl OL transfer function. UGF is 70 Hz. We have also centered SRM OL and copied the servo filters from PRM, gains are set to keep UGF at ~0.1 Hz and 7 Hz
This is a more detailed procedure:
1. Phase: REFL11: 19 degree, REFL55: 50 degrees (25 degrees for PRMI configuration)
2. Input matrix:
3. Servo parameters:
PRCL: gain = -0.02, FM4,5 + trigger FM2,3,6,9
MICH: gain = 0.8, FM4,5 + trigger FM2
SRCL: gain = 0.25, FM4,5 + trigger FM2,3
signal is SPOP22 , levels 40:25
The daytime crew had noticed that there were some ETMX angular shifts happening without any control or intention.
I reseated and strain relieved the bias cable coming into the backplane of the coil driver and now it seems OK.
In the 4-hour-long second trend plot below, the era before 2300 is before reseating. Afterwards, we make a couple adjustments, but so far there has been no un-asked for alignment shifts.
AdS has been run on both arms and offsets saved. Its locked on green/red and the beat frequencies are low and the amplitudes high.
Sun May 18 23:53:37 2014: still OK...I declare it fixed.
We have looked at a few things that do and don't affect the out of loop noise of the ALS X beat, and found that cavity alignment and beatnote RF frequency had the strongest effects.
Possible causes of noise:
1. Air currents from A/C or flowbench. No effect.
* When table lid is on, turning on and off the flow bench air did not qualitatively change the out of loop beatnote time series signal.
2. Scattered light from other beams hitting green PDH PD. No effect.
* There are a few spots of green light that are hitting the case of the PDH photodiode, but when I put an iris in place to block those spots, there was no change in the beatnote spectra. This makes sense to me since none of those spots were close to hitting the diode itself.
* Rana did notice that the beam was not well centered on the PD, so he steered the beam onto the center of the diode. Also, the PD is now tilted a little bit so that the reflection from the diode doesn't go back into the beam path. Neither of these things had an effect that we noticed in the beatnote noise.
3. Oplev laser light getting to PDH PD. Not tested.
* We don't see any red light over by the PDH PD, so we did not try turning off the oplev's laser to see if that had an effect, but we suspect that it is not the cause of our noise.
4. Clipping of main IR / green beam on Xend table. Not tested.
* We should still go have a look at this, but we no longer think that this is the main cause of the elevated noise.
5. Scattered light all over Xend table. Not tested.
* We should still work on dumping extraneous beams on the table, but we do not think that this is the main cause of the elevated noise.
* Rana took some photos so that we can see how truely bad the situation is.
6. Amplitude modulation dip in NPRO. Not tested.
* It is probably still a good idea to check this, in case the dip in the amplitude modulation has changed over the year or two since it was last measured, but we also don't think that this is the main problem.
7. Check PDH servo. Not done.
* I think this is still on Q's long-term todo list, but we should give the PDH servos a once-over.
8. Arm cavity longitudinal motion. No effect.
* While the Xarm was locked with IR, we put a line at 1.7 Hz with 325 counts into the ITMX position. To keep lock, the ETM had to move as well. When we turned on this line (and increased its amplitude up to the final value of 325 cts), we did not see any qualitative change in the beatnote time series noise.
9. Arm cavity alignment. Significant DC effect.
* When the alignment of one of the arm cavity mirrors is changed, the DC value of the beatnote signal changes.
* ITMX moved in yaw, we see a 7kHz/15urad DC shift in the BEATX_FINE_PHASE_OUT_HZ time series.
* ETMX moved in yaw, we see an 8kHz/5.5urad DC shift in the time series. We aren't sure why this is about a factor of 3 times larger effect (same shift for smaller misalignment) than the ITM.
* We want to do a Yuta-style analysis to see what the angle to length coupling looks like, so that we can measure the angular motion of our cavity mirrors and put the expected noise into our ALS noise budget. Perhaps this will help us understand the low frequency difference between our in-loop beatnote error signal and our in-loop PDH error signals (red vs. maroon on the ALS noise budget posted above Pianosa).
* I've asked Manasa to take some transfer functions in the morning, so that we can start to have an idea of what is going on with this.
10. Beatnote RF frequency. Significant broadband effect.
* We have found that when the Xarm beatnote is at low RF frequencies, the noise is high, and when the beatnote is at high RF frequencies, the noise is low!
* Low RF freqs are below about 40 MHz, while high RF freqs are above about 90 MHz. This has not been tested for the Yarm. Also, these are for the case of "temp slider up, beatnote up". I have not checked if the same is true for the other side of the PSL frequency, although I don't have reason to believe that it would be.
* Maybe we are saturating some amplifiers? We need to check this out. One thought that Den mentioned was the harmonics, and that perhaps they are causing trouble in the electronics.
* Den is going to think about implementing a frequency divider so that we can directly digitize the beatnote signal.
* Here are spectra for different cases:
* And here is a spectrogram showing us going back and forth between the high and low noise states:
* A: First noticing that noise is good when RF frequency is high.
* B: Not locked on TEM00 mode, so extra noisy. Disregard.
* C: Bad noise time. Xbeat was 21 MHz (dark purple on DTT spectrum above), Ybeat was 118 MHz (sea green on DTT spectrum above).
* D: Good noise time. Xbeat was 89 MHz (light purple on DTT plot), Ybeat was still 118MHz (turquoise on DTT plot).
* E: Bad noise time. Xbeat was 37.5 MHz, Ybeat was still 118 MHz.
* F: Good noise time. Xbeat was 113 MHz, Ybeat was still 118 MHz.
So, I really should have done this as soon as Manasa measured the arm lengths... I've updated my MIST model with the real arm lengths, but still am using assumed identical losses of 75ppm on each mirror. (I've tried measuring the arm losses for real, but got numbers in the hundreds of ppms, so I need to reexamine things...)
Here's a simulation of the fields in a perfectly locked PRC when CARM is swept (Normalized to input power = 1).
More importantly, here's the latest simulation of MICH vs. PRCL demodulation angle separation in the 3F signals. It seems that we may be getting burned by using REFL33 for the PRC lock. REFL165, on the other hand looks much more robust. We should try this out.
(Some of my previous simulations incorrectly implemented MICH excitations; I only moved the ITMS, not the ETMS along with them, so some other stuff slipped in... )
Here's the angles of MICH and PRCL from the my earlier plot by themselves; this shows that the individual demod angles in REFL165 aren't changing much either.
Since Q has found that REFL165 will be better for holding the PRMI while we reduce the CARM offset, I had a look at locking PRMI sideband locking with both 3f PDs.
I checked the REFL165 demod phase, and changed it from -142.5 deg to -138.5 deg. to minimize the Q signal while driving PRM length.
I found that keeping the MICH and PRCL loop gains the same, and using matrix elements +0.1 for both I and Q for REFL165, rather than +1 for both I and Q for REFL 33.
MICH gain is +0.8, PRCL gain is -0.02. FMs 4,5 on for both, FM 2 triggered for MICH, FMs 2,3,6 triggered for PRCL.
I then locked the PRMI on sideband with REFL 33 and then REFL 165, and measured the other one as an out of loop sensor of the motion. I find that REFL33 and 165 are both comparable, and so we shouldn't have any trouble using REFL165 for locking.
We spent some time tonight looking at locking the PRMI with REFL165 vs. REFL33, while reducing the CARM offset.
We were not able to lock the PRMI on REFL165 I&Q at small CARM offsets. When locking at larger CARM offsets (about 100 counts, which is about 100nm) and then re-adjusting the REFL165 demod phase as I reduced the CARM offset, I saw that I had to significantly rotate the phase. For PRMI only (no arms), the REFL165 demod phase was -138.5 deg. When the PRMI was locked with a -100 count CARM offset, the optimal demod phase was -123 deg. Then at -90 counts the phase was -113 deg. At -70 counts, the phase was -108 deg, at -50 counts it was -98 deg, and at -40 it was -93 deg. We want to go back and look at these more carefully, and in a more continuous way, by watching the sensing matrix calibration lines. It's unclear to me right now why we're seeing this, but it's possible that we're getting some kind of extra 55MHz resonances.
REFL DC looks like it should be good - same slope and gain as sqrtTR, extra 20 or 30 deg of phase margin, so we think that we should be able to transition over to it, and then try engaging the AO path. Tonight we had Den's new 1kHz lowpass engaged, and with this, everything looks nice and stable.
Game plan: Bring CARM in until transmissions are at about 10ish, then try keeping CARM on sqrtInvTrans for the DC part, and engage the AC AO part with REFL DC. We probably just need to try this for a while more to find just the right way to turn it on.
Need to think about demod phase rotation vs CARM offset as well as extra resonances, but this may take a while, and if we can just get the AO path engaged, that would be good.
Attached is the measurement of the transfer function from ITMX oplev error in yaw to the ALSX error signal.
The arm was locked to the IR using POX and the green beat frequency (between X arm trans in green and PSL green) in this case was 27MHz.
The transfer function looks mostly flat between 1Hz - 30Hz at 700-800 Hz/urad. The DC shift that Jenne measured from the time series is ~500 Hz/urad.
So far I have not been able to measure the TF below 1Hz without the arm losing its lock. Updates will follow.
Data xml file can be found in /users/manasa/data/140521/