A small rat / large mouse just ran through the control room. Ugh.
I'll finish up the beat / frequency noise parts of the diagnosis tomorrow later, but I've done some investigation of the AUX X laser RIN.
I placed a PDA255 at one of the rejected beams from the PBS on the downstream side of the IR faraday, making sure the power didn't saturate the PD. I measured the RIN on a SR785, and simultaneously looked at the signal on a 100MHz scope.
The RIN has a very strong dependence on the laser diode current, and no noticable dependence on the crystal temperature or the presence of the PDH modulation / temperature control cables. Here are some traces, note that "nominal" current up until recently was 2.0A.
When adjusting the diode current, a peak beings to appear in the tens of kHz, eventually noticible in the DC power trace on the scope. The point at which this occurs is not fixed.
At all times, I saw a strong intensity fluctuation at around 380-400kHz on the scope whose amplitude fluctuated a fair amount (at least 75mVrms over Vdc=6.5V, but would often be 2 or 3 times that).
I didn't look at the frequency noise while doing this, because the WiFi at the X end was too slow, I'll do more tomorrow in the daytime.
We set out to lock a marconi to the IR fiber beat of PSL + AUX X to measure some frequency noise, and failed.
In short, the Marconi's 1.6MHz max external FM isn't enough oomph to stabilize the PLL error signal. It's actually evident on the Agilent that the beat moves around a few times more than that, which I should've noticed sooner... We could briefly "lock" the PLL for a few tenths of a second, but weren't able to get a spectrum from this.
We also tried using the digital phase tracker temperature servo for some help at ~DC; this worked to the extent that we didn't have to twiddle the Marconi carrier frequency to stay on top of the fringes as the beat wandered, but it didn't otherwise stabilize the beat enough to make a difference in locking the PLL.
I suppose one more thing to try is to lock the PSL laser itself to each AUX laser in turn via PLL, and look for different / excess noise.
The Green and IR beat electronics are a in a little bit of disarray at the moment, but it's not like anyone else is going to be using them for the time being...
Turning on the MCL path (in addition to the MCL FF we always have on) let me lock the PLL for multiple seconds, but low frequency excursions still break it in the end. I was able to briefly observe a level of ~50Hz/rtHz at 1kHz, which may or may not be real. Tomorrow we'll send the PLL control signal to MC2, which should lock it up just fine and give us time to twiddle laser diode current, measure the PLL loop shape, etc.
The new wifi router, a Netgear R6400, has been installed, next to the old one which is disconnected (but not yet removed).
Same SSID, and I've added only the wireless MAC addresses of viviana, paoloa and asia, the three thinkpads inside.
Qualitatively, dataviewer at the X end seems pretty snappy. I'll do some more quantitative comparison of the two routers at some point soon. I will update the wiki, too.
Brief summary of tonights work:
Our "requirement" for the end laser is as follows: We expect to (and have in the past) achieved ALS sensitivity of 1Hz/rtHz at 100 Hz. If the end PDH loop is 1/f from 100Hz-10kHz, then we have 40dB of supression at 100Hz, meaning the free running AUX laser noise should be no more than 100Hz/rtHz at 100Hz.
So, if we expect both the PSL and AUX lasers to have this performance when free running, we would get the green curve below. We do not.
I'll post more details about the exact currents, temperatures and include calibrated plots for the >30kHz range later. Here's the OLG for kicks.
The puzzle continues...
I found some reference for computing "multicoherence," which should properly estimate the potential MISO subtraction potential in situations where the witness channels themselves have nontrivial coherence. Specifically, I followed the derivations in LIGO-P990002. The underlying math is related to principal component analysis (PCA) or gram-schmidt orthogonalization.
This produced the following results, wherein the Wiener subtraction is still below what the coherences predict.
I've attached the data and code that produced this plot.
Here is some of the promised data. As mentioned, changing diode current and crystal temperature didn't have much effect on the frequency noise spectrum; but the spectrum itself does seem too high for our needs.
At each temperature, we started measuring the spectrum at 1.8A, and stepped the current up, hoping to reach 2.0 A.
At 47.5 C, we were able to scan the current from 1.8 to 2.0 A without much problem. At 49.0C, the laser mode would hop away above 1.95A. At 50.4C it would hop away above 1.85A. The spectra were not seen to change when physically disconnecting the PZT actuation BNC from the rear of the laser.
The flattening out at the upper end is likely due to the SR560 output noise. I foolishly neglected to record the output spectrum of it, but with the marconi external modulation set to 3.2MHz/V, the few Hz/rtHz above 20k translates to a signal on the order of uV/rtHz, which seems reasonable.
Data and code attached.
Here are some results from measuring the PSL / AUX Y beat.
With the Y end laser, I was able to lock the PLL with a lower actuation range (1.6MHz/V), and with the PSL in both the free-running and MCL locked configurations. (In the latter, I had to do a bit of human-turning-knob servo to keep the control signal from running away). I also took a spectrum with the marconi detuned from the beat frequency, to estimate the noise from the PD+mixer+SR560.
It looks like the AUX X laser is about 3 times noisier than the Y, though the Y laser looks more like a 10^5 noise-frequency product, whereas I thought we needed 10^4.
Gautam is investigating the PSL / AUX PSL beat with Koji's setup now.
We checked the UGF of the AUX X PDH servo, found a ~6kHz UGF with ~45 degree phase margin, with the gain dial maxed out at 10.0. Laser current is at 1.90, direct IR output is ~300mW.
We recovered ALS readout of IR-locked arms. While the GTRX seemed low, after touching up the beam alignment, the DFD was reporting a healthy amount of signal. ALSY was perfectly nominal.
ALSX was a good deal higher than usual. Furthermore, there's a weird shape around ~1kHz that I can't explain at this point. It's present in both the IR and green beats. I don't suspect the DFD electronics, because the Y beat came through fine. The peak has moderate coherence with the AUX X PDH error signal (0.5 or so), but the shape of the PDH error signal is mostly smooth in the band in which the phase tracker output is wonky, but a hint of the bump is present.
Turning the PDH loop gain down increases the power spectrum of the error signal, obviously, but also smoothens out the phase tracker output. The PDH error signal spectrum in the G=10 case via DTT is drowning in ADC noise a bit, so we grabbed it's spectrum with the SR785 (attachment #2, ASD in V/rtHz), to show the smoothness thereof.
Finally, we took the X PDH box to the Y end to see how ALSY would perform, to see if the box was to blame. Right off the bat, when examining the spectrum of error signal with the X box, we see many large peaks in the tens of kHz, which are not present at the same gain with the Y PDH box. Some opamp oscillation shenanigans may be afoot... BUUUUUT: when swapping the Y PDH box into the X PDH setup, the ~1kHz bump is identical. ugh
The anticlimatic resolution to my subtraction confusion: Spectral leakage around 1Hz. Increasing the FFT length to 256 sec now shows that the FIR WF pretty much achieves the ideal subtraction.
If nothing else, it's good to have worked out how MISO coherence works.
We gave DRFPMI locking a shot, with the ALS out-of-loop noises as attached. I figured the ALSX noise might be tolerable.
After the usual alignment pains, we got to DRMI holding while buzzing around resonance. Recall that we have not locked since Koji's repair of the LO levels in the IMC loop, so the proper AO gains are a little up in the air right now. There were hopeful indications of arm powers stabilizing, but we were not able to make it stick yet. This is perhaps consistent with the ALSX noise making things harder, but not neccesarily impossible; we assuredly still want to fix the current situation but perhaps we can still lock.
On a brighter note, I've only noticed one brief EPICS freeze all night. In addition, the wall StripTools seem totally contiuous since ~4pm, whereas I'm used to seeing some blocky shapes particularly in the seismic rainbow. Could this possibly mean that the old WiFi router was somehow involved in all this?
I hooked up the ALSX DFD output to the fibox, and used the adjustable delay line to set the phase properly. I recorded the noise on pianosa, and have attached it. Of course, this doesn't really capture the low frequency behavior.
Unrelated to this: I found the MC WFS turned off, and the loops ran away when turning them on. I tweaked the alignment, and reset the WFS offsets. Seems stable for now.
Attachment #1 shows the measured AM response. It differs qualitatively in shape from the earlier measurements reported in this elog and on the wiki below the 100kHz region.
It looks like some of the features may have shifted in frequency. The previous measurement results can be found in /users/OLD/mott/PZT/2NPRO, can you plot the two AM measurements together?
Yesterday, I uploaded some EAGLE schematic files and a LISO source file for the green PDH servo electronics to the 40m LISO git repository. In doing so, I realized that the DCC document for the X box (D1400293) was not updated at the end of the electronics work we did in Aug/Sep 2014. This is entirely my fault.
The Y box document (D1400294) is currently accurate.
The missing information is that, as I posted In ELOG 10457, I ended up destroying our original X box, and replaced it with a spare from the ATF. It was restuffed to match the Y end box pretty much exactly. We will update the X circuit DCC page with an accurate schematic and photo.
Gautam tells me that he and Rana were looking at the outdated schematic and thinking about improvements, but at least some of this was already done back in 2014 (specifically, the resistors used to specify the AD8336 preamp gain were changed).
Tonight we embarked on the laser swap. In short, we have gotten ~210mW through the faraday doubler, but no green light is apparent. The laser outputs ~300mW, so it's not exactly a work of art, but I still expected some green. More work remains to be done...
Gautam took numerous photos of the table before anything was touched. One lens was swapped, as per Gautam's plan. The innolight laser and controller are on the work bench by the end table. The lightwave is on the table and on standby, and is not hooked up to the interlock mounted on the table frame, but instead one below the table directly next to the controller. The ETMX oplev laser is turned off.
Chiara reports an uptime of >195 days, so its UPS is working fine
FB, megatron, optimus booted via front panel button.
Jetstor RAID array (where the frames live) was beeping, since its UPS failed as well. The beep was silenced by clicking on "View Events/Mute Beeper" at 192.168.113.119 in a browser on a martian computer. I've started a data consistency check via the web interface, as well. According to the log, this was last done in July 2015, and took ~19 hrs.
Frontends powered up; models don't start automatically at boot anymore, so I ran rtcds start all on each of them.
rtcds start all
All frontends except c1ioo had a very wrong datetime, so I ran sudo ntpdate -b -s -u pool.ntp.org on all of them, and restarted the models (just updating the time isn't enough). There is an /etc/ntp.conf in the frontend filesystem that points to nodus, which is set up as an NTP server, but I guess this isn't working.
sudo ntpdate -b -s -u pool.ntp.org
PMC locking was hindered by sticky sliders. I burtrestored the c1psl.snap from Friday, and the PMC locked up fine. (One may be fooled by the unchanged HV mon when moving the offset slider into thinking the HV KEPCO power supplies need to be brought down and up again, but it's just the sliders)
Mode cleaner manually locked and somewhat aligned. Based on my memory of PMC camera/transmission, the pointing changed; the WFS need a round of MC alignment and WFS offset setting, but the current state is fine for operation without all that.
daqd has indeed continued to be unstable. I found system times had drifted apart again... I think something weird happened in the booting of the frontends. The monit processes were not running on any of the frontends. I ntpdate'd again, and manually started monit on each fronted via sudo /etc/init.d/monit start.
sudo /etc/init.d/monit start
I manually aligned the IMC. Spot positions are all < 1.5mm. PMC trans of ~0.74, MC2 Trans of ~15400, MC Refl ~0.4, which is better than its been for some time now.
Somehow the WFS DC offsets were off, which made it look like it was impossible to center the beam on WFS2. The script for setting these wasn't working so I fixed it, ran it. WFS and MC2 trans offsets were set, WFS are back on and have been holding MC REFL nice and low for ~3 hours.
Arms were dither aligned, wrote the offsets to SDF files. Oplevs need centering. No further daqd crashes.
Taking inspiration from J. Lewis et. al, ITMX has been freed.
PSL Table doors were open, and the laser shutter was closed.
Doors have been closed, laser has been opened.
We went and looked at the monitor plugged into FB. All kinds of messages were being spammed to the screen (maybe RAM errors), and nothing could be done to interrupt. Sadly, a hard reboot of FB was neccesary.
Video of error messages: https://youtu.be/7rea_kokhPY
After the reboot, it just took a couple of model restarts to get the CDS screen happy.
We worked on getting the DRFPMI back up and running, hoping the ALS performance was good enough.
We did succeed in bringing in enough of the AO path to stabilize arm powers > 100, but failed at the full RF DARM handoff.
REFL165 angle was adjusted to -86 to minimize PRCL in the Q signal.
The AS110 signals are mysteriously huger than they used to be. Whitening gain reduced to 15dB from 27dB. Old trigger thresholds are still fine.
The new AUX X laser has a different sign for the temperature-> frequency coupling, so our usual convention of "beatnote goes up when temp slider goes up" meant the ALSX input matrix elements had to change sign.
We think the POPDC PD (which I think is the POP2F PD) may be miscentered, since in PRMI configuration, its maximum does not coincide with the REFLDC minimum, and leaves a sizeable TEM10 lobe on the REFL camera. This was a pain.
Three RF-only locks longer than a minute tonight, out of 5 total attempts.
Last week, I determined that the beam spot on the RF POP PD is too large. This still needs to be fixed. I updated the ASS model to use REFLDC as a PRCL dither error signal; it works.
There seems to be some excess angular motion of ETMY tonight. This is evident in the oplev spectra (as compared to ETMX), and the GTRY camera, and even the retroreflected beam from a misalgined ETMY on the ITMY face when the PRC is carrier locked.
Gautam and I mostly focused on setting up the CAL-DARM_CINV block to produce this (mostly) calibrated spectrum starting from GPS 1143274087. [Darm on unwhitened AS55, DRMI on 3F, one CARM boost]
Here are the control and error signal spectra:
[DTT files attached]
Note to self: archive some of this data
I haven't found any data files for the DARM spectrum of the previous generation of 40m, but with some GIMP-fu, I have plotted Monday's spectrum (green) on top of one of the figures from Rob's thesis.
I've been banging my head against bilinear noise subtraction, and figured I needed to test things on some real hardware to see if what I'm doing makes sense.
I ran the ASS dither alignment on the Y arm, which ensures that the beam spots are centered on both mirrors.
I then drove ITMY in yaw with some noise bandpassed from 30-40 Hz. It showed the expected bilinear upconversion that you expect from angular noise on a centered beam, which you can see from 60-80 Hz below
I looked at the length signal, as the noise subtraction target, and the ITMY oplev yaw signal plus the transmon QPD yaw signal as witnesses.
There is some linear coupling to length, which means the the centering isn't perfect, and the drive is maybe large enough to displace it off center. However, the important part is the upconverted noise which is present only in the length signal. The QPD and oplev signals show no increased noise from 60-80Hz above the reference traces where no drive is applied
I then compared the multicoherence of those two angular witnesses vs. the multicoherence of the two (linear) witnesses plus their (bilinear) product. Including the bilinear term clearly shows coherence, and thereby subtraction potential, at the upconverted noise hump.
So, it looks like the way I'm generating the bilinear signals and calculating coherence in my code isn't totally crazy.
I have copied over the complete frame files from two DRFPMI lock acquisitions + locks to /frames/archive. The data should be safe from the wiper script here.
One, under the subfolder DRFPMI_Mar29_cal is the lock where the CAL-DARM channel is properly calibrated at GPS time 1143274087.
The other lock, under DRFPMI_MAR29_nocal, does not have the calibration set up yet, but was a much quicker acquistion (<2 min from ALS acquisition to DRFPMI) and longer lock (~8min).
Just a heads up that some equipment is hooked up at the PSL table for the repaired AUX laser PLL measurement, I plan to continue with it tonight.
I've taken a few spectra that, along with the PZT coefficient from the repair sheet, that suggest the noise level is ok (incoherent sum of AUX and PSL at about ~3e4 / f Hz/rtHz), but calibrated plots, etc. will follow in time.
The free running PSL+AUX beat frequency noise spectrum has been measured via PLL. AUX laser PZT PM and AM responses were measured too.
Rough notes about these measurements:
Laser -> QWP -> HWP -> PBS -> 10% BS -> Beat
3.4Vpp out of PD, (40% contrast)
20dB Coupler, output to analyzer, coupled output to Mixer (-a few dBm, didn't check specifically)
Mixer: ZP-3+, BLP-5.1 at output
LO: OCXO @ 36MHz 13dBm->5dB Att-> +8dBm LO at Mixer
Got ~65mVpp out of Mixer
Mixer out -> SR560, LP 3Hz, G=500 -> Pomona Summing node -> Laser PZT
~30kHz UGF ~30 deg phase
Spectra, OLG via SR785 taken with free running PSL, anthropomorphic temperature servo. Data sheet calibration used for PZT. SR560 output noise dominates over analyzer, mixer, PD. Spectrum looks ok, I think.
PM measured with AG4395. High impedance probe used for laser PZT, otherwise couldn't lock. PM calibrated via mixer voltage span for fringe-to-fringe.
PSL beam blocked, AUX power increased to read 8.0V, AM measured with AG4395.
AM/PM doesn't look to dissimilar to old measurements on wiki. ~230kHz looks like a fine modulation freq.
Still to be done to AUX laser:
- joint PSL/AUX temperature sweeps
- Output power vs. diode current
- Beam profile
The 2F product out of the mixer is a natural concern when demodulating. However, I think this isn't so big of a deal in our green PDH servos; 420kHz isn't so high of a frequency that the servo amplifiers are bandwidth or slew-rate limited. Furthermore, the amplitude of this line is supressed by the loop somewhat, since it arises from the same field product that the loop is acting on. Measuring the Y end mixer output with a high impedance probe and the AG4395 shows it to be something like -50dBm.
In fact, the main thing that the pomona LPFs are accomplishing right now is filtering the 1F content of the mixer output that arises from the second order sideband creating a signal at 2F, and beating with the LO at (2F-1F)=1F. This line is something like -30dBm (5mVrms) at the mixer output; I can reproduce this amplitude with a back-of-the envelope calculation using a modulation depth of 0.3, 8V out of the PD at DC when unlocked, the mixer datasheet, and the nominal cavity parameters.
The nice thing about this is that we don't need to filter this after the mixer, we can use a [bandpass/lowpass/notch] filter before the mixer (as is done in the LSC demod boards) to filter out the 2F (420kHz) content of the PD signal, which will only introduce some small amount of linear time delay to the PDH loop, instead of the wicked phase loss from the current post-mixer LPF. We can then replace that 70kHz filter with something of lower order or higher corner frequency to win a good deal of phase in the PDH loop.
We can get as much, if not more, attenuation of the 1F line in the mixer output that we get from the post-mixer LPF from using the following passive filter between the PD and mixer RF input:
There should still be some kind of LPF after the mixer, but I haven't yet determined what it should be; this will determine how much phase the PDH loop wins. At most, this should win around 25 degrees at 10kHz.
The filter was designed by referencing the "Handbook of Filter Synthesis" by Zverev, looking for an elliptic filter for matched source and load impedences, 40dB min attenuation in the stopband, a stopband frequency that starts at twice the corner frequency, and minimizing the VSWR between the PD and filter in the passband.
In terms of the tables in the book, this means: n=5, rho=2%, theta=30deg, K**2 = 1.0. The dimensionless component values were scaled by the corner frequency of 200kHz, and reference impedence of 50 Ohm. (The corner is a little lower than the real modulation frequency, since the nonzero resistance of the inductors pushes the frequency up a bit)
The ideal capactior values do not correspond to things we have in hand, so I checked our stock and chose the closest value to each one.Unsurprisingly, due to these component substitutions, and the fact that the coilcraft inductors have a resistance of about 7 Ohms, the predicted TF of the realizable filter does not match the design filter exactly. However, the predicition still looks like it will meet the requirement of 40dB of supression of the 2F line in the PD signal. (Since we have tunable inductors, I've used the ideal inductor values in generating the TF. In practice I'll inspect the TF while I tune them)
[In this TF plot, I've multiplied the real response by 2 to account for the voltage division that occurs with ideal 50 Ohm impedance matching, to make 0dB the reference for proper matching]
The filter's phase delay at the modulation frequency is just about 180, which as a time delay of 5usec works out to 9 degrees of phase loss at 10kHz in the PDH loop. According to some old measurements, the current LPF costs something like 35 degrees at 10k, so this wins at most around 25 degrees, depedent on what LPF we put after the mixer.
LISO source both traces is attached!
We took an OLG measurement of the green PDH loop. It seems consistent with past measurements. I've added a trace for the the post-mixer lowpass, to show its contribution to the phase loss. (EDIT: updated with measured LPF TF)
I used this measured OLG and the datasheet laser PZT conversion factor to calibrate the control signal monitor into the AUX laser frequency noise, it looks consistent with the frequency noise measured via the PSL PLL (300 Hz/rtHz @ 100Hz). Above a few tens of kHz, the control signal measurement is all analyzer noise floor, due to the fourth order 70kHz lowpass after the mixer (the peaks change height significantly depending on the analyzer input range, so I don't think they're on the laser). Gautam will follow up with more detailed measurements of both the error and control signals as he noisebudgets, this was just intended as a quick consistency check.
This morning I poked around with the green layout a bit. I found that the iris immediately preceding the viewport was clipping the ingoing green beam too much, opening it up allowed for better coupling to the arm. I also tweaked the positions of the mode matching lenses and did some alignment, and have since been able to achieve GTRX values of around 0.5.
I also removed the 20db attenuator after the mixer, and turned the servo gain way down and was able to lock easily. I then adjusted the gain while measuring the CLG, and set it where the maximum gain peaking was 6dB, which worked out to be a UGF of around 8kHz. On the input monitor, the PDH horn-to-horn voltage going into the VGA is 2.44V, which shouldn't saturate the G=4 preamp stage of the AD8336, which seems ok.
The ALS sensitivity is now approaching the good nominal state:
There remains some things to be done, including comprehensive dumping of all beams at the end table (especially the reflections off of the viewport) and the new filters to replace the current post-mixer LPF, but things look pretty good.
I've build the filter, and it seems to have the desired TF shape.
I also re-purposed the 70k lowass to a ~120k lowpass by changing the 68nF caps to 22nF caps, since we still want some post-mixer rolloff.
However, putting the ELPF in the chain caused some weird shapes in the OLG. I still need to get to the bottom of it. However, just with the post-mixer LPF modification, here's what the OLG looks like:
As Rana surmises, we definitely still add a boost and maintain a 10k UGF. I still need to look into the state of the remote boost....
As I was looking at filter designs, it seemed difficult to get 40dB of supression at 2F with a bandpass without going to a pretty high order, which would mean a fair number of lossy inductors.
I'll keep working on it. Maybe we don't need 40dB...
ALSX noise is solidly within past acceptable performance levels. The DRFPMI was locked on four out of six attempts.
Some housekeeping was done:
The recombination of the QPD signals to common / differential is imperfect, and limited how well we could keep the interferometer aligned, since the QPD at X has changed. This needs some daytime work.
Some sensing matrix measurements were made, to be meditated upon for how to 1F the DRMI.
As an aside, Gautam and I noticed numerous green beams coming from inside the vacuum system onto the PSL table. They exist only when green is locked to the arms. Some of them come out at very non-level angles and shine in many places. This doesn't make me feel very happy; I suppose we've been living with it for some time.
Our last RGA scan is from February 14, 2016 We had a power outage on the 15th
Gautom has not succeded reseting it. The old c0rga computer looks dead. Q may resurrect it, if he can?
The c0rga computer was off, I turned it on via front panel button. After running RGAset.py, RGAlogger.py seems to run. However, there are error messages in the output of the plotrgascan MATLAB script; evidiently there are some negative/bogus values in the output.
I'll look into it more tomorrow.
I did a quick measurement of the ITMX oplev loops, both pitch and yaw have about the same upper UGF as previous measurements with the previous laser; about 4 Hz.
I think you should use the current actual PRC & SRC cavity lengths as measured, as it would be simplest to simply replace the folding mirror optics without changing the macroscopic lengths / optic positions. (EDIT: Gautam rightly points out that we have to move things around regardless, since our current lengths include propagation through the folding mirror subtrates)
Moreover, the recycling cavity lengths you posted are not the right "ideal" lengths to use, as they do not account for the complex reflectivities of the sidebands off of the arm cavities (I have made this mistake myself). See this 40m wiki page for details.
In short, given our current modulation frequency, the ideal lengths to use would be:
These are the lengths that the recycling cavity optics were positioned for (though we did not achieve them perfectly). If you do a finer PRC/SRC length scan around the DRFPMI resonance of your model, you would presumably see some undesired sideband splitting.
WFS locking point seemed degraded; I hand aligned and reset the WFS offsets as usual.
ITMX oplev recentered. While doing so, I noticed an ETMX excursion rear its head for the first time in a long while :
There was no active length control on ETMX, only OSEM damping + oplevs. Afterwards, its still moving around with only local damping on. I'm leaving the oplevs off for now.
I've been futzing with the common mode servo, trying to engage the AO path with POY for high bandwidth control of a single arm lock. I'm able to pull in the crossover and get a nice loop shape, but keep getting tripped up by the offset glitches from the CM board gain steps, so can't get much more than a 1kHz UGF.
As yutaro measured, these can be especially nasty at the major carrier transitions (i.e. something like 0111->1000). This happens at the +15->+16dB input gain step; the offset step is ~200x larger than the in-loop error signal RMS, so obviously there is no hope of keeping the loop engaged when recieving this kind of kick. Neither of the CM board inputs are immune from this, as I have empirically discovered. I can turn down the initial input gain to try and avoid this step occuring anywhere in the sequence, but then the SNR at high frequencies get terrible and I inject all kinds of crud into the mode cleaner, making the PC drive furious.
I think we're able to escape this when locking the full IFO because the voltages coming out of REFL11 are so much larger than the puny POY signals so the input-referred glitches aren't as bad. I think in the past, we used AS55 with a misaligned ITMX for this kind of single arm thing, which probably gives better SNR, but the whole point of this is to keep the X arm aligned and lock it to the Y-arm stabilized PSL.
I used a Eurocard extension board to peek at the inputs and outputs of each of the gain-ladder AD829s on input B of the CM board in the +31dB configuration with the input terminated. (i.e with the following stages active in this order: +16dB, +8dB, +4dB, +2dB, +1dB).
The voltages I observed imply that the +8dB stage has an input voltage offset of -2mV, whereas all the other positive gain stages show around +-0.5mV. This could explain the shift observed at the +15->+16 transition. (However, since both input channels show a jump here, maybe its something more systemic about the board...)
In any case, it should be simple enough to swap out a new AD829 in place of U9B and see if it improves things, before getting too deep into the muck. (In principle, the AD829 has offset nulling pins, but I'm not sure how to do it in a non-hacky way since the board doesn't have any pads for it.)
I replaced some of the AD829s with other AD829s, but the offset situation didn't improve.
However, I figured that we don't really need the ~100MHZ bandwidth of the AD829, since the IMC loop limits us to a ~10kHz CARM bandwidth. Also, since we don't routinely use IN2 for anything, I felt free to try something else.
Specifically, I replaced all of the positive gain AD829s in the input 2 gain ladder with OP27s (U8B->U12B on D1500308), which should have input offset voltages ~30x lower than the AD829s.
Here is a comparison of the outputs these configurations perform, normalized to the output at the +0dB gain setting - where all of the op amps in the gain ladder are bypassed.
So, most of the transitions now result in an output offset change of less than 0.5mV, which is nice.
The exception seems to be where the +8dB stage is switched in or out. I may try replacing this one, as these transitions cause a lock loss now when trying to lock the arm with high bandwidth using POY.
Some CDS related things:
Keith Thorne has told us about a potential fix for our framebuilder woes. Jamie is going to be at the 40m next week to implement this, which could interfere with normal interferometer operation - so plan accordingly.
I spent a little time doing some plumbing in the realtime models for Varun's audio processing work. Specifically, I tried to spin up a new model (C1DAF), running on the c1lsc machine. This included:
The simple DAFI model compiled and installed without complaint, but doesn't succesfully start. For some reason, the frontend never takes the CPU offline. Jamie will help with this next week. Since things aren't working, these changes have not been commited to the userapps svn.
Barring objections, starting tomorrow morning, Jamie will be testing the new FB code. The IFO will not be available for other use while this is ongoing.
ETMX has been jumping around again lately. Just now, I zeroed the ETMX alignment offsets in the SUS model, and centered the ETMX oplev spot via slow machine sliders. OSEM damping is on, oplev damping is off. Let's see how it moves around in the next day or so.
UTC: Jun 10 2016 23:18:26
Within two hours, it was already all over the place.
I have installed a ZFL-500LN on the RF output of POY11. This should reduce the effect of the CM board voltage offsets by increasing the size of the error signal coming into the board. Checking with an oscilloscope at the LSC rack, the single arm PDH peak to peak voltage was something like 4mV, now it is something like 80mV.
The setup is similar to the REFL165 situation, but with the amplifier in proximity with the PD, instead of at the end of a long cable at the LSC rack.
The PD RF output is T'd between an 11MHz minicircuits bandpass filter and a 50 Ohm terminator (which makes sure that signals outside of the filter's passband don't get reflected back into the PD). The output of the filter is connected directly to the input of the ZFL-500LN, which is powered (temporarily) by picking off the +15V from the PD interface cable via Dsub15 breakout. (I say temporarily, as Koji is going to pick out some fancy pi-filter feedthrough which we can use to make a permanent power terminal on the PD housing.)
The max current draw of this amplifier is 60mA. Gazing at the LSC interface (D990543), I think the +15V on the DSUB cable is being passed from the eurocard crate; I don't see any 15V regulator, so maybe this is ok...
The free swinging PDH signal looked clean enough on a scope. Jamie is doing stuff with the framebuilder, so I can't look at spectra right now. However, turning the POY whitening gain down to +18dB from +45dB lets the Y arm lock on POY with all other settings nominal, which is about what we expect from the nominal +23dB gain of the amplifier.
I would see CM board offsets of ~5mV before, which was more a little more than a linewidth before this change. Now it will be 5% of that, and hopefully more manageable.
With the newly amplified POY signal, locking the mode cleaner to the Y arm at ~30kHz bandwidth was quite straightforward. The offset jumps still happen, and are visible in POY11_I_ERR, but are never big enough to cause much power degradation in TRY (except when turning on CM board boosts, but its still not enough to lose lock). The script which accomplishes this is at scripts/YARM, and is in the svn. The MC2/AO crossover is at about 150Hz with 40deg margin.
For now, I'm using IN1 of the CM board, because I haven't removed the op27s that I put into IN2's gain stages. I believe the slew rate limitations of these prevent them from working completely during the offset jumps. I'll put AD829s back soon.
At first, I had ITMX misalgined to use AS55 as an out of loop sensor, then I aligned and locked the X arm on POX to compare.
Weirdly enough, locking the mode cleaner to the Y arm with 30kHz UGF and two boosts on make no real visible difference in the X arm control signal. This is strange, as the whole point of this affair was to remove the presumably large influence of frequency noise on the X arm signals... Maybe this is injecting too much POY sensor noise?
The workstations' .bashrc is a symbolic link to /users/controls/.bashrc
In it, someone commented out the critical line:
I uncommented it. medm (and all of the other things like cdsutils) work again.
I blame jamie.