Exactly: you'll have to list explicitly what functions those channels had so that we know what we're losing before we make the switch.

Acromag Wireup Update

Channels no longer in use

• From the VME analog output (VMIVME 4116) to the QPD Whitening board (no DCC number on the front), 3 channels are no longer in use
• From the anti-image filter (D000186) to the ADC (VMIVME 3113A) 5 channels are no longer in use (these are the only channels from the anti-image filter, so this filter is no longer in use at all?)
• From the universal dewhitening filter (D000183) to a binary I/O adapter (channels 1-16), 4 channels are no longer in use. These are the only channels from the dewhitening filter
• From a second universal dewhitening filter (D000183) to another the binary I/O adapter (channels 1-16), one channel is no longer in use (this was the only channel from this dewhitening filter).
• From the opti-lever (D010033) to the VME ADC (VMIVME 3113A), 7 channels are no longer in use (this was all of the channels from the opti lever)
• From the SUS PD Whitening/Interface board (D000210) to a binary I/O adapter (channels 1-16), 5 channels are no longer in use. 
• Note that none of the binary I/O adapter channels are in use.

I bet the calibration is out of date; probably we replaced the OL laser for the BS and didn't fix the cal numbers. You can use the fringe contrast of the simple Michelson to calibrate the OLs for the ITMs and BS.

I disabled the OL loops for ITMX, ITMY and BS at GPStime 1194897655 to come up with an Oplev noise budget. OL spots were reasonably well centered - by that, I mean that the PIT/YAW error signals were less than 20urad in absolute value.

Attachment #1 is a first look at the DTT spectra - I wonder why the BS Oplev signals don't agree with the ITMs at ~1Hz? Perhaps the calibration factor is off? The sensing noise not really flat above 100Hz - I wonder what all those peaky features are. Recall that the ITM OLs have analog whitening filters before the ADC, but the BS doesn't...

In Attachment #2, I show comparison of the error signal spectra for ITMY and SRM - they're on the same stack, but the SRM channels don't have analog de-whitening before the ADC.

For some reason, DTT won't let me save plots with latex in the axes labels...

I've incorporated the functionality to generate sub-budgets for the various grouped traces in the NBs (e.g. the "A2L" trace is really the quadrature sum of the A2L coupling from 6 different angular servos).

For now, I'm only doing this for the A2L coupling, and the AUX length loop coupling groups. But I've set up the machinery in such a way that doing so for more groups is easy.

Here are the sub-budget plots for last night's lock - for the OL plot, there are only 3 lines (instead of 6) because I group the PIT and YAW contributions in the function that pulls the data from the nds server, and don't ever store these data series individually. This should be rectified, because part of the point of making these sub-budgets is to see if there is a particularly bad offender in a given group.

I'll do a quick OL loop noise budget for the ITM loops tomorrow.

I also wonder if it is necessary to measure the Oplev A2L coupling from lock to lock? This coupling will be dependant on the spot position on the optic, and though I run the dither alignment servos to minimize REFL_DC, AS_DC, I don't have any intuition for how the offset from center of optic varies from lock to lock, and if this is at all significant. I've been using a number from a measurement made in May. Need to do some algebra...

If you go through this thread of elogs, there are lots of pictures of the SOS assembly with the optic in it from the vent last year. I think there are many different perspectives, close ups of the standoffs, and of the OSEMs in their holders in that thread.

This elog has a measurement of the pendulum resonance frequencies with ruby standoffs - although the ruby standoff used was cylindrical, and the newer generation will be in the shape of a prism. There is also a link in there to a document that tells you how to calculate the suspension resonance frequencies using analytic equations.

Summary:

• SOS solidworks model is nearly complete
  • Having trouble with the design of the sensor/actuator head assembly and the lower clamps
• After Gautam's suggestion, installed Abaqus on computer. Teaching it to myself to eventually do FEM analysis and find resonant frequency of the system
  • Goal is to replicate frequency listed in the SOS documents to confirm accuracy of computer model, then replace guide rods with sapphire prisms and change geometry to get same results

Questions:

• How accurate do the details (like fillet, chamfer, placement of little vent holes), and material of the different SOS parts need to be in the model?
• If I could get pictures of the lower mirror clamp (document D960008), it would be helpful in making solidworks model. Document is unclear. Same for sensor/actuator head assembly. 

Pianosa just crashed and ate my elog, along with all the DTT/Foton windows I had open, so more details tomorrow... This workstation has been crashing ~once a month for the last 6 months.

Summary:

Below ~100Hz, the hypothesis is that the BS oplev A2L contribution dominates the MICH displacement noise. I wanted to see if I could mitigate this my modifying the BS Oplev loop shape.

Details:

• Swept sine TF measurements suggested that the BS A2L contribution is between 10-100x that of the ITM A2L
• The Oplev loop shape for BS is different from ITMs - specifically, there is a Res-gain centered at ~3.3 Hz. The low frequency ~0.6Hz boost filter present in the ITM Oplev loops was disengaged for the BS Oplves.
• I turned off the BS OL loops and looked at error signal spectra - didn't seem that different from ITM OL error signals, so I decided to try turning off the res-gain and engage the 0.6Hz boost.
• This change also gave me much more phase at ~6Hz, which is roughly the UGF of the loop. So I put in another roll-off low pass filter with corner frequency 25Hz. 
• This worked okay - RMS went down by ~5x (which is even better than the original config), and although the performance between ~3 and 10Hz is slightly worse than with the old combination,this region isn't the dominant contribution to the RMS. PM at the upper UGF is ~30degrees in the new configuration. 
• I wanted to give DRMI locking a shot with the new OL loop - expectations were that the noise between 30-100Hz would improve, and perhaps the engaging of de-whitening filters on BS would also be easier given the more severe roll-off at high-frequencies.
• Attachment #1 shows the NB for tonights lock. All MICH optics had their coil drivers de-whitened, and all the LSC PDs were whitened for this measurement.
• I've edited the NB code to make the A2L calculation more straightforward, I now just make the coupling 1/f^2 and give the function a measured overall gain, so that this curve can now be easily added to all future NBs. I've also transcribed the matlab funciton used for parsing Foton files into python, this allows me to convert the DQ-ed OL error signals to control signals. Will update git with changes.

Remarks:

1. MICH noise has improved by ~2x between 40-80Hz.
2. Not sure what to make of the broad hump around 60Hz - scatter shelf?
3. There is still unexplained noise below 100Hz, the A2L estimate is considerably lower than the measured noise.
4. We are still more than an order of magnitude away from the estimated seismic noise floor at low frequencies (but getting closer!).

I've been banging my head against optimal loop shaping, with the OL loop as a test-case, without much success - as was the case with coating PSO, the magic is in smartly defining the cost function, but right now, my optimizer seems to be pushing most of the roots I'm making available for it to place to high frequencies. I've got a term in there that is supposed to guard against this, need to tweak further...

Attachment #2: Eye-fits of measured OL A2L coupling TFs to a 1/f^2 shape, with the gain being the parameter "fitted". I used these value, and the DQ-ed OL error signal in lock, to estimate the red curve labelled "A2L" in Attachment #1. The dots are the measurement, and the lines are the 1/f^2 estimates.

I made a model for our seismometer can using actual data so that we know approximately what the time constant should be when we test it out. I used the appendix in Megan Kelley's report to make a relation for the temperature in terms of time.

 so  and 

In our case, we will heat the can to a certain temerature and wait for it to cool on its own so 

We know that  where k is the k-factor of the insulation we are using, A is the area of the surface through which heat is flowing,  is the change in temperature, d is the thickness of the insulation.

Therefore,

We can take the derivative of this to get

, or  

We can guess the solution to be

 where tau is the time constant, which we would like to find.

The boundary conditions are  and . I assumed we would heat up the can to 40 celcius while the room temp is about 24. Plugging this into our equations,

, so 

We can plug everything back into the derivative T'(t)

Equating the exponential terms on both sides, we can solve for tau

Plugging in the values that we have, m = 12.2 kg, c = 500 J/kg*k (stainless steel), d = 0.1 m, k = 0.26 W/(m^2*K), A = 2 m^2, we get that the time constant is 0.326hr. I have attached the plot that I made using these values. I would expect to see something similar to this when I actually do the test.

To set up the experiment, I removed the can (with Steve's help) and will place a few heating pads on the outside and wrap the whole thing in a few layers of insulation to make the total thickness 0.1m. Then, we will attach the heaters to a DC source and heat the can up to 40 celcius. We will wait for it to cool on its own and monitor the temperature to create a plot and find the experimental time constant. Later, we can use the heatng circuit we used for the PSL lab and modify the parts as needed to drive a few amps through the circuit. I calculated that we'd need about 6A to get the can to 50 celcius using the setup we used previously, but we could drive a smaller current by using a higher heater resistance.

I characterized the black Ithaca 1201 pre-amp that we had sitting in the racks. It works fine and the input referred noise is < 10 nV/rHz. I also checked that the filter selection switches on the front panel did what they claim and that the gain knob gives us the correct gain.

For comparison I have also included the G=100, 1000 input referred noise of one of the best SR560 that we have (s/n 02763) in the lab. Above a few Hz, the SR560 is better, but for low frequency measurements it seems that the 1201 is our friend.

As with the SR560, you don't actually get low noise performance for G < 100, due to some fixed output noise level.

Steve:  sn48332 of Ithaca 1201

BOOM SHAKA-LAKA

PSL shutter closed at 6e-6 Torr-it    

   The foreline pressure of the drypump is 850 mTorr at 8,446 hrs of seal life

V1 will be closed for ~20 minutes for drypump replacement..........

9:30am dry pump replaced, PSL shutter opened at 7.7E-6 Torr-it

  Valve configuration: vacuum normal as  TP3 is the forepump of the Maglev  & annuloses are not pumped.

http://www.supermicro.com/products/system/1U/5015/SYS-5015A-H.cfm?typ=H

This is the machine that Larry suggested when I asked him for his opinion on a low workload rack-mount unit. It only has an atom processor, but I don't think it needs anything particularly powerful under the hood. He said that we will likely be able to let us borrow one of his for a couple days to see if it's up to the task. The dual ethernet is a nice touch, maybe we can keep the communication between the server and the DAQ units on their separate local network.

Jamie pointed out that the compile and install instructions are different for c1dnn:

See also: https://nodus.ligo.caltech.edu:8081/40m/13383.

I think these build instructions have to be run on the c1lsc frontend - in the past, I have been able to compile and install models on any computer with the shared drive mounted (including the control room workstations), but I'm guessing that something has changed since the RCG upgrade. Jamie can correct me on this if I'm wrong.

I wanted to use the foton.py utility for my NB tool, and I remember Chris telling me that it was shipping as standard with the newer versions of gds. It wasn't available in the versions of gds available on our workstations - the default version is 2.15.1. So I downloaded gds-2.17.15 from http://software.ligo.org/lscsoft/source/, and installed it to /ligo/apps/linux-x86_64/gds-2.17.15/gds-2.17.15. In it, there is a file at GUI/foton/foton.py.in - this is the one I needed. 

Turns out this was more complicated than I expected. Building the newer version of gds throws up a bunch of compilation errors. Chris had pointed me to some pre-built binaries for ubuntu12 on the llo cds wiki, but those versions of gds do not have foton.py. I am dropping this for now.

Gautam and I measured the noise of the ADC for channels 17, 18, and 19. We plan to use those channels for measuring the noise of the temperature sensors, and we need to figure out whether or not we will need whitening and if so, how much. The figure below shows the actual measurements (red, green and blue lines), and a rough fit. I used Gautam's elog here and used the same function,  (with units of nV/sqrt(Hz)) to fit our results. I used a = 1, b = 1e6, c = 2000. Since we are interested in measuring at lower frequencies, we must whiten the signal from the temperature sensors enough to have the ADC noise be negligible.

We want to be able to measure to  accuracy at 1Hz, which translates to about  current from the AD590 (because it gives ). Since we have a 10K resistor and V=IR, the voltage accuracy we want to measure will be . We would need whitening for lower frequencies to see such fluctuations.

To do the measurements, we put a  BNC end cap on the channels we wanted to measure, then took measurements from 0-900Hz with a bandwidth of 0.001Hz. This setup is shown in the last two attachments. We used the ADC in 1X7.

There hasn't been a big glitch that misaligns MC1 by so much that the autolocker can't lock for at least 3 months, seems like there was one ~an hour ago.

I disabled autolocker and feedback to the PSL, manually aligned MC1 till the MC_REFL spot looked right on the CCD to me, and then re-engaged the autolocker, all seems to have gone smoothly.

We've been talking about increasing the series resistance for the coil driver path for the test masses. One consequence of this will be that we have reduced actuation range.

This may not be a big deal since for almost all of the LSC loops, we currently operate with a limiter on the output of the control filter bank. The value of the limit varies, but to get an idea of what sort of "threshold" velocities we are looking at, I calculated this for our Finesse 400 arm cavities. The calculation is rather simplistic (see Attachment #1), but I think we can still draw some useful conclusions from it:

• In Attachment #1, I've indicated with dashed vertical lines some series resistances that are either currently in use, or are values we are considering.
• The table below tabulates the fraction of passages through a resonance we will be able to catch, assuming velocities sampled from a Gaussian with width ~3um/s, which a recent ALS study suggests describes our SOS optic velocity distribution pretty well (with lcoal damping on).
• I've assumed that the maximum DAC output voltage available for length control is 8V.
• Presumably, this Gaussian velocity distribution will be modified because of the LSC actuation exerting impulses on the optic on failed attempts to catch lock. I don't have a good model right now for how this modification will look like, but I have some ideas.
• It would be interesting to compare the computed success rates below with what is actually observed.
• The implications of different series resistances on DAC noise are computed here (although the non-linear nature of the DAC noise has not been taken into account).

So, from this rough calculation, it seems like we would lose ~25% efficiency in locking the arm cavity if we up the series resistance from 400ohm to 1kohm. Doesn't seem like a big deal, becuase currently, the single arm locking

The Oplev trace is missing for now, as I have not re-measured the A2L coupling since modifying the Oplev loop shape (specifically the low pass filter and overall gain) to allow engageing the coil de-whitening.

The averaging for the white noise TFs plotted is computed using median averaging - I have used a python transcription of Sujan's matlab code. I use scipy.signal.spectrogram to compute the fft bins (I've set some defaults like 8s fft length and a tukey window), and then take the median average using np.median(). I've also incorporated the ln(2) correction factor.

It seems like GwPy has some in-built capability to compute median (and indeed various other percentile) averages, but since we aren't using it, I just coded this up. 

why no oplev trace in the NB ?

also, this method would work better if we had a median averaging python PSD instead of mean averaging as in Welch's method.

I tried measuring the coupling of PSL intensity noise by driving some broadband noise bandpassed between 80-300Hz using the spare DAC channel at 1Y3 that I had set up for this purpose a couple of weeks ago (via a battery powered SR560 buffer set to low-noise operation mode because I'm not sure if the DAC output can drive a ~20m long cable). I was monitoring the MC2 TRANS QPD Sum channel spectrum while driving this broadband noise - the "nominal" spectrum isn't very clean, there are a bunch of notches from a 60Hz comb and a forest of peaks over a broad hump from 300Hz-1kHz (see Attachment #1).

I was able to increase the drive to the AOM till the RIN in the band being driven increased by ~10x, and saw no change in the MICH error signal spectrum [see Attachment #1] - during this measurement, the RFPD whitening was turned on for REFL11, REFL55 and AS55, and the ITM coil drivers were de-whitened, so as to get a MICH spectrum that is about as "low-noise" as I've gotten it so far.

I tried increasing the drive further, but at this point, started seeing frequent MC locklosses - I'm not convinced this is entirely correlated to my AOM activities, so I will try some more, but at the very least, this places an upper bound on the coupling from intensity noise to MICH.

I hadn't re-locked the DRMI after the work on the AS55 demod board. Tonight, I was able to recover the DRMI locking with the old settings.

The feature in the PRCL spectrum (uncalibrated, y-axis is cts/rtHz) at ~1.6kHz is mysterious, I wonder what that's about.

Wasted some time tonight futzing around with various settings because I couldn't catch a DRMI lock, thinking I may have to re-tune demod phases etc given that I've been mucking around the LSC rack a fair bit. But fortunately, the problem turned out to be that the correct feedforward filters were not enabled in the angular feedforward path (seems like these are not SDF monitored). Clue was that there was more angular motion of the POP spot on the CCD than I'm used to seeing, even in the PRMI carrier lock.

After fixing this, lock was acquired within seconds, and the locks are as robust as I remember them - I just broke one after ~20mins locked because I went into the lab. I've been putting off looking at this angular feedforward stuff and trying out some ideas rana suggested, seems like it can be really useful.

As part of the pre-lock work, I dither aligned arms, and then ran the PRCL/MICH dithers as well, following which I re-centered the ITM, PRM and BS Oplevspots onto their respective QPDs - they have not been centered for a couple of months now.

I'm now going to try and measure some other couplings like PSL RIN->MICH, Marconi phase noise->MICH etc.

Some days ago, I had tried to measure the SRCL->MICH and PRCL->MICH cross couplings using broadband noise injected between 120-180 Hz, a frequency band chosen arbitrarily, in hindsight, I could have done a more broadband test. I've spent some time including the infrastructure to calculate "White-Noise TFs" in the noise budgeting code, where a transfer function is estimated by injecting a "broadband" excitation into a channel of interest, and looking at the resulting response in MICH. I figured this would be useful to estimate other couplings as well, e.g. laser intensity nosie, oscillator noise etc.

I estimate the transfer function of the coupling using the relation (MICH is the median ASD of the MICH error signal in the below expression, and similarly for PRCL)

Attachments #1 and #2 show the spectra of the MICH, PRCL and SRCL signals during 'quiet' times and during the injection, while Attachment #3 shows the calculated coupling TFs using the above relation. These are significantly different (more than 10dB lower) than the numbers I reported in elog 13367, where the measurement was made using swept sine. As can be seen in the attached plots, the injected broadband excitation is visible above the nominal noise level, and I calculated the white noise TFs using ~5mins of data which should be plenty, so I'm not sure atm what to make of the answers from swept-sine and broadband injections being so different.

Attachment #4 shows the noise budget from the October 8 DRMI lock with the updated SRCL->MICH and PRCL->MICH couplings (assumed flat, extrapolated from Attachment #2 in the 120-180Hz band). If these updated coupling numbers are to be believed, then there is still some unexplained noise around 100Hz before we hit the PD dark noise. To be investigated. But if Attachment #4 is to be believed, it is not surprising that there isn't significant coherence between SRCL/PRCL and MICH around 100Hz.

Nov 8 1600: Updating NB to inculde estimated Oplev A2L.

This is the current procedure to start the c1dnn model:

Then to shutdown:

The daqd already knows about this model, so nothing should need to be done to the daqd to make the dnn channels available.

Eurocrate key turning reboots today morning for and c1susaux, c1auxex and c1auxey. Usual precautions were taken to minimize risk of ITMX getting stuck.

The IFO hasn't been aligned in ~1week, so I recovered arm and PRM alignment by locking individual arms and also PRMI on carrier. I will try recovering DRMI locking in the evening.

As far as MC1 glitching is concerned, there hasn't been any major one (i.e. one in which MC1 is kicked by such a large amount that the autolocker can't lock the IMC) for the past 2 months - but the MC WFS offsets are an indication of when smaller glitches have taken place, and there were large DC offsets on the MC WFS servo outputs, which I offloaded to the DC MC suspension sliders using the MC WFS relief script.

I'd like for the save-restore routine that runs when the slow machines reboot to set the watchdog state default to OFF (currently, after a key-turning reboot, the watchdogs are enabled by default), but I'm not really sure how this whole system works. The relevant files seem to be in the directory /cvs/cds/caltech/target/c1susaux. There is a script in there called startup.cmd, which seems to be the initialization script that runs when the slow machine is rebooted. But looking at this file, it is not clear to me where the default values are loaded from? There are a few "saverestore" files in this directory as well:

• saverestore.sav
• saverestore.savB
• saverestore.sav.bu
• saverestore.req

Are the "default" channel values loaded from one of these?

Our new Agilent Technology TwisTorr 84FS AG rack controller ( English Manual pages 195-297 )  RS232/485, product number X3508-64001, sn IT1737C383

This controller, turbo and it's drypump needs to be set up into our existing vacuum system. The intake valve of this turbo (V4) has to have a hardwired interlock that closes V4 when rotation speed is less than 20% of preset RPM speed.

The unit has an analoge 10Vdc output that is proportional to rotation speed. This can be used with a comperator to direct the interlock or there may be set software option in the controller to close the valve if the turbo slows down more than 20%

The last Upgrade of the 40m Vacuum System  1/2/2000 discribes our  vauum system  LIGO-T000054-00-R 

Here the LabView / Metrabus controls were replaced by VME processor and  an Epic interface

We do not have schematics of the hardware wiring.

We need help with this.

Eurocrate key turning reboots today morning for c1psl and c1aux.c1auxex and c1auxey are also down but I didn't bother keying them for now. PSL FSS slow loop is now active again (its inactivity was what prompted me to check status of the slow machines).

Note that the EPCIS channels for PSL shutter are hosted on c1aux.But looks like the slow machine became unresponsive at some point during the weekend, so plotting the trend data for the PSL shutter channel would have you believe that the PSL shutter was open all the time. But the MC_REFL DC channel tells a different story - it suggests that the PSL shutter was closed at ~4AM on Sunday, presumably by the vacuum interlock system. I wonder:

1. How does the vacuum interlock close the PSL shutter? Is there a non-EPICS channel path? Because if the slow machine happens to be unresponsive when the interlock wants to close the PSL shutter via EPICS commands, it will be unable to. The fact that the PSL shutter did close suggests that there is indeed another path.
2. We should add some feature to the vacuum interlock (if it doesn't already exist) such that the PSL shutter isn't accidentally re-opened until any vacuum related issues are resolved. Steve was immediately able to identify that the problem was vacuum related, but I think I would have just re-opened the PSL shutter thinking that the issue was slow computer related.

 IFO pressure 1.2e-5 Torr at 9:30am

Atm. 1,  This was the vacuum condition  this morning.

               IFO P1  9.7 mTorr, V1 open,  V4 was in closed position , ~37 C warm Maglev at normal 560Hz rotation speed with foreline pressure 3.9 Torr because V4 closed 2 days ago when TP2 failed .....see Atm.3

               The error messege at TP2  controller was: fault overtemp.

I did the following to restored IFO pumping: stopped pumping of the annulose with TP3 and valves were configured so TP3 can be the forepump of the Maglev.

closed VM1 to protect the RGA,  close PSL shutter .....see Gautam  entry

aux fan on to cool down Maglev-TP1, room temp 20 C,

aux drypump turned on and opend to TP3 foreline to gain pumping speed,

closed PAN to isolate annulos pumping,

opened V7 to pump Maglev forline with TP3 running at 50Krpm, It took 10 minutes to reach P2  1mTorr from 3.9 Torr

aux drypump closed off at P2  1 mTorr, TP3 foreline pressure 362 mTorr.......see Atm.2

As we are running now:

IFO pressure 7e-6 Torr at Hornet cold cathode gauge at 15:50  We have no IFO CC1 logging now.  Annuloses are in 3-5 mTorr range are not pumped.

TP3 as foreline pump of TP1 at 50 Krpm, 0.24 A, 24 C, it's drypump forline pressure 324 mTorr

V4 valve cable is disconnected.

I need help with wiring up the logging of the Hornet cold cathode gauge.

Udit Kahndelwal received 40m specific basic safety traning on Friday, Oct. 27

Backed up all the wikis. Theyr'e in wiki_backups/*.tar.xz (because xz -9e gives better compression than gzip or bzip2)

Moved old user directories in the /users/OLD/

None of the 3 dd backups I made were bootable - at boot, selecting the drive put me into grub rescue mode, which seemed to suggest that the /boot partition did not exist on the backed up disk, despite the fact that I was able to mount this partition on a booted computer. Perhaps for the same reason, but maybe not.

After going through various StackOverflow posts / blogs / other googling, I decided to try cloning the drives using ddrescue instead of dd.

This seems to have worked for nodus - I was able to boot to console on the machine called rosalba which was lying around under my desk. I deliberately did not have this machine connected to the martian network during the boot process for fear of some issues because of having multiple "nodus"-es on the network, so it complained a bit about starting the elog and other network related issues, but seems like we have a plug-and-play version of the nodus root filesystem now.

chiara and fb1 rootfs backups (made using ddrescue) are still not bootable - I'm working on it.

Nov 6 2017: I am now able to boot the chiara backup as well - although mysteriously, I cannot boot it from the machine called rosalba, but can boot it from ottavia. Anyways, seems like we have usable backups of the rootfs of nodus and chiara now. FB1 is still a no-go, working on it.

Valve configuration: Vacuum normal

Note: Tp2 running at 75Krpm 0.25A 26C has a  load high pitch sound today. It's fore line pressure 78 mTorr. Room temp 20C

Lompoc 4.3M and 3.7M Avalon

Here are the demodulation phases and rotation matrices tuned for the network. For the matrices, I am assuming that the input is [I, Q] and the output is [I,Q].

POP22
phi = 153 degrees
[[-0.894, 0.447],
 [-0.447, -0.894]]

REFL11
phi = 93 degrees
[[-0.058, 0.998],
 [-0.998, -0.058]]

REFL55
phi = -90 degrees
[[0.000, -1.000],
 [1.000, 0.000]]

AS55
phi = 7 degrees
[[0.993, 0.122],
 [-0.122, 0.993]]

In order to try to isolate CPU6 for the c1dnn neural network reconstruction model, I set the CPUAffinity in /etc/systemd/system.conf to "0" for the front end machines.  This sets the cpu affinity for the init process, so that init and all child processes are run on CPU0.  Unfortunately, this does not affect the kernel threads.  So after reboot all user space processes where on CPU0, but the kernel threads were still spread around.  Will continue trying to isolate the kernel as well...

In any event, this amount of isolation was still good enough to get the c1dnn user space model running fairly stably.  It's been running for the last hour without issue.

I added the c1dnn channel and testpoint files to the daqd master file, and restarted daqd_dc on fb1, so now the c1dnn channels and test points are available through dataviewer etc.  We were then able to observe the reconstructed signals:

We'll need to set the phase rotation of the demodulated RF PD signals (REFL11, REFL55, AS55, POP22) to match them with what the NN expects...

[ Gautam , Steve ]

c1susaux & c1iscaux were rebooted manually.

[alex, gautam]

Alex is going to have an undergrad work on a calibration optimization project on the 40m RTCDS system. For this purpose, we wanted to setup a "Simulated DARM loop". Today, Alex and I set this up. I figured we can use the c1tst model for this purpose. We basically copied the topology from Figure 2 of the h(t) paper. Attached are screenshots of the MEDM screens of the system we setup, and the simulink block diagram - the main screen can be accessed from the "SIM PLANT" tab in the sitemp.

It remains to setup the appropriate filters in the filter banks, and an EPICS channel monitor for monitoring the single excitation testpoint in the model. We also did not set up any DQ channels for the time being, as it is not even clear to me what channels need to be DQ-ed.

Do not leave organic trash or food boxes in the 40m to attrack ants !

[steve, jamie, gautam]

The machine that now serves as out Frame Builder, FB1, was sitting on top of megatron. I decided that this wasn't ideal, and asked Steve to get some alternative mounting solution. Today, he procured some shelves to put FB1 on. Jamie suggested looking for the slider-rail that came with the machine, and using that instead, as it will allow us to slide FB1 out of the rack as we do megatron and the old FB. But as luck would have it, the distance between the rack vertical posts is 26 inches, but the rail is 27 inches. So we had to accept using the less ideal solution of putting FB1 on two shelves, with no sliding option. Photo to be uploaded shortly.

For this work, I had to shutdown FB1 for about 1 hour between 3pm and 4pm. It seems to have come back up fine now.

Gabriele's PRCL/MICH reconstruction neural network is now running on c1lsc.  Summary:

• front-end model is called c1dnn, and is running as an experimental user-space process
• c1dnn is getting most of it's needed inputs from existing SHMEM IPC outputs from c1lsc
• none of the output signals from the network are being sent anywhere yet (grounded)
• c1dnn has not been integrated in any way, into the DAQ etc.  it is being run manually by hand, and will be completely shut down after this test

Simple MEDM screen I made to monitor the input/output signals:

The RTS process seems to run fine, but there is quite a bit of jitter in the CPU_METER, at the 50% level:

It's not running over the limit, but it is jumping around more than I think it should be.  Will look into that...

cpuset for cpu isolation for user-space model

The c1dnn model is running on CPU6 on c1lsc.  CPU6 was isolated from the rest of the system using cpuset.  The "cset" utility was used to create a "system" CPU set that was assigned to CPU0, and the kernel was instructed to move all running processes to that set:

controls@c1lsc:~ 2$ sudo cset set
cset:
         Name       CPUs-X    MEMs-X Tasks Subs Path
 ------------ ---------- - ------- - ----- ---- ----------
         root        0,6 y       0 y   343    0 /
controls@c1lsc:~ 0$ sudo cset set -c 0 -s system --cpu_exclusive
cset: --> created cpuset "system"
controls@c1lsc:~ 0$ sudo cset set
cset:
         Name       CPUs-X    MEMs-X Tasks Subs Path
 ------------ ---------- - ------- - ----- ---- ----------
         root        0,6 y       0 y   342    1 /
       system          0 y       0 n     0    0 /system
controls@c1lsc:~ 0$ sudo cset proc --move -f root -t system -k
cset: moving all tasks from root to /system
cset: moving 292 userspace tasks to /system
cset: moving 0 kernel threads to: /system
cset: --> not moving 50 threads (not unbound, use --force)
[==================================================]%
cset: done
controls@c1lsc:~ 0$ sudo cset set
cset:
         Name       CPUs-X    MEMs-X Tasks Subs Path
 ------------ ---------- - ------- - ----- ---- ----------
         root        0,6 y       0 y    50    1 /
       system          0 y       0 n   292    0 /system
controls@c1lsc:~ 0$ sudo cset proc --move -f root -t system -k --force
cset: moving all tasks from root to /system
cset: moving 50 kernel threads to: /system
[==================================================]%
cset: **> 29 tasks are not movable, impossible to move
cset: done
controls@c1lsc:~ 0$ sudo cset set
cset:
         Name       CPUs-X    MEMs-X Tasks Subs Path
 ------------ ---------- - ------- - ----- ---- ----------
         root        0,6 y       0 y    29    1 /
       system          0 y       0 n   313    0 /system
controls@c1lsc:~ 0$

I then created a set for the RTS process ("rts-c1dnn") on CPU6, and executed the c1dnn model in that set:

controls@c1lsc:~ 0$ sudo cset set -c 6 -s rts-c1dnn --cpu_exclusive
cset: --> created cpuset "rts-c1dnn"
controls@c1lsc:~ 0$ sudo cset set
cset:
         Name       CPUs-X    MEMs-X Tasks Subs Path
 ------------ ---------- - ------- - ----- ---- ----------
         root        0,6 y       0 y    24    2 /
    rts-c1dnn          6 y       0 n     0    0 /rts-c1dnn
       system          0 y       0 n   340    0 /system
controls@c1lsc:~ 0$ sudo cset proc -s rts-c1dnn --exec /opt/rtcds/caltech/c1/target/c1dnn/bin/c1dnn -- -m c1dnn
cset: --> last message, executed args into cpuset "/rts-c1dnn", new pid is: 27572
sysname = c1dnn
....

When done I just hit Ctrl-C.

I left the cpusets as they are, with all system processes in the "system" set.  This should not pose any problems since it's the identical configuration as would be if a normal kernel-level model was running in CPU6.

The c1dnn process and it's EPICS sequencer were shutdown after this test.

This happened again just now - it was roughly this time when this happened last night as well.

There was certainly an EPICS freeze of the kind we were used to seeing prior to replacing the martian wireless router sometime in late 2015 (or early 2016?). I was trying to run the dither alignment servos on the Y-arm at this time, and all the StripTool traces flatlined.

I took the opportunity to try accessing testpoints from the iscey ADCs - specifically C1:SUS-TRY_OUT, and it seemed to work just fine. However, I couldn't ssh into c1iscey.

Looking at the dmesg once I was able to ssh in eventually (~2 minutes deadtime tonight, I feel like it was longer yesterday but can't quantify), I see the following: not sure if there are any clues in here, or whether this is the correct log to check. But there are many instances of the nfs server related message in the log. Note that the system time-stamp corresponds to when this freeze happened.

Last night, I collected ~30mins of data for the vertex seismometer channels and the POP QPD PIT/YAW signals with the PRMI locked on carrier (angular FF OFF). The ITM Oplev loops weren't DC coupled, as they are in the full IFO locking sequence, but I feel like the angular FF filters can be improved - there are frequent sharp dives in the AS110 signal level which are correlated with large amplitude motion of the POP spot on the control room CCD monitor.

Repeating the frequency domain multicoherence analysis using BS_X and BS_Y seismometer channels as witnesses suggest that we can win significantly (see Attachment #1).

I've never really implemented feedforward filters - I was planning on using ericq's latest entry on this subject as a guide. From what I gather, the procedure is as follows:

1. Pre-filter the target (POP QPD PIT or YAW) and witness (BS_X, BS_Y) channels
   • Downsample the 2k target data and 256Hz witness data to 32 Hz (how to choose this?)
   • Detrend (linear?)
   • Apply elliptic low pass filter (previously, a 3rd order Elliptic Low pass with 3dB ripple, 40dB stopband attenuation, corner at 5Hz was used).
2. Filter the target signal (i.e. POP QPD PIT/YAW) by the inverse actuator TF.
   • This "actuator TF" is a measurement of how actuating on the angular DoFs of the PRM affects the POP QPD spot.
   • So by pre-filtering the target signal through the inverse actuator TF, we get a measure of how much the PRM angular motion is.
   • The reason we want to do this is to give the FIR filter that produces optic motion (output) given ground motion sensed by the seismometer (input) fewer poles/zeros to fit (?).
   • The actual actuator TF has to be measured using DTT, and fit - is there anything critical about this fitting? Seems like this should be just a simple pendulum transfer function so a pair of complex poles should be sufficient?
3. The actual Wiener filter is calculated by the function miso_firlev.m. There are many versions of this floating around from what I can gather.
   • This function requires 3 input parameters.
     • Order of filter to be fit
     • Witness channels (can be multiple)
     • Target channel (has to be single, hence the "miso" in the function name).
   • Today, at the meeting, we talked about weighting the cost function that the optimal Wiener filter calculator minimizes.
   • The canonical wiener filter minimizes the mean squared error between the output of the filter and the desired signal profile (which for this particular problem is the angular motion of the PRM, calculated by dividing the target signal by the actuator TF, knowing which we can cancel it out).
   • But as seen in Attachment #1, the main reduction in RMS comes below f=5Hz.
   • So can we weight the cost function more heavily at lower frequencies? From what I can find in previous calculations, it looks like this weighting happens in the pre-filtering stage, which is not the same thing as including the frequency dependent weighting in the calculation of the Weiner filter? The PSD and acf are F.T. pairs per the Wiener-Khinchin theorem so intuitively I would think that weighting in the frequency domain corresponds to weighting on the lags at which the acf is calculated, but I need to think about this.
   • What kind of low-pass filter do we use to prevent noise injection at higher frequencies? Does the optimal filter calculation automatically roll-off the filter response at high frequencies?
4. As I write this, seems like there is another level of optimization of "meta-parameters" possible in this whole process - e.g. what is the optimal order of filter to fit? what is the optimal pre-filtering of training data? Not sure how much we can gain from this though.

Some notes from Rana from some years ago: https://nodus.ligo.caltech.edu:8081/40m/11519

If anyone has pointers / other considerations I should take into account, please post here.

Summary:

The signal path for the ASDC signal is AS55 PD --> D990543 (interface board) --> D990694 (whitening board) --> D000076 (AA board) --> ADC Ch 31. Everything in this signal chain should be able to handle signals in the range +/- 10V, which should correspond to the full range of our +/-10V, 16bit ADCs. But the ASDC signal seems to saturate at ~2000 counts (i.e. turning up the analog whitening gain doesn't make the signal get any bigger than this). I investigated this a little more today.

Details:

• The ASDC signal is derived from the AS55 photodiode. According to the schematic, the Op27 that supplies this voltage is powered by +/- 15V, so the output should be able to swing between at least +/- 12V.
• The DC signal goes from the DB15 connector on the side of the PD to the LSC electronics rack, 1Y2, where it is interfaced with an LSC PD Interface Card, D990543. Again, per the schematic, the Op27 driving this voltage is powered by +/- 15V, and so the available output voltage swing should be greater than +/-12V.
• The D990543 output is to its backplane connector. There is an adaptor board hooked up to the backplane that makes these outputs available to a LEMO connector. A LEMO-SMA cable then pipes this output to a D990694.
  • I decided to test the functionality of this board.
  • Disconnected the SMA ASDC input signal (CH8 on the board).
  • Drove that channel with an SR function generator and gradually turned up the Vpp of the input signal (sine wave at 145Hz).
  • Monitored the ASDC channel on dataviewer while doing this.
  • Saw that the ASDC signal saturated at ~2000 counts. Turning up the signal amplitude did not have any effect.
• From the whitening board, the signal goes through an anti-aliasing module (D000076). The final stage LT1125s on these boards should also be supplied with +/-15V.

So the problem lies somewhere downstream of the D990694. There are other anomalous behaviours of this channel - e.g. engaging the analog whitening filters changes the DC offset of the signal. I am going to pull out this board to check it out.

Why does this matter? I want to calibrate the ASDC level (and eventually the other PD DC signals as well) into Watts. This is useful for IFO diagnostics, noise budgeting the shot noise level etc.

According to the AS55 schematic, the DC transimpedance is 66.7 ohms. I claim that the DC power on the AS55 photodiode during a DRMI (no arms) lock is ~1mW. The C30642 photodiode (InGaAs) responsivity is ~0.8 A/W. So I'd expect ~50mV to be the signal level into the ADC (assuming gain of all the other electronics in the signal chain at the start of this elog is unity). This corresponds to ~163 counts (since the ADC conversion factor is 2^16 counts over 20volts). The DC signal level I observed is ~200 counts. So things seem roughly consistent.

*Note: Despite my above statement, I don't think it is true that the AS110 PD has more light on it - the BS splitting the light between

AS55 and AS110 PDs is a 50-50 BS, and using the crude method of putting an Ophir power meter in front of both PDs and

monitoring the power while the Michelson was swinging around freely showed roughly the same maximum value.

The units were still off in my previous post. Here's the corrected, sanity-checked version:

I estimated the uncertainties based on a linear fit to the data I recorded with 75nW incident on the CCD and assumed a 5% uncertainty in that number. This is just an upper limit, to be safe. I had calibrated the power reading placing the Ophir power meter where the CCD would otherwise be and comparing it to the PD voltage of a picked off beam. In my previous figures the axes were mislabeled, so I reproduce them here:

Using the current camera position I recorded 50 exposures both with and without beam (XARM locked vs PSL shutter closed) and averaged the images to see how much the reading fluctuates. The exposure time was 10 ms, which left the maximum reported pixel value in all exposures below 3800 out of 4096. The gain setting was 100, which is what I used to calibrate the CCDs.

(*) I calculated the lens positions to focus at a plane 65cm from the front lens. We're pretty close to that, but I can't confirm the actual distance easily, so I assumed a 5cm error on the distance, which is where most of the error is coming from. This is also assuming uniform scatter.

(**) This is assuming 10W of circulating power

I tried moving the model to c1ioo, where there are plenty of free cores sitting idle, and the model seems runs fine.  I think the problem was just CPU contention on the c1lsc machine, where there were only two free cores and the kernel was using both for all the rest of the normal user space processes.

So there are two options:

• Use cpuset on c1lsc to tell the kernel to remove all other processes from CPU6 and save it just for the c1dnn model.  This should not have any impact on the running of c1lsc, since that's exactly what would be happening if we were running the model in kernel space (e.g. isolating the core for the front end model).  The auxilliary support user space processes (epics seq/ioc, awgtpman) should all run fine on CPU0, since that's what usually happens.  Linux is only using the additional core since it's there.  We don't have much experience with cpuset yet, though, so more offline testing will be required first.
• Run the model on c1ioo and ship the needed signals to/from c1lsc via PCIe dolphin.  This is potentially slightly more invasive of a change, and would put more work on the dolphin network, but it should be able to handle it.

I'm going to start testing cpuset offline to figure out exactly what would need to be done.

We don't really have to calibrate the lens, just the CCD, which we've done. It's more about knowing the true aperture size to know how much solid angle you're capturing to infer the total amount of scatter. For our custom lens tubes this is the ID of the retaining ring.

The Edmund Optics lens tube looks tempting, but itcomes at a price. Thorlabs sells lens tubes that offer a more flexibility than what we have right now, so I bought a few different ones, and also more 150mm 2" lenses. This will allow for more compact solutions and offer some in-situ focusing ability that doesn't require detaching the lens tube like now. Should be here in a couple of days, then we'll be able to enclose the GigE camera in the viewport can with a similar field of view we have now.

I also bought a collimation package for the AS port fiber stuff so we can move ahead with the ringdown measurements and also mode spectroscopy.

An mx_stream glitch would have interrupted data flowing from the front end to the DAQ, but it wouldn't have affected the heartbeat.  The heartbeat stop could mean either that the front end process froze, or the EPICS communication stopped.  The fact that everything came back fine after a couple of minutes indicates to me that the front end processes all kept running fine.  If they hadn't I'm sure the machines would have locked up.  The fact that you couldn't connect to the FE machine is also suspicious.

My best guess is that there was a network glitch on the martian network.  I don't know how to account for the fact that pings still worked, though.

I was looking at the ASDC channel on dataviewer, and toggling various settings like whitening gain. At some point, the signal just froze. So I quit dataviewer and tried restarting it, at which point it complained about not being able to connect to FB. This is when I brought up the CDS_OVERVIEW medm screen, and noticed the frozen 1pps indicator lights. There was certainly something going on with the end FEs, because I was able to ping the machine, but not ssh into it. Once the 1pps lights came back, I was able to ssh into c1iscex and c1iscey, no problems.

Could it be that some of the mx processes stalled, but the systemctl routine automatically restarted them after some time? 

So this wasn't just an EPICS freeze?  I don't see how this had anything to do with any of the work I did earlier today.  I didn't modify any of the running front ends, didn't touch either of the end station machines or the DAQ, and didn't modify the network in any way.  I didn't leave anything running.

If you couldn't access test points then it sounds like it was more than just EPICS.  It sounds like maybe the end machines somehow fell of the network momentarily.  Was there anything else going on at the time?