Tonight we noticed that there were significant improvements to be had in the predicted DARM Wiener filtering FF performance by using weighting filters and more taps in the FIR filter.
The plots below tell the story:
The first one shows the improvement in the residual (black & blue) by applying a weighting filter. The weight filter tilts the spectrum up at HF and applies and all real pole BP from 10-20 Hz.
The second plot shows the improvement gotten by using 3000 instead of 2000 taps for the Wiener filter. With the larger number of taps we not only get the big improvement at LF, but also some beefy reduction in the higher frequency stack modes and the LOS roll mode.
I'm not sure why we haven't run across this before; the weighting filter was arrived at today by just iterating by hand on the placement of poles and zeros until the trace looked nice.
Jenne is going to run this new filter on the S5-month that we have been using for stationarity testing.
* Some notes:
** this Wiener stuff is faster, by far, on rossa than either megatron or rosalba or my laptop. More than a factor of 3.
*** there is a bug with Macports/Matlab - if you get fftw3 with Macports, it sets itself as the right version to use. This confuses matlab in some cases.
if you get the error about libfftw3.dylib, whe trying fft in matlab after installing macports, then you can fix it by setting the Matlab lib/ path with the fftw libraries to be ahead of /opt/local/lib in the LD_LIBRARY_PATH in your .cshrc.
I removed the rsync backup from nodus' crontab temporarily so as to not have multiple backup jobs running. The job I started from yesterday was still running. Hopefully the backup will finish by Monday.
The line I removed was:
0 5 * * * /opt/rtcds/caltech/c1/scripts/backup/rsync.backup
I swapped over to a 3x longer cable (old 65 ft. Pasternak cable from ancient 40m days). The old one was 6m, the new one is 18.2 m. It was already coiled up so I put it into a tupperware box to shield it somewhat from the HVAC wind.
The noise went down nearly proportional to the length (after I recalibrated the DAQ channel for the ~3x higher phase->voltage gain). With this length, the peak-peak mixer range is 5.5 MHz, so still enough to go an FSR here.
I give credit to the low frequency improvement entirely to Tupperware for their excellent containers. The current noise limit is most likely the SR560.
I noticed that the RMTEMP channel was spiking myteriously when Kiwamu opened the PSL door. We found out that the LEMO connectors would intermittently short to the case and cause ~1 deg steps in the temeprature.
We have removed the case and examined it. Not only were the connections to the box intermittent, there was a cold solder joint inside on an unsecured flying add-on opamp. The whole thing is a giant hack.
PK was the last person to work on this box, but I'm sure that he wouldn't have left it in this state. Must be gremlins.
The LEMO connectors on the front are the ones touching. The LT1021 is the badly soldered part.
375 mW is way too much light. We must never put more than 100 mW on any of these diodes. We don't want to blow up more diodes like we did with the WFS. The InGaAs diodes often show an excess dark noise before they finally let go and completely fail. This one may show excess during the shot noise testing.
We should ensure that the beam paths are engineered so that none of these new detectors ever sees such high light levels.
The DC path should be made to let us see a 10V from the differential EPICS readout when there is 100 mA of photocurrent (i.e. an effective 100 Ohms transimpedance):
0.1 A * 10 V/A * 5 V/V * 2V/V
The last factor of 2 is from the single to differential conversion.
If we really only get 15 mV from 375 mW, then this diode or the circuit is broken.
This is just a listing of CDS problems I still notice today:
MC2-MCL button was left ON due to BURT failure. This, of course, screws up our Green locking investigations because of the unintended feedback. Please fix the BURT/button issue.
Although Joe and Kiwamu claim that they have inserted the correct DAQ names for the OPLEVs (e.g. PERROR and YERROR) back in Jan. 11, when I look today, I see that these channels are missing!
All of the SUS used to have only 1 filter module for SIDE. They now have 3 filter modules for SIDE just like the other DOFs.
Today I moved the filters around so that the sensor filters are in SDSEN, the servo filters are in SUSSIDE, and the dewhitening for the coil is in SDCOIL.
I noticed along the way that the bounce/roll mode notches for all of the suspensions are still set for the frequencies of the previous suspensions. Suresh has 'volunteered' to find the new frequencies and make the new bandstop filters by looking up the seminal work on this by Dan Busby / Sam Waldman.
I forgot to elog that last night I touched up the MC2_TRANS QPD setup. I was perplexed by it always going out of alignment so I investigated.
I found that the fork clamp for the steering mirror for the QPD was not tightened. Shame. The beam diameter was equal to the aperture of the QPD and was clipping. Double shame.
I added a lens and tightened the mounts and centered the beam at ~9 PM yesterday. You can see in the attached trend that the measured power went up by ~10%.
Later, there's a big gap where Valera and Steve change out the PMC. You can see that the MC REFL voltage goes from 4.5 V to 5 V (10% increase in the power delivered to the MC).
There's essentially no change in the total transmission - this indicates that although the PMC transmission is now higher by ~10%, the matching to the IMC has been degraded by an equivalent fraction.
Needs some mode matching work.
This is the 140 ft. MFD measurement of the VCO phase noise. It is open loop and so should be a good measurement. The RMS is 30 Hz integrated down to 2 mHz.
I don't know why this doesn't agree with Suresh's measurements of the same thing which uses the PLL feedback method.
In BLUE, I also plot the frequency noise measured by using a Stanford DS345 30 MHz func. generator. I think that this is actually the noise of the FD (i.e. the SR560 preamp) and not the DS345. Mainly, it just tells you that the PINK VCO noise measurement is a real measurement.
I calibrated it by putting in a 5 kHz_pp triangle wave on the sweep of the DS345 and counting the counts in DV.
DTT stopped working for recent data. An 'ls' in the frames/full/ directory reveals:
drwxr-xr-x 2 controls controls 258048 Feb 3 12:26 9807
drwxr-xr-x 2 controls controls 258048 Feb 4 16:13 9808
drwxr-xr-x 2 controls controls 262144 Feb 5 19:59 9809
drwxr-xr-x 2 controls controls 258048 Feb 6 23:46 9810
drwxr-xr-x 2 controls controls 258048 Feb 8 03:33 9811
drwxr-xr-x 2 controls controls 262144 Feb 9 07:19 9812
drwxr-xr-x 2 controls controls 253952 Feb 10 11:06 9813
drwxr-xr-x 2 controls controls 266240 Feb 11 14:53 9814
drwxr-xr-x 2 controls controls 266240 Feb 12 18:39 9815
drwxr-xr-x 2 controls controls 266240 Feb 13 22:26 9816
drwxr-xr-x 2 controls controls 262144 Feb 15 02:13 9817
drwxr-xr-x 2 controls controls 253952 Feb 16 05:59 9818
drwxr-xr-x 2 controls controls 241664 Feb 17 09:46 9819
drwxr-xr-x 2 controls controls 28672 Feb 17 12:22 9820
drwxr-xr-x 2 controls controls 32768 Feb 17 15:06 6663
drwxr-xr-x 2 controls controls 73728 Feb 17 23:39 6664
controls@fb /frames/full $ date
Thu Feb 17 23:39:27 PST 2011
Frank put his low noise preamp info here.
I suggest that we build these (using Altium) but replace the cheapo transistors with the high class MAT03 matched BJT pair from Analog Devices.
This will allow us to have a pre-amp better matched to the noise of the mixers down to low frequency.
Looks like there was a mysterious loss of data overnight; since there's nothing in the elog I assume that its some kind of terrorism. I'm going to call Rolf to see if he can come in and work all night to help diagnose the issue.
When Koji and I were massaging the MC, we noticed that the oscillations were at 48.5 kHz. They were pretty huge and are probably what you're seeing on the beat. My guess is that they are the PZT resonances of the PSL 2W NPRO; we need to put a notch in the FSS box - it still has the notch from the old NPRO.
Its been well noted in the past that sweeping the PMC at high power leads to a distortion of the transmitted power curve. The explanation for this was coating absorption and thermo-elastic deformation of the front face of the mirrors.
Today, I did several sweeps of the PMC. I turned off its servo and tuned its PZT so that it was nearly resonating. Then I drove the NPRO via the HV driver (gain=15) with 0-150 V (its 1.1 MHz/V) to measure the PMC transmitted light. I adjusted the NPRO pump diode current from 2A on down to see if the curves have a power dependent width.
In the picasa web slideshow:
There are 3 significant differences between this measurement and the one by John linked above: its a new PMC (Rick says its the cleanest one around), the sweep is faster - since I'm using a scope instead of the ADC I feel free to drive the thing by ~70 MHz in one cycle. In principle, we could go faster, but I don't want to get into the region where we excite the PZT resonance. Doing ~100 MHz in ~30 ms should be OK. I think it may be that going this fast avoids some of the thermal distortion problems that John and others have seen in the past. On the next iteration, we should increase the modulation index for the 35.5 MHz sidebands so as to get a higher precision calibration of the sweep's range.
By eye I find that the FWHM from image #4 is 11 ms long. That corresponds to 300 mV on the input to the HV box and 15 V on the PZT and ~16.5 MHz of frequency shift. I think we expect a number more like 4-5 MHz; measurement suspicious.
There are 3 standard techniques to reduce this effect:
1) Stabilize the end laser by sensing the green light coming into the PSL before recombination and feeding back with SR560 (this is the only one that you should try at first).
2) Moving to the center of the MFD fringe via ETM steps.
3) Auto-alignment of the beam to the arm.
Ridiculous and hacky. Digital stabilization removed as well as the old "leave a pile of equipment on a stool" strategy.
We used a a BNC cable to send a pickoff of the beam before the recombination to the end via an SR560.
I did some work on the ETMY real and Sim.
It seems like there is still a problem with the input whitening filters. I believe the Xycom logic is set such that the analog whitening of the OSEM signals is turned ON only when the FM1 is turned OFF. Joe has got to fix this (and elog it) so that we can damp the suspension correctly. For now, the damping of the ETMY and the SETMY require different servo gains and signs, probably because of this.
Using 2 dBm for a Level 7 mixer is so bogus, that I will dismantle this as soon as I come over.
Its going to need some kind of way to locate the PMC on the top. In the previous design, we had the 3 balls to decouple the body from the base. That design was flawed due to the roughness of the holes in the PMC body.
Also probably need some kind of relief on the bottom. Its possible that it would be OK like this, but I am unsure if the shop can maintain the flatness we want over the whole length and/or the flatness of any given (OLD) optical table over ~8". Its probably not a good idea to have to torque this (aluminum?) to make it conform to the optical table's shape.
Ooh. Can you explain the purpose of the resistors which are connected to the (+) inputs? It looks like some real electronics ninjitsu.
I am a little concerned about using these low pass filters so close to the band edge. Recall that there is no on-board preamp for the RF input to the mixer.
So, if the input impedance of the filters is not 50 Ohms, we will get some unwanted reflections and sensitivity to cable length.
I think its worth while to check the impedance or S-parameters of these things with the LO activated to find out if we need to remove them or not.
One way to avoid some of the bad stuff in there is to take the 1 dBm input and amplify it to ~21 dBm before splitting and sending in to the Level 17 mixers.
One way to do this is by using the A3CP6025 from Teledyne-Cougar. Its an SMA connectorized amp which can put out 25 dBm and has a gain of 24 dB. We can just glue it onto the demod boards. Then we can remove the ERA-5 amplifiers and just use the broadband splitter as Kiwamu mentioned.
The performance plots for POX_11 in the wiki are horrendous and the schematic is missing.
I opened up the box and found all kinds of horrors. There were multiple tunable parts and a flurry of excess nonsense.
The top 2 worst offenders:
1) The main tunable inductor was busted. I removed it and found that the coil was open. Too much indelicate soldering in its vicinity had melted the wire. Someone had put extra inductors and capacitors around it to make it seem as if the PD was working fine, but the noise performance was off by a factor of ~100.
2) The MAX4107 had a 1.4k series resistor. This make the output go through a 1450/50 voltage division which is not nice for the SNR. I removed it.
I then struggled for awhile to get a sensible response. It turned out that the TEST IN input was not giving me a sensible TF. Jenne and I fired the Jenne laser at it and found that the 11 MHz main resonance is there. In the morning I'll finish this off and post more results. I think its going to end up being fine.
I used the Jenne AM laser to tune up the PD (used to be POX_11 but now is called REFL_11). In addition to the notch at 22 MHz, I have also put in a LC notch at 5*f = 55.3 MHz. The transfer function below shows the RF OUT of the PD v. the drive to the laser. I didn't divide out by the 1811 because its not on the EE bench.
I worked on AS_11 today. Its ready for its noise / optical gain calibrations. I have left it on Suresh's desk.
This was one of the 24.5 MHz Black Box (Ben Abbott) style RFPDs rescued from LLO. The tunable inductor that was installed was too small to get the frequency down to 11 MHz and so I swapped in one of the shielded, ferrite core ones from our '7mm' CoilCraft kit. It had a range of 1.2 - 1.8 uH according to the datasheet.
I wasn't able to simulataneously get the peak at 11.06 MHz and the notch at 59.3 MHz and so I took Koji's advice and tuned the peak best. The plot above shows how the notch is slightly off. I think its not a problem; to get it better we would have to change out the inductor for the "2-omega" notch, but I was too lazy. The thinking is that its more important to have the gain be symmetric around the signal readout frequency so as to not imbalance the audio sidebands.
Since this one is going to be AS_11, we think that the 22 MHz signal will be tiny: the transmission of the 11 MHz sidebands to the dark port is small. If we later want to put in a 22 MHz notch anyway, there is space to do this via the 'active notch' pads around the MAX4107.
For the above plot, I used the Jenne laser. The DC output of the PD was ~30 mV (~0.6 mA). The RF drive to the laser was -10 dBm: no saturations. I have calibrated out the cable responses, but not using the 1811 setup, so the absolute calibration has yet to be done.
Also, it needs some new stickers. It would be handy if someone can figure out how to get some sheets of stickers that we can put into the printer. Then we can laser printer all of the data onto the stickers and stick them to the RFPD box.
* Farfalla, a lab laptop, seems out of network.
If you look at the real host table instead of the misleading host table in the wiki, you will see that someone has deleted Farfalla from there. She needs to be re-added.
I found that the MC autolocker was OFF. Kiwamu says he turned it off because its slow. Suresh says that he has some feelings that maybe something is wrong. I'll let them describe what they know about the MC in an elog.
I checked the trend of the MC and PMC transmissions for the past 30 days:
Looks like the alignment has been drifitng. PMC was corrected recently by Koji, but the alignment of the input beam to the MC or the MC itself has to be fixed. Has someone been twiddling the MC SUS alignment biases??
I think that the gain ramping time (_TRAMP) should be set to 1 second for all filter modules by default. We don't want them to switch instantaneously except in a few special cases.
So Jamie and I wrote a script (in scripts/general/) which sets all of these fields to 1 for a given system. The name of the system is an argument to the script. e.g.
> setTRAMP LSC 1
The idea is that we set it once and then from then on, its captured by the autoBURT. Of course, we have to run this script each time we add new filter modules to a model.
It seems like the best option would be to make the MCASS just adjust the SUS biases and center the beams on the suspended optics. Is this not possible somehow?
I looked at the PSL temperature box. It started out as D980400-B-C. Then it was revised by Peter King as per the LHO mods E020247.
There are some more things to do to it to make it useful for us:
** Frank reminds me that we don't use the TIdal or VME external inputs anymore since we moved to the EPICS/Perl PID control. So all we have to do is make sure these inputs are hardware disabled/disconnected.
I re-installed the box (@ ~8:15) after reflowing some of the solder joints. I will observe it over night and then remove the 1K resistors. Attached is a 8 hour minute-trend.
I compared this 24 hour trend with the one from this day exactly one year ago. Seems the same, so now I can make the resistor change.
I found that a He-Ne laser which has been used for ETMX_OPLEV was NOT giving the light.
Since I didn't find the switch key for it I have no idea if the laser is simply off or dead.
The dead laser was replaced by new JDSU 1104P of 2.6mW. The return beam is big ~5 mm diameter of 0.3 mW, 1400 counts
Whenever replacing any Oplev laser, please also put into the ELOG when it was installed so that we have an electronic record of the laser lifetime.
Rana also advised that we must use the boards which have the piggy-back amplifiers on those signals which are most useful. We referred to Alberto's thesis and chose POY55 (MICH and SRCL), REFL11(PRCL) and AS55 (DARM) as the most useful signals. We currently have these amps on AS11, REFL11 and AS55. We need to convert either AS11 or REFL11 into a POY55. Since we need to troubleshoot REFL11, I thought we might as well modify that and in the process also fix its Q output. So I renamed AS11 as REFL11 and will convert the old REFL11 into POY55 tomorrow.
I think we should leave them as is; the AS11 was made by taking into account the SB levels at the AS port and should not become REFL11. We should instead convert one of the old 25 or 33 MHz diodes into a POY55.
We decided to rename the Input Beam channels (while keeping temporary backwards compatible aliases) as:
C1:ASC-IB_POS_X, C1:ASC-IB_POS_Y, C1:ASC-IB_ANG_SUM, etc.
Looks like 2 different MEDM Snapshot functiions (at least) are broken.
The regular update of the screens here as well as the usual "Update Snapshot" and view "previous snapshot" button on all of the auto-generated screens.
Also, how do we add the snapshot button to the custom made screens?
There has been some input matrix diagonalization in the past by Yuta and Kiwamu, but I find the automation to be not totally satisfactory.
It would be better if we could automatically fit the data to find the Suspended optic eigenfrequencies and then use that to get the matrix. So I wrote a peak fitter to get the matrix.
It gets the data from mafalda with NDS2, then it makes the PSDs, and then starts with some initial guesses (based on looking at the plots) and them uses fminsearch to get the peak frequencies and Q's.
Using the output of this, we can use Yuta's method and take the passive transfer functions with the free swing data (from April 30, so we got do do it quick) to get the input matrix.
Doing the SUS input matrix is nice for having good damping (as long as we remember to include SIDE), but my motivation is to produce a good null stream from the 4 face sensors so that we can estimate the sensor noises at all times.
Just FYI, the 3113 is a 12-bit ADC, so we won't get very good resolution out of these. We should get one of those purple boxes.
Also, I corresponded some with Dave Barker. The 40m Wiki at LHO is down because of disk errors. He's working on it and will let us know. But suggests that we move it to Caltech since he'll turn the box off at the end of the year.
Mon May 16 13:57:49 2011 Wiki is back up.
This is OK....but, the input matrix should come from the same place as the regular input matrix: i.e. it should be just another row like CARM, DARM, etc. rather than have its own screen.
Also, I think a nice mod to all the matrices would be if the ORANGE triangle was only visible when there's a signal flowing through it.
Aka, from a hotel in Pisa.
Restarted Thu May 19 00:21:49 2011 to recover from Jenne's Italian terrorism.
I've moved all of my SOS peak fitting stuff into the scripts area so that Leo can make it better:
findPeaks.m gets the data and makes the fitted spectra that I put in the previous entry.
findMatrix.m is the barely started script that ought to take the TF data and output the matrix to the MEDM screen.
I've finished tuning POY11 and it is now sitting on top of the analyzer waiting for Koji to test its noise.
Some weeks ago, Joe, Jamie, and I reworked the ETMY controls.
Today we found that the model rebuilds and BURT restores have conspired to put the SUS damping into a bad state.
1) The FM1 files in the XXSEN modules should switch the analog shadow sensor whitening. I found today that, at least on ETMY and ETMX, they do nothing. This needs to be fixed before we can use the suspensions.
2) I found all of the 3:30 and cts2um buttons OFF AGAIN. There's something certainly wrong with the way the models are being built or BURTed. All of our suspension tuning work is being lost as a consequence. We (Joe and Jamie) need to learn to use CONLOG and check that the system is not in a nonsense state after rebuilds. Just because the monitors have lights and the MEDM values are fluctuating doesn't mean that "ITS WORKING". As a rule, when someone says "it seems to work", that basically means that they have no idea if anything is working.
3) We need a way to test that the CDS system is working...
I focused these lenses so that we could get a clean image of the mirrors and the OSEMs.
Our goal is to have an image where the optic diameter almost fills the entire monitor. We want the focus to be adjusted for the YAG beam (which is almost the same as focusing for the OSEMs). This will NOT produce a nice image of the cage using visible light, but that is just fine.
When Justin Garofoli was here he found a nice lens combo that did the job, so if anyone can find his old email or elog, lets just go back to that.
For now, we do not need a camera/lens system to focus very tightly on the center of the optic.
While answering a quick question by Kiwamu, I noticed we only had second trends going back to 99050000 GPS time, May 27th 2011.
Trends (I thought) were intended to be kept forever, and certainly longer than full data, which currently goes back several months.
Jamie will need to look into this.
Our concept is to keep second trends for 1-2 months and minute trends forever. The scheme that Alan had worked out many years ago had it such that we could look back to 1998 and that the minute trends would be backed up somehow.
If its not working, we need to get Alan's help to recover the previous configuration.
Koji and I found 2 RFPD boxes to send to LLO. We've put them onto Steve's desk to be overnighted to Valera.
One of them is our old 21.5 MHz gold box RFPD from the FSS (which we don't use). The other one is a 2mm gold box one which was previously tuned for 66 MHz.
They shipped out on Friday
There was a rogue, undocumented, gateway process running on NODUS since ~4 PM. This guy was broadcasting channels back into the Martian and causing lockups in the IOO controls. I did a kill -9 on its process.
Someone will pay for this.