From peter@physics.columbia.edu Thu Sep 8 09:05:17 2005 Date: Thu, 08 Sep 2005 09:05:05 -0400 From: Peter Kalmus To: Gabriela Gonzalez , Michael Landry , Rick Savage , Brian O'Reilly <" irish"@ligo-la.caltech.edu>, Doug Lormand , Corey Gray , Szabolcs Marka , Daniel Sigg , Keith Riles Subject: summary of Pcal known work, and conversation with Keith Riles Hi everyone - Yesterday I talked to Keith to get a list of potential photon calibrator "show stoppers" and start to make a plan for dealing with them. I would expect some further discussion on this subject during detchar on Friday. Hopefully today will see progress towards a comprehensive Pcal plan of action. Please comment. As expected, Keith was concerned primarily with noise issues. There are three kinds of noise: Pcal laser glitches, laser intensity noise, and harmonics. 1. Intensity noise. Here's a link to an elog addressing this issue: http://blue.ligo-wa.caltech.edu/ilog/pub/ilog.cgi?group=detector&task=view&d ate_to_view=08/17/2005&anchor_to_scroll_to=2005:08:18:10:07:38-pete If this was done correctly, it looks like the laser intensity noise is always at least a factor of 100 below AS_Q (larger factor at higher frequencies). Please consider this carefully and let me know if there's anything you're not comfortable with. We would also like to check coherence between the photodiode readout and the GW channel. Finally, a nice to have is more data points (especially at lower frequencies, e.g. 50Hz) and then a log graph showing the noise floor ratio as fctn of frequency. 2. Glitches. John Z. has a tool that monitors glitches. Choosing run frequencies is a prerequisite for starting this monitoring. It might be sensible to start monitoring now, and update the monitoring when the frequencies are finalized. We should gather as much information as possible about glitches before the start of S5. 3. Harmonics. Keith thought that if we were to run at 1300Hz, we could probably tolerate the harmonics at ~2600, 3900, and 5200Hz (well, he said "one or two harmonics") so long as they are "clean". Clean means "occupying only one bin and not wandering". Here's an elog where the harmonics are visible, so everyone has a sense of what we need to deal with. The blue plot in the top graph is what you want to look at: http://blue.ligo-wa.caltech.edu/ilog/pub/ilog.cgi?group=detector&task=view&d ate_to_view=08/16/2005&anchor_to_scroll_to=2005:08:17:08:34:48-pete Input from awg is a pure sine wave. In this plot in particular, the AOM is being driven fairly hard (probably more than half of its total dynamic range - I apologize for not knowing this more precisely), which I suspect is the cause of the relatively large harmonics. I would like to perform an experiment where harmonic strength (ratio of harmonics to fundamental) is observed as a function of modulation depth (portion of AOM dynamic range used). At the least, harmonics should be observed at the S5 run frequencies and amplitudes. Another possiblity suggested by several people to me is to tinker with the input waveform so as to acheive a more pure sine wave output. Another concern is laser stability. Here's a link to the most recent elog showing a ~30 day trend from the photodiode readout channels:http://blue.ligo-wa.caltech.edu/ilog/pub/ilog.cgi?group=detector&task=view&d ate_to_view=08/20/2005&anchor_to_scroll_to=2005:08:20:15:20:53-pete MidX is currently sitting in the optics lab. It would be prudent to check to see if the polarization has changed, and this is the cause of degradation. We'd like to use LineMon to monitor laser stability and line amplitude. Once we finalize run frequencies, Sergei can add channels to LineMon. We would like to monitor the excitation, the photodiode readout, and the lines in DARM_ERR. So I guess at LHO that means 12 new entries to LineMon. Plan of action. Please let me know if something needs adding, removing, or changing of priority. I've attempted to make this list comprehensive 1. Essential tasks (dealing with show stoppers) 1. Choose our run frequencies. Let's work to dissolve this task into smaller subtasks. 2. Coherence between PD readout channel and DARM_ERR or AS_Q from 0 to 7kHz. 3. Contact John Z. about glitch monitoring. 4. Check harmonics line widths. 5. Convince ourselves that Pcal lines / harmonics will not wander. (Not sure how to do this. I doubt they would wander. I assume awg is very stable, and tested, and I suppose there's no more reason for the Pcal lines to wander than the official cal lines?) 6. Contact Sergei and get LineMon going. 7. Plumb together our Pcal processing pipeline - e.g. taking the measurements (once per minute, or possibly a longer integration) and calculating the calibration factor (there's already a Matlab program that does the calculation). 8. Need to have excitations written to frames. 9. LLO status and Pcal plan. 2. Second priority 1. More data points on laser intensity noise, graph of noise floor ratio as fctn of frequency 2. Experiment observing harmonic intensity as a fctn of AOM modulation depth (% of dynamic range used) 3. Resolve state vector issue (should there be a photon calibrator bit or not) 4. Remeasure photodiode calibration factor until we understand its drift and random error. 5. Resolve disagreement between Scientech and Ophir power meters. 6. SciMon photon calibrator tasks must be written down 7. Reduce harmonics by tinkering with input wave form 8. Check MidX laser polarization to see if it's the cause of power degradation (do this before re-installing MidX Pcal box) 9. Servo line amplitudes? I look forward to hearing from all of you. Thanks, Peter -- _________________________________________________________ Peter Kalmus Department of Physics Columbia University lab: 212-854-5762