TMS Meeting Minutes, February 18, 2021

Attendees: Andrew, John, Matthieu, Trenton, Olga, Yang (Zoom Meeting)

  • Andrew believes we may have a sign error somewhere in one of our focus correction offsets, which is causing donuts early on in the night
  • Note, the FPIA matrix is not optimal, especially since we see oscillation
  • We are driving in focus when we run Yang's code, which apears to be causing the focus divergence
  • When the FPIA integrated offsets are applied Z4, Z11, Z22, see good correlation between TMS conrrections and FPIA controls
  • Good correlation between TMS data and FPIA data is very encouraging
  • Proposes engineering time on sky, running FPIA continuously and controlling with the new TMS software to rapidly isolate this error
  • Andrew suggests commanding 500 nm Z11 and 250 nm Z22 with the PSF GUI while running new TMS software, and show that we only have tiny changes on the order of 10 microns

Yang's Update
  • First, sent an email to the German engineers a couple of days ago (FWD to Andrew, Heejoo)
  • Asked if it would be fine for us to take measurements continuously in a loop without any pause
  • They did not specifically address the question, but Yang things it might be okay
  • Just make sure to receive the Etalon reply before sending another measurement request
  • The conversation veered into a parallel issue (TCP/IP disconnect issues)
  • Etalon has diagnosed the issue (had to do with the MATLAB kernel they run in their software)
  • Etalon does not seem to have a solution
  • Appears that we shouldn't leave our system running for a long period of time
  • Andrew to rephrase the question to Kai to get a specific answer (Do we do any harm by continuously polling measurements?)
  • Yang has not been able to spend too much time on TMS because he released new software for AO
  • Regarding averaging temperature telemetry, Yang thinks there may be some problems with doing this through DMS
  • Andrew notes that applying the standard deviation filter to the atmosphere corrected lengths will allow us to sense some of the variability in temperature
  • Yang plans to set aside the temperature telemetry work for now
  • Other than that, Yang just needs to tweak some error messages, should be ready in a very short time
  • Yang would like some daytime testing time after the software is released
  • Currently, still lacking a manual channel configuration setup (which is only a problem when we have an intermittently failing channel)
  • Andrew suggests that this could be implemented by dropping a specific channel from the next reference if a single channel measurement failed repeatedly, say 10 times

Andrew's Update:
  • Analyzing 4 hours of passive observing LBT data with his GMT analysis code
  • Fit raw LBT data with high-order polynomial to take out the real motion and isolate the noise
  • Once noise is isolated, have about +/- 3 microns
  • By subtracting mean deviation of nine different channels from their beginning channel length, the data collapses to a much tighter noise pattern
  • This shows that there is correlated noise and uncorrelated noise
  • Andrew strongly suspects there is a structural vibration, such as within the M1 cell, or the LBC hub
  • A good test to confirm this suspicion would be to look into the diameter lines
  • Obtained mean correlated error and standard deviation for each channel
  • Also obtained mean uncorrelated noise, with about 150 nm magnitude (in the right ballpark)
  • Constructs a distributions of correlated error and uncorrelated error to use with Monte Carlo simulation
  • Andrew finds that the uncorrelated noise is a problem, while the correlated noise is no big deal
  • It appears the uncorrelated noise gives most of the error
  • Expects that the reference measurements do not hold up to evolving uncorrelated noise over time
  • Planning to look into the history of the evolution of noise on LBT
  • Looks like long term references may be limited b/c of uncorrelated noise
  • Andrew notes that there could be some hysteresis in the collimator mounts, temperature equilibrium of the telescope structure, etc.
  • Andrew planning to set up a model like the one presented for the LBT truss and look at the change in reference over time, the growth of correlated and uncorrelated noise
  • This is a question of how well the system sets up at the start of the night from a historical reference (this is Andrew's main concern, at present)
  • Andrew is almost certain that we will be able to set the system better than the open-loop models
Next Steps:
  • Day testing of Yang's new software
  • Obtaining some engineering time
    • Olga notes that there are half nights coming up starting Wednesday of next week and going through the following Monday which we might make use of
  • Fiber routing on SX side is mostly done
    • As of yesterday, only a few more tasks to finish regarding lengthening in some places and coiling up in others

-- TrentonBrendel - 18 Feb 2021
Topic revision: r1 - 18 Feb 2021, TrentonBrendel
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback