TMS Meeting Minutes, October 14, 2021

Attendees: Andrew, Breann, Christian, Heejoo, John, Olga, Trenton, Yang (Zoom Meeting)

Updates:
  • Recently, had a few days for passive mode data acquisition
  • John sent out a very interesting plot recently TMS Corrections for /tmp/tms_left_20211011_telescope.txt
    • John broke down the plots into individual hour blocks
    • John broke out the TMS corrections into accepted measurements and rejected measurements (big shape vs. small shape)
    • The last measurement on the 8 to 9 UT plot were rejected b/c of FPIA running
    • Whenever the green line, instrZ drops to -0.8 mm, FPIA is running
    • John thinks the TMS-measured focus is changing b/c FPIA/PSF is doing things to Z11 and Z22 (though the system should be fixing this in offsets)
    • During periods of continuous operation, we have one or two periods where we drop measurements
    • In the 10-11 UT left side plot, there is funny stuff happening in focus just before 10.5 UT (see mode3Z big square jumps) on the order of 20 microns
      • This 20 microns is well within our resolution range (we are resolving at 1 micron)
      • There don't appear to be any filter changes happening during this time, and guiding corrections are small
      • Andrew doesn't see any way where our TMS setup could allow a pure Z change with no other RX, RY, RZ changes
      • It may be caused by some guider issues, because the DIMM seeing changed substantially at this time
    • Switching over to right side, we look for points where all X Y Z RX RY RZ jump
      • Andrew believes that any instances where we have a dropout in all pose elements at once is the most interesting
      • There is one point at 9.96 UT where there is a pointing offset
      • It looks like TMS calculated a correction when the telescope is already moving
      • The python code should have been able to prevent the measurement with the telescope moving
      • John to look into this
    • For the 10-11 UT right side plot, we get some funny business around 10.8
      • The TMS dots should be immune to FPIA focus offsets
      • The first uptick in instrZ is for the filter focus offset applied at the start of FPIA, followed by the nosedive down to -0.8 mm focus offset for defocused pupils
      • Overall, over long periods of time, TMS is showing good stability
      • The excursions of TMS are on the order of 1 pixel, around 15-20 microns
    • John was hoping to see more of the effects of TMS sensing the necessary correction to get in focus and collimated
    • There is a bit of an offset between the FPIA best solution and the TMS best solution which is greater than John would have expected
  • 1014 passive data
    • After 12, small rotations made and TMS makes corrections for each one
    • Near 9 UT, the corrections are going up and down all over the place (oscillatory but smooth)
      • This is right near zenith, where there have been some issues noted with the LBC hub
    • John suggests tracking across the meridian a couple times on a bad weather night to dig into this issue seen around zenith
    • It looks like we will need to switch off Z4 corrections while using active TMS corrections
  • As a note, the way that Yang's code works, it will append TMS data to the end of the file rather than overwriting; Olga notes that it could be useful to automate this process with a script
  • tms should copy everything needed from the /tmp directory, but it would be useful to make this process more streamlined
  • Yang plans to improve this data handling in the next version, including retaining the last used reference
  • Currently, the software doesn't know the difference between a good reference and a bad reference (there could be a criteria for determining this in future versions)
  • Particularly on nights when the temperature has changed a lot, the amount of time spent on converging FPIA can be quite substantial; it will be interesting to see how TMS helps to improve this

Confusion of Seeing
  • Confusion of seeing when you measure the pupils in very bad seeing conditions
  • When seeing is very poor at the beginning of the night, there is a threshold of seeing beyond which Z11 and Z22 corrections are not sent
  • Olga notes that there are conditions under which FPIA can sent spurious Z11 and Z22 when in very bad seeing and below the threshold (if the threshold is too high)
  • There can be confusion when you've got radially symmetric terms in seeing
  • Oscillations happen when you are below the threshold and send Z11 and Z22 with Z4 and start to see strangeness at the edges of the pupils
  • On this particular night of interest, FPIA was behaving really badly, with 10000 nm of focus
  • Steel was 7-8 degrees, glass was around 4 degrees, and ambient was around freezing; very bad conditions for control
  • FPIA seems to fail when there is a mix in focus and radially symmetric terms that are far apart
  • In order to break this ambiguity, you would really need to take images inside and outside of focus to break the ambiguity in spherical aberration (and really any radially symmetric aberration terms)

Engineering Nights:
  • Possible engineering nights on the 25th and 26th
  • Andrew proposes making up a sort of science campaign that would run TMS through its paces
  • Normal cadences, reasonably extreme cases, with the goal of trying to break TMS in not too crazy of a way
  • Running actively with TMS, using FPIA only sporadically, and get some idea of what seeing is doing, all with the goal of seeing how TMS fares in a real-world situation
  • To John's mind, the question is how long do you go without running FPIA? 1 hr, 2 hrs?
  • Andrew proposes that we would think about the specific night, and how often you would normally run FPIA, then give it a factor of 3 to demonstrate TMS giving a 3X improvement in performance
  • Andrew thinks it may be time to consider using TMS to set the optics at the start of the night using the last good reference from FPIA
  • This could save a lot of time in operations, and John states that they could do this tonight

Chance for TMS Tonight:
  • Andrew planning to cover the latter parts for Heejoo
  • Dave, the OSA, does not yet know how to run TMS in passive mode (Steve can do this, as passive mode doesn't require any coordination with the observation)
  • We'll run passive mode without (re)setting the reference

Fiber Inspection:
  • Leroy just received the tool to test all of our fibers
  • It looks like we can do most of the inspection visually, but for fibers that are broken on the telescope structure, we will purchase an accessory to figure out exactly where the breaks are
  • Two modes of operation: one which shoots a very strong laser beam, enabling visual inspection of breaks; second which uses a pulse to measure continuity
  • Recall, the TMS fibers have a customized fiber tip angle and coating; Andrew suggests using fiber box on the observatory to make the fiber system more modular for easier replacement upon inevitable breaks
-- TrentonBrendel - 14 Oct 2021
Topic revision: r1 - 15 Oct 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