TMS Meeting Minutes, November 18, 2021

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

Andrew's Mathematica Analysis:
  • The change in the change of pose should be zero when you drop channels
  • It is when the numbers are perfect, but the pose acceleration diverges when noise is added
  • When one channel is dropped, the pose acceleration is not obnoxious, but there is about a 1 arcsec jump
  • Andrew suggests this kind of jump could be minimized by correcting with an offset vector, as discussed previously (see Andrew's correspondence with Yang)
  • When two or more channels are dropped, the pose acceleration gets progressively worse
  • When three channels are dropped, things get quite bad, where the error gets into the 100's of microns error range
  • So why is this happening?
    • GMT taking into account the thermal expansion of the mirrors, where the change in mirror size will move the collimators
    • Able to back out delta length from temperature separately from other delta L's
    • This effect exists, but it does not appear to be causing this pose change
    • The inhomogeneities in thermal expansion of the mirror are not at the level of the pose errors seen in this analysis
    • Now, Andrew taking into account uncorrelated length error
    • We have an uncorrelated error on the order of a few microns evolving slowly throughout the night
    • Slow decorrelation error in channel lengths on the order of a few microns
  • Conclusions:
    • If we use an offset vector upon dropping channels, we might be golden
    • Andrew thinks that the current collimator mounts have insufficient stability for the necessary stability of the trusses
  • Do we want to do this kind of offset vector correction while a shutter is open? If so, should we consider a damping factor to minize jumps?
Heejoo's Plots Indicating TMS Jumps on 20211115.080923,080948:
  • At 08:09.23 UT, there appears to be a TMS jump (which shouldn't happen with the std. dev. filter)
  • Group X and Group Ry are very far above and below, respectively, where they should be; this could be an optics pointing change in the X direction, perhaps a rigid body pointing change -Andrew
  • Why would TMS suddenly inject an X-pointing change at two discrete times?
  • Olga noted that the trail in the blurred image was about 30 arcsec in length
  • There are some unanswered questions, so we will likely return to this discussion
Heejoo's Plots Indicating Mode3 Pointing Discrepancies on 20211117:
  • TMS does not appear to show any jumps, but the Mode3 pointing corrections have some odd jumps
Things that would be helpful to automate:
  • When to record a reference (such as after running FPIA) could be automated
    • Only when we successfully run FPIA do we want to set a reference, NOT when it fails
    • It is possible to put a call in the FPIA IDL code to run a Python script to take a new reference, but John thinks this is rather ugly
    • Olga notes that simply operating the LBC is a lot for the operator, so it would be nice to automate some things
    • Remember, it doesn't matter whether the reference is taken with the -0.8 mm offset or not, as TMS ignores this
    • Also, Olga notes that it is important for the operator to be aware of the needed pause between setting a reference and beginning to take data
Upcoming Engineering Time:
  • We took pointing data with TMS running
  • Andrew contends that we should create a new pointing model with this new TMS pointing data
  • He suggests that we will mainly see errors in borosilicate mirror tilt which we can account for with a pointing model
  • Having better-controlled pointing and co-pointing with the most recent TMS pointing data would be very useful
  • Andrew thinks a new co-pointing model could provide quite a big improvement (though we should remember that TMS ignores the pointing model when applying corrections)
  • Olga notes that there were fairly large pointing and co-pointing offsets seen last night
  • John believes that science observations appear to work better when TMS is running than when it isn't
No meeting next week as it is Thanksgiving!

-- TrentonBrendel - 18 Nov 2021
Topic revision: r1 - 19 Nov 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