TMS Meeting Minutes, February 16, 2023

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

Andrew's Mathematica Processing of Thermal Effects on Zernike Modes:
  • Imports thermocouple data and forms triplets sets after std. dev. filtering and fits Zernikes to points on the 36 point array
  • Given X, Y, difference in temp. able to fit Zernike's and obtain a table of coefficients
  • Imports cumzern data file (sparse data), makes decimal hours out of timestamps (creates time axis)
  • Plots fitted coefficients of Turnikits (thermal Zernikes) vs. decimal hours
  • On blue, there is a jmp in focus, which could be attributable to changing filters, active optics stuff on the TEC chips, etc. according to John; however, Andrew hasn't compensated FPIA in this bit, not yet
  • Then, filters out FPIA windows
  • Obtains discrete Z4 points by filtering out the cumzerns; basically looking at the change in Z4 for each new run of FPIA
  • The new cumzern number at the next filtered step shows the accumulative total change in Z4 (have to assume that FPIA takes the focus error to zero)
  • Now, time evolves, after assuming a perfect focus correction, then looks at the accumulated corrections for the next step
  • When you see a 0 nm Z11sum, that usually means that FPIA has breached the seeing limit (1.5 arcsec, I think) and doesn't apply the Z11 correction
  • At the exit of each FPIA run, with it's association cumzern Z4sum, we set the TMS focus point to the FPIA focus
  • Andrew is trying to go from Thernikes to Zernikes using data tables for fitting
  • For each change in Z4 seen, we are considering three things, Thernike 4, Thernike 11, and Thernike 22 (focus, SA, higher order SA)
  • We will have to begin with a linear assumption, assuming the change in Z4 is proportional to a coefficient change in Thernike 4, T4, T11, and T22
  • ACTUALLY, T0 (piston) gives us Z4 (focus) at this point; hence T4 will give us some differences to second order
  • John contends that T0 will give us some contribution to Z11
  • Andrew contends that T4 will give us Z4 and Z11 (referring to the GMT FEA thermal deformation analysis run by the RFCML)
  • Andrew is fitting Thernikes to glass temperature differences
  • T0 is thermal piston (frontplate to backplate dif results in focus, Z4)
  • We expect to not perfectly fit, but to have some correction factors
  • Andrew's thermal model currently has some problems where it undershoots under highly dynamic thermals in the mirror, where higher order terms are likely large contributors
  • Andrew currently building large data tables constructed of Delta T0, T4, T11 and fitting to Delta Z4, Z11, and Z22 (likely going to use a least squares approach and look at residuals)
  • Looking at LBCr on 1/27/2023, there is a large excursion in T0 (piston) and a bit more slope in T4 and T11 terms
  • When the dome is open, the T0 piston term drops by around -0.6 C within 2 hours
  • Olga to make a datetime cumzern vector for Andrew to use with his customized cumzern
  • Olga notes some concern about the fitting of crosstalk terms with the Thernike to Zernike fitting
  • Andrew notes that the focus residuals are quite a bit larger than the others, and of most concern to thermal correction (dominant problem)
  • Andrew also wonders what error budget term we should attribute to glass corrector (LBC optics) error terms
  • Olga notes, that we are just considering the faceplate to backplate, but what about absolute gradients in faceplate to backplate, e.g. C-ring is warm, and side of mirror close to C-ring is warm, might this contribute in some way?
  • Andrew would like to standardize a new format for cumzern in digital hours
  • Olga to update the scripts that save these logs and do some of the data combination
  • Andrew to make his analysis work for an adjustable number of terms to fit
  • Next step is to produce data entries for a big least squares fit

Nights on Sky with TMS:
  • Arizona block coming up on Saturday, and it's possible they will use the LBC's and TMS
  • INAF has bright time, so they don't have LBC's in their plan
  • There is some engineering time, but Olga needs to look at the disposition of that time
  • There is some ECD time coming up on 2/26, after the AZ time ends, and it doesn't appear to be spoken for, so there may be some possibility to use the LBC's
  • Andrew would really like a block of four hours to run zfast-extra on a fairly stable Northern field that doesn't change elevation much and drive termperature into the primaries
  • Last time Andrew did this, he got a whole sine curve in 8 hours, so 4 hours should be enough to get half
  • Better to do this at the beginning of the night before natural thermals become a problem; possibility that natural thermal may prevent collimation at the start of the night
  • We should have things taper towards ambient in the last two hours after driving hard away for the first two hours, leaving the mirror in a stable thermal state when we are finished
  • This should give a pretty good thermal excursion within the acceptable 5 Kelvin limits
  • Will be great to have essentially continuous wavefront measurement with vfast-extra
  • Seeing less than 1.4 arcsec desired
  • Last time we tried this test, we tried to push the mirror thermal excursion to go above the thermal slew rate
  • When Andrew last ran this successfully, he got a nice curve of +/- 2 deg C around ambient

-- TrentonBrendel - 16 Feb 2023
Topic revision: r1 - 16 Feb 2023, 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