TMS Meeting Minutes, August 19, 2021

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

Update on Fiber Switch Conversation with Mark (eTalon):
  • Mark provided recommendations on the fiber switch
  • Mark updated us on the software testing
    • Mark was not able to replicate the problems we were having on their system
  • Heejoo is planning to buy a high-quality, shielded USB cable for the MultiLine unit and will replace the cable at the next visit
  • Andrew remarks that it could be the USB cable, or it could be something internal to the MultiLine unit
  • If it turns out we have a system that randomly freezes and requires in-person reset, we could consider using the system without the shutter and just switch off the laser after each session
  • This may be a satisfactory solution for now until the multiplexer is fixed
  • Once we are done with commissioning TMS, we may start the discussion on purchasing an extended mutliplexer to increase our channel capacity
Updates to TCP/IP
  • New version handles TCP/IP requests more robustly
  • Mark asks to have time to install the update and test with the laser on
Visiting LBTO at the end of August:
  • Trenton and Heejoo to visit and reinstall the TMS collimators and conduct the laser tracker survey
  • Leroy to assist with routing fibers to the mirror cells
  • Andrew would like to request engineering time for daytime and nighttime checkouts
  • Sept. 1 is the first day of startup operations
System Checkouts/Demonstration and New Pointing Model
  • Andrew is confident we have enough working to be able to successfully demonstrate the operation of the system in a dynamic environment without having to run FPIA again and again
  • Able to do daytime testing over the full range of elevation to see which channels are flaky
  • We don't have to fully correct thermal effects on the fly, but we must be able to demonstrate that we can correct for thermal effects offline and eventually in real time
  • It would be very useful to use the TMS to develop a new dataset for the pointing model
  • Taking pointing data is exactly the sort of TMS testing that Andrew would like to conduct
  • Consider a scheme like trying to take defocused FPIA images while conducting pointing analysis
  • Usually, when preparing a pointing model, a WCS is fit to the images in post, so the FPIA defocus may be outside the capability of the system
  • 28 star images are required to make a new pointing model
  • Perhaps getting out of focus images at every second or third point would be enough
  • LBC pointing models just require the centroid, they don't require running FPIA at each point
  • With TMS, the motion and flickering of fluctuations from each point could be removed
  • If we were to collimate and obtain wavefront info at every fourth star, we could develop a very effective model
  • Co-point at every star (or not, according to John), collimate at every fourth star should work quite well for the pointing model
Mirror Thermal
  • Andrew claims it would be quite prudent to take mirror thermals into account in the pointing model
  • Until we have mirror thermal coma and the associated tilt, we may not have a full picture of the pointing model
  • This may be easiest to achieve in the second half of the night, which may be better than the first half of the night to minimize thermal effects
Moving Trip Up
  • Heejoo and Trenton will go up the mountain from 8/25-8/27
Andrew's Thermal Analysis
  • Faceplate and backplate data from 06/07/2021
  • Computed avg. dif. between faceplate and backplate
  • Avg. difference on Sx of about 2000 nm of focus
  • Quite different results on Dx, with very different magnitude
  • Steps to go through for thermal analysis to make sense:
    • UT Time aligned with TMS data collection time (patchy data)
    • Temperature data placed on time access next to TMS data
    • Compute linear fit for thermal Z4 and TMS Z4 to compare (TMS Z4 * TMS delta Z)
    • Zernikes from FPIA outputs (in passive TMS mode)
    • Upfront loading of focus, -875 nm defocuson Dx, do some correctinos, then backout at end, effectively doing what FPIA does
    • Able to see periods of data in between runs of FPIA
    • Looking specifically at changes between periods of FPIA runs
    • Extrapolating to end of TMS run, may compute change of TMS Z4 made up of a thermal plus a Z change
    • Thermal Z4 is very linear, so able extrapolate backwards to fit
    • Predicted change in FPIA may then be obtained
  • Andrew's model tells us that, at any point in time, we can look at the change in delta T for the faceplate and backplate to predict the change in thermal focus
  • Breann suggests producing another model to test the current model, and to consider using machine learning to use much more of the available data
  • Andrew notes that the residuals left in the model are consistent with the residuals observed on the actual TMS system
  • Andrew requests having the entirety of the faceplate and backplate thermocouple data for all passive nights, provided by Trenton
  • Possibility of hiring a New Zealand student to work on a machine learning project for TMS
  • If we get a good thermal model for focus, coma, and spherical, we can go back over years and years worth of data to fit structural thermal data and learn the gradients
  • If you can back out the mirror contributions reliably, you can fit residuals to thermal structure deformation with high certainty
  • The TMS enables us to separate mirror contributions from structural contributions
  • Recalling the mirror lab doc on GMT primary mirror thermal modes, we could potentially fit these same thermal modes onto the LBT primary mirror data (Trenton working on this)
  • Trenton constructing a solid model of the LBT primary mirrors

-- TrentonBrendel - 02 Sep 2021
Topic revision: r2 - 16 Sep 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