TMS Meeting Minutes, March 10, 2022

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

Problems with TMS 3/9:
  • Complete failure of TMS on right side
  • Half the channels are missing (gone to zero signal strength); unprecedented in 5 years of TMS and def. not a laser failure
  • Most likely effect is a misalignment of all channels, possibly by a mirror drop
  • Andrew suggested it could be a swing arm misalignment, but John notes that it would only misalign by 100 um
  • It is possible that the mirror hub was moved significantly
  • Olga notes large, ~2000 nm, coma correction was used last night
  • Did run WRS, commanded 2138 nm Z8 and -321 Z11 on blue side
  • Then FPIA ran, and undid most all of this (removed 1500 nm Z8), leaving a net correction of 700 nm
  • Olga will send dohybrid + dofpia data from start of the night
Optical Model:
  • Olga remarks that the giant focus term comes from what's calculated of the estimated external diameter for an estimated focus
  • If FPIA miscalculates spherical, then it could wrongly attribute this error to focus (Olga notes that this is partially corrected in the cross talk matrix)
  • Olga's by eye calculation of the pupil diameter matches with the FPIA calcuation of pupil diameter (should be 64 pixels)
  • Andrew's non-sequential model, sourced from Trenton, doesn't seem to give the right size pupil images (should be 63 pixels across with 300 nm Z11)
  • Andrew plans to delve into model and see why it doesn't make sense
  • John reminds everyone that, on the blue side, L2 is displaced, so we have to compensate with 1000 nm Z11
  • FPIA seems to struggle most when there are curling edges in the wavefront from higher order spherical
  • Heejoo is a dad!
  • Ratio between inner and outer on FPIA pupils should be 0.22, that's including the extra 1050 nm of spherical
  • Checked the pixel scale, at 13.5 um pixels, and the geometry, seems fine
  • Andrew plans to dig deeper into the model and go back to the sequential model
  • John remarks that there could also be as-built radius variation in the mirror that we are not capturing in the optical model
  • John also reminds Andrew of the thing known as "The Squirrel", a polishing error due to lack of testing at the center of the lens, which causes a second central hole at off-axis fields
  • With "The Squirrel", there could be a particular type of higher order spherical that increases the size of the central hole
  • The change in position of the central hole is field-dependent, as is the size
  • Andrew to check how much spherical aberration would be required to create a hole of the wrong size by this much of this size, and then what is the effect on the PSF
  • Believes he may have gone through this before and determined that it was hopelessly complex and unsolvable
  • Andrew has hope that we could now push through this impass, as we have more knowledge and more people looking at this
  • This is essential for developing a good modeling tool to determine what FPIA is doing with higher order spherical, what is going wrong, and how to proceed with fixing or replacing it with something better
  • Olga agrees that this needs to be addressed
  • TMS won't fix the problem with FPIA not converging
LBT + GMT Workshop:
  • Andrew plans to put together a proposed agenda for discussions
  • Requesting input from TMS
  • Christian notes that it will be possible to be together, but 20 may be too many
  • Christian booking a room
  • John can take 4 others up the mountain, Christian able to take others up
  • Andrew would like to go up, and he thinks 4-5 others could
  • Would also like to have Patricio able to access the telescope to look at the collimator mounts
  • At least a few hours required for playing around the telescope
Plots from Last Night:
  • Red channel had atrocious FWHM/seeing at around 10 UT
  • Red side: At the beginning of the night, ellipticities are mostly up and down, and at the end of the night they are mostly L/R in the telescope coordinate frame (rotations taken care of)
  • Blue still looked elongated, but the IQ was much, much better than ride
  • When temperature changes, structure on both sides should be very similar, but the glass can be radically different
  • BK7 (LBC Red) is more sensitive in expansion (higher CTE) than Fused Silica (LBC Blue)
  • Fused Silica (blue) is more sensitive in dN/dT than BK7 (red), though not enough to make it overall worse
  • Temperature terms in the collimation model should account for this... but there are some issues
  • Andrew states that we MUST understand the thermal properties of the primary mirrors
  • We need to bore down on the problem of thermal mirror shape change with a particular emphasis on mirror focus, then thermal coma, then thermal tip/tilt, and of course thermal spherical aberration
  • What's encouraging is that the Mirror Lab has very nice thermal experiments done in the Lab with just the borosilicate and no other confounding factors, which make really good predicitive models
  • The TMS provides us with a controlled environment for the structure, but we still need to dig into what the mirrors are doing
  • Christian proposes running TMS in passive mode for now, as this is super useful for thermal modeling
  • With TMS and a thermal model of the primary, we can take what the Mirror Lab was doing and apply it to a real telescope
  • The symptoms of mirror thermals, rather than the cause, may be more useful right now, and Andrew plans to take more of a lead on this
  • When stars show up as a ring, this appears to be consistent with thermal focus on the mirror
  • The thermal conditions don't appear to be that bad, though, according to John
  • 0.5 degrees is enough to give 20 microns of focus on the mirror
  • dofpia didn't do a good job on the right side in the middle of the run, because the red performance was terrible
  • Andrew remarks that there are three main tools that make this analysis make much more sense
    • Remove structural positional misalignments with TMS
    • Look at thermal fits of Zernikes front to back
    • And then look at FPIA cumulatives
  • Olga notes that this behavior is not a one-off, where red images degrade to donuts and blue does not
  • Andrew wants to start work on buidling up a good data set to fit Zernikes to the front and back of the mirror
  • Once the Zernikes are fitted, can refer to the Mirror Lab patterns and determine the appropriate corrections
  • Andrew argues that it would be very useful to use the TMS pointing model that was produced
  • With this, the residuals would come from thermal tilts, which we could sense and analyze
  • Andrew to let the group know what results he finds with Pat over the weekend
  • Hoping they find a fairly simple fix

-- TrentonBrendel - 04 Apr 2022
Topic revision: r1 - 05 Apr 2022, 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