TMS Meeting Minutes, May 6, 2022

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

Issues w/ MultiLine PC TeamViewer:
  • There may be some issues with multiple accounts connecting to TeamViewer at the same time
  • Andrew suggests putting a "known issues" section in the wiki to describe the possible failure modes and operations involving TeamViewer
Channel Dropping Test:
  • 7 operational channels on both sides
  • Dropped channels on both sides and didn't find any noticeable jumps in the pose correction vector upon drops
  • Andrew proposes that we look at the duration of the goodness of the solution
    • Drop a channel, wait a half hour, then try to re-add the dropped channel and see how far off the correction is
    • As of now, Yang's code stores dropped channel configs and uses that until a new config is obtained with more/less channels
    • When Yang's code sees that a channel is dropped, it remembers the lengths at that instant to obtain a last good pose
  • During last Friday's test, we did run this sort of an analysis while only running for a few minutes with the dropped config before returning to the standard set
  • Last night, 5/04, there were cases where the Sx side dropped channels and the pose changed smoothly and then it picked channels up again gracefully
  • Request to include information on when and which channels are dropped on John's magnificent plots
  • Andrew remarks that it is strange for the Dx side to show more issues than Sx, but Dx dropped more channels than Sx
2022.05.05 TMS Run:
  • Approximately every third point, the corrections on Dx are cycling upwards as the channels drop and things are out of whack; not observed on Sx
  • John and Andrew agree that we should turn off channel dropping in the meantime as we sort out this bug in the algorithm
  • We need to go on sky and do some more engineering to sort this out
  • The std. dev. filter is causing the drops here, and the channel dropping algorithm is not causing this effect
  • Around 6:05 UT, there were some dropped channels on Dx that resulted in some wonky corrections being sent
  • 06:03:39 UT, the channel dropping caused a jump in the correction sent
  • When channel dropping occurs, the pose that should be sent to John's script should be the PREVIOUS pose from the last measurement, NOT the pose from the reference that may have been taken an hour ago before filter offsets were applied
  • To send a zero correction at the point of a drop, you must send a no change in incremental pose, meaning you send the last measured pose
  • So, we can't use the original "reference" but rather the most recent pose
Active Night, 2022.05.03:
  • Red was performing very poorly
  • In our TMS plot, we can see that TMS was not driving this poor image quality
  • From 04:00 to 04:40, the LBC TEC chip was sending large focus corrections which TMS was cancelling out
  • We are unsure as to whether the active focus corrections being applied by the TEC were good or bad
  • Olga notes that there was an issue collimating at the start of the night and there was lots of extra negative Z22, which may have caused some issues
  • John remarks that whether or not you have crappy images, TMS should just hold that over time; also, John states that the mirror should have a half-hour time constant over this sort of temperature range, so he thinks that the thermal contributions are not a likely culprit and that you actually couldn't even apply this much thermal defocus
Active Night, 2022.05.04:
  • Collimation at 10 UT, with a reference taken, the TMS correction is very continuous over this period
  • If the reference had been skipped, it's possible TMS could have provided a nice, continuous correction over the focus period
  • The net Z4 over the FPIA run was 630 nm, which TMS may have been able to do instead of using FPIA
  • No filter changes, so that can't be the cause
  • There was a jump in the correction after changing to a new target on the left side, with a smaller jump on the right side
For Tonight:
  • Channel dropping has been disabled by Yang
  • Also, Ch. 15 needed to have it's reference length adjusted; it has been fixed

-- TrentonBrendel - 05 May 2022
Topic revision: r1 - 06 May 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