TMS Meeting Minutes, March 18, 2021

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

Future Software Updates:
  • Yang believes the current engineering version of the TMS software is stable and most features are completed
  • Planning to move on to the next phases (two phases planned)
  • Immediate plan: abandon all the files that we are using and switch to a database on the mountain (we don't yet have a databse on the mountain yet)
    • Goal is to have the program record everything to the database on the mountain
    • Yang plans to work with Matthieu to develop this database
    • Increased convenience and integration with the current web DMS system
    • Simply need to add a channel in the DMS system
    • The current web DMS doesn't do much, just generates plots over time
    • Yang is fairly certain the capability should grow in the future
    • Pretty sure that a realtime DMS option is on the horizon (as of now, only able to look at historical data ranges)
    • Would enable us to watch the TMS corrections and the truss lengths in real time
    • Next move is to push the software output into a database
    • John has some reservations about this, as LBT has had a very poor experience with developing new databases
      • Need someone to maintain the database, and need some tools to interface with the database
    • As a reminder, most of the files do not come directly from the database but are reimported from HDF5 files from the TCS system
    • In preliminary discussions with the software group, no one objected to this idea
    • Matthieu concurs that there shouldn't be too much of an issue with setting up TMS data in a database
      • Also able to set up replication sets and other useful tools
    • In Yang's idea, this shouldn't cause any significant problems, and it's much more efficient to access data in a database
    • With time, Yang expects the database to grow with capability to meet everyone's needs
    • Also, we can develop a tool to extract data individually, as needed
    • The underlying infrastructure setup and management is primarily an IT issue
  • Next plan: need to have a unified application that is easier to operate than the current system of scattered scripts with various configurations which are kind of confusing
    • Yang intends to build a streamlined GUI (not like TCS)
    • 95% of TO's who will be operating this need a few big buttons
      • START (PAUSE): could automatically take a reference for you
      • STOP: erases all history
      • PAUSE: does not erase any history (e.g. std. dev. history, etc.)
      • RESUME: continue operating with the same reference generated at START
    • In John's opinion, you'll need more buttons to manage all of the operations of TMS (start, stop, take reference, launch TMS software, start TCP/IP, etc.)
    • Etalon has told us that it is not possible to remotely turn on the TMS software
    • Some disagreement on what we want the TMS GUI to do
    • From John's perspective:
      • At start, use TMS to create a reference, and continue operating until a pause is needed
      • On restart, need to be to move to
      • John argues that it is very important to have the ability to take a reference on command
      • The user must decide when to take a reference
      • Might use the reference from the previous night to pre-position the optics
      • Advocates for buttons with the all necessary features to run TMS well
    • In Yang's opinion, most of the time, the TO/OSA need not think about when it's necessary to take a reference
      • Yang's goal is to abstract the inner workings of the TMS for the TO's
      • Simplify the process to run and stop the TMS system using START, PAUSE, RESUME, STOP
      • Yang does intend to have a configuration panel that is mostly hidden to the top-level operator
        • Planning to enable functionality to change channels, drop channels, rename channels, take references, change the filter parameters, etc.
      • Yang believes that the main operating environment should be fairly simple for the TO to use
    • John recalls all of the ways TMS is used:
      • First case: use TMS to pre-position the optics at the start of the night using a previous reference, run for only a couple of iterations, then stop/pause
        • Then go off, find your star, run FPIA, obtain good collimation, and take a TMS reference
        • Then, once you're ready to observe, run TMS with the reference set earlier
        • There are reasons why you may need to pause TMS operation, then resume with the same reference at a later time
        • Finally, ability to stop TMS at the end of the observing run
          • Yang concedes that a CALIBRATE button could be used to take a new reference as needed
          • Consider using words like REFERENCE
    • Yang not planning to do this immediately, as this will likely take some time
    • This is the step after creating a database
    • John also notes that the software should be callable from a script
      • Have the ability to talk to the GUI from the command line, quivalent to GCS client
    • John asked about the possiblity of displaying TMS results while running, especially in the case of an error cropping up
      • Most errors arise from dropping a channel or TCP/IP failures giving garbage measurements
        • Yang notes that channels which are dropped will be disincluded automatically, with no operator input necessary
      • Problem with the scripts we are using now is that they have a firehose of numbers flowing by, and it can be challenging to see the errors
      • Yang notes that you should be able to filter out only what you want (John notes the LSS filter operation as an example)
      • Yang inclined to display the most relevant information on the screen
      • Each message will be tagged so that you can choose which to filter (works like the event logs)
      • In the case of a major error, Yang suggests that TO's just call support for assistance with restarting the TMS
    • Best case, able to just keep TMS running constantly, though this hasn't proven to be a stable operating mode
    • John notes that the GUI shouldn't be so simple that it isn't usable


  • There was one night when TMS was used passively this past week
  • Engineering time was last because of external issues
  • There may be a chance to run TMS on the 29th

-- TrentonBrendel - 18 Mar 2021
Topic revision: r1 - 18 Mar 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