TMS Test, August 13, 2020

Attendees: Andrew, Heejoo, John, Trenton, Yang (Zoom session)

TMS Test Summary:

We did not have a normal meeting this time. Instead we did a TMS comprehensive test from 4:00pm until 10:00pm local time. We monitored our test from 4:00pm until about 6:00pm, and then we ran the test in a loop collecting data on its own until about 10:00pm local time. This was our first TMS comprehensive test in recent time. We mainly tested the following items:
  • We moved our software into a 64-bit Fedora Linux machine with a pure Python 3 environment. This was the first time we exercised this environment. It is an important step to move the software into a modern environment that is supported and being properly developed.
  • We reorganized our software quite significantly (other than the Python 3 upgrade). We reused part of our older TMS software, but we also put in newer components dealing with newer TMS measurements and pose calculations. So this is the first time we ever tested this configuration.
  • We have now switched to take direct measurement from the eTalon Multiline TCP server for raw laser length measurements and now perform our own vacuum length correction as well as the pose calculation. Verifying the correctness and functionality of this piece of software was a main part of the test.
  • Other misc. items include integration with current TCS (telescope control), PSF and so on.

TMS Test Results:

The test was successful overall:

  • The 64-bit/Python 3 environment setup was working fine so far (after several tweaks during the test). So this is an encouraging result. We will be developing our software fully in this environment from now on.
  • The new software organization and structure worked well. However the current organization is just a temporary solution meant to test certain things quickly. In the long run, we will be doing a redesign on most part of the software.
  • The new TMS measurement on raw data as well as the basics of pose calculation worked as expected (after some slight tweaking during test).
  • The interaction with TCS/PSF etc. worked well too.

We did encounter a few issues:

  • The change of pose calculation when we have missing channel information did not work as expected. Initially we thought there were bugs in the program. However further investigation showed we have some misunderstanding on some of the properties of this problem. We have had a somewhat in-depth discussion post our test on this topic and will have a new design that works correctly as well as efficiently.
  • During the middle of the unmonitored test phase (approximately around 8:30pm to 9:15pm local time), we encountered an issue where the TMS corrections drove the LBT primary mirror to its collimation limits and we had to manually recover and reset the primary mirror from that state, and then resume the data collection loop. We had some clues why this happened. But we are still investigating this problem at this time.

Next Steps:

  • We will need to do some analysis and mining on the data we collected during the test, including extraction of all relevant LBT telemetry of interests from that time period.
  • We will need to discuss and prepare the second TMS test (time and scope).
Topic revision: r1 - 20 Aug 2020, YangZhang
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