Thoughts on the 2025 test beam

Mechanics

• 20x20 chambers:
  • Too difficult to add or remove. Some of the on-chamber HV patch panels (found in GDD while changing HV schemes) conflicted with the absorbers and required removing the absorbers. Can it be less tight?
• 50x50 chambers:
  • Mounting them would have been easier with a bottom support plate.
  • The HV cables were easy to break when soldered directly on the PCB and the HV pads were easily damaged, especially because of space conflicts with the HRS connectors. They should be replaced by SHV connectors on the PCBs like for the 20x20. This would have also made it unnecessary to solder the HV cables on the HV patch panel on top of the mechanics. 

Services

• We changed gas connections very often, probably wasting gas in the process. Some T connectors with handles could have been useful to stop the gas flow during these interventions.
• The cables of the VMMs were too tense and the VMMs detached from the detectors very often, sometimes also taking other VMMs away with them. Looser cables would have helped. The mechanics could also be adjusted to accomodate for that. For a similar reason, it was very difficult to make checks on the HV because the PCB and the HV cables were very difficult to reach.
• The VMM cables were not labeled, making it difficult to put them back if removed by mistakes.

DAQ

• In most of the stable conditions, we operated the detectors with quite high thresholds (~3.3 fC) without much control about the ideal threshold by each detector. We should take a threshold scan as soon as the conditions are stable to understand for each detector when we start losing efficiency and the acceptable noise rate.
  • We should also check the grounding to be sure the noise is the lowest possible we can get in this setup.

Operations & data taking

• For most of the runs we used the nominal working points found in the 2024 test beams with APVs, without guarantee they would work well with the VMM, especially with the thresholds applied. The efficiency scans show we were quite at the onset of the efficiency plateau.
• There was no automated system for keeping a log of the runs (ID, datetime etc.). In some cases the run numbers had to be provided manually which was very error prone (in some cases run were overwritten by mistake). The run conditions (detectors in, HV status, mapping etc.) were also not saved automatically.

Data quality

• In some of the runs I could see two or more peaks in the difference between the detector and the trigger arrival time, but we had no time to check better.
• The efficiency was flat in the drift field scan, not sure why.

General organization

• We did not always keep in control the gas bottle pressures.
• Confluence is very good for writing logs of the operations and documentation of the procedures, but terrible for keeping tables that need to be updated periodically (e.g. the list of runs). A much better approach is using Google sheets and import relevant tables to Confluence if needed.
  • Confluence does not have a database