.png)
Test beam agenda, materials, minutes here: https://indico.cern.ch/event/1682537/
Detectors under test: 4 50x50 detectors (2 µRWELL + 2 MicroMegas), 2 RHUM MIcroMegas (1 small pad + 1 small pad with capacitive sharing readout)
Extra detectors: tracker (3 10x10 triple GEMs, 250 µm pitch, 3x3 cm2 read out area)
Started almost from scratch (detectors not previously tested, first experience with VMM setup built from zero, new mechanics, new firmware with triggered mode to be tested). We spent most of the time debugging and took data of good quality after much effort tuning the details. We took good data with the one HCAL chamber that was working well and with the RHUM chambers.
We prepared the setup during 4-5 days before the TB in the RHUM mechanics. We assembled the tracker in a separate mechanics, mounted on a vertically movable table.
We installed the setup in a vertical automatic platform in the experimental area. The setup turned out to be too tall for the beam, so we had to dismount the top part (containing the detector) and place it on a separate manual table next to be bottom part (containing the crates).
We read out the setup with two FECs (one for the tracker + trigger signal, 7 VMMs in total, one for the DUTs. At most we read one RHUM chamber with 4 VMMs + two HCAL chambers with 2 VMMs at a time). We used 5 and 12 V fans for the tracker and 24 V fans for the DUTs (capped to 12 V because the LV power supply kept getting in error). For each FEC we had one powerbox.
During the last days of the test beam we had to use two separate SRS crates, each with one FEC, to prevent continuous disconnections.
We left the DAQ PC in the experimental area and connected to it remotely from the control room because none of the ethernet cables available in the area were long enough to reach the setup or were not transmitting data properly.
We used the FEC triggered (gated) firmware, apparently for the first time with two FECs. We realized after a while that in this mode we needed to send the trigger signal to the CTF and not to the two FECs separately (same as in the APV setups basically), otherwise the two FECs have non-synchronous data). We also did not notice from the beginning that vmm-sdat requires a different data format option (-df TRG) to reconstruct the data taken with triggered mode.
We took runs using a wrapper around tcpdump that allowed automatically transfering the runs to eos and connected to the HV mainframes for HV scans. This made it easier to run long scans (e.g. HV and drift scans at the same time) but we found out often that the VMMs from one FEC would lose communication after a few runs. In the future a monitoring of the connectivity, as well as the temperatures, and being able to change parameters programmatically (e.g. the thresholds), would be helpful.
Grounding took much of our energies. The noise of the HCAL chambers "as-is" was of the order of a few MHz, while for comparison for the RHUM chambers it was enough to connect copper braids to the corners to run with ~ 3fC thresholds. After several trials and errors in the experimental area we moved the 50x50 chambers to the GDD lab with one FEC and did the grounding there. Different grounding schemes seemed to have an impact, not always being reproducible. The most common grounding scheme seemed to involve:
After some effort we got to a configuration with about ~10 Hz per VMM. However some tuning after mounting the chambers in the setups during the accesses was still needed. The fact that we had an unstable beam with a lot of dedicated MDs was helpful in debugging.
At the end we took most of the data with ~ 3fC thresholds. We also realized quite late that we were not applying the threshold calibrations by-channel, which instead is helpful to reduce the thresholds further. In triggered mode we also ran a threshold scan (yet to be analyzed).
In super short: