Timing System
Currently, no problems reported with timing endpoints that can't receive clock/timestamp (that haven't been worked around)
List of endpoints at NP04 at https://cernbox.cern.ch/s/OOIoDZuX3FvdiXA
Status for timing endpoints on NP04:
- WIBs using out of date endpoint firmware which is believed to be the cause of problems accessing endpoint registers from the master. The good news though is that the "downlink" works and that "uplink" transmission is possible from most of the WIBs in a crate. The fact that the problem only affects certain WIBs and all traffic goes through the same PTC indicates that this is a WIB/WEIC/PTC issue, not a central timing system issue
- The DAPHNEs timing endpoints work ( @Manuel Arroyave has reported issues with the timing endpoints in the DAPHNEs. However, if they don't reset the boards the link is stable. I suspect this might cause problems in the future but I don't think there is anything we can do at the moment ( suggestions welcome) )
- The control units for lasers P1 and P2 can receive clock and timestamps. Return path has been tested.
- Timing hardware in CRT rack has been updated. Delivering 62.5MHz and 1PPS.
- CTB is receiving clock and timestamps. Issues with return path fixed earlier today ( will need a minor addition to a future patch release on CTB side)
- Fanout-1 ( delivering timing signals to NP04 ) locked to 10MHz clock from beam instrumentation White Rabbit network. Will attempt to install an additional Charon (10MHz/1PPS) endpoint in NP04 barracks and check that the 1PPS from the DTS is aligned with the 1PPS from a White Rabbit endpoint.
We will run the endpoint monitoring service when the system is stable ( to avoid filling Grafana with timing system error messages about dead timing endpoints ).
Trigger
Various fixes and updates related to np04 running:
- Online latency monitoring finalised, PR pending.
- New TC types for CTB. Can run CTB successfully with MLT merging, logic works without issues.
- New monitoring for the trigger: buffer occupancies, zipper occupancies. Implemented, Grafana update pending.
- Bugfix in the TC & TA adapter / buffering that would prevent retrieving all TCs and TAs from the buffer if they have identical start time.
- New ability to run TPG without software triggers.
- Making naive TPG consistent with AVX.
- Changed trigger's thread names to be all unique: will allow CPU pinning if needed.