Choose timezone
Your profile timezone:
Developer's meeting
→
Europe/Zurich
Present: Phil, Enrico, Roland, Giovanna, Milo, Frank
Readout:
- checking internals using probes (Andrea suspects data corruption arises because of FIFOs filling up)
- today: running with 10 links connected to BNL711 -- good for debugging as it leads to problems quickly
- will switch back to fewer links later, to take noise data
- we are using APA #6: this went in untested, making it more urgent/relevant to measure noise
- installed 712 card (can split between 711 and 712)
- firmware exists for 712
- suspicion that somehow the 709 is not really in a good state [would be good to determine what is lacking]
- checking mostly JBSC settings (Super chunks and Jumbo blocks): this is what we should strive to use
- for now, giving up on multi-DMA (may remain problematic because of too shallow FIFOs)
Offline:
- in BR, we are now appending 8 extra words for each artdaq fragment (issue pointed out by Tom Junk)
- suspicion (Giovanna): difference between single and container fragments
- cannot assign different fragment types to account for this (as artdaq does this behind the scenes)
- so will need to fix software to be able to deal with both kinds of fragments
AOB:
PCIe QAT cards now in np04-srv-018 and machine in Roland's / Enrico's office
- Milo can check together with Thijs
- newer cards than used before: should be more performant; need different driver
There are minutes attached to this event.
Show them.