Page History

SVT DAQ meeting agenda

Discussion points:

* Dev board: quick tests of new firmware, new dev board status
* Status on getting the Test Run DAQ up and running in group C
* Hybrid production plans/discussion
* Proposal
* JLab DAQ integration status
* Status of APV25 trigger buffering
* Status of Any news on RCE and COB production for HPS
* Hybrid production plans/discussion (schedule )
* AOB

Minutes

...

5/

...

3/2013* Status and plans of the firmware on dev board in group C

Ben has updated the firmware on the dev board to be consistent with what was used in the Test Run. The changes are the FIR filter and threshold blocks of the firmware.
Ben mentioned that chip 4/5 doesn't work while 4 worked under the old firmware. He had verified that the output from the APV25 was incorrect and thus didn't want to change the firmware to compensate for this. It was mentioned that chip 4 had a different gain than the others as well.

We should run on the gimp and also on a different good spare hybrid and compare it to "QA" data to make sure that the new firmware is ok.

There is a new board that has the same data path as the RTM which will be completed and tested in the coming week or two (to check that it was loaded similarly to the RTM). This will become the new development board. When ready, we should run tests with spare hybrid. We'd also want to compare it to the RTM; unfortunately we only tested hybrids that are in the detector with the RTM so we'd have to run one of those with this new board. We have to think what the best solution to do that is (adapter on flange is one possibility).

There is a new config file that should be used: devboard_new.xml. Ryan/Ben would change so that a file supplied on the command line can be set as default file.

* APV25 trigger buffering

Not done yet but Ben has a plan on how to do it. The main issue is that on the dev board there is no space to do this for all channels. It was suggested to do this for fewer than 5 buffered trigger and/or single apv25 channel(s) to circumvent that problem and allow for testing this feature.

The trigger rate also needs to be controlled in order to test this. At the moment the trigger rate is set to an average value. It should be possible to send multiple triggers for every trigger it normally sends to test the burst trigger acceptance?

There was also some discussion on how the trigger handshake between the SVT DAQ and JLab DAQ will be handled in relation to the trigger buffering. Ryan was gonna think about this.

* Status on getting the Test Run DAQ up and running in group C

We have booked wednesday as a tentative day to start setting up the DAQ.

* JLab DAQ integration plans/discussion

Postponed.

* Status of RCE and COB production for HPS

Going well and target is production boards ready in fall. Currently COB has been sent out and RCE will be sent shortly. We should expect boards to start being built in May.

* Collaboration meeting talks

Ben and Ryan will give talks. Might be an overview talk by Pelle if needed; depends on schedule.
There will be a 2-3 hours breakout session for DAQ discussion on Tue, Wed or Thu. Sergey has promised to organize time (Pelle will check).

* Proposal

DAQ section in good shape. Pelle will update the section describing the RCE configuration.

More info: there are 8 RCE's per COB and tentative plan is to handle 4 hybrids per RCE (there is a 10Gbit limit per RCE and a ~4Gbit rate from each hybrid). One of the things that was mentioned was the need for data reduction at high rates due to limited memory to read out full frames. Only time we expect to read out full frames are during calibration where manual low rates will be set.

Thus we need only 2 main boards in a single ATCA crate to handle 36 hybrids including the trigger interface module.

* Budget

Updated budget sent to Marco with specified nr of required/spare/prototype for each item.

- Changes: slight update on connector counts.
- Added PS to DAQ budget; incl. 1w of sho work for cabling

* AOB

Jlab integration
Pelle is working slowly in trying to compile sergey's code on a linux machine at slac. 
When pelle is back we should try to setup the backup solution with a separate processor blade and work on that.

Board production status
Two boards are out for production, COB and RCE, and boards are expected to be produced by end of may. The firmware is progressing well. The RTM will go out for production slightly later, probably end of May. 

Hybrud and FE board schedule
Pelle showed the current schedule for hybrid and FE board which starts out this week. Initial reaction was that the schedule as it is now has good flexibility and Ryan didn't see any problem with any of the milestones. The FE board firmware work has also to a large extent been done already by Ben.

Hybrid production
The hybrid requirements was discussed and Tim showed some sketches on the mechanical design. There was a question on the mechanical stability on the silver epoxy joint planned to seat the CF frame onto the "lip" of the hybrid. Pelle and Tim glued two pieces of FR-4 with the epoxy for a quick mechanical stress test after the meeting. Tim mentioned the joint will have to take 5lbf (?).

Ryan said that the hybrid schematics is only ~1 day of work. Basically removing the unused components and adjust space for a connector. He would make an initial block-out of the board next week to get a better feeling for how it looks. In Tim's sketches the hybrid is aporox. 49.5x45mm. The current hybrid is 69mm long.

The main challenge with the connector is the HV specs. We could solder the connector on the board directly and/or have a separate HV connector on the board. Should be thought about some more. 

There was a discussion about what aspects of the features we've seen could have something to do with the hybrid or what we could improve. 
One of the outstandig issues was the problem operating cold. SLAC will try to get the cold setup from SC and re-run those tests w/ the experts close-by. Ryan's hypothesis is that it was a trigger alignment problem.

Pelle will ask around for day (or perhaps just use the next mtg) to have a dedicated meeting in understanding the edge channel noise problem. To test Ryan's hypothesis of a power distribution problem there will need to be some more work done, like operating it with a separate board for debugging.

Ryan will go ahead and blick out the new hybrid as if no special changes would be needed and if we learn anything new go back and mitigate that in the design.
GIMP old
Image Added Image Added
GIMP new
Image Added Image AddedPelle will setup a TI discussion meeting with William at JLab during collar. meeting.