FA125 Update January 28, 2016

From GlueXWiki
Jump to: navigation, search

Location and Time

Room: ?

Time: 2pm.

Remote Connection

You can connect using BlueJeans using the meeting number 589 693 655 .       (Click "Expand" to the right for more details -->):

(if problems, call phone in conference room: 757-269-6460 might not be the correct number)

  1. To join via Polycom room system go to the IP Address: 199.48.152.152 (bjn.vc) and enter the meeting ID: 589693655.
  2. To join via a Web Browser, go to the page [1] https://bluejeans.com/589693655.
  3. To join via phone, use one of the following numbers and the Conference ID: 589693655
    • US or Canada: +1 408 740 7256 or
    • US or Canada: +1 888 240 2560

More information on connecting to bluejeans is here

Specific instructions for connecting via polycom:

  • Turn polycom on if necessary
  • With the polycom, place a call at 199.48.152.152
  • Press # to enable the polycom keypad, then enter the meeting id: 589693655 and #
  • You may have to unmute the microphone: #*4
  • Turn the computer on if needed

Agenda



Minutes

Present: Naomi, Curtis (CMU), Beni, Cody, Lubomir, Eugene, Luke, Chris, and Simon (JLab). This followed on from the tracking meeting.


Naomi has continued with comparisons of the fadc pulse and raw data using emulation software. FA125_firmware_check The vast majority (1e6-30 per 1e6) of these are in complete agreement. There are a small number of known discrepancies applying to hits before the start and after the end of the hit search window due to different methodology. There are also a small number of unexplained discrepancies where the amplitude is 1 bit different from expected, this could be due to timing problems. Cody and Naomi will correct these, but the current firmware version will be used for the upcoming run.


There are also some rarer but more dramatic problems where data disappear (they are missing from the EVIO file) - CDC pulse data (one or both words) or some of the window raw data samples - or become corrupted (different channel numbers in the pulse/window data pairs). Run 4731 has 3 words missing in 59 data files.


Beni has been looking into this with Cody, particularly at run 4711. They found data missing in events after an event where all channels reported data, and then they discovered that the max buffer size set for DMA transfer in the CODA readout file was not large enough to hold all the fadc data and following that event the data from that particular ROC was corrupted for the rest of the run. This should be fixed in the CODA readout file, and protection could be added to the fadc by inhibiting later triggers until all data has been read out.


Run 4003 was taken with the CDC in long readout mode and the trigger blocklevel=1 and bufferlevel=4. Bufferlevel=4 allows up to 4 triggers to be accepted but the long mode readout is too slow for that, it should be used with blocklevel=1 and bufferlevel=1 (1 event per block and one trigger at a time).


Beni and Cody found that the missing words are from the first channel reporting back from a chip, this could be linked to a time discrepance in the write enable arrival. The missing samples - could be a block level problem or caused by the buffer size when all channels report data.


Changes to be made to the firmware logic (next version) - start the peak search with the sample containing the leading edge time. If that sample is after the end of the hit search window, return integral=0 and amplitude=last sample in hit window.


Useful Info