1 / 19

GSI Aftermath: a TDAQ account

GSI Aftermath: a TDAQ account. S. Biondi, A. Mengarelli , R. Ridolfi , M. Villa University and INFN Bologna Bologna, 17/05/2019. Outline. Runs GSI integration experience What worked What needs improvements f or DAQ-WD and DAQ-VTX integration DAQ-WD interface DAQ-VTX interface

jsummer
Download Presentation

GSI Aftermath: a TDAQ account

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. GSI Aftermath:a TDAQ account S. Biondi, A. Mengarelli, R. Ridolfi, M. Villa University and INFN Bologna Bologna, 17/05/2019

  2. Outline • Runs • GSI integrationexperience • Whatworked • Whatneedsimprovements for DAQ-WD and DAQ-VTX integration • DAQ-WD interface • DAQ-VTX interface • Offline event building • Nextsteps GSI Aftermath: a TDAQ account

  3. Runs • 14 in total for 504 MB (daq) + 223 MB(vtx)+8726 MB (wd): • 7 April – no target – first DAQ runs with beams– changing detector conditions… • 2197 (BM only), 2199 (BM, VTX), 2200 (BM, VTX), • tests: 2202 (BM, WD, VTX), 2203 (BM, WD), 2204 (BM, WD,), 2205 (BM, WD, fragm. Trigger), • Allgood, no target, VTX thresholdslow, BM HV nominal: 2210, 2211, 2212 (+2209 no beam, 58k) • 8 April – BM HV & TOF scans • BM HV scans, with target & high VTX threholds: 2239, 2240, 2241 • TOF scan: 2242 GSI Aftermath: a TDAQ account

  4. GSI integrationexperience • Theselists show the TDAQ point of view (not the data content) • Whatworkedwithoutproblems: • Beam Monitor (VME boardsonly; not a single DAQ problem; all information stored; allthresholdsstored). But… no HV in DB! • Network: everybodywasable to have a cable connection. • Storage: toofew data taken – goodruns in lessthan 50 GB! • DAQ and detectors in standalonemanner….. GSI Aftermath: a TDAQ account

  5. Whatneedsimprovements • Trigger configuration: fixedat GSI; toomuchworkload on the shifter • WD and VTX startingprocedures: needs to be automatic • Several minor detailsat the TDAQ level: allmeant to reduce shifterworkload or to haveautomatic data, configuration and counterssavings • Online monitoring and warnings GSI Aftermath: a TDAQ account

  6. DAQ-WD interfaceproblems - 1 • BCO start, timestamp start (used for eventsynchronization); WD startsearly; TDAQ sendstwoBCOReset: WD uses the first, trigger moduleuses the second (BTW: toofew bits in the BCO#) • Trigger start; • WD provides first triggersbefore the actualrun start (between the first and the secondBCOReset) • Cured with changes in the V2495 firmware GSI Aftermath: a TDAQ account

  7. DAQ-WD interfaceproblems- 2 • WD eventlosses • Examples from run 2212 – 116 k events for daq: • WD evt 0 notmatchable • WD evt 1-12350 matched with daq 0-12349 • From WD evt 12350-41367 oneeverythree WD eventsisnotwritten out (hwtrig # jumps) • No more data after WD evt 41367, hwtrig 51948 (daqhas 64k more events) • Same pattern in otherruns: ONE unmatchableeventat the beginning, ~10k perfect, then 1 over threelost, stop recordingat some point GSI Aftermath: a TDAQ account

  8. DAQ-WD interfacesuccesses • Except for the begin, all WD triggers are correctlyhandled by the DAQ (no losses), the busylogicworks ok, • BCO timing difference are OK (allmatcheswithin+-1 us) • Partialconclusion: • Once solved for start-of-runproblems and eventlosses (need a SSD), the events can be built online reliably and automatically. • Joint work alreadyplanned (june?) GSI Aftermath: a TDAQ account

  9. DAQ-VTX interfaceproblems - 1 • Integration doneonlyat GSI – notideal… • Busy from vertexmissing; decided for a fixedbusylenght (about 2-3 us) costingus a DAQ rate of 300 Hz • BCO notforwarded to VTX - > problems in event building -> Timestampcandidates: framecounter (185 usperiod) internal clock value (0.5 usperiod) • VTX ismissingtriggers (rundependent) GSI Aftermath: a TDAQ account

  10. DAQ-VTX interfaceproblems - 2 • Basicallyallvariablesthat can be used for event building are NOT reliable: • HW triggersrestartsafter the first 100-200 event (approx 11 s VTX black-out) • Sequentialevents with the same frame countervaluehavebeenobserved • Clearlywrong frame countervalues • Out of sequenceinternal clock values and/or frame counters • Buggy HW trigger counters (rare) • Wrongevent-internal clock valueassociation (constantly off by 1 event) GSI Aftermath: a TDAQ account

  11. DAQ-VTX interfaceproblems - 3 • Observed VTX freezing: • About 11 s long, alwaysafter 100-250 events from the start, consistently on allrunsand implying an hw trigger reset • Can happenalsoduringruns and withouthw trigger reset (e.g.: run 2212, VTX evts 39450-39451 matched with 40975 and 41248, 930 msapart) GSI Aftermath: a TDAQ account

  12. DAQ-VTX interfacesuccesses • 7 aprilruns • Long sequences of eventsthat can be matched (DAQ-VTX), starting from event 0 • Notmany VTX fragmentlosses (<3%). • 8 aprilruns (high vtxthreshold) • Clearlydifferent from 7 aprilruns. • No matchingstarts from evt #0 • LOTS of triggers are lost (busytoo short?) egrun 2240: 20004 daqevts, 10238 vtxevts (50%), 8749 matches from evt 128 to 19954 Almostall VTX triggeredevents are recorded (wrt WD, events are small) GSI Aftermath: a TDAQ account

  13. Offline Event building matches Note: WD – allbut first events are matched(in eachrun) VTX - #vtx/#daq 98% first 3 runs, #vtx/#daq 50% last 3 runs, VTX matchesat 99.5% in run 2212, 85.4% in run 2240, 27% in run 2242 GSI Aftermath: a TDAQ account

  14. RUN 2212 RUN 2212 BCOlikedifference: VTX-DAQ BCO difference: WD-DAQ RUN 2242 RUN 2242 BCO difference: WD-DAQ BCOlikedifference: VTX-DAQ

  15. Conclusions • Severalproblemshavebeenobservedat GSI – and can be observedonly with real detectors (simulations are notenough) and enough time • WD seemsworking in a reproduciblemanner -> a path to fixallproblemshasbeenalreadydrafted • VTX shows differentbehaviour in differentdays/runs. Weneed to study and understandall the (TDAQ) problems • For now: wehave7 runs with all detectors in it (3 looks good). It’stime to play with and to seewhat isinside the data Shoe ! GSI Aftermath: a TDAQ account

  16. The end

  17. DAQ-WD matches in run 2212 WD EV HW BCO BCODIFF1 BCODIFF2 0 1 63376 63377 128912 1 2 23084294906228 4467 2 3 4682 2374 2374 3 4 1546810786 10786 4 5 32067 16599 16599 5 6 41071 9004 9004 6 7 44099 3028 3028 7 8 50682 6583 6583 8 9 59076 8394 8394 9 10 81604294916380 14619 10 11 13739 5579 5579 11 12 18962 5223 5223 DAQ Evt HWT BCO BCODIFF 0 0 33393 33394 1 1 35768 2375 2 2 1422810 1387042 3 3 1439408 16598 4 4 1448413 9005 5 5 1451440 3027 6 6 1458024 6584 7 7 1466418 8394 8 8 1481037 14619 9 9 1486616 5579 10 10 1491839 5223

  18. M. Villa - A DAQ system for FOOT

More Related