DØ Calorimeter Data and Code Status

collision.gif (1032 bytes)

Last Update: November 12, 2002 - L. Groer
[Return to Main Task Force Page]

Significant Events or Periods for Calorimeter Data Taking

Major changes   L1 CAL changes
Date Store First Run Last Run Condition/change
September 25, 2002 1795 164963   Timing of experiment moved by two ticks (264 ns). FPGA v30 installed on all Timing and Control Boards to accomodate this.
August 28/29, 2002 1711 163063   08/28: Added a capacitor and changed a resistor on all 1152 BLS boards (except L+M boards in the ECS). This should reduce the "bounce" seen by the L1CAL trigger on a L1 accept.
08/29: Installed FPGA v29 in all 13 T&C cards.
August 15, 2002 1663 161972   Installed FPGA code v28. Only use 41 L1 SCA cells; fix race condition; changed synch between ADC + T&C; fix bug in ADC digitization when gets SCL init.
Greatly reduced out-of-synch problem (mixed events) observed for crates 0x44, 0x47 and 0x4A.
August 14, 2002 09:40 1661 during 161916   Installed FPGA code v28 (in crate 0x47 only) to cure frequent hang ups of this crate. See logbook entry
Aug 1, 2002 161025   Timing adjustment of L1Cal digitization by 56 ns (to sample at peak of pulse)
July 31, 2002 17:34 1613 161020   FPGA v27 fixed and installed. Repairs multiple readouts of 0x4C.
July 23, 2002 1569 160361   L1 CAL reference set #3 fixed, reducing L1 trigger rates
July 5, 2002 13:55       Latest version of PIB-eproms installed: PibRampB6 version 8
- Corrects the update of current pattern number in Pib memory
- Wait function implemented that allows a wait up to 10000 iterations
June 30, 2002 1482 158527   preliminary EM calibration adjustment for L1 CAL (HAD not adjusted because it was fine)
June 30, 2002 1482 158527   Started permanent multi-buffer running.
June 26, 2002 1464 158062   Changed zero suppression threshold online from 2.5 to 1.5
June 21, 2002 1448 157713   Extension of L1 trigger eta to 2.4.
June 18, 2002 1434 157476   First store after shutdown. Changed to FPGA code version 26. Installed all SBC's.
June 3, 2002 1396   155605 June Shutdown.
May 24, 2002   154935   Switched from Run 1 to Run 2 electronics for central Trigger Towers (|eta|<0.8)
May 17, 2002 1332 154569   Turned on readout of TCC in crate 0x4c.

Run Ranges from Bob Kehoe that are good (i.e. do not exhibit event mixing comparing the L1 trigger energy to precision readout). 
>161972 (Aug 15)                  Good
 158527 (Jun 30)-161971 (Aug 15)  Bad
 158041 (Jun 25)-158064 (Jun 26)  Bad
 157573 (Jun 19)-157576 (Jun 19)  Bad
 157478 (Jun 18)                  Bad (single buffered but data looks very wierd)

Special Runs or Data Sets

Date Run Events Conditions Purpose
Nov 12, 2002 11:00-12:00 167956
167957
167958 		29882
2202
31241 		official/special/global_CMT_calnoise-9.20 (Physics) trigger tuned for zero-bias/min-bias/regular triggers of 10/10/10 Hz. Calorimeter readout is unsuppressed. Won't be reconstructed unless specifically requested. Study noise contributions from running with beam and other triggers reading out. Sample for zero-bias/min-bias overlays.
Sep 23, 2002 14:35 164561 50k Zero-bias run, no beam, CAL and SMT reading out, SMT HV on, CAL in pedestal subtracted mode, BLS mode force x8 or unforced(?), all L1 and L2 SCAs used, no other runs going. Study SMT noise contributions. Calibration file.
Sep 19, 2002 ~18:00 164363 30k Zero-bias run, no beam, CAL and SMT reading out, SMT HV on, CAL in pedestal subtracted mode, BLS mode force x1, all L1 and L2 SCAs used, no other runs going. Study SMT noise contributions. Calibration file.
Sep 6, 2002 11:35 163392   Calorimeter special run with online zero suppression at 2.5 Study threshold effect
Aug 25, 2002 11:35 162594   Standard data run but reprocessed with 1.5 and 2.5 offline threshold. Study threshold effect
Aug 23, 2002 19:29 162523   Calorimeter special run with single buffer mode Study event mixing

Reconstruction Code Versions

Status of data reconstruction      Reco status from Algorithms Group

*Note that Dates are for original release of code version, not when they started processing data.
Dates* Version Features
Nov 4, 2002 p13.02.00 Executive Summary: Zero suppression correctly applied (2.5 default) to data and MC using new suppression values; pileup uses individual channel ADC pedestal widths with non-linearity correction for noise simulation; ICD weight increased by 35%, one CC massless gap and two EC inner hadronic layer weights increased by factor of 1.95.
  • Data + MC
    • caltables
    • Layer weights for CCMG5 (ilayer=8, |ieta|=12) and ECMH1 (ilayer=11, |ieta|=13,14) increased by factor of 1.95 to account for 10pF feedback capacitor rather than 5pF
    • ICD (ilayer=9) ADCtoGEV conversion increased from 0.0000694 to 0.0000936
    • Addition of data/cal_zsup_noise.dat version 30-OCT-2002
      Used for zero suppression in d0sim and calunpdata and for noise simulation in pileup. Replaces calunpdata/peds/cal_zsup_limits.dat pileup/calo_noise.rcp
    • calunpdata
    • Correct zero suppression cut (<= rather than <)
    • Modified to use new zero suppression file
  • MC
    • caltables
    • Correct gainpath (x1/x8) swap in MC and add versioning for old MC
    • Reduced MC gainpath switch X1X8_ADC from 4095 to 3100
    • MAX_ADC reduced from 32767 to 28000
    • calunpdata
    • Update versioning for prior MC packing/unpacking for gainpath and previous layer weights
    • Use gainpath as set in pileup package if it exists
    • Use consistent gainpath for NLC and suppression
    • Apply offline zero suppression to MC if turned on
    • Correct small memory leak
    • pileup
    • Modified to use individual channel noise simulation
    • Add persistence for gainpath and set this correctly
  p12.xx.xx
  • Not used for data reconstruction or MC production. Was used for Level 3 triggers and filters.
Oct 12, 2002 p11.13.00
  • No modifications for data reconstruction
  • pileup: replace calo_noise.rcp with new, non-linear corrected symmetrized (in phi) noise parameters per ilayer,ieta for noise simulation in MC.
  • MergeMinBias.rcp defaults changed to take this into account: 
          float Electromagnetic_cut = 1.0;
      float Hadronic_cut = 1.0;
      float rms_scale_factor = 0.65;
      float zs_scale_factor  = 0.56;
    These parameters can be set via mc_runjob
Sep 12, 2002 p11.12.01 No modifications for calorimeter data or reconstruction.
Sep 11, 2002 p11.12.00 Modification of noise simulation in pileup to use single gaussian rather than double gaussian.
Specialized versions s11.00.00-x.x were made to run with different thresholds (x.x = 0.0, 1.5, 2.0, 2.5)
New H-matrix training added to hmreco - will improve electron id efficiency.
Problems with Run 1 channel addressing in the ICR corrected in caladdress.
  p11.11.00
  • Should be good for both MC and data.
  • First release with offline zero suppression (default: 2.5) in calunpdata
    This version has a bug in that the suppression is based on "<" instead of "<=" so effectively is a cut at about 2.0-2.1 sigma for EM layers and 2.4 sigma for hadronic layers.
  • Fixed bug from p11.10 for non-linear corrections in data.
  • Make NLC corrections use "double" instead of "int"
  • Fixed bug in emreco resulting in incorrect HMatrix chisq being returned
  • Protect against negative energies in hmreco
  p11.10 Good for MC, but not for data (MC fix messed up real data)
  • caladdress - Fix phi swap for ICD. Return invalid for Run I physics addresses.
  • caltables - fix non-linear corrections for Monte Carlo processing
  • Fix thumbnail bugs for EM reco
    • em_analyze - No pointer to CaldataChunk is removed. Road info is also available without the SEMparticleChunk
    • em_evt - EM id is saved in thumbnail chunk (EM part)
  • Fix MET block in reco_analyze. The reco_analyze root tuple block called OLDMET is now called METMET.
  p11.09 first release with NLC, good for data but not for MC
  • cal_nlc - Fix bug in component test
  • caltables - Fix ICD ADC to GEV constants. Changes needed for non-linearity corrections in pileup
  • calunpdata - Fix ICD electronics map.
Feb 2002 p10.15 and prior
  • p10.15 - Fix of bug where ilayer = 8 (CC MG) was empty in the MC for all previous p10.xx versions (xx=05-14). Fix memory leak in NADA.
  • p10.14 - Enable NADA in dynamic killing mode
  • p10.11 - New H-matrices tuned with plate MC (no preshower). Enable only officially supported jet algorithms.
  • p10.08 - Fix EM calorimeter weights
  • p10.07 - Apply correction factor for resistor swap in cal_weights
  • p10.04 - Correct FPE in kinem:y for CalData
  • p10.01 - Fixed incorrect handling of CAL weights for events generated with p09.08.00. This caused all calorimeter-related objects (jets, electrons, etc.) to have very incorrect energies. Fixed incorrect calculation of ICD/MG fractions in jetreco. Corrected min Et in jetanalyze.

Timing and Control FPGA Versions

Version Installed Features
30 Sep 25, 2002
  • Move timing to be compatible with 264 ns shift of Trigger Framework timing to accomodate L1 tracks to L1 Muon.
29 Aug 29, 2002
  • Fix ADC busy and reset ADC on ADC crate timeout (which should not happen)
28 Aug 14-15, 2002 The major changes in this version are (from Dean's log entry August 14, 2002 11:43:27 PM,
sequence number: CAL/Log - CAL_Log - 114969
  • We now use only 41 "peak" cell adddress (6-46), instead of the 42 previously used.
  • Fixed a race condition in 4 crates where the timing counter used to measure the length of time charge is held until it is allowed to be readout was not being reset to zero on a new trigger. Thus those crate incorrectly reported an error.
  • Changed slightly they way I sense that the ADC is not in sync with the T&C card. This might help the "0x51" hangs...might not.
  • Fixed a bug which would sometimes cause the ADC to get confused if an SCL init accidently happened while digitization was in progress.
27 July 31, 2002
  • Added timer for L2 hold (in 133 usec units)
  • L2 state bits should be working
  • Crates 0x44, 0x47 and 0x4A have mixed events when run in multibuffer mode.
26 June 2002 during shutdown
  • First version for new T&C boards 3.0
  • Crates 0x44, 0x47 and 0x4A have mixed events when run in multibuffer mode.

Notes

Date Created: Wednesday, August 14, 2002 9:20:49 AM CDT
Date Saved: Wednesday, August 14, 2002 11:30:08 AM CDT
Category - Topic - sequence number: CAPTAIN_Shift/Log - CAPTAIN_Log - 114758
Operator(s): Gene Fisk
Keyword(s): :CAPTAIN:

09:40 We have paused Run 161916 at 299499 events to allow D. Schamberger to down- 
load some new CAL timing readout control FPGA code in crate 47 that will test 
a fix to a "too close timing" problem. Also downloaded software to correct an 
SMT bias voltage problem with x62. 

09:42 Run 161916 resumed. 

10:40 End Run 161916 with events to change pre-scales as we are now at L = 9 E30. 
---------------
Date Created: Wednesday, August 14, 2002 9:38:55 AM CDT
Date Saved: Wednesday, August 14, 2002 9:40:21 AM CDT
Category - Topic - sequence number: CAL/Log - CAL_Log - 114745
Operator(s): Marc Hohlfeld
Keyword(s): :CAL:

Dean changed FPGA code for Crate 47 to version 28. This should cure the 
0x51 and 0x8010 problems in the STATUS field.
[Return to Main Task Force Page]
This page maintained by Leslie Groer
Last modified: Wed Dec 18 11:57:28 CST 2002