Online database

1
Online database
filled for a full set of runs
filled for each run or after 1st step
filled for each step?
filled after validation
2
Offline database
3
Designer

• change online/offline tables in designer
• extend error_id to 3 digits
• add pref, hard constraint in CAL_COLLECTIONS
• change ext_tool inval to hard
• check comment on cal_pul_set tables
  ramp_stepstep_numberstep_size
• put invalid and cut channels into database and
  create MAP
• evaluate limits for mean/sigma/diff to kill
  channels in pedestal runs
• check INCREMENT BY value on sequences
• modify CAL_HARDS table, check design
• put cal-hards into offline
• space defintion in offline ddl
• test fully latest design from online to offline

4
calib procedures
PREF 10k events gain 8 run ?commit gain 1 run
?commit collection extraction ? ref download
files, bad channels
PEDS 500 events gain 8 run ?commit gain 1 run
?commit collection extraction ? download files
from ref bad channels
PULS run 1 gain0 3/5000 ? commit collection extr
action ?bad channels?
GAIN run 1 gain8 80/20 ? commit run 2 gain8
80/200 ? commit run 3 gain 1 80/1600 ?
commit collection extraction ?file with values
TIME run 1 gain1 20000/50/5 ?
commit collection extraction ?file with values
ASCII file index? or date?
5
Hot Cell Killer

• generate list of Channel to be killed
• writes downloads and into database
• unkills all hot cell killed cells
• allow to modify ped/limits for channels
• give list of cells killed by hot cell killer
  with original ped values
• bug in KILL and UNKILL bad channel table not
  correct
• put Runnumbers (date) extraction id, choice of
  run to recover?
• bug in channel retrieve
• message box/balloons
• develop automatic recovery
• error if channel is not a number!

6
HotCellKillerGui

• Ursula
• new version exists for Shifter mode,
  corresponding to the new database
• put instruction in Help menu and add balloons

7
Hot Cell Killer GUI

• allows to modify pedestal and limits for a set
  of channels, crates new download files and does
  correct book-keeping in the database
• add measured and ref values
• optimize queries
• put error labels
• allow for validation of channels
• read file from Examines and kill channels
  directly?

8
trigger files

• Ulla Blumenschein
• pedestals calib_all_x8/1 calib_ref_x8/1
• BLS-mode (x8/1), ADC-mode (unsuppressed)
• PULSER(off)
• pulsers calib_gain8_20/gain8_200/gain1_1600/puls_
  2/time_5
• BLS-mode (x8/1-2ticks), ADC_mode(pedsub),
• TC(single buffer, trigger-bit) ? comics files
  to be changed
• PULSER pickle files for all pulser modes
  produced (to be tested)
• in /online/data/coor/configurations/commissioning
  /cal/
• How to download ICD pulser? in TC crate not
  working
• How to download PIB reset?
• How to put pulsers off after coor download?
• ? runtype DATApib reset pulser shutdown?
• trigger file ramp mode only working prefix
  pulsar-cal-
• trigger file with calib mode to be tested with
  new calib manager
• number of steps for Taka in trigger files?
• calibration types peds/gain/time/puls?

9
l3fCalCalibTools onwards

• Robert Zitoun
• fine for pedestals
• pulsers
• remove 1st event in each ramp step
• ramp_step step_number step_size
• pulser config for Taka?
• Robert/Stephanie/?
• produce standard scripts to determine nlc/gain
  coefficients, delay parameters
• derive correction factors and programs for their
  determination

10
Calibration Manager CAL_Validation

• Taka Yasuda
• pedestals code for pedestals from L3 to
  database ok
• pulser imminent
• Ursula
• access-methods to database ok
• new pedestal validation ok, pulser ok
• gain/time validation to be tested a.s.a.p
• forward status/mess to calibration_manager_gui
• see Wade for implementation in cal-display

to run validation and make download files
this lights will go blue-green or red
status messages
11
Peds Validation

• mean8 and mean1/8 lt 400 gt 700 kill
• mean 8 and mean1/8 /- 60 (3 sigma) warn
• mean8/mean1
• sigma8 gt 100 kill
• check sigma for L/M cards
• sigmas dependence on capacitance, input capa
  (5/10pF) put into DB
• mean8-mref8 sqrt(N)/sigma ? plot without absgt 5
  kill
• sigma-sref/sigma
• plot of error correlations

12
ERROR codes
can be extended if necessary will be kept in
unrelated table
1 'invalid' 2 'cut' 10 'ped' 11 'ped gain8'
12 'ped gain1' 13 'sigma' 14 'sigma gain8' 15
'sigma gain1 16 'mean' 17 mean gain8' 18 mean
gain1' 19 'ped drift 20 ped drift gain8 21
ped drift gain1 22 sigma drift 23 sigma
drift gain8 24 sigma drift gain1
30 'pulser' 31 'evts diff' 32 'reported gain1'
33 'reported gain8 40 'gain' 41 'gain8' 41
'gain1' 50 'delay' 60 'correction' 70 hot
cell' 71 modified cell 90 'else'
add 100 for warnings? add 200 for recovery
invalid/cut channels are now taken from DATABASE!
13
database transfer

• Taka Yasuda
• in principal no problem
• to be tested with development version of
  database
• produce top level tables in principal no
  problem
• Ursula
• put old pedestal values in new offline database
  when in production
• provide scripts to insert derived calibration
  coefficients
• nlc/gain, opt. delay values, correction factors

14
offline database in reco

• Ursula
• final tests transfer of data, server,
  reco_access
• cut production version
• generate new dictionary files for database
  server
• produce server kit
• complete access methods from reconstruction
  program
• Leslie/Robert
• use the access to the database

15
Pulser hardware

• change DAC for postive offsets in all pulsers
• spare PIB to be tested
• take picture of PIB
• PIB EPROMS
• pattern ramps implemented properly
• wait function value to be determined
• pulser shutdown on PIB-reset
• measurements
• delay ramps for optimal delay values
• DAC-offset measurements
• X-talk measurement from pattern ramp

16
Pulser GUI

• Marc Hohlfeld
• pattern ramp implemented in special version
  v2_6b_pulser.py
• to be added
• DAC offset/delay offset enable toggle
• wait value
• RunII pattern 8 patterns only
• TICK in TC ctrl gui
• check pulser readback

17
and then

• try out in real life
• produce some documentation
• implement L1cal (Bob Kehoe)