Next: Macros for pedestal/pulser (obsolete) Up: MiniDAQ instructions Previous: Tracking and hit finding

Code for Pedestal/Pulser Studies

The reformat.c routine translates/processes raw data in the MiniDAQ format described by SN0282.

A series of reformat-based .kumac's, fmt_ped, fmt_geom and fmt_track, find pedestals, make geometry maps, and track cosmic rays for these data tables. The .kumac's exist in ~staronl/Staf/wrk/tfc.

Binaries (tfcStaf) for irix and Solaris exist in ~staronl/Staf/bin and can be run via the script ~staronl/Staf/wrk/tfc/Staf .

The running mode of this code is determined by switch settings in the fmtpar switch table. There are 6 distinct running modes, as distinguished by the value of fmtpar.event_type; these are listed below, together with their flags:

Flag   Event Type                      Description
----   ----------                      -----------
-3     Gain data         Fill "gain" table with those calculated online
                         and included in the BEGIN_RUN event

-1     Geometry data     Fill "valid" and "native_map" tables from incoming
                         Structure/Data tables (pending only_type value)

 0     "Normal" data     ped/gain-correct data (optionally) while
                         repackaging it into pixel tables

 1     Pedestal data     Calculate pedestal mean/RMS over many events

 2     Pulser data       Calculate pulser mean/RMS over many events

 3     Gain calculation  Subtract pedestal means from pulser means
                         and calculate the gain via the integrated area
                         of the pulser peak


There are two possible ``data sources.''  One can extract data
from ByteData (8-bit/1-byte) or ShortData (10-bit/2-byte) tables.
The source is selected via the switch fmtpar.data_size:

data_size  Operation
---------  ---------
   -1      Use ByteData; do NOT decompress to 10-bit data
    1      Use ByteData; decompress
    2      Use ShortData

For fmtpar.event_type=0 (normal data), one can also choose to apply
pedestal and/or gain corrections, or neither, via the switch
fmtpar.ped_subtract.  Flagging conventions are:

Flag   Operation
----   ---------
  0    Apply NEITHER pedestals NOR gains
  1    Apply BOTH pedestals AND gains
  2    Apply gains ONLY

Besides these flags, several other entries in the fmtpar table control
aspects of reformat.c's behavior.  Following is the .idl file used to
create fmtpar, with the remaining fields explained:

/***********************************************************************
*	Reformat switch settings
***********************************************************************/
  struct tfc_fmtpar {
     long   nped_samples[6];  /* # pedestal events in mean ped */
     long   npuls_samples[6]; /* # pulser events in mean pulser */
     long   event_type;       /* event type flag: 
                                 -1=geom, 0=data, 1=ped, 2=puls, 3=gain */
     long   ped_subtract;     /* 1 if ped should be subtracted from data */
     long   thresh;           /* pedestal-subtracted adc threshold */
     float  times_rms;        /* quant subtracted=ave+(times_rms)*rms */
     float  ave_gain;         /* gives absolute scale for gain */
     long   first_bucket;     /* first bucket used in gain calculation */
     long   last_bucket;      /* last bucket used  "   "     "         */
     long   only_type;        /* only index type to be processed (-1=all) */
     long   data_size;        /* size of data tables to be used (bytes) */
     long   printf;           /* output flag */
  };

Variable           Description
--------           -----------

nped_samples[6]	   The number of pedestal events used to calculate the
                   mean/RMS values in the output pedestal table.
                   The index corresponds to the time-ordered sequencing
                   of ordered pairs (sector,RDO) that reformat ``sees.''

npuls_samples[6]   The number of pulser events used (etc...)

event_type         controls the operating mode (-1,0,1,2,3 details above)

ped_subtract       selects application of pedestal/gain tables 
                   (0,1,2 details above)

thresh             lowest value written to output pixel tables

times_rms          (pixel ADC) = (raw ADC) - pedestal - RMS*(times_rms)

ave_gain           Average value of integrated pulser signal (averaged
                   over all pads).  This value is also used as the 
                   normalization factor for applying gains calculated 
                   online.

first_bucket       Lowest bucket number to be used in either gain 
                   calculation translation of raw data

last_bucket        Last bucket number (etc...)

only_type          Setting this to any non-negative value will result
                   in processing only rows of the Index table having
                   this data_type

data_size          Size (in bytes) of tables from which to extract raw
                   data (1 for ByteData, 2 for ShortData)

printf             If printf==1, then diagnostic output will be generated
                   during run-time.


Written: 10-JUN-1997 by Nathan Stone (edited by R.Bossingham)


Page maintenance:

Roy Bossingham, LBNL
RRBossingham@lbl.gov