Electron Detector DAQ
- Nomenclature Conventions
- Pin Plan of v1495
- Library of functions
- Signal Flow Layout
- Signal Flow Timing
- Output data structure in CODA
- Generic info: v1495
The signals received from 96 strips of each detector plane are processed to form a single trigger in each trigger window. Signals from strips 1-4 are OR-ed to form the first sub-cluster, strips 5-8 are OR-ed again to form the 2nd sub-cluster and so on. The first and second sub-clusters are OR-ed to form the signal called plane-1's cluster-1 output. The second and third sub-clusters are OR-ed again to form the signal called plane-1's cluster-2 output. Following this pattern, there is a total of 7 signal clusters formed from each plane. The output from corresponding signal clusters of all 4 planes are checked for the trigger condition in the coincidence time window set by the pulse-width of Delay-line(parameter pwdl).
For instance, if the parameters
event-trigger = 2 accum-trigger = 3 PWDL = 100 ns
then the master board will generate the event trigger output only if at least 2 of the 4 clusters(one coming from each plane) are found high during the rising edge of the inbuilt 25ns clock. The occurrence of above condition will be checked 4 times at intervals of 25 ns because the coincidence time(PWDL) is set to 100 ns (4x25)
The flexibility of 8 channels (every signal cluster has 8 channels) was allowed because we were not sure if we would be able to install our detectors at a precision of more than a millimeter. The pitch of the detector is 200 um, hence a 8 strip flexibility will allow an overall position offset of 1.6mm (=8x200um) across the 4 planes.
v1495 Firmware Revision History
All the documented firmware revisions that are updated here, can be found in the SVN mentioning their revision numbers.
- Revision # 2005: fully working and checked with Readout.
- Revision # 2006: [updated on wiki on Jan 29 '10] added an added feature of minimum width rejection in the Reconditioning module of the firmware
- Revision # 3210: After moving to the 32 bit version for slave boards, the firmware # starts from 3201 onwards. As of Oct 20, we are using Revision number 3210 on the slave boards. Notice that Revision version 320F (the just previous version) had a bug.
- Revision # 3212: The firmware was modified to be able to read out scalars. This revision expects an external signal at an exact 1Hz rate at F(3) and the firmware counts the number of rising edges that came out of the reconditioning circuit for each strip (irrespective of the trigger condition fulfillment)
- Revision # 3213: In this revision, the trigger cluster size was made to be 4 strip rather than the earlier 8 strip cluster.
- Revision # 3214: [this version was loaded in December, this entry made on Jan 25, 2011] It was noticed that the accumulation registers are holding in memory the old counts(from previous run). In this revision we hence need to device a way to flush the registers every time a coda 'GO' is executed. The previous REG_RESET didn't seem to be doing that correctly. This new version has been tested for consistency in both event and accumulation mode with the previous firmware using 8 strips trigger cluster algorithm.
- Revision # 3216: In testing mode as of Mar 30, 2011. This has the 2 strip signal at Compton edge needed to calibrate the photon detector. This signal only uses plane 3 and plane 4 for forming this signal.
- AA07 : unlike v.AA06, the buf_busy signal does not affect the generation of trigger signals from the master board.(needs to be investigated further)