More FED FE Testing Results

1
More FED FE Testing Results

• Mathew Pearson Gareth Rogers

2
Plan for testing FE(status at last FED meeting)

• AIM
• Test the FE firmware.
• Test processed raw data mode.
• Test zero suppression mode.
• This involves
• Loading pedestal registers in FED, and reading
  back.
• Loading cluster threshold registers, and reading
  back.
• Loading common mode median registers and reading
  back.
• Re-ordering and pedestal subtraction (processed
  raw mode)
• Common mode median calculation and subtraction
  (zero suppression mode).
• Common mode median override (zero suppression
  mode).
• Disable strip function (zero suppression mode).
• Cluster finding (zero suppression mode).
• Complement data word (frame finding).
• FED error flags.

3
Plan for testing FE(current status)

• AIM
• Test the FE firmware.
• Test processed raw data mode.
• Test zero suppression mode.
• This involves
• Loading pedestal registers in FED, and reading
  back.
• Loading cluster threshold registers, and reading
  back.
• Loading common mode median registers and reading
  back.
• Re-ordering and pedestal subtraction (processed
  raw mode)
• Common mode median calculation and subtraction
  (zero suppression mode).
• Common mode median override (zero suppression
  mode).
• Disable strip function (zero suppression mode).
  Partly tested. Small problem.
• Cluster finding (zero suppression mode).
• Complement data word (frame finding). Small
  problem.
• FED error flags.

4
Cluster finding

• Clustering algorithm
• All hits above threshold 1 (T1) are accepted,
  except single channel clusters which must be
  above threshold 2 (T2). T2gtT1.
• All clusters must be at least two strips away
  from every other cluster. Any clusters violating
  this must be joined together.
• Example

Using step 1 only
Using both step 1 2
T2
T1
4 clusters
2 clusters
5
Cluster finding

• Testing Procedure
• Run FED in zero suppression mode.
• Set pedestals to zero.
• Enable CM calculation and subtraction.
• Set T1 and T2, where T2gtT1.
• Send various hit patterns into FED from FED
  tester.
• Vary T1 and T2 to see effect on clusters.
• The hit pattern used for these tests

T2
T1
6
T150 T260
7
T150 T2200
8
T1100 T2110
9
T1100 T2200
10
T1100 T2250
11
T1160 T2170
12
T10 T2250 Not expected! T2 has no effect
if T10.
13
T1255 T2255
Setting both T1 and T2 to 255 (max
possible) excludes those strips from cluster
finding. (expected from documentation). In fact,
one only needs to set T2 to 255 to exclude strips
from cluster finding.
14
Cluster finding

• Conclusions
• Cluster finding algorithm working as expected.
• Setting T2255 exclude strips from cluster
  finding (even if T10).
• Observations
• T2 has no effect if T10 (except when T2255).
• Inter-APV cluster finding does not exist in FED.
  But we can lower thresholds at APV edge if we
  want to cater for this (for example, in ORCA
  cluster finding).
• Gareth has implemented some restrictions in the
  Fed9USoftware (in Fed9UVmeDevice) which prevent
  users making undefined choices
• Both T1 and T2 must be non-zero, or,
• Both T1 and T2 must be zero.
• T2gtT1

15
Disable strip function

• This function will allow us to ignore a silicon
  strip.
• Any disabled strips will be excluded from
• cluster finding.
• common mode calculation.

• Initial testing procedure
• Run FED in zero suppression mode.
• Set pedestals0, cluster thresholds0, disable CM
  calculation.
• Send in flat data (in effect one big cluster).
• Disable strips.
• Then
• Set cluster thresholds to non-zero (lowlthigh).
• Send in hit pattern.
• Disable strips and see if they are excluded from
  clusters.
• (sorry, no plots of this).

16
(No Transcript)
17
Disable strips

• Results
• Can disable groups of neighbouring strips.
• The disabled strips are excluded from cluster
  finding.
• Can disable the 1st and 128th strip on an APV
  (with no neighbouring strips).
• Unable to disable any other isolated strip (from
  2nd to 127th).
• Not an urgent problem, but does need to be fixed.
  Disabling isolated strips is something we need to
  be able to do.
• Conclusions and To-Do
• Disable strip function is working only for
  continues groups of strips.
• Not yet established if common mode calculation is
  affected by disabling strips.

18
Complement Data Word

• This command complements the 10-bit data word in
  frame finding mode
• 0-gt1 and 1-gt0
• (has no effect in scope mode).
• This will be needed if we turn off the APV
  inverter.
• Conclusions
• Complement is working, but....
• Problem with channel ordering The channel
  actually set is not the one we intended to set!
• (not sure if problem is software or firmware
  only discovered yesterday).

19
To-Do

• Find out where channel ordering error is for
  complement command.
• See if disable strip excludes strips from common
  mode calculation.
• Will need to re-test this command from scratch
  anyway once the problem with single strips is
  fixed.
• Test the FED error bits.
• Skew clocks on 1 (or a few) channels so APV frame
  arrives too late/early.
• Send in wrong APV pipeline address.
• Send in APV error bits.
• Any more testing ideas very welcome!