Improvement of the Beetle Readout for the MaPMTs

1
Improvement of theBeetle Readout for the MaPMTs
J. Bibby, S. Eisenhardt, F. Muheim, N. Smale

• Achievements since last LHCb week
• Common Mode correction
• Light scan
• High Voltage scan
• Conclusion

Stephan Eisenhardt University of Edinburgh
RICH MaPMT meeting, 17.06.2003
2
Achievements since last LHCb week

• New Beetle chip
• visible noise factor 2-3 lower compared to the
  old chip which had died
• (another sign that the old chip was dodgy)
• Tuning of existing setup
• noise reduction of factor 2
• Noise reduction session (2 1/2 days together
  with John and Nigel)
• rerouting of power supplies !!
• shielding of LED pulser !
• new coax cables between MaPMT and HD board !!!!
• new GND routing layout for unused channels at PMT
  base !!!!
• new star point for GND at front-end !!!!
• massive (1/2 inch) Al GND plate below HD mother
  board !!!
• all-in-all a noise reduction of factor 5
• Common mode reduction
• problem 12 connected channels move more than the
  others
• solution split into different groups and treat
  independently
• noise reduction of factor 3
• problem individual channels have 10 weight
• ? artificial dip at cut position in corrected
  spectrum

3
Data Frame from Ntuple

• FED takes
• 160 samples / event
• header samples
• show double peak
• also visible on oscilloscope
• structure comes from the front-end of the system
• 12 channels bonded
• 1 analog TestOut
• 1 outside the sampled frame due to mismatch in
  timing
• 10 channels visible
• bent base line
• Beetle1.2 feature
• signal inverted for better match to FED
• gives positive charge signals

w/o CM correction
128 ? 1 data frame
header bits ?200mV
bonded channel
4
Signal w/o CM correction

• Uncorrected pedestal width ? 4.5 ADC
• Double peak structure variable with runs
• (two peaks, shoulder or broader flank)

pedestal run
signal run
? 2
5
Signal with CM correction

• Uncorrected pedestal width ? 4.5 ADC
• Corrected pedestal width ? 1.4 ADC
• artificial dip in spectrum from CM correction on
  few channels

w/o CM correction
with CM correction
? 2
6
LED Light scan

• Varied amount of light to determine a working
  point with ? 1

7
HV scan I - corner pixel

• HV -750-825V
• Uncorrected pedestal width ? 4.5 ADC
• Corrected pedestal width
• ? 1.4 ADC
• clear single photon signal

8
HV scan II - corner pixel

• HV -850-925V
• Uncorrected pedestal width ? 4.5 ADC
• Corrected pedestal width
• ? 1.4 ADC
• clear single photon signal

9
HV scan - center pixel

• HV -750-900V
• this pixel has a significantly higher gain

10
Things to do

• calibration of signal
• proper S/N calculation
• straighten out shifts in timing
• check LED pulse width with analog TestOut channel
• get proper BoardBeetle
• the whole issue about digital readout is not yet
  touched...

11
Conclusions

• Great progress since last LHCb week
• level of noise using the HD board is down by 2
  orders of magnitude
• setup much better understood
• this knowledge is put into the design of the
  BoardBeetle
• analog readout of 8-stage MaPMT with Beetle1.2 is
  possible
• proper S/N figure still to be determined
• Binary readout option (no CM correction) still to
  be evaluated with BoardBeetle
• Readout of 12-stage MaPMT with Beetle1.2MA0
  (Nigel) looks very promising as well ? report at
  MaPMT workshop
• a great Thank you! to John and Nigel!