LHCb RICH HPD Photodetector Test Facilities

1
LHCb RICH HPDPhotodetector Test Facilities

• Infrastructure / DAQ System
• Software
• Test Program
• Outstanding Issues

Stephan Eisenhardt University of Edinburgh
PDTF PRR, Edinburgh, 18.07.2005
2
Aims and Strategy

• Verify specs of 550 HPD tubes and characterise
  them for selection over 18 months
• delivery from DEP to Edinburgh Glasgow
• rate of 30 HPD per month
• max. 8 weeks from delivery to provisional
  acceptance of HPDs
• Photo Detector Test Facilities (PDTF)
• two PDTF sites Universities of Edinburgh
  Glasgow
• two test stations per site (i.e. 4 DAQ chains in
  total)
• one test station/site for mandatory tests for
  provisional acceptance
• second test station/site for extra
  characterisation and redundancy
• test one HPD / work day / site
• automation is essential

3
Infrastructure / DAQ System

• based on CERN HPD DAQ
• similar to system at DEP
• hardware status
• 4 full DAQ chains
• 2 spares for custom electronics
• usable for prototype HPDs
• (2 stations) and
• production HPDs (4 stations)
• 4 setups operational
• 1pAmeter only manual (Glasgow)
• (further SW implementation needed)

• improved automation wrt. CERN system
• no more hand-steered/recorded
• parameters
• Silicon Bias Voltage steering
• Silicon Bias current read back
• LED delay setting wrt. DAQ clock
• Temperature monitors

4
Photo Gallery I
bias pick-up
dark box
HPD
pA meter
flat pointing light source
DAQ PC
bias supply
electronics supplies
MB/FB supply
LED trigger
PP JTAG
fibre
to ZIF board
LED pulse
MB FB
PCI bus
5
Photo Gallery II
HV interlock
light sources
HV supply
HV distribution
HPD
I/O board
independent PMT
HV interlock
HV interlock
sealed divider chain
LED
HV terminal
fibre mount
prototype HPD
6
single LED/fibre light source I

• moved light fibre relative to HPD
• default position homogeneous light distribution
  on HPD
• off-axis positions show enough margin

HPD response to LED light (430nm) 10000 hits
each
display true HPD orientation
centre position (default)
-5cm, to the left of HPD (towards back of box)
5cm, to the right of HPD (towards front of box)
7
Further Infrastructure

• Test jigs for mechanical dimensions
• base existing, diameter jig to be made (1month)
• Movable Pin-hole Mask
• prototype exists
• final design imminent
• mask moved by solenoid, manual switch outside
  box
• 2 months to have 4 systems
• Backpulse Measurement
• hardware available, tested and working ? see Edi
  1 setup in lab
• still needs optimisation for amplification and
  shape of light pulse
• time scale for completion 1 week 2 months for
  first system
• Quantum Efficiency
• prototype of Cherenkov light candle exists
• simulation of prototype exists
• 90Sr sources and Quartz windows to be purchased
• 1 month to have orders and mounts (2 systems)

8
Software Outline

• DAQ based on Labview 7.1 und WinXP
• commercial software Labview, Maestro-32
  (backpulse)
• custom software DAQ (PTS), monitoring, task
  flow,
• Data analysis C/ROOT code based on cygwin
• PDTF software outline
• PDTF main control task flow, logging, summary
  data
• monitors independent, safety features
• automated tasks do the work, use PTS for DAQ

HV control
BiasV control
LED delay control
T sensors
9
Software Layout
PDTF Taskflow
Temp monitor
LED monitor
HV monitor
Bias V monitor
Initialisation
Power ON
10
Online Display Elements I
HPD pixel hit display (true orientation)
scan parameters
sub-task steering
actual scan parameter
Non Zero Hit events
DAQ analysis status
11
Online Display Elements II
raw hits on chip / event
analysis results on scan points
clusters on chip / event est. from Poisson
clusters on chip / event form alg.
results of last point
Ion Feedback est. from clusters gt 3 pixels
Ion Feedback est. from clusters gt 8 pixels
distributions of last point
12
single LED/fibre light source II
delay scan
tail in LED light spoils Backpulse measurement as
it integrates over longer times ? concept
for improvement exists
delay_control1 gives no hits from pixel chip ?
move timing at least by 400ns
Strobe scan
Ion Feedback scan
13
Required Tests

• with HPD preparation

monitored parameters
DAC scan
threshold scans
IV scan
HV settling scans
long LED run
HV settling long dark count run
strobe scan
QE measurement
(10 of HPDs)
Backpulse meas.
14
Test Program I

• JTAG register check
• sets and reads back registers
• stored of failures (should be 0)

• DAC linearity scan
• scans through selected DACs
• comparison by eye with reference
• stored DACs scanned and status

• IV scan (5-90V)
• characteristic I-V curve of anode
• human judgement on shape of curve
• stored scan data, I at 50V 80V

15
Test Program II

• Threshold scans (Alice mode)
• a) scan of main threshold using noise
• to determine optimum setting
• b) scan of testpulse
• to determine noise levels of pixels
• (S-curve)
• stored pre_VTH, scan profiles,
• bad TP map, noise distribution,
• 50 point slope of S-curve
• analysis of S-curve not yet automated

• Threshold scans (LHCb mode)
• a) short scan to compare
• b) scan to compare S-curve
• stored as above match
• not yet automated

16
Test Program III

• HV ramp dark count settling
• ramp in steps while monitoring the HPD
• for dark counts, sparks, LED light
• at 20kV take dark count settling for 30 min
• stored profile of dark count rate settling
• not yet automated

• Strobe scan
• scans LED delay in 4ns steps across 50ns strobe
  pulse
• stored profile of photoelectron response,
• optimum timing setting

integrated as one task

• Ion feedback scan
• scans chip delay in 25ns steps up to -400ns
• stored profile of Ion feedback response,
• maximum rate and position

17
Test Program IV

• Bias Voltage scan (5-90V)
• determine photoelectron response
• stored profile of photoelectron response, 50
  90 points

• High Voltage scan (1-20kV)
• determine photoelectron response
• stored profile of photoelectron response, 50
  90 points

18
Test Program V

• Long LED run (10 hits/pixel)
• photoelectron response
• dead pixel map, position and diameter of PC image
• stored photoelectron response, hist. of cluster
  sizes,
• image position size, demagnification, dead
  pixels
• not yet automated

• Long Dark Count run (5M events)
• photoelectron response
• stored photoelectron response, hist. of cluster
  sizes
• not yet automated

• LED run Dark Count run in LHCb mode
• less statistics, to compare with Alice mode
• stored photoelectron response, hist. of cluster
  sizes,
• match of results
• not yet automated

19
Test Program VI

• Distortion map
• pin-hole array 5cm in front of HPD (every 4th
  pixel)
• to check regularity of pattern
• stored rotation angle, pattern regular
• not yet automated

• Backpulse measurement
• photoelectron response via DAQ / via Backpulse
• fitting of backpulse spectrum
• stored 3x photoelectron response, efficiency
• not yet automated

(10 of HPDs)

• Quantum Efficiency
• integrated measurement with Cherenkov light
  candle
• stored photoelectron response, QE
• not yet automated

20
Outstanding Issues

• Hardware
• FLIC/Slink unreliable, several failures so far
• have only 4 each, no spare ? failure halts extra
  characterisation!!
• LED light source still gives too long pulses
• OK for HPD DAQ, but jeopardises Backpulse
  measurement
• Software
• some tasks not yet integrated (need to operate by
  hand ? much higher human error probability)
• automation of Glasgow pAmeter needs special
  coding in BiasV monitor