Status of HPD Testing

1
Status of HPD Testing

• Status of Test Stations
• Test Results
• Summaries on DEP visit 15.12.2005
• Conclusions

Stephan Eisenhardt University of Edinburgh
RICH meeting, CERN, 18.01.2006
2
Status of Test Stations

• HPD qualification testing cross-checks
  re-tests
• done on 3 test stations Edi1, Edi2, Gla1
• in operation since Nov. 2005
• station Gla2 progress set-back
• hardware commissioned
• software installed
• HPD tested up to few kV (cautious for possible
  light leaks)
• cross-calibration started (comparison with Gla1)
• ? this week disk crash of system disk
• still running with DAQ SW released in 11/05
• upgrade planned for 02/06
• two updates of the analysis SW and the summary
  sheets since 11/05

3
Visual Inspection I

• anodized holes
• few HPD without anodization removed
• taped holes
• dowel holes
• scratches on windows
• 5x 1-2cm scratches found
• (H516020, H518003, H524004 (2x), H539007,
  H540005)
• 3x surface damage found (H518001, H518006,
  H518007)

0.5cm
2cm
H516020
H518007
4
Visual Inspection II

• 2cm gap in Silicon
• H530001, H527007, H541013
• hole at cathode cable
• H539003, H542001, H540005, H540006, H540009,
  H546003
• gap in indium seal
• H530010
• side hole
• H518003

5
DEP Visit Summary I

• visit at DEP on 15.12.2005, attending FM, AP,
  TG, NH
• visual inspection and mechanical precision
• scratches and damages on windows
• ? DEP to check procedures to avoid
• window chamfers
• ? all in Scotland polished, DEP polishes frosted
  edged
• anodized holes
• ? DEP checks procedures to remove anodization,
  LHCb fixes existing anodization
• silicone holes from potting
• ? DEP gives reference to material, LHCb fills
  holes
• hole in indium seal
• ? DEP watches/improves indium homogeneity

6
New Tower of Pisa Issue

• found end of last week (in last shipment of 9
  HPD)
• 2 HPD which failed the mechanical test in the jig
• both lean by approx. 0.4mm towards 3 oclock
• (i.e. in same direction)
• ? not mountable in RICH (jigs cannot be closed)
• DEP is informed we request the HPDs to be
  re-potted

H541004
H542020
7
Quantum Efficiency

• for all results minimum and typical as defined
• in technical specifications of contract

maximum dark count rate 5 kHz/cm2
ltQEgt 26.2
minimum QE 20
typical QE 23.3
PDTF
PDTF dark counts vs DEP red sensitivity
DEP - QE at 270 nm blue sensitivity
minimum QE 20
QE blue vs red sensitivity
DEP - QE at 800 nm red sensitivity
8
DEP Visit Summary II

• quantum efficiency
• QE in the blue (370nm) is excellent
• there are a few HPD with a large red sensitivity
  (2-4_at_800nm)
• there is a negative correlation between red and
  blue sensitivity ( -2 blue for 3 red) and a
  positive between the red sensitivity and the
  dark-count
• ? DEP to try to keep the red sensitivity low
  without compromising the blue sensitivity

9
Anode Leakage Current _at_ 80V
PDTF Bias I nA
Bias current PDTF vs DEP
5 HPDs tested OK 16,15,14,13,12mA 1 HPD
rejected 25mA
Typical 1 uA
PDTF Bias I nA
T difference PDTF chip on DEP chip off
DEP Bias I nA
DEP Bias I nA
PDTF Bias IV scans
normalised
Typical 1 uA
6 HPDs with I 10 mA
low I class
PDTF Bias I ?uA
medium I class
10
High Anode Leakage Current
H530001
H539001
H539012
of bake outs 3
of bake outs 3
of bake outs 1
16.4 uA at (80 216.4) V 47 V
13.7 uA at (80 213.7) V 53 V
25.1 uA at (80 225.1) V 30 V
light level increased by factor 100 !!
provisionally accepted
provisionally accepted
rejected
11
DEP Visit Summary III

• anode leakage current
• measurements at Scotland give larger values
  likely due to temperature (chip off at DEP)
• 6 HPD at Scotland have very high (gt10mA) leakage
  current
• 5 have lt20mA, get fully depleted and operate
  normally
• (bake outs 2x three times, 2x two times, 1x one
  times)
• 1 has 25mA, does not get fully depleted and has
  a very small electron detection efficiency
• ? rejected (three times baked out)
• all (3) HPDs which are three times baked out have
  large or very large leakage current
• DEP stops after two bake outs and puts anode to
  the side
• whenever possible DEP measures leakage currents
  between bake-out cycles

12
Bump Bonding Degradation

• only one example H518005
• 2 bake outs
• O(200) pixels detached lt 95 limit (400 pixels)

Long LED run 200k triggers
Long LED run, LHCb mode 200k triggers
13
DEP Visit Summary IV

• bump bonding
• H518005 shows detached bump bonds, gone while
  bake-out, likely due to non-flat assembly
• two bake-outs ? LHCb responsibility
• ? DEP from now on to test HPD with LED light
  source

14
New Dark Count Status

• Dark Count problem was due to a light leak
• diameter of Lemo cable smaller that Lemo plug
• after fix all (12) HPD with Dark Count above
  5kHz/cm2 were re-measured
• two classes
• homogeneous
• ? high red sensitivity
• central IFB
• ? micro-discharges?

15
Dark Count Rates
PDTF Dark Counts kHz/cm2
Dark Counts PDTF vs DEP
maximum - 5 kHz/cm2
PDTF
DEP

• 15 out of 72 HPDs above limit
• 6 above 10 kHz/cm2
• two patterns
• homogeneous
• centre ion feedback (micro-discharges)

light leak in DEP box
16
Ion Feedback
PDTF Ion feedback strobe scans
PDTF Ion feedback 5M trig
max. 1 _at_200C
max. 1 _at_200C
micro-discharges suspected


DEP Ion feedback
max. 1 _at_200C
note scale (signal x 100) usually very low

17
DEP Visit Summary V

• dark-count and ion feedback rates
• 50 of the DEP dark-count rates above 5kHz/cm2
  spec due to light leaking test box
• ? light tight box delivered by CERN in November,
  now installed and in operation
• 15 out of 72 HPD show dark-count rates gt5kHz/cm2
  (5x gt10kHz/cm2, max. 17.9kHz/cm2)
• ion-feedback rate with LED typically below 0.1
  ? good vacuum
• ion-feedback rate due to dark-count typically 1
  ? points to micro-discharges
• DEP to strive to deliver HPD within specs
• DEP to additionally measure dark-count with LED

18
Status of HPD Test Program

• as of Fri 13.01.2006
• to be retested 4 high leakage current minor
  technical fault (for last two deliveries)
• 3 QE study of high leakage current HPD (on
  provisionally accepted HPDs)
• rejected 1 insufficient depletion (three
  bake-outs) caused insensitive chip
• 1 indium blob on window caused huge noise
• 10 point (50HPD tested) reached on 20.12.2005

19
HPD Test - Time Profile
20
Plans for HPDs

• maximum HPD test rate
• we can test 30 HPDs/month, i.e. 15
  HPDs/month/site
• if we need to test at a higher rate
• planning and additional resources are required
• time required for HPD testing
• mandatory tests one day is sufficient
• 6 hours or less if all goes according to plan
• re-testing will occupy 2nd test station when
  required
• QE and Backpulse measurement only at one site
• HPD delivery to CERN (for system test)
• this week 38 provisionally accepted fully
  tested
• remaining at Scotland (for QE backpulse
  measurement retest LC)
• 10 of HPD 6 provisionally accepted, i.e.
  1/delivery
• 3 provisionally accepted, but to be retested
• (high leakage)
• expect next delivery from DEP early next week
  (26 HPD)

21
DEP Visit Summary VI

• logistics
• 27 HPD delivered on 31.10.2005 provisionally
  accepted on 21.12.2005
• 1 HPD delivered on 31.10.2005 rejected on
  21.12.2005 (depletion problem)
• 14 HPD delivered on 9.11.2005 provisionally
  accepted on 11.1.2006
• batch 4 completed with shipment of 9 HPD arrived
  on 16.12.2005
• expect shipment of 26 HPD in 3rd week of January
• ? DEP to try to provide electronic versions of
  the data sheets

22
Summary Sheets Data Base

• summary sheets
• two updates since 11/05
• off-line suppression of noisy pixels
• analysis plot updates
• re-analysed old data
• ASCII output file for DB under way (1 week)
• database
• waits for ASCII input file
• then test submissions
• first query implementation recreate summary
  sheet from DB
• subsequent implementation of further standard
  queries

23
Conclusions

• despite of the issues reported
• default test result is an excellent performance
  of HPD
• our issues are well addressed by DEP
• iron out some start-up problems