CDF R5900 MAPMT Experiences

1
CDF R5900 MAPMT Experiences

• Outline
• CDF usage
• Testing
• Performance
• Comments
• Not sure what to cover in this talk threw in a
  bit of everything
• Been 8 years since wrote req. I have
  forgotten much

2
Usage in Calorimeter - Overview

• Mechanical Structure
• Reuse existing Hadron steel
• Replace EM detector

• PES Detector Elements
• Scint. Strips WLS readout

PPR
(PES)
3
Shower Maximum Detector Elements

• A position detector embedded at 6 l deep in
  the EM calorimeter measures the position
• of es and gs
• Sub-millimeter position resolution for energies
  gt 100 Gev
• 6400 channels of 5mm strips
• 416 MAPMTs

Hamamatsu R5900-M16 in CDF mount
Strips
4
Endplug Preshower

• First layer of EM section read out separately
• MAPMTs are the same as used in the SMD(R5900-M16
  Hamamatsu)
• 72 Tubes used
• High Eta - use only 8 pixels/tube
• 1 cm thick tiles
• Same tower structure as EM(960 channels)

5
Calorimeter Frontend Electronics
SMD crate - 21 MAPMTs Read out by each
crate Green is RG-174 cabling From tubes

• Schematic of Calorimeter and
• and Front end electronics readout
• path

Schematic of Calorimeter and front end readout
chain
ADMEM
On CAFE
SMXR
(SMXR)
6
Location PMT boxes

• Common enclosure w/ Cal PMTs
• Temp. controlled
• dark
• Laser - One pixel pulsed per MAPMT of PES and PPR
• Monitors tube gain - but only one pixel
• Assumption is that tube gain will drift for all
  pixels
• Impractical to do more than one pixel/tube in
  mixed system

7
Numbers

• CDF bought 470(PES) 80(PPR) tubes
• Use 416(PES) and 72(PPR) on detector
• 16400 6400, so using 416 means we had some
  wasted channels
• PES cost - 1115/tube- bare
• Bases/mounts built by CDF
• Total cost for that testing 200K
• Custom mount, CNC built to align well with 16
  optical fibers delivering light

8
Testing/performance

• Tubes generally received in good shape from
  Hamamatsu
• Rejected 12 tubes of 470, bulk had lost vacuum
• Bulk of tubes produced in 1996
• my memory is not so great
• Hamamatsu tubes may be different now
• Tested
• Gain/QE
• Cross talk
• Linearity
• Stability
• Dark Current

9
MAPMT Gains - Summary

• 31 specification maxmin gain for pixels in a
  tube
• Average gains show no systematically hot or
  cold pixels

10
Quantum Efficiency - Summary

• Preshower tubes show 3 higher QE
• Produced 2 years after Shower Max
• Specification is QE gt 5
• Good uniformity

11
Cross Talk and Relative Gain

• Cross Talk
• Signal seen in one pixel from light on another
• Each histogram pixel color shows crosstalk
  for that pixel for light incident on one of the
  other 15
• Require lt 5

• Relative Gain
• Plots show pixel gain on MAPMT face
• 3 HV values(2 shown)
• Gain shown relative to highest gain pixel in tube
• Require lt31 variation

Quite sensitive to fiber type delivering light,
if there is any gap
12
MAPMT Linearity

• 505 nm LED used
• Light intensity varied
• Stable, Linear Monitor Tube
• 400 counts 100pC
• Require lt10 dev at 100pC
• Tubes were quite marginal in this regard many
  barely met the spec

13
Background Light Gain Stability

• Gain Stability for high and low DC background
  light
• Mimics high and low eta sections of det.
• Require lt 5 gain shift in 24 hours
• Running experience is that they are quite stable
  in gain over the first two years radioactive
  sources show minimal shift

14
Power Cycling Gain Stability

• Require lt5 gain shift when cycling HV power

15
Gain - Summary

• Correlation - Ham. Gain and UCLA
• Gain spread and Abs. Value within spec.

16
MAPMT Dark Current

• Dark current provides floor for radioactive
  source current measurements
• lt 100 pA dark current makes negligible
  contribution to measurement error

17
Comments

• Good ¼ tubes not the same as larger ones
• Linearity in our tubes was not great
• Lots of connections at rear of tube
• CDF chose to make the connector at the tube
• Cheap connectors ? numerous single channel
  problems after several recablings
• Soldered cables to transition boards
• Might have been better (more costly also) to
  solder at the tube end and put a connector at the
  electronics
• Need to control the fiber/tube face gap carefully
• Reflections can cause cross talk
• Integration is important
• Minos probably has much more current data