All DØ Meeting, 12/01/00 Central Fiber Tracker Light Guide Performance Thomas Nunnemann Fermilab 90 days before RunII … (a.k.a. 12/01/00)

All DØ Meeting, 12/01/00

Central Fiber Tracker

Light Guide Performance

Thomas Nunnemann

Fermilab

90 days before RunII …

(a.k.a. 12/01/00)


Optical Readout of Scintillating Fibers

Scintillating FiberOptical Connector

Waveguide Fiber

Mirror

Photodetector CassetteElectrical Signal Out

Cryostat


discrimination of signal from Visual Light Photon Counters (VLPC) with SIFT chip:

provides trigger information threshold of 1 photo electron (p.e.) signal anticipated (mayor

challenge) average p.e. yield measured in Lab3 cosmic ray test:

7 p.e. (with 11m light guide) contribution of 0 p.e. and 1 p.e. peak (Poisson), i.e.

inefficiency of single layer at threshold of 1.5 p.e 7 p.e. average: 0.7% 5 p.e. average: 4.0% Note: CFT has doublet layers!

We cannot afford any additional light attenuation! Optical transmission from SciFi to VLPC has to be optimal!

Light Yield and Trigger Efficiency


Central Track Trigger (CTT) requires a hit in all 8 axial CFT layers (8 out of 8)

for highest p-t bin (p-t > 10 GeV): 7 out of 8 MC simulation based on

7 p.e. yield, not including compromised fibers (Alan Bross):

CTT Single Track Efficiencies

(-0.5<<0.5)

Threshold Cut (PE) 8 out of 8 7 or 8 out of 8 1.0 100 (100) 100 (100) 2.0 97 (94) 100 (100) 2.5 95 (88) 100 (99) 3.0 90 (77) 98 (96)

But: Hit efficiency (doublet layer) in cosmic ray test was estimated to be approx. 98% (Brian Smith). Additional inefficiencies in CR test?

Some naïve estimates:

single layer eff. 98.0% 99.0% 99.5% 99.87%

8 out of 8 85.1% 92.3% 96.1% 99.00%

7 out of 8 99.0% 99.7% 99.93% 100.00%


152 Light Guides on Each Side of CFT

CFT installed in central cryostat

Light guide with curved connector


Light Guide Routing 1

2-FRONT

CFWCFE

CENTERBEAM

FORWARD DETECTORS

FORWARD PRE-SHOWERICD

LEVEL ZERO

WAVE GUIDEPATCH PANEL

VLPC CRYOSTAT INSTALLATION

SEPT 98

WAVE GUIDE FIBERBUNDLE INSTALLATION

VLPC CASSETTE

VLPC CRYOSTAT

SILICON TRACKER POWER SUPPLIES

VLPC CASSETTE

VLPC CRYOSTAT

TRACKER CABLING SLOT

ECS






Light Guide Testing 1

Scintillating fibers can be illuminated with LED panels. Each ribbon (256 fibers) is covered by 3 panels on both ends

respectively. LED panel designed for calibration purpose (no absolute calibration

possible, coupling LED panel – SciFi non-uniform and unknown) provides a measurement of light attenuation (LG quality)

Photo peaks in the ADC spectrum of single fibers:

Poisson fit (smeared by Gaussians) provides a measure of the average p.e. yield for a fiber channel (scintillating fiber + clear fiber + VLPC chip [quantum efficiency]).

p.e. yield with LED system: 0.5–2.5 p.e. depends on light coupling into scint. fibers


CFT Front-End: 4 production cassettes, any 2 integrated in the read-out (limited by electronics)

LED pulser data taken with primary DAQ (L3) CFT is currently primary DAQ user

hundreds of runs with millions of events provides thorough test of DAQ and data logging system

online/offline analysis using examine and stand-alone kumac/fortran code

online: direct feedback to installation team mapping SVX channels – fibers not provided in DØ offline framework

(stereo boards) LG are tested with 2-3 different LED settings.

provides a partial cancelation of LED panel effects P.e. yield can be determined with 7% accuracy on average.

But non-uniform coupling LED panel – SciFi can fake losses. classification of LGs based on uniformity of p.e. yield distributions

can be compared to ribbon and LG QC with radio-active source (Lab3)

Light Guide Testing 2


QC: Light Guide with Uniform Response

dead electronics LED panel edge

LED panel edge


QC: Light Guide with Crack Structure

The worst LG so far!

Even-odd structure indicates crack in curved connector in between doublet layer.

To minimize cracks curved connectors (LG to ribbon) are now glued with temporary clamps. Glue is strong enough to stop evolution of cracks (tested independently).


CFT Curved Connector


QC Summary: LG on Axial Ribbons

Problematic LGs were detected during QC: easy: wrong labels, wrong pinning of connectors coupled to

cassettes severe: several LG reglued, 5 LG entirely replaced (cracks)

Classification of the 152 LG (axial side), multiple class. allowed: low gain means < approx. 75%, LED effects not entirely cancelled,

performance of whole system including scintillating fibers good uniformity: 43 edge fibers with low gain: 25 drop at edge (<20%): 30 drop at edge (>20%): 16 Indication of crack: 17 several (>3) low gain fibers: 9 structure (period 32): 21 bad: 1 (shown in QC example) marginal: 20


QC Summary: LG on Stereo Ribbons

82 out of 152 LG installed and glued, QC completed for 69 LG On average LGs on stereo side have a better performance than those

on axial side: due to less handling? initial routing now close to final route only 1 replacement candidate up to now Ironically optimal performance is much more crucial for LGs on axial side,

since stereo layers are not included in CTT.

Installation and QC anticipated to be finished 12/11 2 additional checkouts of optical/electronic read-out system needed

before RunII after LG connection to final cassette after installation of AFE boards

What I haven’t mentioned: only limited QC of CPS LGs possible due to non-uniform LED response

(by design to account for large dynamic range) FPS LGs not yet produced

Documents

All DØ Meeting, 12/01/00 Central Fiber Tracker Light Guide Performance Thomas Nunnemann Fermilab 90 days before RunII … (a.k.a. 12/01/00)