June 1, 2000 Daniel Levin -- University of Michigan

U. Michigan Tube

QA/QC

Daniel Levin University of Michigan

Ann Arbor, June 1, 2000

420 Mod 0 (EMS 5) tubes made from 7/26 – 9/3, 1999 • Pre-assembly checks

– End-plug specs, tube survey, tube length

• Wire tension

• Wire position

• Tube to end-plug contact resistance

• Leak detection

• Dark current

June 1, 2000 Daniel Levin -- University of Michigan

Endplug

preparation

700 endplugs surveyed:

OD/ID of body,

inner hole concentricity,

twister concentricity & size

===> 99% within spec.

Endplug

assembly

ultra-sonic cleaning, twister/stopper insertions, O-ring mountingEndplug

problems

5 failed: precision go/no-go check

3 failed: threaded bras s pin bent

2 failed : crimp pin hole too small

June 1, 2000 Daniel Le vin -- University of Michigan

Tube preparation350 tubes: no precleaning

53 tubes were cleaned. In 50% small metal chips were removed

All but 40 inner wall ends & endplugs cl eaned.

Tube Survey: 15 tubes: burrs at ends

10 tubes: scratc hes on inner wall

2 tubes : inner diameter too small

6 tubes: residues visible

1 tube: pin hole in wall

2 tubes: same ID number

June 1, 2000 Daniel Levin -- University of Michigan

Tube assembly parametersEM crimping: C=10 uF HV=11.2 kV

crimping (if leaking): HV=11.5 kV

O-ring compression about 20%

Wiring pre-tension= 475 g for 15 sec. Final tension = 359 g

Assembly time= wiring, tens ion/length measurement ~ 8 min

June 1, 2000 Daniel Le vin -- University of Michigan

Tube assembly problems4 broken wires: during tensioning

2 broken wires: due to bend in endplugs

1 broken wire: operator error

1 tube: improperly clocked in endplug holder

69 tubes : leak at endplugs required reswaging

all but 6 tubes were sealed by reswaging

June 1, 2000 Daniel Le vin -- University of Michigan

Status of Tube QA/QC Measurements

(as of 9/99)

Wire tension : mean 359, rms=1.7 g

1 tube out of tolerance

Wire Position (X-ray) : mean offset = 10 +/- 5.2 microns

1 tube out of tolerance

Tube Length: rms < 100 microns

1 tube out of tolerance

Tube to End-Plug contact resistance:

all tubes in itially at 5 +/- 2 x 10

-5

Ohms

Leaks :

4 tubes 1.2-3.7 * 10

-5

cc*bar/sec1 tube 1.0 *10

-4

cc*bar/sec1 tube 1.0 *10

-6

cc*bar/sec

June 1, 2000 Daniel Levin -- University of Michigan

Leak Checking

leaks are possible from end-plug face !

All Mod 0 tubes rechecked

• 420 Tubes have been rechecked at 3 bar He

• New leak test encloses face of end-plug & is sensitive to leaking

via plastic/Al interface

Results:

• Previous 6 tubes flagged in first pass

• A dditional 2 tubes leak out the end-face:

1.3 * 10

-5

cc-bar/sec

1.0 * 10

-4

cc-bar/sec

412/420 tubes pass, < 2% rejection

June 1, 2000 Daniel Levin -- University of Michigan

Resistance Measurements: Nov 99

Cleaning Method Number of

Tube Ends

Number with >

1.0 mOhm

Number with >5.0

mOhm

None

32 3 0

Alcohol

652 9 0

Scotchbrite on

Recycled,

unchamfered end-

plugs

100 7 1

Scotchbrite on

chamfered end-

plugs

32 0 0

Scotchbrite

Aggregate

132 7 1

• All tubes cleaned with alcohol have at least 1 end < .05 mOhm

• Highest measured resistance =8 mOhm

June 1, 2000 Daniel Levin -- University of Michigan

Old & New Contact Resistances

408 tubes

Nov 99

# with > 1 mOhm =

# with > 5 mOhm = 1

Largest resis. = 8

May 00

June 1, 2000 Daniel Levin -- University of Michigan

Contact Resistance:

EM crimp with & without steel spring

Without

steel spring after 9 months:

Mean of 816 contacts = 0.17 +/-0.7 m ?

With

steel spring after 1 week

Mean of 32 contacts = 1.2 +/- .9 m ?

June 1, 2000 Daniel Levin -- University of Michigan

100

150

-20 -10 0 10

Entries

Mean

RMS

388

-0.4681

1.081

? Tension (g)

Change in tension

since fabrication:

9 months

June 1, 2000 Daniel Levin -- University of Michigan

Dark Current Results

Co mplete report: ATL-COM-MUON-99-038

• 384 tubes subject to long duration testing:

– tested 8 batches of 48 tubes/ batch

– perform initial HV polarity reversal on high current tubes

• all tubes start near zero current

– continuously monitored for 30-66 hours

Over some part of a cquisition interval : 33 tubes @ DC> 5 nA

Over final 60 minutes: 14 tubes @ DC > 5 nA

Over final 60 minutes: 7 tubes @ DC > 10 nA

June 1, 2000 Daniel Levin -- University of Michigan

Dark Current Measurement Observations1. Most initially high DC tubes sh ow diminished currents

over hours or days.

2. About 70% tubes with high DC responded dramatically

to the initial HV reversal.

• Most high current tubes improve dramatically with

burn-in.

• Some tubes spontaneously conducted high currents

after many hours of operation. Of 37 tubes with DC >

5 nA:

35 identified within ~ 1 day

June 1, 2000 Daniel Levin -- University of Michigan

Observations cont. 4. A protocol in which a second HV reversal is applied if

DC> 5 nA is often effective.

5. A low humidity environment reduces external leakage

paths.

June 1, 2000 Daniel Levin -- University of Michigan

-10

010

203040

506070

-50 0 50 100 150 200 250 300 350

Minutes

nA and % Rel. Humidity

-2

810

50 55 60 65 70 75 80

% Rel. Humidity

Dark Current nA

Typical Result of Humidity Tests

June 1, 2000 Daniel Levin -- University of Michigan

Dark Curren t Testing: Conclusion

Goal: Efficiently reject high DC tubes

CERN test beam (U. Bretzler) shows they yield noisy TDC spectra.

Constraints:

1.) Don’t impact production flow

2.) Keep rejection rate at a low level (< 2%)

Test protocol:

1) Conduct measurement over 1 day of at least 48 tubes. (Single

day’s production quota.)

2) Apply initial HV reversal to all high DC tubes.

3) Perform second HV reversal after several hours high DC persists.

– About 3.3-3.5 KV, drawing ~.3 mA for ~ 1-2 minutes

4) Reject tubes if Dark Current:

a) > ~1.7 nA/m for final 2 hours.

b) is not reduce d by second HV reversal

Mod 0 result: 2 tubes rejected

99.5% acceptance rate

June 1, 2000 Daniel Levin -- University of Michigan

QA/QC Test timing

leak test ~ 6 min/tube

x-ray tests ~ 8 min/tube (no longer applicable)

EMMI test: (estimat ed) < 2 min/tube

dark-curre nt (48 tubes per batch )

gas & electrical connections, leak check ~ 2 person-hours

scan 48 tube IDs ~ 5 min

gas flow and pressurization: 1 hour

data acquisition: <24 hours

dismantle setup, move tubes: 1/2 hour

resistance measurement < 1 min/tube

June 1, 2000 Daniel Levin -- University of Michigan

0.5

1 1.522.533.54

2600 2800 3000 3200 3400

Tube Voltage

Tube Dark Current vs Voltage

June 1, 2000 Daniel Levin -- University of Michigan

Tube QA/QC Summary

Leak checking method effectiv ely identifies leaks Al/plastic interface.

8/420 tubes rejected

Contact resistance:

After 9 months no appreciable change

Additional measurements will provide better understanding of long term effects.

At minimum alcohol cleaning of tube ends is desired .

Wire position: within 25 microns

Wire tension:

initially all tubes 359 +/- 1.7 g, 1 tube out of tolerance

new measurements of 388 tubes, all wi thin spec.

387 within +/- 4 gr, 1 @ -7 gr

Dark Current:

Monitor & burn-in for ~ 1 day with polarity reversal when needed.

low rejection rate. (< 1%)