Minutes of the meeting on module production Heidelberg 11.-12.3.04
======
participants: Adam Navrot, Marek Szekowski, Antonio Pellegrino, Henk
Schujlenburg, Sebastian Bachmann (thursday), Ullrich Uwer, F. Eisele
------
agenda:
1. results of test measurements on modules
- wire tension, HV tests, dark currents, dark pulses (in air and Ar:CO2)
- gain, current measurements
- equipment for tests ...
- agreement on common tests procedures
- acceptance criteria for wires,...
- cleaning procedures fro twisters, PCB's etc....
( all 3 groups: Marek, Sebastian, ?)
- review of test equipment (wire tension, HV tester(Marek))
2. Module production: critical points
- template and panel adjustment
- glueing steps /amount of glue etc.
(all 3 groups)
3. straw confectioning, crimping,..
change of endblock/
Henk, .....
4. material procurement and delivery schedules:
a) existing materials in stock (Jan)
b) existing orders and expected deliveries (Jan)
- panels/ sidewalls (update leszek) ( GTS foil, Cu-tape, labels, solder)
- straws/twisters/endblocks
- wire
- PCB's (end and middle)
- small parts
- Al spacers/ plastic spacers: decision on order in Crakow
-
5. planning for module production and material procurement:
a) for 10% of modules - (march-june 04)
b) 40% deliveries ( --> end 2004)
6. steps to common production instructions and test procedures
documentation
7. module tests under real conditions and necessary equipment
a) verification of present construction - update of tests agreed on in Amsterdam
(tests with counting gas, HV stability,...)
b) routine tests for all modules
c) survey of equipment
8. station assembly
a) NIKHEF fall 04
b) Plans for CERN, tentative time scale for CERN start
- infrastructure, manpower..
------
1. experience with module production
======
S. Bachmann reports from Heidelberg experience based on 2 1m modules and the first two
5m panels.
a) glueing panels to straws:
At the first attempt with a 5m panel, the amount of glue (+25% glas bubbles) was too much
(ca. 300 g/panel)==> the straws were glued to the template at many places.
The second panel glueing was successful.
Glueing of straws:
======
The following parameters were used:
- ca. 130 g /panel (Al surface only) without glass bubbles
- extra glue with 50% glas bubbles in the PCB area
- the glue was distributed to the panel with a foam rubber roller
- the glue polymerised for 1 hour at 23 degrees before glueing it to the straws.
It was rather viscous then. The inspection of the glue at that time shows that it
has formed small droplets everywhere on the panel such that the thickness of th glue
varies locally between nearly zero and ca. 150 micron.
NIKHEF has recently used ca. 220 g of glue /module but including the glue in the
PCB area. They added 25% of glass bubbles.
Both at HD and NIKHEF the straws were not glued to the panel in rather small local areas
which could be repaired after the panel was removed from the template by adding additional
glue.
The following conclusions (and recommendations) are reached:======>
i) glass bubbles should not be used in the straw area.
ii) using a rubber roll to dispense the glue gives satisfactory results (better than using
a metal tool with tooths).
iii) waiting for polimerisation for ~1 hour is recommendable
v) the amount of glue should be around 130 - 160 g in the straw area.
------
distance panel/straws:
======
Both NIKHEF and Heidelberg observed that the feets of the support tool were not touching
the template at all 8 places leading locally to too large distances between panel and straws.
This may be due to some tolerances in the template or of the support tool.
conclusions:======>
a) reduce the nominal distance between panel and straw from 100 micron to 50 micron
b) gauge every support tool by adding thin plates to the feet such that all feets
touch the template.
------
side adjustments and tolerances:
======
all groups report that PCBs stick out sometimes on one side of the panel creating problems
during module assembly (sidewall glueing)
===>
a) for 5m panels the panel will be pushed in the middle (at 2.5 m lenght) to the correct
position at the time when the panel is fixed to the support tool.
b) the width of the middle PCB should be reduced by 200 micron (machining of existing
pieces)
------
Dowel pins , tolerances ans spacers:
======
==>Henk and Adam will work out a solution which prevents that the dowel pin is glued to the
template.
The dowel pin tolerances are defined very strict at present which creates problems when
the module is assembled because we have three holes in 2 panels and one spacer which have
to be perfectely aligned w.r.t. each other.==> Henk will define a modification.
It is agreed that there is NO strict tolerance requirement for the Al (and plastic ) spacers,
there the holes can be larger because the precision in positioning is defined by the panel
inserts.===> Henk will think of 'more reasonable' technical tolerances.
------
wiring
======
a) the wiring speed looks reasonable at all places ( about 5 to 6 hours) for a 5 m panel.
b) Heidelberg reports that many twisters were moving during wiring due to air pressure.
This affected ca. 20% of twisters which moved by large distances such that HV stability is
lost. Another 20% moved by small amounts ( up to 1cm).
==> the crimping (done by HD technicians at NIKHEF) is not reliable enough. NIKHEF saw this
problem only at much reduced level.==> conclusions: we need more safety=>
==>see straw confectioning
c) soldering of wires: Heidelberg uses HERAB solder tin with very small amount of flux resin.
NIKHEF and Warsaw use normal solder with flux resin. They report problem with resin
spraying over larger distances.
===> NIKHEF and WARSAW got samples of HERAB solder tin from Heidelberg and they will try
it out. Soldering is clearly more difficult but there is agreement that large amounts of
resin have to be avoided.===> they will switch to another solder
------
wire tension and HV tests
======
all group show their measurements on the first modules or panels.
wire tension:
-it is agreed to use 70 g /wire for the time beeing
- for 5 m modules the wiring should be done separately for both straws using independent
weights in order to minimize the danger that tension is lost by friction at the 4 twisters.
- HD uses one 70 g weight applied directly. NIKHEF and Warsaw use pulleys with 140 g weights.
The wire tension measurements of Warsaw show sometimes high tensions up to 100 g/wire.
The origin may be incorrect use of the pulley -- to be checked.
** in general the wire tensions are within narrow bands of +- 10% of the average. There
are however individual straws which significantly differ from the average
--> wrong twister position, wrong use of pulley, etc..HD exchanged these wires in the 1m
module. Not done at the other places.
**
acceptable range of wire tension -when do we replace a wire?
------
There are two limits: - tension too high near the breaking limit
- tension too low giving HV problems.
The lower limit is unknown -- it depends on the outcome of HV tetss which are not yet done
systematically.
===> HD will use the next panel to strech half of the wires with only 60 g in order to
establish a lower limit on the tension based on HV tests ( this will need > 2 weeks).
tentative ideas how to define acceptance limits:
a) one needs an absolute lower limit on the tension (comes out from HV tests?)
e.g. ~ 60 grams
b) upper limit < creeping limit of ~ 120 g e.g. <85 g.
c) define acceptance cuts always relative to the average of all straws in one module
(to get rid of absoulute tension measurement)
------
HV tests:
======
all groups reported results - see transparencies . Comparison is done in air and at a HV of
1.6 kV.
Heidelberg reports very few wires with higher dark currents -- most wires are below 5 nA.
The few wires which are above have been exchanged and the high values disappeared.
The measurement is done with a HV current meter developed for HERAB -rather refined.
There is one group of straws which show significantly enanced dark currents in average -
this is traced back to using the rest of a wire roll where the wire was in contact with
the plastic roll==> could be indication that cleanliness of wire plays a major role.
The two other groups see a large fraction of wires with high and sometimes very high
currents (exceeding 100 nA). The currents are partially observed to get reduced with time -
Warsaw reports that after flushing with Ar:CO2 many noisy wires have much lower dark currents.
The evidence is however not so convincing at present.
==> we definitely need a better defined program of measurements and we have to understand the
differences
Known differences:
- measurement method: Heidelberg measures current with HV current meter
the others measure voltage drop across a resistor.
===> Heidelberg will also measure the dark current of the 1m module of Warsaw (now at HD)
- Warsaw and NIKHEF use solder with flux resine, HD does not
===> we want to switch to solder tin with little flux resine
- cleaning of material (panels, twister, PCB,....) before every working step with
Aethanol is done at HD, not at the other places
===> all materials have to be cleaned with Aethanol before every new step
- endblocks and twisters in US bath
PCB's and panels by wiping with Aethanol
------
HV tests with counting gas to evaluate the robustness of our solution:
======
First tests of modules are reported by NIKHEF and Warsaw. Grass Appeldorn has measured
currents with a Strontium source for one module systematically. Warsaw reports first
measurements for the 2.5 m module.
The following tests will be done now (starting week 12):--> see test program sheet
(these test were already discussed and agreed on end of january but not yet started.)
1. look at noisy wires with a scope, measure dark pulses and F55 (or Co 57) pulses,
measure dark current.
goal: find out origin of problems and define acceptance limits.
==> Heidelberg has adapter cards and (discrete) analog amplifiers to do this measurement.
Both Nikhef and Warsaw will take along one adapter card with 8 amplifiers. --> everybody
will do the same measurements.
==> U. Uwer is in charge to write up a clear procedure for theses measurements defining
ther operation point (prob. 1450 V for Ar:CO2 70:30) and threshold for dark current rate.
2. measure HV behaviour in horizontal and vertical position. Repeat by deforming the module
goal: see if distance of twisters is small enough to avoid HV problems. These tests
will be later repeated at HD with wires which are strung with only 60 g.
3. mistreat the module by carrying it arouns and bending it. Repeat gas tightness test
and HV test.
------
3. straw confectioning
======
The straw confectioning machine was inspected and all steps of straw confectioning demonstrated
including proposals for improvement. HD will update a small memo on experiences and proposals for
improvements and distribute it.
Main areas of necessary improvements are:
a) The twisters need better crimping to have more safety against movement during wiring and transport.
It is proposed to use 4 crimping points per twister insted of 2. HD will design a suitable
soldering head.
b) the risk of damaging the straw must be reduced: - the straws should be supported/protected
over a larger length especially at the end.
- the straws have to slide to the cutting tools without the need to guide them by hand
- movement of the straw along ist axis should be avoided if possible
---> see the separate memo on straw confectioning
Modification of endblock to allow crimping of tounges:
The plans how to modify the endblock exists. Phillips has to build new inserts for the mold
which may take at least 5-6 weeks. A crimping tool is at NIKHEF to test the method.
===> while this modification is supported it is decided to order another 10000 endblocks
of each type and 20000 twister at Phillips with the present mold before it is modified.
4. material procurement and delivery schedules
======
It is agreed that we presently care for 2 production steps:
a) the production of 10% of modules in the period april to june ( about 30 modules in total)
b) another 40% of all modules from july to end of 2004.
The possible distribution of modules among the 3 production sites and the required amount
of material is listed in EXEL tables which are updated by F.E. ===> see the WEBsite
status and decisions:
10% production goal:
we miss assembled feedthrough boards , Al spacers and small parts for the time being.
This blocks production at all places.
** feedthrough boards:
55 PCB's for both up and low side are available and components for the full OTR.
20+20 assembled PCB's are promised to be delivered before march 26. (this is good for 10 modules).
another 35+35 PCB's are expected to be assembled starting next week. Delivery is planned for
week 17 (before april 23).
==> since both HD and Warsaw cannot continue it is agreed that 5 upper +3 lower PCB's and
corresponding components
are sendt to HD on monday for assembly there,unless the electronic workshop at NIKHEF is able to
assemble them within 3 days. Two assembled upper boards will then be sendt from HD to Warsaw
as fast as possible.
** small parts: the material is availble now at NIKEF but no parts are produced.
==> HD will evaluate on monday if its workshop can produce small parts for a preseries
of 10 modules fast. If YES then NIKHEF will send corresponding material to HD.
==> HD will also try to find a company which is able to produce all small parts. It
will give an answer up to march 30 if it is able to take over the responsibility to procure
these parts completely.
** spacers: A company in Crakow has given a good offer to produce the AL spacers but asks
for a change of the flatness tolerances from 25 micron to ~ 100 micron.
==> Henk will produce new drawings and send them to Crakow (Leszek) via E-mail.
Antonio will place an oder of 60 pieces of a preseries during his visit to Crakow next week.
Delivery is expected and needed up to april 15 ( arrival at the institutes).
** plastic spacers for short modules:
Jurek has produced 6 prototypes which look good and will be used for the first short modules.
His offer looks acceptable. ===> Antonio will place an order for the full amount.
---
other material partilally missing for the preseries:
** STraws: production of another 10000 straws by LAMINA will be ordered immediately (so far
only 3000 straws are available).Next delivery should already arrive april 15 at NIKHEF.
( straws will be needed e.g. at HD to start continuous straw confectioning by students)
** short panels: Crakow is asked to make next panel delivery end of april. This delivery
has to include at least 20 S1U panel (and 10 additional plastic spacers).
------
next large orders for 40% panels or full amount:
======
** all remaining PCB's for long modules have already been ordered.
the middle PCB's for short modules will be ordered after inspection of the delivered preseries.
** 10000 + 10000 endblocks and 20000 twisters with pesent mold
** 60 rolls of wire
** all AL spacers
** all plastic spacers
These orders should be placed with minimal delay such that the material is available begin of july.
------
Material for panel production:
======
a) the rest of the GTS capton/Al laminate is ordered and expected end of june.
b) Cu strips: Crakow got 13 rolls a 20 m each. This is sufficient for 650 panels e.g. for the
whole production.
c) solder Al123: they used 30 g for 60 panels and have 30 g left. So they need another 300 gramm
e.g. 10 rolls. Next delivery is needed within 1 month.
===> to be delivered by NIKHEF
d) labels: the delivery of labels by NIKHEF so far was lacking behind production--> handwritten
labels are glued to the panels. It is proposed that all labels are produced in one go now by NIKHEF
All institutes are willing to replace the handwritten labels by barcode labels as soon as they
are delivered to them.
e) PEEK material for plastic inserts and spacers: Leszek is asked to get this material
now as soon as possible.
------
5.Working instructions, test documentation ..
======
Working instructions have to be made commonly availavle. We cannot wait for a final document,
which will still take long time. ===> the present version of instructions should be
available on the WEB and should be updated as soon as new experiences require it.
So Jan and Sebastian are asked to do that next week.
-examples to be written up are:
- amount of glue, distribution and glueing procedure
- cleaning instructions
- test instructions for wire tension and HV on air..
We need rather soon a stable system which allows to document test results and material
lists for a module etc....
===> a dedicated meeting will be held at NIKHEF to discuss these issues and decide on
procedures. tentative date is: thursday april 20. To be checked with JAN.
topics: - instructions to build a module
- test equipment and test documentation
- program and data base to document module assembly and test
** Ulli proposes that the responsibles of module production have a weekly telefone conference
to exchange news on production , material procuremnet etc. This proposal is welcomed
by everybody. F.E. is asked to make the necessary steps and eventually also participate.
------
6. test instructions and procedures:
======
- gas tightness and gas flow
==>Sebastian is asked to make a short instruction for test of gas tightness.
--> the test setup should be robust and simple. NIKHEF will need a sniffler for Argon.
==> can be provided by HD.
- HV tests
==> Ulli will write a short instruction how to do HV tests in air (HV setting, limits ...)
==> he will also write up tets instruction for test with counting gas
==> he is responsible to provide a simple and effective hardware to look at analog signal
pulses for every straw.
- HV tester (Warsaw)
The tester needs still improvements:
a) the noise level is still too high
b) Arek has to include a diagnostic step which allows to find shorted wires and disconnect them
( it is porosed to start with a first step at ~ 10 V which will lead to 1 microAmp for a shorted