Summary of ATLAS SCT links meeting, CERN, 26 June 1996.
Conclusions and actions from SCT links meeting of 26 June 1996.
---------------------------------------------------------------
1. LED and PIN DIODE performance and radiation hardness.
- presentations by Remi Mommsen and John Dowell, comments on PIN's
from Tony Weidberg.
- Remi gave a detailed discussion of radiation level and location
a) 56 ABB LED's irradiated at ISIS to neutron dose of 10**14,
3.10**14 and 10**15. In progress. Measurements from end July.
b) 56 ABB LED's irradiated to protons at PS, 3.10**13-5.10**14.
Serious damage.
All measurments, including annealing, at room temperature, and
period of 2-3 days between irradiation and measurements.
All except one diode alive, and this perhaps the result of
mechanical stress. Light output after annealing typically
40-80% of initial, depending on radiation dose.
c) Annealing strongly depends on current - slow at 10mA, fast
at 30mA. Raises the possibility of a special high-current
DC annealing mode for use in ATLAS.
d) Substantial uncertainty on flux measurement.
e) There was discussion on safety factor of operation. No clear result.
f) 26 LED's are now being irradiated at 3.10**14, with pulsed
forward bias.
g) Future measurements- 320 LED's from GEC expected in July
- 5 16-fold arrays of low threshold VCSELs from
Sandia in July; 15 extra arrays in September
- a major question is how to keep control of fluences. In
current measurements, separate monitor for each array. Note also
Birmingham measurements using a rotating table to equalise fluences.
- John Dowell described measurements of 4 ABB's at the PS, including
2 measured online. Total dose 6.10**13 over 6 days, and operation
at 20 mA (10 mA used in case of Bern irradiations). The LED's had
already been irradiated with neutrons after which they annealed
back to full light yield. Birmingham see a light yield of ~60%
(not incompatible with the Bern measurments given the uncertainty
in the overall fluence).
- John described future expected measurements.
a) The dynamitron at Birmingham will give up to 10**14 n/day,
and expect radiation of ABB LED's there Monday 1st July.
Plan tests with proton + neutron with GEC packages; 10 LED's
and 14 PIN's
b) Will do longterm tests with 100-odd PIN's
- Dave Campbell asked about VCSELS, and radiation status of that.
Werner Langlans responded that Stefanini had tested with neutrons
to 10**14. Possibility of VCSELS for the PIN diodes will be
discussed by Werner and Kurt.
- Proton irradiation of PINS. October?
2. Placement of readout hybrid.
-----------------------------
- Robert Kowalewski presented 3 options.
A) One optical hybrid/module, near module
B) One per local support, at end of local support.
C) Optical conversion outside radiation (in crack??) with a 10m
twisted pair to that point.
- Option B was disfavored following discussions at Munich. Signal and
noise tests using shielded twisted pair cables will be performed
in collaboration with the TRT group. A detailed plan of the tests
proposed will be circulated for comment.
- Options A, B, and C were costed (see transparencies).
The question of fibre costs is important. Note also the fibre
costing given below by John Dowell.
- Question of patch panel position was raised, and regrouping of
fibres and cables. DAQ have asked to group into ROIs.
- Importance of reaching a decision on options was stressed. The
agreed timescale for such a decision is spring 97.
- Option D, where twisted pair transmission is used all the way to
the ROD (after switching to thicker cables and introducing a
repeater), will be included in the revised costing to be done.
3. Clock and LVL1 trigger distribution.
------------------------------------
- Dave White gave a detailed presentation of the pros and cons of
pulseheight modulation, as compared with the biphase mark scheme.
- Details are shown in transparencies. He judged that each scheme
had strong points, and in particular noted that the decoder of the
biphase mark scheme was not trivial. Further, there was concern that
with the biphase mark one could lock to the wrong bunch crossing.
However, AC-coupling is an advantage, and a single pulseheight enables
a doubling of the S:N relative to the PHM scheme. Dave concluded
that the biphase mark scheme was more robust at the system level.
- The system differs from TRT because they use twisted pair, with
separate clock, trigger, hard reset etc lines. It was noted that a
+- 0.5ns jitter exists on the RD12 clock, and it is necessary to
be able to make delay scans. This is implemented in the ABC chips.
4. Update fibre costs. John Dowell/Ian Dawson.
-------------------------------------------
- Updated costs, including connectors, were obtained as estimates
from E u r o-O p t i c s (non-disclosure). Using this, John updated
costs for the full optics chain.
- The total estimate for 1996 estimated prices becaomes 1.7MCHF, and
1.38MCHF in 1999. This updates the estimate of 1.05MCHF previously
quoted by John in May. The fiber costs estimated by Robert did
not include assembly into ribbons, and were based on an estimate
of 0.2 CHF/m for non-radhard fiber from Robert McClaren.
- Oxford had a price for a special radhard fiber from F u j i k u r a
for 0.25 CHF/m, which is sharply lower than the 1 - 1.5 CHF/m
quoted by John for radhard fiber - this needs to be pursued.
5. Actions summary.
----------------
a) R.K to rediscuss with Robert McLaren wrt costings, and revise
accordingly, including new E u r o-O p t i c s numbers.
b) Richard asked that an overall view of the TRT, SCT and Pixel
systems should be prepared, to ensure complementarity at each
step. In particular (Alex Grillo), the pixel community needs
to provide a strong case for a 80Mbit/sec clock speed. Do they
need it?
In particular a detailed evaluation and test plan for options
A and C were requested for the next meeting. RK and DC noted
that for this, a specification of the voltage compliance, and of
the transmission line operation (differential +-4mA current
level as in TRT?). It was agreed that this test should be for 15m
of cable.
c) The biphase mark scheme was adopted. Dave W. will work on
a specification for the decoding IC (Doric).
d) A. Grillo noted requirement to
- include possibility of a 10m twisted pair transmission
in the ABC spec (switch?)
- provide specification for power supply to the optics
hybrid (in crack or as an individual optic module). This
is delicate for questions of common mode etc. He noted
an attempt to group all supplies of a module in a single
'box'. He requested an ~1 month action.
6. AOB.
---
a) The meeting requested Robert Kowalewski to convene and
coordinate the readout activities in the coming year. Robert
agreed to this task.
b) Next meeting: SEPTEMBER 12
14:00
c) Richard Nickerson suggested naming the group without making
reference to "optics", as twisted pair transmission may be
the final choice. Robert proposes "SCT links" - feedback
requested.
(Minutes prepared by Allan Clark and Robert Kowalewski)