Minutes

(1) Fibre Routing (J. Dowell)
R. McLaren has started to study fibre routing from the calorimeter crack to the
RODs. He has started to look at LAr and will consider SCT next. The concept for
a low mass plastic fibre clip to replace MPO connectors was shown. R. Mclaren
had discussed this with Molex with the idea of Molex doing the development of
this. John Dowell made a proposal to use 12 way ribbon fibre everywhere. For
the barrel this would mean data fibres from 6 modules would be grouped into one
12 way ribbon. For the TTC fibres this would require a 12 way ribbon to be
split to provide 6 fibres to each of two rows of cable harnesses. 
There are two major questions: does this work for the forward  and is this ok
for the RODs ?

(2) System Test Preparation

(a)  BER tests of TTC links (I. Mandic)
At the last SCT week, problems with getting low BER TTC links to work were
presented. These problems were thought to be due to the jitter in the BiPhae
Mark optical signal and in the extreme sensitivity to jitter of the current
DORIC design. In order to verify this a very low jitter BiPhase Mark vernier
board has been built. This board allows the separate adjustments of all the
edges of the BiPhase mark signals. This board works well and will be used to
understand the performance of the current DORIC chip.


(b) Cross-Talk measurements (I. Mandic)
I. Mandic reported on measurements to study possible cross-talk effects on the
GEC opto-package. The cross-talk noise was measured by sending "0"s in the TTC
stream whilest simultaneously sending data via the LDC. The conclusion was
that the ratio of signal to cross-talk noise was better than 30:1. This should
be greatly improved with the new DORIC design optimised for the larger light
levels from VCSELs.

(c) Barrel Dog-leg cables and Optohybrids (P. Jovanovic)
6 barrel dog-leg Cu/kapton cables to connect the opto-hybrid and module hybrid to the low
mass power tape have been fabricated at CERN. The ceramic opto-hybrids have
been delayed and should now be produced in two weeks. Different grounding
schemes will be easy to implement. A simple test board to allow BER tests,
cross-talk measurements etc has been made.

(d) Forward Dog-leg Cables (T. Smith)
A design is being pursued for Cu/kapton dog-leg cables which would allow the 
mounting of the GEC opto-package directly on the dog-leg cable. These cables
will be produced by end of August.

(e) Opto Receiver Tests (A. Rudge)
An 8 way opto-receiver module based on the AME 8 way PIN array has been
designed. The PIN array allows an MT-8 fibre ribbon with an MT-8 connector to
be plugged into it. Breadboard level tests showed excellent noise performance:
noise~ 1.3 pA/sqrt(Hz) and very clean eye diagrams with LEDs as the optical
source. A 3U Europa card PCB version will be fabricated next week and it will
be used to perform BER tests.

(3) Radiation Tests (G. Mahout)
(a) PIN diodes.
Dark currents and responsivity of 96 PINs on ceramic tiles have been measured
with the scanning machine. These tiles are currently being irradiated in ISIS
up to a fluence of 2 10**14 n/cm**2. After irradiation they will be
re-measured. If they are ok then half of them (6 tiles) will be exposed to a 
further 6 10**14 n/cm**2
to verify their suitability for the pixel detector. An aging test facility is
under construction which will be used to perform long term aging tests
after irradiation (at SCT and pixel levels).
(b) Fibres
A 12m length of the Fujikura 50/60/125/250 fibre is being exposed to a Co60
gamma source at Birmingham. No significant attenuation has been observed up to
a dose of 5 MRad. The exposure will be continued up to 10 Mrad. If the results
are good, then we will try to purchase the production quantity of this fibre.

(c) LDC and DORICs
4 LDC and 4 DORICs are currently being exposed to a low dose rate gamma source
at Birmingham. All chips are correctly powered. One LDC has suffered mechanical
damage, the other chips show unchanged currents. The dose so far is 1.8 Mrad.

(4) VCSEL tests (I. Mandic)
Detailed DC and AC measurements have been performed on VCSELs produced by
Honeywell, MITEL and Truelight. Very clean waveforms were obtained for the GEC
packaged MITEL VCSELs driven by LDC. The main conclusion is that lead lengths
should be as short as possible to minimise stray inductances. The LDC current
mirror type circuitry is more appropriate for driving VCSELs compared to the
CMOS switch used in BiLED.

(5) ASICs (R. Wastie)
(a) BiLED
The jitter of the BiLED chip will be reduced. The design will allow for
adjustment of the mark:space ratio to be 50%. The current driver output stage
will use the same circuitry as LDC (see item (4) above).
The chip will be submitted on 29th October.
(b) DORIC4 (R. Wastie reported for D. White)
The main problems with DORIC3 are that it doubles the clock jitter of the input
waveform and that it has a very narrow acceptance gate (~ 2ns) for the data
edge of the signal. In the new design the output clock jitter will be equal to
the input clock jitter. The acceptance gate for data will be 12.5 ns wide.
The input stage will be re-optimised for the larger optical signals expected
from VCSELs. The submission date is 9th July.


(6) Project Plan
This item was delayed until the next meeting



Actions 
(i) (T. Weidberg) Ask G. Tappern to provide accurate estimates for
the lengths of rad-hard fibre required for the SCT.
(ii)(Birmingham) Continue the gamma exposure up to 30 Mrad to confirm
suitability for pixel detectors
(iii) (T. Weidberg) Arrange exposure of Fujikura fibre to 8 10**14 n/cm**2 at
Ljubljana
(iv) (T. Weidberg) Discuss with pixel community if they wish to purchase the
fibre at the same time