Summary of ATLAS optical link meeting, CERN, 14 May 1996.
Conclusions and actions from OPTICAL LINK meeting of 14 May 1996.
-----------------------------------------------------------------
1. Electronics architecture.
- presentations by Tony W, Philippe F, Francois V.
- options of C&C architecture must be carefully evaluated, but
after assessment of position of optical hybrid (see below).
Both of the current options (clock modulated by amplitude or
frequency) have advantages and disadvantages.
- it was emphasised and agreed that redundancy of the C&C against
point failures was important.
- (see below) there was significant discussion on optimal choices for
respectively VCSELS and LED's.
2. LED and PIN DIODE performance and radiation hardness.
- presentations by Tony, Kurt B., and Ian D.
- LED radiation hardness currently uncertain; tests continue and
superficially inconsistent data neeed to be understood. Studies
with and without in-situu powering and monitoring continue.
- Driver module being considered. LT pair preferred wrt current
mirror by RAL/Oxford. KB noted possible advantages of the
(modified) current mirror architecture. These suggestions will
be studied. Power consumption was considered important (see
the VCSEL and twisted pair options below). Kurt emphasised also
the larger signals from VCSELS => implying different receiver
design, and possibly an optimal choice of C&C. He noted that
twisted pairs do not require a receiver.
- John D. emphasised the need for CERN-based effort on the receiver
circuit. This was agreed to be studied, but also wrt other
help required for the project.
- VCSELS remain a viable option. Full twisted pair readout must
remain an option. The VCSEL option to be studied after receipt
of units later this year.
- Tony noted the new GEC diodes, and first results (before radiation
testing). It was noted that these are the only
units so far packaged to be suitable for single hybrids for each
module (2 LED's, 1 pin diode). Given that these devices passed the
specs, they could be considered a final package for the SCT.
- There was an initial discussion on the location of the optical
hybrid (see below). GEC offer possibility of a single LED package-
allowing a package with one pin diode and 2 LED's per optical
hybrid. The package is 10*5.6mm. ABB, and the VCSEL option, require
8-fold packages.
- Tony noted status of the pin diodes. Nominal radiation studies
are encouraging, but full radiation testing is necessary. This
includes full electronics specs as a function of time following
irradiation.
3. Optical fibres
- Presentation by John D.
- Summary of the meeting of April 4th. Various items noted:
* radiation - not yet proton radiations, silicon core required
=> cost increase.
* connectors MT look promising but spring clip unsatisfactory
* cost needs review: his estimate, without labour and without
fibre assembly, is a factor 2 higher than CORE.
4. Action and summary discussion.
-----------------------------
* It was agreed that all aspects, would be reviewed as best as possible,
with full consultation to other activities, to reach the best possible
solution.
* It was agreed that redundancy against point failures should be maintained.
* It was agreed to concentrate first on agreeing on aspects of the design
and architecture having implications on the front-end chip, readout,
module design, and mechanics. In this context it was decided to meet
at or near the next ATLAS week (end June) and to have reports on the
following:
A) A further understanding of costing for the fibres (John Dowell
and Ian Dawson)
B) Investigation of options for the receiver low-noise amplifier
(Allan Clark)
C) A report from a working group (Robert Kowalewski and Tony Weidberg)
detailing key issues concerned with the placement
and packaging of the optical hybrid.
In particular it was requested to understand the options of
a seperate optical hybrid per module, as compared with grouped
LED/PIN's at the end of each local support, and connected to
individual modules via twisted pair. This study was requested
to include mechanical, electronic and cost issues.
Allan C. was requested to ask implications on the FE drivers, from
the electronics group; that is driving a (say) 80cm twisted pair
rather than a 10cm hybrid line and/or twisted pair.
* It was agreed that the question of C&C architecture, and the study
of LED's cf. VCSELS or twisted pairs, should follow the above study.
A majority considered it premature to make a base line too early.
* I was agreed that radiation studies of the LED's and the PIN's was
still a key issue, and that ongoing work was crucial (LED's) and
that similar studies were necessary for PIN's (Ian Dawson noted
possibility of this study, to be confirmed ).
NEXT MEETING.
------------
CERN, Bat 304-A
9:00 - 13:00
-------------
(Minutes prepared by Allan Clark)