Tuesday 8th September Bat 40 -S2-C01
9.00-13.00 Optolinks (1)
14.00-18.00 Optolinks(2)
Minutes 8th September
(1) T. Weidberg reviewed the schedule which requires final prototypes
of links for the system tests in summer '99. This required a decision on
the optocabling scheme for the barrel and the forward.
(2) R. Nickerson reviewed the opto harness scheme designed for the
system test '98. Optohybrids and dog-leg cables have been designed by Birmingham
and produced at CERN. GEC LED/PIN packages have been used with this
system in successful tests of the links (see talk by I. Mandic). There
were two major areas of concern with this scheme. Firstly the power tape
is on the barrel and is not directly cooled and secondly it was not clear
how to make a permanent connection from the power tape to the dog-leg (the
system test '98 uses ZIF connectors). Fluid dynamic calculations showed
that the temperature increase in the worst point of a silicon detector,
for the worst case of the end module in the barrel would be about 0.5 degrees
C. This calculation was pessimistic because it ignored the heat flow along
the barrel which would be significant. A solder connection between two
single sided kapton cables and a double sided kapton cable was designed
in Oxford. A version with copper cables was successfully produced and a
version with aluminum to copper kapton connection will be made. The connection
problem therefore appears to be solvable. There was a consensus to accept
this scheme.
(3) T. Jones reviewed the current scheme for the forward opto-cabling.
For the forward detectors it is not possible to produce a separate opto-harness
which could be produced , tested and installed before the modules were
placed on the disks (unlike the barrel case). The fibre cabling was very
complicated and involved threading fibres under the cooling pipe. A proposal
was shown which involved an opto-package with the VCSEL/PIN package
and an electrical connector. The electrical connector would plug into the
hybrid and the DORIC and VDC chips would sit on the module hybrid. This
scheme is not optimal from the point of view of noise but will probably
be acceptable because of the larger light levels available from VCSELs.
This will be checked with the existing opto-packages and the current DORIC3
chip in the near future. With this scheme the fibre cabling is greatly
simplified and there is no need to thread the fibre under the cooling pipe.
There was a consensus to accept this scheme assuming that the noise tests
were successful.
Minutes 9th September
(1) I. Mandic showed system tests of data and TTC links for a system
based on the GEC LED/PIN packages mounted on the opto-hybrid
plus dog-leg cables. Four such systems were tested. A special venire board
was constructed to allow for the adjustment of the clock to data edge for
the Biphase mark encoded TTC signal. A scan of this time showed that DORIC
correctly decoded the TTC data within a 2ns window. (n.b. this window has
been increased to 12.5 ns for the new DORIC). With the timing optimised
very low BER were achieved and the DORIC showed a reasonable dynamic range.
Low BER were achieved for the data links if the optical power was
greater than 4 uW. Cross-talk measurements were made by pulsing the LDC
with a 20 MHz clock whilst sending TTC data to DORIC. Cross-talk
was observed at low optical power for the TTC link but there was still
a clear region where low BER could be maintained. The cross-talk will be
greatly reduced in the new DORIC which has lower gain as it is optimised
for use with VCSELs.
(2) A. Rudge showed results with the prototype MITEL PIN array. The
performance was very poor, probably due to the PINs being damaged. Mark
Pearce will show these results to MITEL and try to obtain a new array.
(3) A. Rudge showed some studies with his receiver array using the
AME PIN array. Two versions had been constructed, a high gain
one for use with LEDs and a low gain one for use with VCSELs. He showed
results of a simple BER test using non-random data and NIM logic.
These showed much lower noise than seen in the system tests. This discrepancy
should be understood.
Results of time jitter vs optical power for the two versions were shown
and a very clear correlation between jitter and optical power was observed.
With VCSELs a time jitter as low as 10 ps was observed ! Studies
of the use of VCSELs in an analogue system were shown (see transparencies
for details).
(4) G. Mahout showed the status of the high statistics PIN radiation
studies. 96 PINs were irradiated at ISIS and the fluence was measured to
be 1.45 +/- 0.05 10**14 n/cm**2. The unirradiated PINs showed a drop of
54+/- 1% instead of the 30% observed from the test last year. The response
for the unirradiated (irradiated) PINs was 0.51+/- 0.01 A/W (0.24
+/- 0.01 A/W). There was some discussion if the PINs were correctly biassed
during these tests and it was agreed that this would be checked before
the results were shown at the ROME LEB workshop. The PIN diode aging
test set-up is now functional and one tile has been tested for one month
without showing any significant change.
(5) I. Mandic showed results of radiation tests of the Honeywell HFD3013
PIN diode . Three such devices in ST packages were tested. The devices
were irradiated to 4 10**14 n/cm**2 at the Ljubljana reractor. The responsivity
decreased by 32% after irradiation in agreement with last years tests with
the Centronics PINs but less than that observed with the PINs on
tiles (see above).
(6) R. Wastie reported on the ASIC status. DORIC4 was submitted 17th
July ;98 chips back 9th October. VDC submitted for the next AMS run
back 18th December. BPM will also be submitted for this AMS run.
For comments, please contact Tony
Weidberg
Last update 26 May 1997