SCTlinks 
 

 Minutes of SCT links meeting

 14:00-18:00   22nd September

 Bat 40-RD-10 
  

John presented the results of gamma irradiations, using the Birmingham Cobalt 60 source, of production samples of the main delivery of Fujikura step-index multi-mode pure silica core fibre. There is no indication of degradation up to 350 kGy which is higher than the Pixel level of 300 kGy. More Details are avaiable here. John described the results of lifetime studies of 96 epitaxial silicon photodiodes in four sets of 24 irradiated to different neutron fluence levels up to 10^15 n/sq.cm./s. The tests are carried out at an elevated temperature of 60 deg.C to accelerate the ageing and each photodiode is under constant illumination from a VCSEL at ~800 microW with a reverse bias of 10V. Allowing a factor of 1/3 for the on-time of the LHC, the total ageing time is 3978 hrs equivalent at the LHC. For operation at -7 deg.C, the enhanced time is 261 yrs assuming an activation energy of 0.6 eV (or 133 yrs at 0 deg.C). Given that no photodiode has failed, using Poisson statistics, the mean-time-to-failure is estimated to be >2720 years for a batch of 24 at -7 deg.C with 90% confidence, assuming a constant failure rate. The corresponding upper limit on the number that would fail in e.g. the barrel section of the SCT (2112 modules) in 10 years is 8. Although the reverse bias required to achieve full depletion increases with increasing fluence, there is no indication that this changes with time after irradiation. While no photodiode has failed, there are five cases of detached wire bonds, four of which are not attributable to external causes. This point is being investigated. More details are avaiable here. 20 DORIC4s have been irradiated with neutrons and gammas and work correctly after. More details available here. 9 VDCs irradiated to 3 10**15 n/cm**2 at Ljubljana work correctly.
These VDC chips will be irradiated to 10 MRad at Birmingham. 7 MITEL 1A444 VCSELs have been irradiated to 5 10**14 p/cm**2 in the CERN PS. All VCSELs were alive at the end of the tests and gave acceptable light outputs after 48 hours of annealing.
More details are available here. Measurements of SEU with 14 MeV neutrons showed a significant BER with fluxes equal to that expected in the Pixel B layer. These erors are due to large energy depositions
in the active volume of the PIN diode. The problem can be greatly reduced by operarting the links at higher light levels. Further tests will be carried out next year with pion beams at PSI. More details are avaiable here. 2 of the phase 2 Marconi packages will be irradiated in the PS in November to a fluence of 4 10**14 p/cm**2. These tests will be done unpowered. Further tests of 4 powered packages
will be performed next year to a fluence equivalent to the Pixel B layer. KK Gan reported on the work to produce a ceramic opto-package based on a simple cap and base scheme. Many problems have been encountered and nearly all have been solved. It is hoped to produce working VCSEL packages in the near future. More details are available here.

The first two of the Marconi phase 2 packages are being assembled and tested this week. We will mount them on dog-legs and perform the standard link tests.

Honeywell have provided us with a 12 channel VCSEL array couple to 12 way ribbon fiibre.  The package works well and has been irradiated to 5 10**14 n/cm**2 at ISIS and all channels are still working. Discussions are in progress about how to produce an ATLAS style 2 VCSEL and one PIN package. More details are avaiable here.

Taiwan have joned ATLAS and are keen to produce an opto-package for SCT and Pixel detectors. They will work to the agreed ATLAS specs. They will produce first prototypes before the end of the year. We will collaborate with integrating them into our system tests and on radiation tests. If this stage is successfull more detailed radiation and enviromnetal tests will be performed next year. A FDR for the opto-packages will be held in April 2000.

4 channel prototype DRX chips have been successfully tested. A 12 channel final prototype will be submitted in September. DORIC4A has been tested and it solves the 20 MHz lock in problem that was present with DORIC4. More details are available here. The first prototype 4 channel BPM chips ahve been tested. The main problem was that the output current was always fixed to the maximum value. This has been understood as due to a problem with an i/.o pad. This has been confirmed by cutting lines wiith an ion beam for some channels on 3 chips. The output pad has been fixed and the 4 channel chip has been resubmitted to AMS. These chips will be used in the harness tester. The final version will be a 12 channel chip and it will improve the mark:space ratio adjustment logic which is not fully correct in the correct chip.More details are avaiable here. The first system tests with the new Marcini package will be performed with the existing style of dog-leg cables. Mechanical porototype studies wi;ll be performed to define the details of the final system. Parts are available for tests of the solder connection between the low mass tape and the opto-flex. Rob Millea (RAL) will get samples bonded up by a company (Unitex) using a pulsed thermode.
The pieces will be measuered electrically and using an IR camera. If the tests are successfull we will try to purchase a pulsed thermode machine to do the bonding. More details are available here.  A fibre routing via the vousoir is acceptable for the latency budget whereas the routing around the outside edge of the muons is not acceptable. TC are aware of this and think that the vousoir routing is acceptbale. An analysis of the fraction of dark fibre showed that we should use routing via quadrants. Decisions on which patch panels we should use,  will require prototyping work. More details are avaiable here. A harness test system has been defined and the modules required for it are being designed. More details are avaiable here. The specifications for the optics was discussed and an agreed set ofspecs will be made avaiable after discussions with the Taiwan group and the Pixel opto group.
  Measurements of noise with and without the use of redundant TTC links have been performed at the system tests. An indication for an increase in noise was reported but more data will have to be taken with better quality modules. More details are avaiable here.  F. Hartjes reported on the status of the plug-in opto-packages that could be used for the forward system tests before the Marconi packages are avaiable. More deatils are avaiable here. The current version of the BPM chip can only drive common anode VCSELs. The current MITEL VCSEL array is of the common cathode type. We hope that common anode VCSELs will be avaiable on the near future but the use of intermeditae circuitry between BPM and a common cathode VCSEL array has been ionvestigated as a backup option on case we can not obtain common anode VCSEl arrays. Circuits based on  a current mirror and a balanced driver both worked well when used with high speed transisors. More informations is avaiable for the first and second options.