Minutes of the 24 April 2006 phone conference

1. Approval of minutes of the previous meeting.
   No comments have been made.
2. News on bare modules and components assembly and deliveries:
   1. deliveries from AMS
      20 new modules were delivered on 11 April and 16 new modules + 5 reworked have been delivered on 19 April.
      The situation with VDDA shorts was particularly bad for the 11 April batch, with 5 shorts, but it improved in the next batch (1 short out of 16 modules).
      The modules are enough to feed the lat 10+10 needed in Dormund and Siegen.
   2. deliveries from IZM
      IZM production concluded with 1143 accepted modules, (more than the 1139 specified in the contract), against a total of 1178 delivered (97% yield). That included FEI 2.1 modules. The total number of accepted FEI3 is 1123.
      Final version of spreadsheet has been uploaded.
   3. distribution and flow of components (type 0 cables, shipping frames, FlexMCC)
      • distribution of AMS bare modules
        Martin has performed a comparison about the ranking of AMS modules in different labs (description, data). Since Dortmund is not significatly worse than average, module assembly there will proceed as scheduled.
        The current tested modules in Milano are enough to complete the initial production quotas in Germany and will be distributed.
        Since AMS flow will go on for a while, as soon as laboratories have available flexes to assemble additional modules they should make their request to Attilio.
      • update of B-layer threshold
        At a SQTF meeting there was the discussion if the running b-layer threshold suggested at our MTF meeting was the proper approach. Somebody feel that giving a more equilibrated share between AMS and IZM modules would be beneficial in distributing the risk that one specific technology may fail for an unknown region. Anyhow this point of view was not generally shared and a request was to check the actual worsening of b-layer quality modules after loading on stave.
        Preliminary results from Claudia shows that "good" AMS and IZM modules get on average a worsening of 40 points after loading on stave, indipendent of bump bonding (plots).
        Attilio promised to look in better detail to the specific "disconnected pixels" failure.
        Since not many new modules were available, there is no need to update the b-layer threshold.
      • distribution of flex modules
        Not all the shipments decided at last meetimg have been performed. This is especially urgent for modules needed at CPPM.
        Joerg answers that partial shipments from Dortmund were done and the final ones are ready to go.
        Joerg also claims that many modules which should appear as ranked in the spreadsheet do not look as such. Joerg should send a list of affected modules to cross check with Attilio.
        As to Bonn, Attilio's table show more modules than actually available in Bonn (after the meeting this was traced to some modules sent back to Bonn from Wuppertal and Siegen in order to repair MCC bonds).

        Maurice remarks some modules marked as "spares" in Attilio's ranking may be usable. He suggests that for modules between the loading threshold of 1000 and the "one chip dead" value of 2880, a case by case choice should be done.

      • other items
        Claudia asks to bond and return as soon as possible the modules she sent to Bonn for MCC bonding.
3. Inventory of hospitalized modules:
   A final list of barrel module problems in LBL have been received. It seems difficult they can be repaired, but it is agreed to send the three usable modules to CPPM and all the discarded one to Genova, in order to be stored/further debugged.
   Attilio's tried to go through the "not-ranked" entries in the ranking spreadsheet to disentangle "Hospitalized" modules, which may be not recovered, and real modules waiting for testing.
   Approximately he expects that 60 of the 160 not-ranked modules may actually be not usable modules.
   Update of wiki page
   No progress has been done in the description of the debugging procedure.
4. Updates on stave testing/ranking
   Markus would like to know if the next inserted stave will receive a quick test before being shipped to Bonn for the STAVE test.
   As to the source test, to reduce testing time in Bonn, the STAVE test is performed requiring 50k triggers. With respect to the standard 100k triggers scans, that results in about 20 more fake "dead" pixel below the passive components. This is a systematic effect in the ranking, but should not have any special implication, since these stave are already assigned to Layer2. It is therefore accepted to have this reduced test statistics, in order to keep STAVE testing in Bonn at the current rate.
   He also found two modules in a Genova stave with much less disconnected after the insertion than before. From the original data it looks like that in the stave test they got a smaller than usual statistics. After the meeting Claudia has suggested that this problem can be due to few slighlty noisy pixels which reduce the hit rate on the other ones.
5. Lab tour
   Bonn: stave test setup has been moved in order to improve the thermal performances.
   Dortmund: there are some problems on the module test setup, TPLL or TPCC hangs up during tests (mostly source test), improves a bit with cooling, but increased in the last week.
   Genova: they started to attach pigtails and recovering hospitalized flexes. New bare modules are needed.
   Siegen: Iskender reports on some problematic modules
   Module 511080 has a high leakage current, accompanied by localized discharge and noise (report). It is decided to assign a 3000 user penalty, so that this module will not be loaded in experiment.
   Module 511819 seems to have a too restricted FDAC range, therefore it cannot be tuned without setting manually the IF tuning. The affected chips had also problem at wafer testing 511819, but were below threshold to be flagged. Manual tuning would be problematic in ATLAS running. Discussion will proceed off-line in order to asses whether this module is acceptable or not.
   Finally there is one module for which the MCC FIFO test fails for FE 4, but ASSY test is fine. At a closer look it seems that bits 24 and 25 within the FE4 FIFO are stuck to 0. It may affect ToT measurement. It is decided to proceed with full test and check if this defect has any implication.
   
6. AOB.

   The RDB web server in Bonn is on again and all data have been restored!

   Due to 1^st May holiday, next meeting will be held on Monday 8^th May