Site Qualification of SPI

2004/1/14

Five SQ modules were assembled: first 4 modules with <100> sensors and the last with <111>. The last one with <111> was to make sure that nothing was out-of-control with the series production sensors. For the <100> sensors, the z profile of <100>, available from the barrel module page, was applied. The QA was made for the steps: IN - initial after the assembly, TC - after thermal cycling, LT - after long-term test (warm), and LTL - after long-term test (cold). The spreadsheet summaries in below have those sheet of IN, LT, etc. A brief summary of the results are

3D: Critical parameters such as midyf and loCoolingFacing_b were well within the tolerances of GOOD. Only some of OptimalMaxZ parameters were out of GOOD but within PASS limits in the modules with <100> sensors.

summary

In series modules with <111> sensors, the OptimalMaxZ parameters are very much within 50 um, as same as used to be.

series modules

IV: Leakage current behaviour was as good as one we can hope for, in IN and LT. The larger currents of <100> modules were due to the larger currents of the <100> sensors.

summary

Dead channel: ASIC's being used were those with 1 ana_bad channels for the first 4 modules, and an old ASIC-hybrid (registered after repairing mechanically damaged wire-bondings) for the last one.

579 - 1 low and 1 high gain channels appeared, seemed intrinsic in ASIC's
580 - no new channels appeared
581 - no new channels appeared
582 - 12 low gain channels appeared, seemed intrinsic in ASIC's
583 - 13 channels were the shorts in the sensors, 1 noisy channel appeared

One noisy channel in the module 583 could not be identified due to the ASIC's. Therefore, this was the only channel out of 1536*5 = 7680 wire-bonds which might be damaged slightly by wire-bonding; a rate of 0.013%.

summary

In series modules with perfect chips, there is basically no dead channels as expected.

An example of RC plot

S-curves: Noise occupancy was measured after LT both in warm and cold. The plots below are of the cold. Generally, the s-curve behaviour was good. There was no severe wiggles, i.e., wiggles in both positive and negative thresholds in the link1.

Assembly sheets