Contamination examples CMS borrowed modules CMS modules showing

Contamination examples – CMS borrowed modules • CMS modules showing decreased RSSI current signal as well. • Various hypothesis, two modules affected by the issue both show very dirty end-plates where the fiber connect. VTRx 200214 (MM) from CMS HCal VTRx 200199 (SM) from CMS HCal ALICE ITS UPGRADE 1

Contamination examples – ITS observed same behavior of CMS • Contamination seems very sensitive to temperature (could move over time), especially around 38 – 40 C (only indicative values from early testing). • Spatial orientation of the modules MAY play a role VTRx 301350 (SM) lens BEFORE heating VTRx 301350 (SM) lens AFTER heating at 40° ALICE ITS UPGRADE 2

ITS considerations about cooling • Patching a thermal contact between the VTRx modules and the cold plate on the ITS Rus is difficult: the boards are complex and extremely thigh in mechanical tolerance. • Adding a patch will require REMOVING the cold plate, and adding an extra pad, likely bolted to the mounting holes of the cold plate itself. While doable, this will pose at risk the components beneath, which are connected to the cold plate by (very sticky) thermal pads. • Multi-pin components (1500+ pins for the main FPGA only) are especially susceptible to bending of the PCB • Each board should be re-tested before installation • Conclusion: we would resort to such an approach only as very last solution. ALICE ITS UPGRADE 3

ITS considerations about cooling • Plus, our cold plate is a very cheap one: engineers are not truly convinced on how effectively we may drain power. ALICE ITS UPGRADE 4

ITS considerations about cooling 0. 5 mm plastic + glue • We are running our own assessment of the cooling effectiveness we may reach, however: at the end, even with perfect (i. e. zero impedance) thermal path, the impedance of the ROSA assembly remains • Before any further consideration, we therefore need a design from EP/ESE team about a possible standard solution (identical for all experiments) and its thermal properties ALICE ITS UPGRADE 5

ITS current stand • Rework of the 200+ RUs considered so far too risky, and with doubtful benefit as long as the ROSA stays unchanged. • TPC measurements did confirm that most of the affected MM VTRx drop the optical signal down to about 50%, i. e. they keep the link on. Disturbing, but the experiment would survive. • For the remaining 5% of lost links, we can substitute the VTRx one by one. • Our real issue are the SM modules used for the trigger tree, where we expect a 50% link loss due to reduced power margin. • The preferred solution for this case is increasing the optical power via an external system, likely placed in the CTP bunker. We foresee a 4 x maximal increase, which will require so re-wiring at PP 1, considered much less risky than touching the RUs. • Considered the latest findings, a “cleaning” campaign of the VTRx modules, if possible, would nicely complement our present approach. ALICE ITS UPGRADE 6