KLA-Tencor uses Mellanox Infiniband switches in our tools; the tools are installed in semi-conductor fabs around the world. Because our tools are installed in semi-conductor fabs, smoke resulting from electronic failures is huge no-no. Recently, a Mellanox Infiniband switch (model IS50XX) overheated in one of our tools, generating enough smoke that it was detected by our smoke detector and the tools was automatically shut down. A later inspection of the switch revealed that the power supply cooling fan had stopped rotating and was the likely root cause of the overheat (the power supply was actually still functional, but would overheat quickly on power up).
Has anyone else encountered such failures with these switches?
How difficult would it be for Mellanox to improve the thermal protection so that it riggers prior to smoke generation?