Running Computers Cold

bath of liquid nitrogen. :-) Of course they were (mostly) designed for that. The CMOS logic we used ran with max clock cycles of about 14nS at room temp but could easily go down to 5nS when in the LN2. Some interesting things were learned (about physics) when signals crossed the boundary between the LN2 and air... as the memory subsystem was not in the LN2. There were significant impedance changes in the tiny coax lines that ran between the two environments... so we ended up with signal integrity problems at that boundary that nobody had anticipated. Was fun stuff. Another thing to consider if you bring a machine into the warm from a cold environment is condensation. The thermal expansion that Ethan pointed out is important but you will also often see water condensing on surfaces and components as the thing warms up. You definitely want to wait for that moisture to evaporate before you apply power. Chris -- Chris Elmquist