In my first year at the University of Washington, I was an undergraduate research assistant for the DAMIC
collaboration. There, I designed and built a 100 Watt linear (yes, linear) heater controller for keeping
the dark matter detectors at the right temperature during testing. The system was incredibly sensitive to
noise, so no switching components could be used.
Later, I build a ramp generator which would allow a user to program the slew rate of the bias lines that
powered the detector. The ramp generator had to consume absolutely no current when it was on or off, and
the bias voltage had to be able to change between 40 and 200V.