We have been testing the new lamps. The changes are mostly configuration of the photodiode signal to require before turning off the HV, and of the fan level to set after that. It looks like we will settle on a warmup period close to the hardware limit of three minutes – that nearly doubles the measured Cd flux over two minutes. Still working on the fan level, but we suspect that will need to be chosen based on what we see on the detectors with the lamps on the PFI. Might need to be adaptive.

The best performance as tested off the telescope comes from warming up to a threshold of Cd=0.6 (in internal units), which takes the full 180s before the hardware cutoff. Then running the fan at 15%. Warmups take about the same amount of time if done again after powering the lamp down, so operationally I'm guessing that shutter-controlled illumination is the best choice.

If we need to make the Cd lines brighter it should be possible by adding active fan control. As it stands if the fan is run too slowly (~10%) the flux increases uncomfortably after a few minutes.

Tagged 1.3.2