Dead koki Motherboard

On 17 October 2000 the ESI control computer koki died. Apparently the CPUs on the motherboards are known to have a heat-dissipation problem, and the koki CPU destroyed itself. Koki is an Ultra 140, and the closest CPU we had to this was an Ultra 170, which was pressed into service by Jon Chock and Bob Kibrick. Experimenting with different configurations, one was created that allowed operation of ESI that night with only moderate loss in capability.

The links below are e-mails that describe the process of turning the Ultra 170 into a replacement for the dead koki:

• Bob Kibrick describes the main steps involved in converting the Ultra 170.
• A follow-up by Bob Kibrick describing possible alternatives to try and get the replacement koki working more as the original koki was.
• Jon Chock describes how to get koki to use kanaha as its ypserver.
• Kevin Tsubota describes changes to TDC and AUX to get DCS communications with the ESI computers.

An entire observing run ran with this configuration, with one major problem. The timing of the CCD readout, disk writing, etc. was not normal, with an extra pause at the end of each exposure which eventually reached a hundred seconds or more.

After the run Jon Chock replaced the original koki motherboard and converted the repaired koki back to normal operations.

• Jon Chock decribes reverting to the normal operation after the koki repair.