The new mirror support cell air compressor is now fully installed and operational. It is set to provide 12 inches of water (0.43 psi), a little higher than the nominal 10 inches of water (0.36 psi), and appears to be holding well. There is a refrigeration desiccant system on the compressor (to keep moisture out of the lines) that is not currently installed. It should be installed in the next few weeks, as soon as the proper fitting can be installed.
The problem with the dome encoder does not appear to be with the encoder itself, rather with the encoder input channel to the pmac. The pmac has 2 cards with 16 possible channels, of which 10 (or so) are in use by the various encoders. One card has only two encoders, the focus and the dome azimuth, and it appears that the input channel for the dome azimuth (#6) is corrupted. Brevin was able to remap that channel to an open channel (#8), which appears to have fixed the problem. The dome is now making repeatable rotations to the same commanded azimuth positions, without significant failure.
However, it now appears that the dome is no longer synced with telescope. The telescope points correctly, but the dome azimuth is significantly off, by as much as 90 degrees depending on location. This is repeatable for the same telescope positions, indicating that it is not a loss in encoder counts. Further, its correctable by commanding the dome to the telescope azimuth. This, unfortunately, does not provide a fix as tracking forces the dome to update its position, immediately rotating back to its previously determined, incorrect position.
It is not clear how the dome receives its commands, from the telescope azimuths or from its own pointing model, but in either case something is not responding correctly. I’m in discussions with Treffers to see what can be done.