February, 2012

These notes describe the software in use at the MMT for wavefront sensing. The software was originally developed by Steve West, then was worked on heavily by Tim Pickering. It has most recently been worked on by Tom Trebisky. The WFS software began supporting only f/9, support for f/5 was added later. There were two variants of the wfs software (with a huge amount of shared code), the two software packages were unified, then support for MMIRS was added. F/15 is an entirely different beast, outside the scope of this discussion.

Things to do

It would be nice to improve integration between the wfs and the cell (primary support) computer when sending force sets to the cell. At this time it is pretty much a russian roulette situation - a force set gets send and it is either applied or rejected wholesale. One improvement would be for the cell to apply whatever portion of the correction it could and indicate in some way to the WFS software what it did (I applied all of what you asked for, none of it, or some percentage). Typically there is one (or a few) actuators that limit what can be done.

Better than this would be to allow wfs software to interrogate the cell and learn what the current force levels and limits are and make its own decisions about whether to ask for a full correction or to scale down what it asks for. The current safety checking and limits in the cell would be retained of course, but this would allow the wfs to be a lot smarter about what it does and to entirely avoid pushing the primary into limit situations.

Tim and Tom have talked for a long time about communicating directly to the secondary via the mmt remote protocol rather than via telserver. Using telserver and MSG was very tempting and easy in the early days and a good way to get on the air. Now it sometimes becomes a point of failure. Communicating directly would be, well, ... more direct.