10.26.10

Yesterday I finished the code to compute the differential slope measurements. I was able to get a decent looking poke matrix out of it after finding the right size perturbations. Of course, since the new WFS grid is only 9x9, the modal images are roughly a quarter the resolution, so some of them look more like crumpled paper without squinting. I'm probably bumping up against the spatial Nyquist with the higher modes, but I'm far to lazy to quantify that. I suspect that some of the higher DM61 modes are unobservable, although its uncertain what use those are anyway.

With that done I managed to get a stable integrator running, but of course since all the measurement are differential right now, there was basically nothing to compensate. Today I finished rewriting the code to generate disturbance commands matching the output from the state-space models, so the next step is to try out the LTI and adaptive loops (first in silico obviously).

I also tried using the pixel sub-sampling mode on the WFS, which halves the image resolution but can quadruple the frame rate. Sure enough, I was able to get nearly 450fps out of it using the current beam size; huge considering just a couple weeks ago I was struggling to get 20 fps. Obviously reducing the image resolution decreases the sensitivity of the WFS, but this might not be a problem for my purposes. The advantage of doing differential measurements is that I can basically use this mode with few code modifications. I would just have to define a new reference grid with this smaller resolution.

The next (hopefully final...please let it be final) hardware step is to get the camera triggering working so I can actually use these superfast frame rates. I've set the ball in motion on this, but it looks like its going to be pretty involved, requiring a new cam driver and modifications to the DM driver so I can use the existing NI board to generate the trigger. Having this running by the end of next week will be a tall order, but not outrageous if I can get the software in time.

Make it so.