I retried a new mask based on the nice web generator and it worked MUCH better than the one I'd designed myself in Photoshop. Even so at f/2.8 its very difficult to discern best focus - but much better than any Hartman style mask for sure. I use the mask to get close and then run the real-time CCDInspector curvature tool to determine when I've optimized FWHM across the entire frame.
What I found interesting was how the size of the pattern appears to shrink with wavelength. I don't understand the physics behind this phenomena but it was a surprise. The difference between my R and B filters was significant.
I wrote a small Maxim script to continuously grab frames and run an unsharp mask on the frame which aids centering as well. I also found using a line profile tool across the spikes let me determine with a high degree of accruracy when I'd centered the line.
Now I need to rig something up to swing the mask over the end of the scope remotely so I can stay in my office during the winter and still focus using this tool. I was thinking of using a gear retract servo for an RC plane which flips 180 degrees between two positions.
Previous
Index
Next
Flat
Re: Revolutionary new way of focusing! No less!
Edit
Reply
Quote
