Quote:
Originally Posted by Shiraz
looks like a useful performance gain Peter.
Did a quick analysis of some of your unsaturated stars using Pixinsight (assuming the image data is not stretched). If I have the right pixel scale, the FWHM for the best fit Moffat PSFs went from about 6.2x5.5 arcsec down to 5.7x5.3 arcsec (ie AO mainly tidied up the star roundness). The peak signal went up by ~10%, as you found, so I guess no stretching.
But 5+ arcsec seems to be really bad seeing - I assumed 1.6 arcsec pixel scale, is that right?
|
Sorry, that's not correct. (I've since uploaded CCDinspector's FWHM analysis of over 130 stars in each)
The web (visible) star field data is 8 bit .jpg.
I took all measurements using raw 16 bit (flat/dark corrected) .fits files.
Symmetry certainly was improved a tad, but the raw intensity curve is quite different to the 8 bit data.
This is my take on the physics as to why it works:
Waves of air are always passing by the telescope aperture. Sometimes they are slow and smooth, fast but still smooth, slow but rough and lastly fast and rough.
If the dominant wavelength slowly moving and is more than the telescope aperture, AO tip-tilt systems can make a correction for the angle which the wave it is causing the image to shift (up/down/left/right). It is here where amateur AO works.
Nights of bad seeing make the frequency of the wave increases.
Add in higher order harmonics to make things worse.
Amateur AO (read tip/tilt) is unlikely to have either the correction rate, or will likely follow a higher-order harmonic of the wave in the wrong direction.
But on nights where the atmosphere is rolling along at a smooth 5-10Hz, as was likely on the night I grabbed the data here, both resolution and stellar intensities show significant improvement with AO than without.