Re: ATM Active Optics

["John A. Maxwell" <jmax@toad.net>]

Andy Saulietis <iss@pvtnetworks.net> wrote:
>   Antonio writes,
>   
>   > Has anybody challenged the idea of buildin  an active
>   >optics system. It's been going round my head for some time and I  don't 
>   >too ambitious project. Does  anybody  have information?      Benoit
>   >Schillings and  Brad Wallis page (developers  of SBIG AO7 system):
>   >http://voltaire.csun.edu/ao.html There is also another unit by Stellar
>   >Products  in: http://wwwstellarproducts.com     But both are omercial
>   >sites, so the have  just the pretty pictures  but no info :-(     Does
>   >anybody have further  information or knows of someone who has tried it?    
>   
>   Benoit's system is actually a rapid response tip-tilt mirror
>   which corrects for guiding errors and the average movement
>   of the seeing disk & thus improves the apparent resolution
>   of CCD/photo images as these are now limited only by the
>   size of the seeing disk.
>   
>   These systems are not actually 'active' optics in the sense usually
>   applied, in which a movable optical surface is used to modify the
>   wavefront to compensate for atmospheric aberrations, and actually
>   reduce the size of the star image by improving the wavefront. These
>   systems are very expensive & a long way from being within the reach of ATM's
>   currently, but some of the requirements are getting easier, such as
>   the availability of Ghz processors, and lasers. I've been toying
>   with ideas related to using a CCD camera to record a Harmann test
>   in ~ real time to control the figure of a large thin mirror. 
>   
>   

OK - I haven't even played with building thing one relating to
adaptive optics, but I have done a moderate amount of educated
fantasizing about it, and I've played with the math a bit.  (i.e. John
places his hat firmly over his mouth, so as to talk through it).

That said, some points:

-If you're at all serious about this, get your hands on a copy of JOSA
 March 1977. It's a special-topic issue on adaptive optics. Good Stuff.

-I'm not at all sure that a CCD camera and digital frame processor is
 the best way to go for this. The required spatial resolution for your
 wavefront sensors is quite low; basicly, the same resolution as your
 mirror deformation system. In other words, a 4x4 grid of mirror
 actuators only requires a 4x4 "pixel" wavefront error sensor. This
 being so, I'd be inclined to try photomultiplier tubes (if I could
 pick 'em up cheap) or good quality photodiodes (Burr-Brown makes some
 nice ones with integrated op-amps for first stage signal buffering)
 for sensors, and an analog computer to generate the actuator
 signals. As Andy notes, the required update rate is ~100Hz; this is
 cake for even cheap op-amps.

-If I did try a digital processor, I'd look at low-end DSP chips;
 there are some evaluation boards out there in the $100-$200 range
 that could probably handle the job. A DSP feels better matched to the
 task; only moderate amounts of RAM, but gobs of computational power.

-The primary headache, from what I can see, is the deformable mirror,
 with the wavefront sensors running a distant second. The deformable
 mirror is actually a moderately formidable construction project,
 depending on how you tackle the problem.

--
John A. Maxwell (jmax@toad.net)

Customer: I'm using Windows 98
Tech Support: Uh-huh
Customer: My computer isn't working
Tech Support: Yes, you said that