Gleb wrote:
I consider designing a low-cost AO system for a small telescope. It
seems we have all necessary resources: we produce deformable mirrors and
wavefront sensors, we also produce closed-loop AO systems with up to 59
channels, so integration wouldn't be too complicated as all structural parts
are available.

On the other hand we have little experience with astronomy and therefore any
advice would help.

The preliminary tech requirements:

1. To be mounted in 1.25 inch ocular socket (2" socket??). To be used with
telescopes with diameter in the range 25cm to 1m.

2. Aberration free afocal mirror system, transparent in the visible and
near IR and fully operational even with AO switched off. To achieve this,
we'll use a system with a field of a couple of mm (in the primary focus)
for a foacl ratio of 1/10. The field and F# are compromized to reduce the
complexity of the optics, but the field will be limited anyway by the
anisoplanatism of the AO and the F# must be small for a HR imaging

3. To have al least 19 degrees of freedom (depending on the seeing can be
good to correct up to ~13 Zernike terms to about 10% of the uncorrected
value). 37 degrees of freedom is also possible but I'm not sure a small
scope really will collect enough light to correct that many terms in real
time.

4. To operate on a natural star with magnitude of at least 4 (with a 25cm
telescope), using 50% of light for running the AO and 50% for registration.

5. To be easy in setting up and running. To use single +12V power supply
and three cables connecting the system with the deformable mirror
controller and the dedicated control laptop PC.

6. The total weight of the optical correction unit mounted to the telescope
not to exceed 1kg. Mirror controller incl power supply - also 1 kg, add some
extra for cables and laptop.


The system is supposed to provide a diffraction-limited imaging in a rather
bad seeing conditions. It will allow stable imaging of bright objects such
as stars, double stars and planets. Another advantage of using such a
system is that it will correct the aberrations of the telescope, improving
the quality of optics, for instance making the period of mirror cooling also
available for observations. In fact, correction of the static aberrations
can be done on a bright star once, and then the system can be used in
static correction mode.

The project is technically feasible (although quite expensive in its
development stage), but I still have my doubts regarding its usefulness:

1. Small field and ability to work on only bright objects will limit the
usability to very bright double stars and planets. Are (amateur) astronomers
really interested in this?

2. Although we plan to have it transparent, the system will limit the field
of view and reduce the amount of light available for observation. The light
loss will be compensated by the resolution gain, but the effect can be
limited or even negligible for a small telescope.

3. The system will require an additional laptop computer to run the AO and
will add to the complexity of the telescope setup.

4. It can be quite expensive, especially in the beginning, though if there
is a market, the price can be very acceptable.

I would appreciate any comments on the above mentioned topics.

It might sell to some wealthy amateur astronomers for visual use but
more likely to university level spectroscopists with intermediate size
instruments (1-2 meter) where image size limits spectral resolution. It
might also sell to target shooters (and perhaps military snipers) trying
to resolve targets at kilometer ranges through sun warmed atmosphere.
That has an order of magnitude more seeing distortion than looking
vertically at night but plenty of light would be available. Does your
software require a point source reference or can it optimize on the
structure of a random field?

The unit you propose would only be useful if the corrections are fast
enough to keep up with typical atmospheric scintillation. Static
corrections alone would probably be cheaper to correct by buying better
optics. Doing both would be of advantage.

Unless you give a dollar value "quite expensive" doesn't mean much. At
$1k (unit and software, user supplied computer) they'd sell to the
masses. (I'd buy one just to play with). At $10k some would sell. At
$100k few if any who would want it would have the money.