active vs adaptive optics

Don Bruns wrote:

The SBIG AO-7 is a good example of an active optics system, but some
people might also consider it an adaptive optics system. In use, it
might correct a small amount of atmospheric "tilt", but I believe most
of the improvement comes from correcting slower drive error.

Adaptive optics is useful only over a very small field of view -
typically a few arcseconds in the visible spectral region, for full
adaptive optics, and up to an arcminute for tip-tilt only. Since most
users of the SBIG AO-7 use a much wider field of view, true adaptive
optics would actually degrade the image more than an arcminute from the
guide star. Since the real improvement seen with this device is evident
over the entire image, the improvement is due to telescope drive errors,
not atmospheric turbulence. With an update rate of 10 samples per
second, the bandwidth is close to 1 Hz; this is too slow for most
atmospheric turbulence correction.

This depends on your definition of "active" vs "adaptive" - I tend to
use the more restrictive definition of adaptive. A fast visible
tip-tilt system is a true adaptive optics system, if it can keep up with
the atmosphere and is used only over a field of view less than an
arcminute. The Stellar Products AO-2 did exactly this.

Don Bruns

Hi Don,

Thanks for sharing your views. To get another experienced view I
decided to break out the "bible" of AO -- Hardy's "Adaptive Optics for
Astronomical Telescopes". I couldn't remember if he defined the terms,
"adaptive" and "active", separately or not. Here are some things he has
to say that seemed relevant to the question:

In Chapter 2, "Adaptive Optics in Astronomy", which is largely
descriptive and introductory following Chapter 1, "The Short, Eventful
History of Adaptive Optics":

"The function of adaptive optics in astronomy is to remove aberrations
from the optical path between a celestial object and the imaging
device."

"Adaptive optics is capable of improving the performance of most
optical instruments used in astronomy, including spectrographs and
interferometers as well as imaging devices. It is not limited to
compensating atmospheric turbulence, although this is usually the most
serious problem in ground-based observations. Adaptive control can be
used to correct figure errors in the primary mirror as well as optical
errors arising from any random disturbance, ranging from slow variations
in temperature and the gravity vector, which may have a time scale of
hours, up to the rapidly moving turbulence eddies transported by
high-velocity winds, producing bandwidth that approaches 1000 Hz.
The atmospheric disturbances that affect astronomical telescopes
have characteristic frequencies above 1 Hz. Although, in principle,
adaptive optics will work down to the lowest temporal frequencies, it
has proved convenient, in practice, to compensate the lower frequency
disturbances due to temperature and gravity by a separate control system
known as "active optics." The primary mirror is potentially the largest
source of low-frequency wavefront errors in astronomical telescope; the
main task of active optics it to control its optical figure, eliminating
large, low-frequency phase errors. This allows the adaptive optics
system to employ a small deformable mirror to achieve the high temporal
bandwidth necessary to compensate atmospheric turbulence. Another
difference is that active optics systems usually employ a local
reference (which may be optical or mechanical) to control the primary
mirror, while adaptive optics systems must, of necessity employ distant
reference sources to measure the wavefront. Some active optics
installations use a natural star to calibrate the control system."

Although Hardy doesn't raise the same point that you do about the small
field of view of adaptive optics but in a later chapter he does describe
the usefulness of tilt correction according to a telescope's D/r0.

In Chapter 6, "Wavefront Correctors":

Section 6.10, "Tracking mirrors":
"Compensation of overall wavefront tilt, or the angle of arrival,
is the simplest type of adaptive optics, and it has been used in
astronomical telescopes since the 1950s." He refers to Babcock's 1956
paper.

In Section 6.11, "Dirigible Optics":
"Dirigible optics are rigid optical components that are steered or
moved in real time to compensate wavefront distortions. Their wavefront
correction capability is generally limited to low-order modes, such as
tilt, defocus, astigmatism, and coma. A tip-tilt mirror may be
considered an elementary device of this type, but the term is intended
to describe powered components, both lenses and mirrors, that may be
physically moved in five of their six possible degrees of freedom."

He goes on to describe components with optical power and other
corrections that can be obtained with them, such as defocus correction
through axial displacement. He doesn't specifically call these "active"
systems but it seems from what I quoted above that they fit his
definition, primarily because these corrections are slower (as with your
point about the bandwidth capabilities of slow tip-tilt mirrors like the
AO-7).

I guess I don't quite see a "definitive" answer here, either. But it
seems to depend on what characteristics you're referring to --
bandwidth, field of view, etc. It seems like there are different
answers to what defines "adaptive" and "active" control systems
depending on the characteristics you're talking about. They're fuzzy
definitions. Does that seem right to you or am I just confused?

I always feel I'm getting educated when I read Hardy. And I'm always in
a bit of a daze afterward. It's like Star Trek when the ship's
computers are being fed by an alien vessel at a higher speed than the
primitive Enterprise computer can tolerate. So if you have any
clarification please let me know. I've got to go cool my brain off.

Mike Simmons

Mike Simmons wrote:

Don Bruns wrote:
...
This depends on your definition of "active" vs "adaptive" - I tend to
use the more restrictive definition of adaptive. A fast visible
tip-tilt system is a true adaptive optics system, if it can keep up with
the atmosphere and is used only over a field of view less than an
arcminute. The Stellar Products AO-2 did exactly this.

....
Thanks for sharing your views. To get another experienced view I
decided to break out the "bible" of AO -- Hardy's "Adaptive Optics for
Astronomical Telescopes". I couldn't remember if he defined the terms,
"adaptive" and "active", separately or not. Here are some things he has
to say that seemed relevant to the question:
....
I guess I don't quite see a "definitive" answer here, either. But it
seems to depend on what characteristics you're referring to --
bandwidth, field of view, etc. It seems like there are different
answers to what defines "adaptive" and "active" control systems
depending on the characteristics you're talking about. They're fuzzy
definitions. Does that seem right to you or am I just confused?

Hi Mike,

Your quotes from Hardy are right on target. Science moves forward, so
definitions also change. I remember reading an early paper (maybe
Babcock's) that mentioned the distinct possibilty of taking 5 minute
film exposures of Mars (any more would cause rotational blurring) with
adaptive optics, and 2 hour (!) exposures of Saturn, to get really good
signal-to-noise ratios on the finest grain film! Things have changed a
lot since the 1950's...

Don

Don Bruns wrote:

Your quotes from Hardy are right on target. Science moves forward, so
definitions also change. I remember reading an early paper (maybe
Babcock's) that mentioned the distinct possibilty of taking 5 minute
film exposures of Mars (any more would cause rotational blurring) with
adaptive optics, and 2 hour (!) exposures of Saturn, to get really good
signal-to-noise ratios on the finest grain film! Things have changed a
lot since the 1950's...

Ha! Earlier at the very same observatory, colleagues of Babcock's
father Harold (who was also a solar astronomer at Mount Wilson) were
taking multiple night exposures just to get usable spectra with the
world's largest telescopes. Exposures of 30 or 40 hours in some cases.
This is with blue-sensitive (though "sensitive" seems a misnomer now)
glass plates and red science objects.

But I've noticed that with the current techniques of combining multiple
images to reduce the signal-to-noise ratio, total exposures are again
creeping up. The best images again have hours of total exposure times
in RGB and L. Hours with CCDs! Maybe there's no real shortcut to the
"best" results a technology has to offer after all. :-)

Mike Simmons

On Fri, 14 Nov 2003 15:05:07 GMT, Don Bruns wrote:

The Stellar Products AO-2 had a bandwidth over 30 Hz, so it did correct
image motion due to atmospheric turbulence. For Mars this summer, I
used a ToUcam with short exposures, so I did not need to use image
stabilization. Since stacking the images is so simple, that is
equivalent to a longer exposure.

Further, image stacking even simulates higher order correction than simple
tip/tilt, since distorted images are removed. It's a great approach if you can
afford the total loss of signal (not an issue with bright planets).

_________________________________________________

Chris L Peterson
Cloudbait Observatory
http://www.cloudbait.com