Requirements to Do Interferometry

When doing astronomical imaging via interferometry, is it necessary to know
the distance between the apertures with absolute precision, thus making
it impossible to build a portable system that would be assembled in the
field.

Or, with visual interferometry, say, would it be a simple matter of
projecting the various images onto the same spot?

TIA.

John Schutkeker wrote in message . ..
When doing astronomical imaging via interferometry, is it necessary to know
the distance between the apertures with absolute precision,

"Absolute precision" is meaningless. It is necessary to know the
distance to very high precision. I've heard a quarter of a wavelength
as a stated requirement. Visible light is on the order of 0.5 micrometer.
Imagine one quarter of that.

thus making
it impossible to build a portable system that would be assembled in the
field.

NASA and the European Space Agency both have deep-space interferometers
on the drawing board. That means fleets of spacecraft launched
far out into interplanetary space and calibrating themselves to
the necessary precision.

I think most of those missions are in the infrared, wavelengths
on the order of 1-2 micrometer.


Or, with visual interferometry, say, would it be a simple matter of
projecting the various images onto the same spot?

I think what you're asking is "is it possible to do some sort of
autofocus procedure to calibrate your distances to the necessary
precision". I think the answer is yes. At least when I was reading
a radioastronomy textbook recently there was a section on autofocusing.

The more bright references you have to calibrate on, the better
a job you can do. A portable system might use not only stars but
some sort of beacon from home base.

You could also use some sort of laser ranging scheme to measure
your telescope distances relative to each other.

- Randy

In message , John Schutkeker
writes

When doing astronomical imaging via interferometry, is it necessary to know
the distance between the apertures with absolute precision, thus making
it impossible to build a portable system that would be assembled in the
field.

You need to be able to hold the path lengths equal to within a fraction
of the wavelength of light if you are hoping to see visibility fringes.
Imaging via interferometry is a tricky business as you only measure one
Fourier component of the sky brightness distribution at a time with each
pair of mirrors.

Michelson was the first person to use a 20' rigid beam with mirrors on
to do stellar interferometry using the Wilson 100" scope for the
combining step. Good enough to infer diameters for several red giants,
but requiring an experimental genius to make it work.

Or, with visual interferometry, say, would it be a simple matter of
projecting the various images onto the same spot?

It is way more complex. Take a look at COAST or one of the other optical
interferometer designs - it is pretty much right on the limit of
practicality even on a permanent optical bench in a temperature
controlled bunker.

Regards,
--
Martin Brown

Martin Brown wrote in
:

Imaging via interferometry is a tricky business as you only measure one
Fourier component of the sky brightness distribution at a time with each
pair of mirrors.


I can think of two ways to handle this task.

I could use a computer program to make FFT's of the image signal and
combine them according to whatever algorithm is traditionally used. I'm
not an experimental genius, but I am an expert scientific programmer, and I
can code any algorithm for which I have a properly written spec.

And if I can't do it with FFT's, I'd need to buy a tunable detector and
build a circuit to step through the Fourier components before feeding them
to the algorithm. That doesn't sound too hard, either.

So far it doesn't sound too bad. Am I missing something, or am I being too
optimistic in assuming that I can just open a textbook to get the
algorithm? Thompson's "Interferometry and Synthesis in Radio Astronomy,"
seems like a good place to start, although $110 is a lot more than I like
to pay for a textbook.

Do you think that it might be possible to get the synthesis software that
somebody else has already written?

"John Schutkeker" wrote in message
...
Martin Brown wrote in
:

Imaging via interferometry is a tricky business as you only measure one
Fourier component of the sky brightness distribution at a time with each
pair of mirrors.


I can think of two ways to handle this task.

I could use a computer program to make FFT's of the image signal and
combine them according to whatever algorithm is traditionally used. I'm
not an experimental genius, but I am an expert scientific programmer, and
I
can code any algorithm for which I have a properly written spec.

And if I can't do it with FFT's, I'd need to buy a tunable detector and
build a circuit to step through the Fourier components before feeding them
to the algorithm. That doesn't sound too hard, either.

So far it doesn't sound too bad. Am I missing something, or am I being
too
optimistic in assuming that I can just open a textbook to get the
algorithm? Thompson's "Interferometry and Synthesis in Radio Astronomy,"
seems like a good place to start, although $110 is a lot more than I like
to pay for a textbook.

Do you think that it might be possible to get the synthesis software that
somebody else has already written?

Useful info from a recent workshop on interferometry but heavy reading at
http://www.atnf.csiro.au/whats_on/wo...2003/prog.html

The software is readily available to synthesise images. I have only used
MIRIAD but others exist and all are free.
http://www.atnf.csiro.au/computing/software/


--
Terry B
Moree
Australia

"MB" == Martin Brown writes:

MB In message , John
MB Schutkeker writes

When doing astronomical imaging via interferometry, is it necessary
to know the distance between the apertures with absolute precision,
thus making it impossible to build a portable system that would be
assembled in the field.

MB You need to be able to hold the path lengths equal to within a
MB fraction of the wavelength of light if you are hoping to see
MB visibility fringes. Imaging via interferometry is a tricky
MB business as you only measure one Fourier component of the sky
MB brightness distribution at a time with each pair of mirrors.

I wouldn't say that the imaging per se is difficult. As you say, you
measure a bunch of Fourier components, transform them, and voila,
that's the image. The tricky part is calibration: Making sure that
your visibility measurements are not corrupted by any number of
effects such as atmospheric propagation, time keeping, etc.

--
Lt. Lazio, HTML police | e-mail:
No means no, stop rape. | http://patriot.net/%7Ejlazio/
sci.astro FAQ at http://sciastro.astronomy.net/sci.astro.html

"JS" == John Schutkeker writes:

JS When doing astronomical imaging via interferometry, is it
JS necessary to know the distance between the apertures with absolute
JS precision, thus making it impossible to build a portable system
JS that would be assembled in the field.

JS Or, with visual interferometry, say, would it be a simple matter
JS of projecting the various images onto the same spot?

You don't quite say explicitly that you are interested in optical/IR
interferometery. In radio astronomy, there are a number of
transportable interferometers, in which the antennas are moved between
a limited number of predetermined spots. (Notable such
interferometers are the Westerbork Synthesis Radio Telescope, the
Australia Telescope National Facility, and the Very Large Array.)

At least in the case of the VLA, after moving the antennas, *baseline
calibration* is done. The antennas are moved to predetermined
locations, but the antenna never really returns to exactly the same
spot. Baseline calibration is the process of correcting for the
difference between the assumed and actual antenna position. One
observes a series of sources whose positions are considered known,
measures the responses from the sources, and then derives a series of
corrections to the assumed antenna positions.

I'm not as familiar with the case of optical/IR interferometry. I
would think that the same thing should be possible, though of course
the tolerances are much tighter.

--
Lt. Lazio, HTML police | e-mail:
No means no, stop rape. | http://patriot.net/%7Ejlazio/
sci.astro FAQ at http://sciastro.astronomy.net/sci.astro.html

Joseph Lazio wrote in
:

You don't quite say explicitly that you are interested in optical/IR
interferometery.

Both.

In radio astronomy, there are a number of
transportable interferometers, in which the antennas are moved between
a limited number of predetermined spots. (Notable such
interferometers are the Westerbork Synthesis Radio Telescope, the
Australia Telescope National Facility, and the Very Large Array.)

I'm designing an entire system, on a small scale, using commercially
available radio dishes, that is transportable to *any* location of the
users choice, via, say, a box truck or a few large pickups.

So each dish would be much, much smaller that the dishes used in, say, VLA,
as would the spacings. But cost would fall by a quantum increment, and
user convenience would increase similarly. All this should make the system
valuable to astronomers.

"JS" == John Schutkeker writes:

JS Joseph Lazio wrote in
JS :

In radio astronomy, there are a number of transportable
interferometers, in which the antennas are moved between a limited
number of predetermined spots. (Notable such interferometers are
the Westerbork Synthesis Radio Telescope, the Australia Telescope
National Facility, and the Very Large Array.)

JS I'm designing an entire system, on a small scale, using
JS commercially available radio dishes, that is transportable to
JS *any* location of the users choice, via, say, a box truck or a few
JS large pickups.

JS So each dish would be much, much smaller that the dishes used in,
JS say, VLA, as would the spacings. But cost would fall by a quantum
JS increment, and user convenience would increase similarly. All
JS this should make the system valuable to astronomers.

At the risk of sounding snobbish, I'm doubtful. Note that the
interferometers I've named above have 25-m antennas. You don't say
what size antennas you have in mind, but it takes a lot of smaller
antennas to equal the collecting area of a 25-m antenna, and radio
astronomy is all about collecting area.

I also see a couple of other difficulties. If you do use smaller
dishes, you may decrease the cost of the antennas, but you push a lot
of cost into the back-end processing. Also, it's not obvious to me
why a transportable system makes sense for astronomers. I'd like my
systems to be far away from humans and stay far away (because other
equipment used by people tends to generate lots of interference).

For reference, the US concept for the Square Kilometer Array (SKA)
involves 12-m antennas. In order to get the required collecting area,
4500 such antennas are required. There is no serious discussion of
making the antennas movable, and there are serious concerns about
whether the back-end processing is possible, even in 2020.

The Allen Telescope Array involves 6-m antennas. In order to get the
required collecting area, the eventual goal is to have 350 antennas.
Again, these are not going to be movable.

--
Lt. Lazio, HTML police | e-mail:
No means no, stop rape. | http://patriot.net/%7Ejlazio/
sci.astro FAQ at http://sciastro.astronomy.net/sci.astro.html

In message , John Schutkeker
writes
Martin Brown wrote in
:

Imaging via interferometry is a tricky business as you only measure one
Fourier component of the sky brightness distribution at a time with each
pair of mirrors.


I can think of two ways to handle this task.

You misunderstand the point that I was trying to make. A single Fourier
component is not a whole lot of information about the sky brightness
distribution. You need many days observations made at different baseline
lengths to be able to build a synthesis map.

If you had said you were an experimental genius optical bench engineer
with a decades of experience of making large holograms you might just
stand a chance of seeing meaningful fringes. Doing optical
interferometry for aperture synthesis is extremely difficult - so hard
that only a handful of groups world wide have managed it.

OTOH Speckle interferometry is probably tractable but AFAIK no amateur
group has yet succeeded in using that (much easier) technique.

I could use a computer program to make FFT's of the image signal and
combine them according to whatever algorithm is traditionally used. I'm
not an experimental genius, but I am an expert scientific programmer, and I
can code any algorithm for which I have a properly written spec.

The FFT is the least of your problems. Plenty of packages exist to
reduce radio astronomy data and make synthesis maps.

And the amount of useful data you are likely to generate in optical
interferometry even if you did by some miracle get fringes is so small
that ordinary slow Fourier transforms would be fast enough. I am not
sure what the largest optical aperture synthesis image is at present,
but I would be surprised if it contained more than about 1000
independent pixels.

To get anything like a decent phase observable in the optical you must
have 3 separate mirrors and simultaneously combine light from all
possible pairs to get their interference fringes (it gets harder with
each antenna you add).

So far it doesn't sound too bad. Am I missing something, or am I being too
optimistic in assuming that I can just open a textbook to get the
algorithm?

The algorithms are pretty trivial (though there are plenty of nasty
experimental gotchas that every practical implementation has had to deal
with)

Thompson's "Interferometry and Synthesis in Radio Astronomy,"
seems like a good place to start, although $110 is a lot more than I like
to pay for a textbook.

Kraus, Radio Astronomy 2nd edition (paperback) is cheaper and might be
an easier introduction. As are the IAU symposia on Indirect Imaging.

If you are seriously going to try and do this you will need
Thompson,Moran&Swenson and also whatever new stuff has been published on
experimental optical interferometry techniques. You may be able to find
some of the latter on ADS abstracts.

Do you think that it might be possible to get the synthesis software that
somebody else has already written?

Or play with the online toy aperture synthesis demo in Oz. That will
give you a feel for what sort of "images" a toy interferometer setup
will produce.

Regards,
--
Martin Brown