electronic telescope primary mirror

[Jamie Morken]

Hi all,

Would it be possible to substitute a 2d CCD array for the 3d parabola
mirror in a telescope if each CCD pixel only grabbed light with a phase
delay proportional to its "virtual" position on a parabolic mirror?


__ __
''-_ __-''
'-__ _--' virtual primary mirror
''-_ _--''
'''
--------------------------------- CCD array

(created by AACircuit v1.28.4 beta 13/12/04 www.tech-chat.de)

So if a pixel in the CCD array was 1meter away vertically from the
virtual mirror, you would delay sampling the pixel about 3 nanoseconds
(approximate time it takes light to travel 1meter) compared to the pixel
at the root of the parabola, and you would sample the pixels at a rate
proportional to the frequency of the light you are receiving. I think
the light would also have to be polarized so that it is only travelling
vertically to the CCD. Also the CCD array would have to be custom in
that each pixel would have to be able to be triggered seperately and
very quickly.

I have always wanted to be able to build a telescope using just a planar
sheet of material and avoid having to have a precision mirror so this is
just a wild idea about that!

cheers,
Jamie

 

[Don Pearce]

Hi all,

Would it be possible to substitute a 2d CCD array for the 3d parabola
mirror in a telescope if each CCD pixel only grabbed light with a phase
delay proportional to its "virtual" position on a parabolic mirror?


__ __
''-_ __-''
'-__ _--' virtual primary mirror
''-_ _--''
'''
--------------------------------- CCD array

(created by AACircuit v1.28.4 beta 13/12/04 www.tech-chat.de)

So if a pixel in the CCD array was 1meter away vertically from the
virtual mirror, you would delay sampling the pixel about 3 nanoseconds
(approximate time it takes light to travel 1meter) compared to the pixel
at the root of the parabola, and you would sample the pixels at a rate
proportional to the frequency of the light you are receiving. I think
the light would also have to be polarized so that it is only travelling
vertically to the CCD. Also the CCD array would have to be custom in
that each pixel would have to be able to be triggered seperately and
very quickly.

I have always wanted to be able to build a telescope using just a planar
sheet of material and avoid having to have a precision mirror so this is
just a wild idea about that!

cheers,
Jamie

This would work if the CCD pixels responded to the actual waveform of
the light - giving an AC signal out at however many Teraherz light
goes at. Unfortunately they don't - they integrate and give an output
simply proportional to intensity, so phase information is lost. A nice
idea, but a non-starter, I'm afraid.

d

Pearce Consulting
http://www.pearce.uk.com

 

[Rene Tschaggelar]

Hi all,

Would it be possible to substitute a 2d CCD array for the 3d parabola
mirror in a telescope if each CCD pixel only grabbed light with a phase
delay proportional to its "virtual" position on a parabolic mirror?


__ __
''-_ __-''
'-__ _--' virtual primary mirror
''-_ _--''
'''
--------------------------------- CCD array

In addition to what was said : 1 in^2 of aCCD is much
more expensive that 1 in^2 of a mirror.

Rene

 

[Gareth]

Hi all,

Would it be possible to substitute a 2d CCD array for the 3d parabola
mirror in a telescope if each CCD pixel only grabbed light with a phase
delay proportional to its "virtual" position on a parabolic mirror?

In theory something like this may be possible, but in practice no.

If you could record the waveform of the light at each point and maintain
the phase information it would be possible, in theory, to process the
data to produce an image. However, in order to avoid grating lobes, you
would need a receive element spacing less than the wavelength of light,
i.e. < 500nm. A 1m square array would therefore need >400,0000,000,000
pixels, each sampling at thousands of THz. Note that each image pixel
will be the sum, with appropriate phase, of all the receive elements,
not just the data from a single receive element. It quickly becomes
obvious that even if you could build the array (which you can't) there
is no way you could process that amount of data.

With longer wavelengths and lower frequencies it is possible. For
example phased array radars use a similar idea to produce an
electronically steered beam.

Gareth.

--

 

[[email protected]]

Hi all,

Would it be possible to substitute a 2d CCD array for the 3d parabola

mirror in a telescope if each CCD pixel only grabbed light with a phase
delay proportional to its "virtual" position on a parabolic mirror?


__ __
''-_ __-''
'-__ _--' virtual primary mirror
''-_ _--''
'''
--------------------------------- CCD array

(created by AACircuit v1.28.4 beta 13/12/04 www.tech-chat.de)

So if a pixel in the CCD array was 1meter away vertically from the
virtual mirror, you would delay sampling the pixel about 3 nanoseconds
(approximate time it takes light to travel 1meter) compared to the pixel
at the root of the parabola, and you would sample the pixels at a rate
proportional to the frequency of the light you are receiving. I think
the light would also have to be polarized so that it is only travelling
vertically to the CCD. Also the CCD array would have to be custom in

that each pixel would have to be able to be triggered seperately and
very quickly.

I have always wanted to be able to build a telescope using just a planar
sheet of material and avoid having to have a precision mirror so this is
just a wild idea about that!

This isn't going to work for even more reasons than have been listed
already.

The two that I'd like to throw in are first, that you can't gate your
CCD on and off in the 3 nanosecond required to get the 1 metre distance
discrimination in your example, where telescope mirrors are figure to
submicron accuracies corresponding to attosecond discrimination, and
second, that if you could, your window would only be open for
attoseconds, and with current technology you'd be hard pressed to get a
mark to space ratio of one in a million, so you'd be rejecting all but
one in a million of all the photons hitting your detector.

This is that sort of idea that gives brain-storming sessions a bad
name.

 

[Guy Macon]

1m square array would therefore need >400,0000,000,000 pixels,
each sampling at thousands of THz.

Gareth, marketing just called; they need that CCD array by
Friday, at a retail cost of less than $10 per array...

 

[Nicholas O. Lindan]

Would it be possible to substitute a 2d CCD array for the 3d parabola
mirror in a telescope if each CCD pixel only grabbed light with a phase
delay proportional to its "virtual" position on a parabolic mirror?


__ __
''-_ __-''
'-__ _--' virtual primary mirror
''-_ _--''
'''
--------------------------------- CCD array

Close, but no cigar: Look up 'long baseline interferometer' and 'phased
array antenna' w/ google.

 

[Yukio]

I suspect that That Planar Sheet of Material would still have to be Figured
to the same Degree of Precision as the Primary Mirror it is replacing, so
there is no simplification . eg Flat to a fraction of a wave-length of
Light ! Very much harder than making a precision mirror !!



Did you know that the Hubble mirror could have been QC inspected and tested
with only a Pin-hole Light source and a Knife-edge. Idea was overruled as
too primitive, like building the Panama Canel with a pick and shovel.!

Yukio YANO

An old amateur telescope maker !

 

[PaulCsouls]

I don't think you understand how a lense works. You would get the same
image with a parabolic array of CCDs as you would with a flat array.
With a flat array you need a lense to image onto the array. With a
parabolic CCD array you would need a different kind of lense to image
onto the array. You can image onto a CCD using a pinhole lense. The
pinhole breaks up the image into little pieces putting a different
piece of the image onto different pixels of the CCD by the
relationship of the angle of the object to the pin hole and the CCD.
Just imagine looking through a pinhole and by moving your head around.
As you move your head you glimpse a little different piece of the
scene on the other side of the pin hole. If you move your head around
real fast the little piece will combine and you will see the whole
scene. With a mirror or lense the entire scene ends up on every spot
of the lense. What happens to create an image is that the light is
bent in such a way as to create an interference pattern that generates
the image. With a device as you describe you would only get a blur. No
interference pattern; no image.

Paul C

 

[Rich Grise]

Did you know that the Hubble mirror could have been QC inspected and
tested with only a Pin-hole Light source and a Knife-edge. Idea was
overruled as too primitive, like building the Panama Canel with a pick
and shovel.!

Do you have links to any more information about this, or is it just more
UL?

Thanks,
Rich

 

[Ken Taylor]

Gareth, marketing just called; they need that CCD array by
Friday, at a retail cost of less than $10 per array...

.....because they've sold 200 already!

Ken

 

[Yukio]

do a google search for Focault testing !

I came across some web pages on the Hubble Telescope Mirror tests

I made my own mirror as a teenager fifty years ago

This is what happens when the Paper Pushers over-rule the Grunts that do the
actual work !! PP insisted on Gilding the lily by adding a couple of extra
lenses in an "autocollamator setup" and got it wrong somewhere along the
line ! A simple Focault Test setup would have tipped them off that
something was amiss. But hey this a Electronics Design Newsgroup, but then
again this is how I got interested in telescopes , electronics, CCD's,
Electron Microscopes, and Video.

Yukio YANO

 

[Rene Tschaggelar]

Do you have links to any more information about this, or is it just more
UL?

Rich,
Telescopes can be tested against a point source. In earlier
times one used aluminum foil with a pinhole in front of a
lamp, nowadays, I'd rather use a laserdiode without optics
in 500m distance. The 10um in square are pretty close to a
point soutce. Through the telescope one then sees the bessel
rings of the fourier transformed point source if the telescope
is right. Otherwise one sees the distortions in interferometry
style.

Rene

 

[John Larkin]

In theory something like this may be possible, but in practice no.

If you could record the waveform of the light at each point and maintain
the phase information it would be possible, in theory, to process the
data to produce an image.

But light is quantized into photons. One photon hits one and only one
detector, no matter the size of the detectors.

However, in order to avoid grating lobes, you
would need a receive element spacing less than the wavelength of light,
i.e. < 500nm. A 1m square array would therefore need >400,0000,000,000
pixels, each sampling at thousands of THz. Note that each image pixel
will be the sum, with appropriate phase, of all the receive elements,
not just the data from a single receive element. It quickly becomes
obvious that even if you could build the array (which you can't) there
is no way you could process that amount of data.

There is a single-photon-sensitive high-speed detector: a microchannel
plate multiplier followed by a position-sensitive charge detector,
like a 2-d delay-line detector. That will log the time of hit and x-y
location of photons hitting a surface, up to maybe 1kx1k resolution,
at rates around 10M/second. But it won't image unless fronted with a
lens or something. A photon hit on a surface doesn't tell you which
direction the photon came from.

John

 

[John Larkin]

do a google search for Focault testing !

I came across some web pages on the Hubble Telescope Mirror tests

I made my own mirror as a teenager fifty years ago

This is what happens when the Paper Pushers over-rule the Grunts that do the
actual work !! PP insisted on Gilding the lily by adding a couple of extra
lenses in an "autocollamator setup" and got it wrong somewhere along the
line ! A simple Focault Test setup would have tipped them off that
something was amiss. But hey this a Electronics Design Newsgroup, but then
again this is how I got interested in telescopes , electronics, CCD's,
Electron Microscopes, and Video.

Yukio YANO

There was plenty of evidence that something was wrong, but they
ignored it and stuck with the results of one defective test fixture.
The Hubble mirror was so bad that any competant hobbyist-level mirror
grinder would have caught it.

Read "The Hubble Wars" by Eric Chaisson. After weeks of trying to
focus the scope in space, a big meeting was called, and they figured
out what must have happened. One of the prime optical designers
excused himself, stepped into the hallway, and vomited.

John

 

[Rich Grise]

Rich,
Telescopes can be tested against a point source. In earlier
times one used aluminum foil with a pinhole in front of a
lamp, nowadays, I'd rather use a laserdiode without optics
in 500m distance. The 10um in square are pretty close to a
point soutce. Through the telescope one then sees the bessel
rings of the fourier transformed point source if the telescope
is right. Otherwise one sees the distortions in interferometry
style.

Ah. Thanks for this. All I could think of was, "point source, pinhole -
doesn't that mean "spherical?"

Forgot all about those rings.

Thanks!
Rich

 

[Gareth]

But light is quantized into photons. One photon hits one and only one
detector, no matter the size of the detectors.

I had not considered quantum effects, but microwaves are EM radiation
and phased array antennas work. Obviously each light photon will have a
much higher energy than a microwave photon due to the higher frequency,
but I can't see why the fundamental theory would be any different.

There is a single-photon-sensitive high-speed detector: a microchannel
plate multiplier followed by a position-sensitive charge detector,
like a 2-d delay-line detector. That will log the time of hit and x-y
location of photons hitting a surface, up to maybe 1kx1k resolution,
at rates around 10M/second. But it won't image unless fronted with a
lens or something.

To form an image you would need a sample rate high enough record the
actual EM wave and be able to preserve the phase. If you have that data
then you should be able to do the math to simulate the lens or mirror.
This can be done with microwaves, so I can't see why it would not be
theoretically possible with higher frequency EM waves such as light.

A photon hit on a surface doesn't tell you which
direction the photon came from.

I suppose not, but if you have a sufficiently large number of photons
such that the light appears as a continuous wave I think it should still
work. A phased array antenna can tell which direction a radio signal
comes from, though obviously not from a single photon.

One key difference between light and microwaves is that for microwaves
the energy of a photon is very small compared to thermal noise (hf <<
KT) so quantum effects can usually be neglected. Obviously this is not
the case for light.

Maybe there is some photon density threshold where the phased array
approach breaks down?

Gareth.

--

 

[John Larkin]

I had not considered quantum effects, but microwaves are EM radiation
and phased array antennas work. Obviously each light photon will have a
much higher energy than a microwave photon due to the higher frequency,
but I can't see why the fundamental theory would be any different.

Photons are photons, and all the rules apply regardless of wavelength.
One photon can only excite one atom, so it can hit only one place. The
reason a microwave phased-array detector works is because it detects
huge sloshes of many, many fairly coherent photons. The energy in a
single microwave photon is undetectable. A 1 GHz microwave photon
packs about 7e-25 joules.

For light, the equivalent would be for a fast pulse of lots of photons
to hit the planar detector all at once, like from a pulsed point
source. Then you could analyze the timing of the hits at each pixel
and determine the direction. That is technically feasible. But you
can't image random photon hits using just a planar detector, because
one photon can only hit one detector.

One key difference between light and microwaves is that for microwaves
the energy of a photon is very small compared to thermal noise (hf <<
KT) so quantum effects can usually be neglected.

They can be neglected because you can't detect a microwave photon! So
all you can do is work with huge bunches of them, and they average out
to appear like classical, non-quantum waves.

John

 

[dd]

CCD will be inchoherent detection.
The system could be theoretically possible for single spevtral line
sources.
if the ccd elements were replaced with mixer diodes with response up to
light em waves.
The diodes would be illuminated with coherent local oscillator light and
the mixed down em result detected in a narroe band magnitude/phase
detector.