How do you build an optical array telescope?

[michael.h.f.wilkinson]

Thanks TeaDwarf, I should have known that it wouldn't be that straightforward!

Is anything worthwhile ever straightforward?

[FraserClarke]

Yes, you're absolutely right. Having your telescopes on the same mounting removes all of the large path length differences you'd otherwise have to correct with optical delay lines.

That would work as an interferometer, if the two sub-apertures were in phase.

So I just tried this out, and it seems to work. Well cool

I put a three hole sub-aperture mask in front of the telescope, and could see a nice little triangular fringe pattern superimposed on the airy discs of the individual sub-apertures.

It's unstable of course -- because the phase between the sub-apertures is changing. But it was slow enough for my eye to be able to process it and see the fringes moving about.

Eyepiece interferometry. Nice

I'll try to make a little webcam movie if I can...

[FraserClarke]

Managed to make a little webcam movie of this effect (which mainly went to show my webcamin' skillz aren't up to much!). Attached are two images of Pollux taken with a triangular three sub-aperture mask in front of the telescope (my cheapo focusing mask). What these show are the airy disc of the star (diffraction limited image from the ~5cm subaperture; though you can't see the airy rings in the image they were obvious in the eyepiece), with the fringes caused by diffraction *between* the subapertures superimposed. The fringe spacing should be at the diffraction limit given by the spacing of the subapertures (~0.4" in this case).

In the still image, you can see a moment (~1/30th of a second) when all the subapertures were in phase and interfering with each other. So there are three sets of fringes, at 120 degrees to each other (as the holes are at 120 degrees). The up-down fringes are quite obvious, and the ones going top-left to middle-right are not too bad. The top-right to middle-left fringes are pretty weak, but you can just make them out...

In the animated gif (if it works!?) there is ~0.5 seconds worth of data slowed down, showing how the coherence between the sub-apertures changes very rapidly. Sometimes 1 and 2 are coherent, sometimes 2 and 3, etc... I wasn't running fast enough to track most of the changes smoothly.

So, it seems you can build an interferometer (of sorts) with an amateur telescope and a bit of cardboard

Now, I guess you're gonna ask how you interpret the data... errrr.... not sure yet... gonna have to read up on that.

BTW -- this is a lot easier to see in the eyepiece than on a webcam I found...

[michael.h.f.wilkinson]

Nice experiment.

It shows you can build an interferometer with cardboard, but only one with a smaller aperture than your scope .

[PeterW]

Have a look for the Cambridge Optical Aperture Synthesis Telescope (COAST). Radio astronomers have much lower frequencies to deal with and so it is possible to record data along with a local atomic clock so that you can post-process the interference... with visible light the frequency is much higher and so you can't get the synchonisation. Optical fibres will expand and contract. You need a nice solid optical bench with very stable temperature so that you can maintain the distances and alignment... not something you can do easily at home.

PEterW

[ngc2403]

lol solid optical bench,

that is what i work beside most days

its the optical bench for LISA Path Finder, trust me they are alot of work

taken 5 years ish to make two good ones

[benjohn]

On the thoughts of doing interferometry electronically. I'm afraid it's not possible at optical wavelengths. You need to detect and record the phase of incoming wavefront, and then combine them later. This is how radio interferometers work - but there aren't any detectors that can do this at optical wavelengths (by that I mean anything short of about 100 microns wavelength). That is why optical interferometers are so complex -- you physically need to bring the photons together and get them in phase.

Hi,

This seems to be the key to why radio astronomy can so successfully achieve huge apertures without great problems. Radio telescopes capture the phase of radiation as well as the amplitude. Combining images from as many scopes as you wish is "simply" a matter of processing. It becomes a computation problem, which the world is expending a great deal of effort in commoditising, instead of a horrifyingly exacting technical and engineering problem, being tackled by a dedicated few.

Given the background economics, it would be stunning if an optical imaging system could capture phase, or so I've fantasised a few times. An obvious goal would be an optical square kilometre array, for example :-)

So, as you're the very first person I've found who's brought this up, could you explain why an optical system can't capture phase? I've really no idea about the technicalities, so I'm very intrigued. The possible payback of being able to do optical phase and magnitude measurement would be profound, so if it was a hard but possible problem, it would be worth endeavouring!

A couple of thoughts: optical systems, having a much shorter wavelength, require enormously faster capture and processing. But, the total photon count in an astronomical image is large, but not inconceivably so. The photon event rate for a dim subject (like me) is also not beyond possibility. With good sensors, the photon count for b bits is 2^b per pixel (1 count = 1 photon). An n pixel sensor can "measure" n*(2^ photons before saturation on a completely uniform image. For 16 bits and 1 megapixel (scale out for a larger sensor), this is about 60 billion photons – no problem for a computer to store and process.

In terms of the event rate, assuming uniform events and a 1 minute capture, you're looking at a billion events a second to capture – again, well within achievable  even on fairly mundane hardware.

…So what is it that makes photon phase unmeasurable for light?

Thank you.

[niallk]

Hi -I'm no expert-

Google Keck - the two telescope optical outputs can be combined using an interferometer which does detect phase information.

The boundaries are presently being pushed towards extending RF circuits into the THz domain (long IR). Even a cell phone these days does exploit multiple antenna feeds to enhance signal to noise ratio.