Hacker News Reader

Comment by buescher | Hacker News Reader

Not that inexpensive - that's a $12000 lens with a nearly 6" aperture. The telescope itself is a fairly specialized instrument. It's more like a redundant array of very fast (for a telescope) lenses. From the paper: "Our primary goal is to test predictions that at very low surface brightness levels galaxies display a wealth of structures that are not seen at higher surface brightness levels." (But yes, the lenses are going to be cheaper from Canon than the tubes would be from a specialized telescope maker).

View on HN

The thing to appreciate is that in academic science, $12000 lens is inexpensive; the scale is different from consumer and prosumer pricing. A typical project like this would get at least a million dollaras in funding (see https://www.artsci.utoronto.ca/news/astronomer-roberto-abrah... which shows they got $2M to buy lenses; 120 lenses * 12K = $1.5M.

I work with microscopes that cost $1M and just sit in a lab. That's not atypical for an academic or industrial microscope.

One of the biggest issues in modern science is that to make many discoveries you need to pay very high prices to get the latest and greatest hardware. I've been exploring how to make lower cost telescopes and microscopes (and definitely love this project), that are "good enough" to open up new areas of research/discovery for people with budgets in the $1K-10K range. But it's hard! So far I have mostly been relearning what people already knew in the 1800s and early 1900s, that is easily obtained off-the-shelf tech today.

> I work with microscopes that cost $1M and just sit in a lab. That's not atypical for an academic or industrial microscope.

Ooo, what for!? I always love hearing what people do with optical equipment I can't afford. :)

I'm sure I'd be interested in your pet project as well!

The scopes I work near are typically for applying cutting edge imaging techniques to cellular or organoid growth. Like this: https://www.zeiss.com/microscopy/en/resources/insights-hub/l...

Wow, thanks! That first image is really something else.

I wonder if Dr. Pilipp Seidel has any relation to Philipp Ludwig von Seidel.

The point is it's not that inexpensive for the optics. I've worked in an interferometry lab where we used high-volume photographic lenses where we could, and yeah, they were probably an order of magnitude less expensive than something made for the purpose, and real exotic optics were more than two orders of magnitude more expensive.

But talking apples to apples, a $12000 professional camera lens is closer to something like a $12000 research microscope than you think in terms of the build-or-buy decision. There's also a whole industry of telescopes for amateurs that are made in low volumes to much higher optical standards than photographic lenses that are mostly not that expensive. A top tier 6" refractor might be $15K and blow the camera lens away as a general purpose astronomical instrument, but it would not be nearly as fast, which is important for this application, and if you placed an order for 120 of them, you might get them in five years. Maybe. I'd guess that made-for-the-purpose tubes would come in within a factor of two or three of the off-the-shelf lens if you could find a supplier. They might even be cheaper. The project risk would be larger, and that might be determining also.

There are plenty of much cheaper telescopes in that 6" aperture range. My 152mm was only $1k. My polar mount was $1500. Even adding in a tracking scope/camera to attach is still a fraction of the price of single lens. Tack on a similar SBIG dedicated astro camera and cheap laptop to run the guiding and imaging would still come in under that price tag.

So, is there an advantage of having all of the lenses on the same tracking platform to justify the expense of the single mount? If you place individual 6" scopes in an area where humans could comfortably move between them all pointed at the same object or even slightly different areas to get the wider image, would that not be the same/similar result? Essentially, building the VLA but with commodity off-the-shelf visible scopes.

It's more of a "that they could do it at all" sort of feat. They're coming in somewhat cheaper (probably) and at lower project risk by using off the shelf professional camera lenses than by having tubes made for the purpose.

Your 6" scope is slower, probably much slower, than the telephoto lens they used. There really aren't any amateur telescope tubes I know of that you could directly compare to the 143mm aperture f2.8 Canon lens. The right comparison would be to a 6" apo, which would run $8K-$16K and still be slower.

You of course are correct regarding f-value. My specific scope maths out to an unusual f/4.8 at 731mm focal length. However, I'm not trying to take 1/100th images. I'm doing 30s exposures, so a f/2.8 vs f/4.8 isn't that bad of a trade off.

Even if this isn't doing the same "science", it would be an interesting thing to play with for sciene. Instead of stacking images from the same camera, just stack each image from the array. Or capturing an entire mosaic in one "snap" which is essentially what WASP is doing (mentioned in a post from yesterday).

That's almost certainly an achromat unless you got a screaming deal I want to know about. It's not in the same league optically as the Canon lens, which is more like an apochromat, even though you'd probably be disappointed using the Canon lens as a visual telescope - it will be tough getting and keeping sharp focus on stars with the Canon. They had to use a feedback control system on the focuser in the early paper. They also use a lot of image processing.

Also they're not so much using the speed of the lens for shorter exposure times but for field of view and for high sensitivity.

In general photographic lenses make mediocre telescopes and telescopes make mediocre photographic lenses - try using one of your tubes for some terrestrial photography to see. So it's pretty amazing that the Canon lens performs so well to begin with, that they're able to use it like a fast apochromat, and even more so that they're able to build it out to be roughly equivalent to a really large apochromat. With eight lenses, the early paper claims the instrument is equivalent to a 40cm f/1 refractor. How would you build such a thing? Well, this is how.

From what I understand, reconstructing images from virtual telescopes requires extremely accurate timing so all the wavefronts are in phase: https://en.wikipedia.org/wiki/Aperture_synthesis

Yes, but that’s not what they’re trying to do.

Reply on news.ycombinator.com