PDA

View Full Version : Why use large telescopes rather than ...?



a1call
2007-Sep-20, 12:50 AM
Hi all,

I am wondering why is it better to use large telescopes rather than just electronically amplify a smaller lenses light/image?

Is it because it will give a better signal to noise/sky-glow ratio?

Why is there interest in telescopes with multiple smaller lenses distributed over large areas (can't remember the concept's name)?
Why is that advantageous.

The question does not relate to a specific optic so I thought it would be more appropriate here than the section about telescopes. However please feel free to move it there if applicable.

Thanks in advance for any insights.

DanishDynamite
2007-Sep-20, 01:04 AM
Hi all,

I am wondering why is it better to use large telescopes rather than just electronically amplify a smaller lenses image?Resolution. Amplification only works if you have more info available in the image. Amplifying a pixel of info provides no new information.

Is it because it will give a better signal to noise/sky-glow ratio?It is because it will provide more resolution, more info.

Why is there interest in telescopes with multiple smaller lenses distributed over large areas (can't remember the concept's name)?
Why is that advantageous.
Smaller mirrors are easier to make. No mirror larger than 6 meters in diameter has been made accurately enough.

The question does not relate to a specific optic so I thought it would be more appropriate here than the section about telescopes. However please feel free to move it there if applicable.

Thanks in advance for any insights.
Hope I provided a few.

a1call
2007-Sep-20, 02:20 AM
Hi DanishDynamite,

Thanks for the prompt reply.

I would think resolution of an image is not directly related to the size of the objective lens but rather the number of pixels of it's imaging plate. I think light gathered through any size objective lens can be manipulated with the correct lenses to project on as large an imaging plate as necessary with as many pixels as can be manufactured and wired up.

Saluki
2007-Sep-20, 02:33 AM
Hi DanishDynamite,

Thanks for the prompt reply.

I would think resolution of an image is not directly related to the size of the objective lens but rather the number of pixels of it's imaging plate. I think light gathered through any size objective lens can be manipulated with the correct lenses to project on as large an imaging plate as necessary with as many pixels as can be manufactured and wired up.


A larger diameter means more light collected. More light=better resolution of dim objects. A better CCD can help of course, but it is limited by the amount of light reaching it.

The multi-scope arrays are an application of interferometry (google it). From my meager understanding, it works great with long wavelengths (radio scopes), but requires extreme precision to work well with optical range wavelengths.

a1call
2007-Sep-20, 03:23 AM
Thanks Gents,

I did some research it seems for both issues, very large arrays as well as large objective lenses the resolution (as in resolving power and not number of pixels) is enhanced due to interferometry.
I have a fair understanding of interference but need time to digest the relevance to it here.

Again thank you both.

Grashtel
2007-Sep-20, 03:34 AM
I would think resolution of an image is not directly related to the size of the objective lens but rather the number of pixels of it's imaging plate. I think light gathered through any size objective lens can be manipulated with the correct lenses to project on as large an imaging plate as necessary with as many pixels as can be manufactured and wired up.
Unfortunately the universe thinks otherwise and has you outvoted, the angular resolution (http://en.wikipedia.org/wiki/Angular_resolution) of a telescope (or any other kind of imager) is limited by the diameter of its primary element, you can't just use smaller pixels to get more data (astronomers would love it if you could though).

parejkoj
2007-Sep-20, 03:40 AM
theta = 1.22 * lambda / d

That's where it's at! More mirror means more resolution, means you can make out small details better.

Ok, a bit more detail: lambda is the wavelength of the observation (e.g. 500nm for yellow light), d is the diameter of the primary mirror, giving a resolving power of theta. So, for example, Hubble cannot resolve things smaller than 0.05 arcseconds apart, when looking at yellow light, and only 0.1 arcseconds when looking at near infrared (1000nm = 1 micron).

Thus, the bigger the mirror, the better your resolution.

Also, the bigger the mirror, the more light you can collect. Light collecting power goes up as the area of the mirror, which is proportional to the radius squared! Thus, Subaru (a single mirror 8.4m across) can collect (8.4/2)2 / (2.5/2)2 = 11 times as much light during the same observation duration as Hubble.

a1call
2007-Sep-20, 03:55 AM
Grashtel,

Thanks for the wiki link. Should help with comprehension.





Also, the bigger the mirror, the more light you can collect.

Yes, and I thought that was all there was to it. This part is easy to compensate for with electronic amplification.

Resolving power, I need to think about for a couple of days or so. I guess it ties in to light's wave feature and for it having an actual wavelength.
The funny thing is with very old lens-less cameras which only had a hole as the objective, there the smaller the hole the sharper the image would be.

Added: My main comprehension obstacle is, I do not understand why the blurring which fuzzes the close objects is there. Where did it come from?
Why would the sharpness of a focused image by a single convex lens be influenced by the wavelength?

yuzuha
2007-Sep-20, 10:57 AM
Hi DanishDynamite,

Thanks for the prompt reply.

I would think resolution of an image is not directly related to the size of the objective lens but rather the number of pixels of it's imaging plate. I think light gathered through any size objective lens can be manipulated with the correct lenses to project on as large an imaging plate as necessary with as many pixels as can be manufactured and wired up.

Not really... you are limited by the sensor technology, both for telescopes and common digial cameras. While camera marketers would have you believe that more is better, pushing the number of pixels per given sensor size beyond a certain optimum number for a given technology is actually detrimental. Diffraction begins to introduce image blur and signal to noise ratio goes down as varius noise levels increase when you decrease the pixel size in order to squeeze more pixels onto a given size sensor chip (a lot of consumer cameras already have terrible low light performance, and produce grainy images full of color and luminosity noise because they've crammed too many mpixels onto a chip in order to keep up with the megapixel race... even in dSLRs: just compare the high ISO performance of say, the 6mp Nikon D40 against the 10mp D40X). For an interesting discussion on camera sensors see http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary/

While astronomical sensors are usually cooled to decrease various sources of thermal noise, they play by the same rules. Once you have the optimum sensor to cover your image focal plane, the only way to increase resolution without sacrificing sensitivity or image quality is to increase the diameter of the primary. Then you start having to use adaptive optics to reduce atmospheric disturbance, and then you have to resort to long baseline inferometry to get around the physical limitations of building huge primaries.

Ken G
2007-Sep-20, 11:20 AM
Yes, and I thought that was all there was to it. This part is easy to compensate for with electronic amplification.


A key piece that hasn't been mentioned yet is the concept of "photon shot noise". This is a key limitation for signal-to-noise if you want many pixels of resolution, and it cannot be mitigated by electronic amplification. Basically, light comes in quanta, and the variation in the amount of light you get into neighboring pixels scales like the square root of the number of photons detected per pixel (Poisson noise). So for a fixed pixel plate, you increase the signal/noise by the square root of the collecting area, i.e. by the diameter of the aperture, just by detecting more photons per pixel. Note that electronic amplification doesn't help because it also amplifies the variance, i.e., the noise.

So it would seem the advantages to big apertures come down to reducing blurriness and graininess, the former by reducing diffraction and the latter by reducing photon shot noise relative to the signal strength. There are also issues of electronic background noise that might also get amplified, so the basic problem with using amplification to see dim sources is the potential for amplifying noise.

parejkoj
2007-Sep-20, 01:57 PM
Remember, astronomical objects are way dimmer than anything you see in day to day life. A 2.5m mirror like Hubble might receive one or two photon's per second from a distant quasar. So an 8.4m like Subaru would collect 10-20 photons/second. You can't amplify what ain't there, so the more photons, the better your image. Also, What Ken G said above about noise...

Try getting an image of something dim in the sky with your digital camera (heck, try taking a picture of a darkened room); you won't see much. Small aperture+noisy detector=bad image.

a1call
2007-Sep-21, 01:49 AM
Unfortunately the universe thinks otherwise and has you outvoted, the angular resolution (http://en.wikipedia.org/wiki/Angular_resolution) of a telescope (or any other kind of imager) is limited by the diameter of its primary element, you can't just use smaller pixels to get more data (astronomers would love it if you could though).

Thanks again for the link. Finally got the chance to study it. I think I am getting it now. So the blurring is caused by diffraction and that is why it is a function of the wavelength and aperture. If I am getting it correctly small (compared to the wavelength) lenses/aperture modifies the parallel star light into close centered waves which diffract and blur reducing the angular resolution of the image. :)

a1call
2007-Sep-21, 02:05 AM
yuzuha,
Thanks for the expert insight. What I was trying to get across was that image from any size primary could be projected on any size sensor plate and not that more pixels could be fit on a given size sensor plate. If diffraction and noise to signal ratio were not factors, I still maintain that the size of sensor plate and thus the number of its pixels are not a factor of the size of the primary.

a1call
2007-Sep-21, 02:12 AM
A key piece that hasn't been mentioned yet is the concept of "photon shot noise". This is a key limitation for signal-to-noise if you want many pixels of resolution, and it cannot be mitigated by electronic amplification. Basically, light comes in quanta,.....

Oh yes, Good ol' Quanta again interfering with my linear perception of the universe. :)
As usual I am in complete agreement.

mugaliens
2007-Sep-21, 12:45 PM
In addition to larger telescopes being able to increase the number of sometimes scant photons arriving on the imaging receptor, multiple smaller telescopes can to the same thing, and are much easier to build. This is the same theory behind radio telescope arrays (http://en.wikipedia.org/wiki/Australia_Telescope_Compact_Array). Radio involves photons, just of a different frequency, and the resulting images are better than from a very large radio telescope like Aricebo (http://en.wikipedia.org/wiki/Arecibo_Observatory).

Technically, the technique is called synthetic aperature (http://www.sandia.gov/RADAR/whatis.html). It works with all bands of the electromagnetic spectrum (http://graphics.stanford.edu/papers/confocal/), and can greatly enhance images which are partially occluded.