4.53
#...
·
 |
|---|
|
Hi out there, an often asked question brings up a thought I have for a long time and I'm not sure if I answered it to myself correctly. So I'd like to ask for some explanations to help me understand this better. There is always this recommendation, that the pixel scale of the camera/telescope combination should be between 1 and 2 arcseonds. I understand, that those numbers depend on the seeing and the observing conditions and should be adopted a bit. But lets say, I have the "standard" seeing of germanys darker regions (2 arcsec). For a long time, I dreamed of a scope with a long focal length and it was hard for me not to buy one. This pixel scale explanation finally helped me understand, that this may not be a good choice. So if I summarize my knowledge, I know that focal length and pixel size depend on each other because of the seeing and finally this limits the possibilities of the equipment. If I want to use a longer focal length, I need to choose a camera with larger pixels to compensate for the seeing. That means, the amount of detail I can image is always limited by the seeing. Larger pixels usually means larger sensors and that leads to a wider FOV in the final image. So far, so good... But there are some huge scopes out there one can buy. For example the 14" Edge HD or even a 14" RC or some more fancy stuff. If I adopt the mentioned rules, I have to buy a camera with larger pixels as well to use these scopes under the same seeing conditions. One benefit of these scopes is the aperture, that allows you to see more details, because this is only dependent on the aperture, isn't it?. But lets say, I want to image a tiny galaxy. Then I will never be able to get a better resolution with those larger scopes because of the seeing. So I am at the point, that a shorter focal length with smaller pixel sensors should be comparable to a longer focal length with larger pixels, right? (noise not included in the thoughts) The difference is, that the latter will have maybe a larger aperture which lets me image fainter details but the light will be spread on a few pixels only. So I can't resolve those details. So why are so many people buying larger scopes if there is not much benefit from doing so? (Planetary imaging not taken into account, I concentrate on deep sky imaging.) Thinking about this, I am wondering if I miss something here. What is wrong with my thoughts? Thanks for your help. CS Christian |
7.67
#...
·
·
1
like
|
|---|
| https://www.rc-astro.com/mtf-analyzer/ You might find this page helpful. I think it covers most of the sampling stuff you want to know. Unless you have truly atrocious seeing, the real limit is how well your mount guides. Supposing you can properly make use of sampling at 0.31 arc seconds, unless you have a killer mount the guiding is going to smear it to around to 0.6 arc seconds or higher. Christian Großmann: To this I say, "kinda sorta". Today, for larger pixels you are better off just buying a CMOS and resampling it instead of trying to buy a CCD with large pixels. Whatever hardware binning advantages CCD cameras have over CMOS cameras, newer CMOS cameras more than make up for with their much higher quantum efficiency and lower read noise. For imaging, the REAL benefit of a larger aperture is it allows you to use a higher focal length at a reasonable focal ratio. This allows you to have a finer image resolution without having to spend eons collecting data. Consider that using the same camera, a simple Redcat51 can collect signal faster than a hefty Planewave just by virtue of having the better focal ratio. However, if you found a camera with small enough pixels for the Redcat51 to match the image scale of the Planewave, it would get blown out of the water. Christian Großmann: To further hone in on this, lets check it out!  Here we have two imaging systems with the same resolution at 1.03". So, in that sense they are the same. However, check out the value for ps, the larger telescope with the larger pixels collects four times the signal as the smaller telescope. |
1.91
#...
·
·
1
like
|
|---|
|
Hi Christian, I'll try and explain it as easily as possible, simplifying physics. I had the same problem of not understanding how oversampling could possibly be a problem. I mean, who wouldn't like to have more resolution? But if an image is oversampled, it means you have more pixels for any giving object. But the object you captured through the atmosphere just doesn't have this resolution. There is just nothing for the extra pixels you have to capture. And now these pixels are supposed to get filled with what? No detail available. So everything is a bit blurry and unsharp. Additionally, lower sampling has a better SNR, because you spread the same amount of light on fewer pixels. Larger scopes are better at "averaging out" trouble in the atmosphere over their large aperture. And if we're talking real big scopes, those have adaptive optics that move with the turbulences, minimizing the problem. Also as @SemiPro said, larger aperture gathers more light, so at the same sampling, you collect more light compared to smaller telescopes. Personally, I get better resolution at ~0.7 "/px, with average seeing, compared to 1.61 on my other setup. Hope it's a bit clearer now. |
22.40
#...
·
·
1
like
|
|---|
|
Watch this presentation. Starting at about 32 minutes the effects of sampling and seeing on image quality is discussed in some detail. https://www.youtube.com/watch?v=te4UVYi6n44 John |
16.18
#...
·
·
1
like
|
|---|
|
Lots of great and detailed advice here. I am the new owner an RC12 with 0.8x reducer and a ZWO 6200MM. I live in Bortle 2 with 3-4 arcsec seeing. This was a big investment at 2000mm focal length on a poor seeing site. My solution is to binx2. This gives me 0.75arcsec/resolution element, still oversampled with the seeing. But its not terrible, given the BXT will give me a little back. Do I worry about on-chip binning a CMOS detector? No, not at all. Indeed my 2017 iMac thanks me for it. Sure I lose a little S/R but the readnoise is still small compared to the Poisson noise on typical skies [10% iof sky even in NB and moon-less Bortle 2)]. In the end I get a 4K image with a relatively large aperture over a field that encompasses the largest objects I am likely to image. I use the large aperture as a flux bucket - not for the focal length. Sure I would love a better seeing site - but I can also resort to getting L-band on a remote good-seeing site and adding it into the mix. CS Brian |
 #...
·
|
|---|
|
The way I understand it, the resolution is limited by seeing. Even an 8 inch telescope has better theoretical resolution than the average seeing in our back yards. This doesn't apply for lucky imaging of planets. For DSOs even if bigger aperture won't buy you better resolution it will reduce the total exposure time because you will be catching more photons per unit of time. Pixel size also has a big impact on imaging time. Big pixels catch more photons analogous to big aperture. So in an ideal situation you will have big aperture, big pixels and proper image scale for the seeing. |
11.57
#...
·
·
3
likes
|
|---|
|
Let's say your seeing is 2 arcsec. Then you want a pixel size in the range of about 0.75 to 1 arcsec. This way, you are a little oversampled but not too oversampled. For simplicity, let's say your pixel size is fixed to 1 arcsec. (With today's CMOS technology, you can bin 0.5 arcsec pixels to 1 arcsec with little penalty.) When your pixel size is fixed to 1 arcsec, how much light falls into a pixel would be purely determined by the aperture size. Not by focal length, nor by focal ratio. For example, you can get a scope with a focal length of 1000 mm, and a camera with 5 micron pixels. Each pixel would be 1 arcsec. Or, you can get a scope with a focal length of 2000 mm, and a camera with 10 micron pixels. The pixel scale is still 1 arcsec. In both cases, if the scopes both have aperture sizes of 30 cm (same light collecting power), the pixels would have identical angular resolution (under the assumed 2 arcsec seeing) and identical S/N ratio. (The first system would be F3.3, and the second system would be F6.6.) Assuming you can find the above two combinations, the remaining questions would be: 1. which one is more affordable or of better quality (depending on what you value), 2. which one has better corrected optics, and 3. which camera has more pixels. If pixels are both 1" in both cases, then the FoV of the two systems are determined by the number of pixels of the camera. And to take full advantage of the FoV offered by the camera, you need better corrected optics. So #2 and #3 usually go side-by-side. You can get a long focal length system while still having large FoV, by putting a camera with many pixels on it. But cameras with many pixels (and large pixels) tend to be expensive in today's market. You can also get a short focal length systems and put a camera with small pixels on it. But short focal length (fast F ratio) with good optical correction tends to be expensive, and is sometimes hard to maintain (collimation). So each side of the spectrum has its pros and cons. The above is how I view the problem. Cheers, Wei-Hao |
4.21
#...
·
·
1
like
|
|---|
|
I am confronted with the same question - should I get a larger scope or not...? Well, currently I am using a very good 10" Newton, made by Lacerta. Fused silica mirrors, carbon fiber housing - everything very stable and almost no refocusing necessary due to temperature change. Image circle good enough for APS, which is more than fine for longer focal lengths, because for more FOV, I use my RASA11. The seeing - at my location it is usually around 1.5-2 arcsec, which is quite fine. Sometimes I get around 1.2-1.5 arcsec - which is super. Bortle - well, at my location most of the time a 4, sometimes decreasing slightly, but not often. Wind - unfortunately sometimes a problem because I cannot build a proper observatory dome, so my equipment does not have wind shields - wind above 0.7m/sec creates problems. I am thinking about getting a 14" Truss with f/6.8 from TS, still in a price range, which I can afford - but does it really improve my results? I am not sure. I can use my various QHY cameras to get the best possible FOV in the image plane ( 600, 294, 268, 163, 183) and if I calculate the effective aperture, I am most likely not much better off. For many targets, I would need about 2000+mm FL - with the Newton (1250mm native FL), I use an apo extender 1,5x and get close to 2000mm at 4,5x1.5= about F/6. For longer FL, I can use a 2x extender to get to 2500mm but then the aperture decreases to about F/8.6 or so. Well of course here, the 2400mm Truss kicks in at F/6.8 but not much. At those long FL, the QHY294 at 11MPx mode makes sense - the pixel size is in that mode 4.63micron, which is pretty close to best possible, My mount is quite fine, it is a very reliable EQ8R-Pro and if I want, I could swap my RASA over to this mount and put the Newton on my other mount, which is an iOptron CEM120EC2 - which is almost perfect. So - what is the conclusion - I am following this discussion with great interest and hope to get wiser and come to a valid conclusion.... CS Georg |
