Which CCD is best?

Paul Haese · #1 27-05-2009, 04:41 PM

Not talking about name brands here, I am talking about which CCD will be best for my needs. The actual chip

I am told that 2" per pixel is best, but if I am using a KAF8300 which has pixels at 5.4um and using the formula on Peter Ward's site s= (205 x pixel width)/focal length (mm) I will be getting 0.68" per pixel on the RC 8". Even the STL11,000 has pixels at 9.0um and this would still not give me 2" per pixel.

What I would like to know is this good or bad and why? Some detailed explanation would be preferred rather than short answers. I want to get my head around this issue and at present I cannot seem to understand why it should be a problem.

Alchemy (Clive) · #2 27-05-2009, 05:01 PM

Quote:

Originally Posted by Paul Haese

Not talking about name brands here, I am talking about which CCD will be best for my needs. The actual chip

I am told that 2" per pixel is best, but if I am using a KAF8300 which has pixels at 5.4um and using the formula on Peter Ward's site s= (205 x pixel width)/focal length (mm) I will be getting 0.68" per pixel on the RC 8". Even the STL11,000 has pixels at 9.0um and this would still not give me 2" per pixel.

What I would like to know is this good or bad and why? Some detailed explanation would be preferred rather than short answers. I want to get my head around this issue and at present I cannot seem to understand why it should be a problem.

its to match the seeing, pixel sampling at half the seeing will provide optimum resolution. you can go tighter hoping for good nights , id rather oversample than under.

Paul Haese · #3 27-05-2009, 05:07 PM

So, Clive, this does not present as a problem then? I take it I am over sampling?

Geoff45 (Geoff) · #4 27-05-2009, 05:15 PM

In order to get the full resolution from your image, you need the smallest resolvable detail in the image to spread over two pixels. So it depends what sort of detail you can expect. If the seeing disk is 2", then you want 1" per pixel. However if the seeing improves to 1", then you would go for 0.5" per pixel. If the seeing is perfect, you would have the Airy disk cover two pixels. There are always trade offs. You don't always go for resolution, sometimes you just want a wide field, in which case you might use several arc sec per pixel. If you are doing planetary stuff you would barlow the image so that the theoretical diffraction disk covers two pixels and hope that you get this resoltion in some of the 1000 frames.
Geoff

rogerg (Roger) · #5 27-05-2009, 05:29 PM

My scope runs at 0.86 arc sec/pixel typically. I think it's good to have that level of resolution sometimes, when the seeing is good. The rest of the time it would be better to have larger pixels.

I often think my ideal would be a camera of reasonable megapixel size such that when binned 2x2 it still results in a nice sized image. That way, I would get a camera that results in about 1 arc sec/pixel when at 1x1, and on nights of average viewing use it at 2x2. This theory doesn't work with my current camera (ST7) because the chip is so small at 2x2 I get a tiny image, so I pretty much stay at 1x1 which is the 0.86.

There are cameras with larger pixel sizes than 9um. But the larger the pixel the more expensive the camera. Because of that cost I have been thinking in recent times it's better to get a large megapixel camera which can bin 2x2, so you save cost and probably end up with a larger resulting image anyhow. You'd probably end up with more sensitivity that way too. My theory might be wrong, I haven't done the maths with great care.

Roger.

PeterM · #6 27-05-2009, 05:29 PM

Hi Paul,
Don't know if this is useful, I imagine you have already seen CCD Calc by Ron Wodaski. If you have an earlier version a newer full version can be downloaded for free below. You can use the various scope camera combinations or create your own camera/scope setup - good for your new GSO RC.

"in all version since Version 1.2, you can create your own telescope and camera entries. It also supports the latest version of Automapper - Automapper can read the latest CCD Calc image scale from the registry."
http://newastro.wodaski.com/book_new/camera_app.php
PeterM.

jase (Jason) · #7 27-05-2009, 05:42 PM

Paul,
I'm not going to reinvent the wheel. There is plenty of info online.

"A good rule of thumb to avoid undersampling is to divide your seeing in half and choose a pixel size that provides that amount of sky coverage. For example, if your seeing conditions are generally 4 arcseconds, you should achieve a sky coverage of 2 arcseconds per pixel. If your seeing conditions are often 1 arcsecond, you'll want a pixel size that yields 0.5 arcseconds per pixel."
Reference - http://www.ccd.com/ccd113.html

I would suggest you review the above URL as there is some excellent info available.

If you are looking for an opinion. There is nothing wrong with heavily oversampled data. You simply need to work it differently. Deconvolution is your friend.

Paul Haese · #8 27-05-2009, 11:48 PM

Thanks guys, I think I have got it now. I will be oversampling a long way, which means I have better resolution but dimmer images. I gather it will mean I need to image for longer and use deconvolution more heavily.

Is this right? At least I think this is correct.

Alchemy (Clive) · #9 28-05-2009, 06:42 AM

Quote:

Originally Posted by Paul Haese

Thanks guys, I think I have got it now. I will be oversampling a long way, which means I have better resolution but dimmer images. I gather it will mean I need to image for longer and use deconvolution more heavily.

Is this right? At least I think this is correct.

thats about right.

also you might just whip the camera onto a widefield for a bit of fun so the extra sampling will help as you still have pixels to play with.