The new Atik camera compared against a good Canon

[ollypenrice]

Is Coventry known for it's seeing, Tim?? Maybe it should be!

Olly

[Steve 1962]

Great thread - I'm just horrified by Mr Penrice's appalling misuse of the apostrophe! I've come to expect far higher standards than that!

[Tim]

Is Coventry known for it's seeing, Tim?? Maybe it should be!

Olly

It's known for it's light pollution!

An extra factor in all of this is the quality of the tracking and the guiding. With those long Ha subs, as long as the mount isn't being twitched all over the shop chasing the seeing, fairly reasonable resolution should be achievable. This requires excellent polar alignment and a mount capable of keeping steady over the duration of the exposure. Incidentally, the pics from the 428ex at 2800mm were also taken on a standard EQ6.

The M51 was the surprise for me. I was binning the LUM subs 2x2 because standard logic dictates that. However I decided to try without binning just to compare the two resolutions, and the unbinned subs are definitely sharper, that's 0.48 "/px against 0.96"/px. It doesn't make sense, but it's there in black and white. I don't have the 2x2 image to hand but here's more of the 1x1 option.

[ollypenrice]

Is Coventry known for it's seeing, Tim?? Maybe it should be!

Olly

Ooops, yes indeed. Sorry about that! How many times have I red inked that one?

I think Tim's point about theory and practice is fundamental to imaging. It isn't the theory that goes in the picture frame.

Olly

PS: A quick thought on small sensors; on some objects a little more sky can be added very quickly. There are two methods...

- Make a short mosaic either side of the object, just in RGB for the starfield. Half an hour per channel is plenty and may even be the best way to take starfields. Overlay the object from the main run between the starfiled panels.

- Place the galaxy initially to the right (say) of the frame and do half the run. Then place it in the left side for the second half of the run. The object itself gets the full whack but you still have an extended starfield either side of it. A bit of Photoshop or Registar-and-Photshop will do the blending. I used this semi-mosaic technique on M33 in the TEC. It worked like a charm and added nothing to the net exposure time while widening the FOV.

[pete_l]

If you really want a 'for dummies' quickie here is my attempt.

- About 2 arcseconds per pixel. (Remember you can quadruple your pixel area by binning 2x2 when you bung the camera in the C14!)

Although when you look at the spec of (say) a Tak TSA102N, it claims to have a spot size (centre and edge - impressive!) of 8-12 microns.

Depending on what exactly the term means, is there much point in using a CCD with a smaller pixel size than this - irrespective of image scale and arc-seconds per pixel - if the best amateur scopes are going to have spot sizes in that range?

[Gina]

PS: A quick thought on small sensors; on some objects a little more sky can be added very quickly. There are two methods...

- Make a short mosaic either side of the object, just in RGB for the starfield. Half an hour per channel is plenty and may even be the best way to take starfields. Overlay the object from the main run between the starfiled panels.

- Place the galaxy initially to the right (say) of the frame and do half the run. Then place it in the left side for the second half of the run. The object itself gets the full whack but you still have an extended starfield either side of it. A bit of Photoshop or Registar-and-Photshop will do the blending. I used this semi-mosaic technique on M33 in the TEC. It worked like a charm and added nothing to the net exposure time while widening the FOV.

What a good idea!

[x6gas]

The M51 was the surprise for me. I was binning the LUM subs 2x2 because standard logic dictates that. However I decided to try without binning just to compare the two resolutions, and the unbinned subs are definitely sharper, that's 0.48 "/px against 0.96"/px. It doesn't make sense, but it's there in black and white. I don't have the 2x2 image to hand but here's more of the 1x1 option.

We've been chasing the same targets with similar kit, Tim, and I am also finding that using the 490ex unbinned in my C11 with the Optec Lepus telecompressor at f/6.2 (giving about 0.5 arcsec/px) is sharper. It is also allowing me to use levels of high pass filtering that I would have said were insane...

http://stargazerslounge.com/topic/185684-m81-bodes-nebula-galaxy-l-and-lrgb/page__st__20

http://stargazerslounge.com/topic/183881-m51-whirlpool-galaxy-first-light-atik-490ex/

(black point clipped on the M51 and I erroneously mention a 460 - this is defo the 490)

As Gina mentioned, I think that the theory is that you should be over-sampling by x2 - so if the seeing limited resolution is ~2 arcsec/px, then aiming at ~1 arcsec/px is probably a good thing... but I am oversampled by twice that!

Last night I was shooting M64 with the 490 binned 3x3 for luminance. It's bonkers fast and still gives 1.5 arcsec/px... I'll post it once processed...

[x6gas]

A focal reducer is a non starter, my C11 is the Edge version and a 0.7x FR is an arm and a leg. I have Hyperstar, that takes it down to F/2, but this is almost too much of a jump for a lot of the smaller objects. What I really need is a FL of around 1200. I have a 8" Quattro, but that is FL 800mm, great for wider field shots.

Robin

You could consider the Optec Lepus unit: http://www.teleskop-express.de/shop/product_info.php/info/p5036_Optec-Lepus-0-62x-Kompressor-for-Celestron-Edge-HD-1100.html

It doesn't have the same image circle as the Celestron one but it is half the price. I'll leave others to judge my results with it, but I've been pleased.

The biggest trouble with my HD11 at native f/10 is guiding, but even with an alpine guide graph the results are OK. I guess there is something in this discussion since my guidecam is offering 0.4 arcsec/px so even if it's wobbling across four pixels I am still oversampled...

[perfrej]

I have a Lepus for my EdgeHD 925. It is a good reducer but somewhat limited in terms of imaging circle. It does, however, cover the KAF-8300, albeit just.

As for imaging with over-sampled systems, it is not all wrong. With exceptional optics and perfect tracking, as well as excellent sky conditions, I do not see anything wrong with, say, 0.3"/px. I am baffled by all the binning going on in the world. It gives the false impression of getting the advantages of bigger pixels, something that you really do not get. Items like noise, bias and full well capacity/dynamics do not follow the up-scaling. Given the choice between binning and over-sampling, my bet is on over-sampling.

/p