Comparison of CCD chips

I would want to see the spectral curves for the one shot colour camera. Assuming it is better then the mono (which it seems to be) then the colour camera would be the one I would go for.

Jed - I agree though I'd not look at just the spectral response. I'd try (by looking at actual images) and see how the different colours 'register'. I know most processing software allow re-aligning of colour planes, but I'd be cautious on the accuracy a chip maker can lay down a Bayer pattern on millions of very small pixels.

If the images look good and the spectral response is higher, then the colour ccd would be the front-runner.

Another factor would be 'science' - in order to do science, mono would be by far the best (if not the only) choice. Of course, if you are only interested in imaging objects for their appearance, you can disregard this factor.

Dave

(Al - its rare that any CCD imagers are above the 90% mark - 70% plus is good. 90% plus is rare and very expensive in amateur circles).

Yea, I dont think any have come near amateur price ranges yet unfortunately.

Professional ccd's often can have a peak above 90% though. This is achieved is by thinning the back of the CCD and letting it get exposed from behind. This is more efficient as the photons are not obstructed by any of the gate electronics. This means evenly thinning the back of the ccd down to where the silicon is transparent, at a thickness of <~15microns. The process makes these CCD very expensive as Dave says.

[edit]
Actually just look at the numbers dave posted for front illuminated and back illuminated to see the difference it makes to QE. (Sorry Dave didnt notice that when I posted this)

Michael - here's a list Kodak chips of <10um pixel and at least 2048x2048 with > 50%QE

Kodak KAF-6303E Front-Illuminated 68% 3072 x 2048 9 µm
Kodak KAF-16801 Front-Illuminated 65% 4096 x 4096 9 µm
Kodak KAF-16803 Front-Illuminated 59% 4096 x 4096 9 µm
Kodak KAI-11002M Interline 51% 4008 x 2672 9 µm
Kodak KAI-4021M Interline 55% 2048 x 2048 7.4 µm

If you can go up to 15 µm, options get a little better:
Kodak KAF-09000 Front-Illuminated 69% 3056 x 3056 12 µm
E2V CCD42-40-1-368 Back-Illuminated 96% 2048 x 2048 13.5 µm
E2V CCD42-40-1-383 Front-Illuminated 50% 2048 x 2048 13.5 µm
E2V CCD47-10-1-371 Back-Illuminated 93% 1056 x1027 13 µm
E2V CCD47-10-1-373 Back-Illuminated 73% 1056 x1027 13 µm
Fairchild CCD486BI Back-Illuminated 96% 4096 x 4096 15 µm *
Fairchild CCD3041 Back-Illuminated 96% 2048 x 2048 15 µm

* I think this CCD imager (case plus chip) is somewhere around the $40K mark.

(Al - its rare that any CCD imagers are above the 90% mark - 70% plus is good. 90% plus is rare and very expensive in amateur circles).

Michael - you really need to figure out your average seeing and focal length(s). You can then calculate the best pixel size for your given focal length as half the average FWHM. Of course, you can alter your focal length with a reducer or Barlow.

If you want an imager to cover more than one focal length (and the FL's are very different) you'll have to trade off (or buy two imagers).

As regards the vendor's statements on darks etc, I'd be looking for documentary proof (eg dark frames) that support their statments. Though, I have heard the SE *do* have very low noise.

As for Colour or Mono - take your time and do the research...

FWIW
Dave

Dave,

Thanks for the list of chips.

In terms of pixel size, I'd be looking for around 7um for the Pentax. This would give me 2.8arcsec/pixel. For widefield imaging, this seems to be considered fine. 9um would give me 3.7arcsec/pixel, which I could live with too, but my impression so far is that I should try and go for something in the 7um range if possible. However, if anyone has any views on this issue, I'm willing to listen as I've alot to learn about the whole ccd-->telescope matching process.

One thing I have noticed, and it's apparent in you table, is that you usually get a higher QE if you have larger pixels. Seems to be difficult to get a high QE with small pixels.

To use 7um on the 12" @f/6.3, I would then bin 2x2 for 1.5arcsec/pixel or 3x3 during not so good conditions for 2.3arcsec/pixel.

The only one that would interest me at the mo if the KAI4021, which is the one in the mono camera I'm looking at. The KAI11002 would be the next choice in terms of price, but it's QE graph is virtually identical to the 4021. The 6303 is an interesting one as the QE peaks in the 60something% in the 600-700nm range, which is ideal for H-alpha and SII. Expensive chip though.

One thing I read re the colour camera, it can't be used in colour mode if you bin it. Is this something that hits all color chips, does anyone know? My guess is that it probably does as you are combining the light from adjacent pixels of different colour into one large pixel. However, I'd appreciate if anyone knew for sure.

The other factor against the colour is narrowband imaging, which seems to be the way forward for imaging DSOs, especially if you have light pollution or the moon is hanging around. . With Bayer pattern, you get Red, Green, Green, Blue in a square pattern. So, for H-alpha imaging, which is in the Red, I would effectively have only 1/4 of the pixels operating, thus reducing my effective chip area by 75%. For OIII, I've have half the pixels operating and for SII, I'd only have 25% working. With mono, they'd all be working.

Even if the mono chip had half the QE in a particular frequency, surely it would be better to have 4 times the pixels operating at half the QE rather than 1/4 the pixels operating at twice the QE in a fragmented pattern, would it??? :

Thanks lads for all the info so far. Much appreciated!

Even if the mono chip had half the QE in a particular frequency, surely it would be better to have 4 times the pixels operating at half the QE rather than 1/4 the pixels operating at twice the QE in a fragmented pattern, would it?

Sound logical to me... I'd recommend calling up SBIG or Finger Lakes and ask the lads there - I've found FL very helpful with no sales pressure at all.

Dave

Some sample images taken with the mono chip (SXVF-H16):

Rick Krejci
www.starlight-xpress.co.uk/SXV-H16_image...veil_ngc6992_h16.jpg
www.starlight-xpress.co.uk/SXV-H16_images/900_vdb142_h16.jpg
www.starlight-xpress.co.uk/SXV-H16_images/900_cocoon-h16.jpg

Charlie Warren
www.starlight-xpress.co.uk/SXV-H16_images/Sh2-240%20Ha.jpg
www.starlight-xpress.co.uk/SXV-H16_images/B33-small.jpg
www.starlight-xpress.co.uk/SXV-H16_images/M31-Color-small.jpg
www.astrofx.com/H16/M45H16/m45_master_rgb4_crop_df2.jpg
For more: www.astrofx.com/NewImage/newimage.htm

Ian King
www.starlight-xpress.co.uk/SXV-H16_images/gammahalfsize.jpg

Tommi Worton/Ian King
www.astronomy-uk.co.uk/astro/latestimage...al-web-ian-tommi.jpg

And of course we've seen some of the great images taken by our own Kieran "phoenix" here.

Sweet - the horsehead is a good comparison image since we see it a lot. Nice crisp detail there.

Any images taken with the colour version?