New CCD

[Charles Dunlop]

http://www.optcorp.com/product.aspx?pid=319-320-1542-15881

What you guys think of this? I could scoop it up right now, with a rotator and be super ready to roll and guide off axis. Thoguhts?

[Charles Dunlop]

I went with the QSI. FLI vs. QSI, tough call, but the guy at OPT made the decision final when he said his QSI was his favorite.

[Richard Crisp]

let me clarify my statement about science and engineering: it was in reference to developing a robust method for flat fielding, for camera characterization and other things related to the technology side of the imaging... not target selection, not how to balance a star color and other "art" related things....

for camera performance optimization it is an engineering task. It was to that end that I was making my comments....

[Neil Heacock]

Interesting discussion guys.

I'd have to say that I don't completely agree with the very last statement from Richard that "this is science and engineering: not art".

Personally, for me it's more about the fascination and beauty of what's out there rather then the science of what's out there. Although, of course, the science of describing what we are seeing (assuming that science is correct) is part of the fascination, but when I shoot an object I'm concerned about the color, the composition, the processing to try and not throw out any of what was captured, and the desire to present it in as natural as possible a form... but not for the science of it so much as for the beauty (art) of it.

I'm just saying that in my particular case - and I think the case of many as interest in astrophotography is growing so incredibly fast among amateur astronomers - it's more about the art then the science. Kind of like when I see a picture of a flower or a mountain or a creature I'm not thinking about the science displayed in the image so much as the beauty of what's being photographed.

Things like learning the best ways to calibrate the image are important and I like to learn what I can, but again, it's to try and make the final presentation as beautiful as possible as opposed to as scientifically accurate as possible.

From my perspective...

-Neil

Oh, and Charles, if you haven't got it yet, get the imager and the rotator and let us know how you like it!

[Richard Crisp]

how much lighting variation do you see over the EL panel? If you haven't measured it then you ought to...

what does the spectrum look like. how does that compare to the spectrum of the sky?

lets see measured data and a description of how the measurments were made.... handwaving doesn't get it... quantitative data is what I am requesting... this is science and engineering: not art

[Richard Crisp]

the spacing of the sensor to the filter is what matters; to get them close you have to put the sensor close to the window....hence the problem

you can take twilight flats or you can take sky flats mid day... i do the latter when my ota can allow it, for the truss tube cass, I have to rely on the former. fast downloads make possible what would otherwise be impossible.

you may not realize it but flats also work into the nonlinear region if it is v/v nonlinearity (that associated with the source follower). your camera ought to be linear for a hell of a lot more than 50% of the dynamic range.

OK I'll take the bait on your noise claim....for what exposure duration? how about a one hour exposure?

You need to grasp a fundamental concept" there's a maximum practical exposure time and it is defined as that time for which the dark shot noise is equal to the read noise... if you take a longer exposure you are adding noise to the image beyond the minimum attainable....

then there's the issue of dark signal non-uniformity and dark spikes.... if your dark spikes saturate from too warm of an operating temperature and too long of an exposure... then you cannot correct that.

If Haworth hasn't included a Photon transfer characterization then he hasn't done much in the way of a meaningful comparison...

here's such a comparison:

http://www.narrowbandimaging.com/incoming/u8300_ml8300_qsi583_ptc_d...

most of the "comparisons" I have seen are focused on stupid things like look and color and fit and feel... what matters is the parametric performance, the orthogonality of the sensor and the mechanical integrity... that is seldom if ever analyzed in a non-ambiguous and objective way....

[Greg Marshall]

I skimmed through the paper - don't have time to read it all now, but I'm familiar with the concepts. To address your points:

1. How does the close spacing of the filters affect dewing of the window? That was the original contention. I'm sure that the thermal design of the QSI is not the best it could possibly be - product design always involves tradeoffs. But I have never had any problem with dewing. Frost is a small issue: In sub-freezing conditions (assuming you ever want to be imaging in such conditions) it is necessary to let the camera warm up (to warm the cover glass) before cooling the sensor. Not a big deal and rarely comes up.
2. Ahh, you're talking about twilight flats. I've never used that technique, preferring to use a flat light source that allows me to do flats any time.
3. My 15 second H-alpha is 50% of full well. The camera is linear up to about 40K ADUs (due to anti-blooming feature). I do know very well how to make proper flats.
4. Umm, AFAIK the 8300 is not back thinned. If you are concerned with spectral difference anomaly, you can always use a broadband source such as incandescent bulbs.
5. With the QSI 583 there is little difference in noise between -15C and -25C, but I use -25C when I can. It also correlates to some degree with the kind of imaging: In winter I do mostly narrowband imaging from my home observatory, while in summer I go to dark sky sites for broadband imaging. Thus, the lower temperature goes with longer exposures.

You are assuming a lot when you say I am thinking too simplistically. Your images are excellent and I respect that, but this is not a contest - we're just advising Charles on a camera choice. Having seen very little data on the SE model he is considering, I can't adequately compare it to others, but I am very happy with my QSI 583 and think that it is more than worthy of consideration. I would refer you to David Haworth's paper in Astro Photo Insight comparing various cameras based on the 8300 sensor.

[Richard Crisp]

el panels are at best a hack and suffer from a large number of serious deficiencies. Unfortunately there is a lot of commercial momentum behind them from those that are in the business of selling them but they don't work very well in the general case..... if that is the only way you can make your flats then it is better than nothing but it is making many compromises....

did you ever wonder about the interaction of the wavelength dependency of the QE curve, artifacts in the sensor and the lack of correlation of the spectral curve of the EL versus the night sky? and then there's the failure of them to be suitably uniform.... they are really a poor choice for flats but better than no flats...

but this hobby is rife with ignorance being taught as if it were truth so it doesn't surprise me to see them being pursued.... if you have a light tight OTA try making flats in the daytime using this method...

http://www.narrowbandimaging.com/taking_flats_in_the_daytime_page.htm

you can test the integrity of them using a FLAT FIELD PHOTON TRANSFER CURVE... I teach how to do that in the flats tutorial.

this method makes perfect flats every time provided you fix all light leaks....

It always amuses me watching the excuses people make for their cameras.... 20 seconds download doesn't cost me any time during imaging (ignoring the real issue" flats) or "I find -15c ok...:" really? would you prefer to operate at -25C?

I can operate my ML:8300 at -30C on a 100F ambient..... 2.2 seconds to download and store.... No worse noise than others...
10x faster than QSI, 60C from ambient easily....no excuses, simply the best parametric performance across the board and at a good price too....

[Richard Crisp]

finishing: my ML8300 downloads and stores each image in 2.2 seconds versus the 23 seconds it takes for the same sensor in that molasses-slow QSI583 that I claim is the worst value proposition in the business.
3) Well there are flats and there are flats made correctly.... sure you can take a flat using an Ha filter in 15 seconds and get to 10-15% of full well. You will have noisy flats too... you need to take flats at high signal level to minimize the number needed to get to a particular SNR for flats.... most do it wrong and most have no idea to KNOW what signal level is appropriate and how many flats are needed...do you have a good answer to that question? I do and will share: this is work in progress but deals with the issues in a definitive and dispositive way: http://www.narrowbandimaging.com/incoming/flats_part1_part2_part3_e...
4) how do you deal with the spectral differences of the EL panel versus the sky? If you have a wavelength dependent anomaly, which is common with back thinned sensors, and you use broadband filters your spectral difference can result in artifacts remaining that will not calibrate out.
5) if -15Cis good enough in the summer then it is no advantage shooting colder in the winter unless you are using longer exposures in teh winter.... there's a relationship between maximum exposure time and dark shot noise... when the dark shot noise exceeds the read noise, then you have passed the maximum practical exposure time...

for optics:; the dependency is huge.... I have a 36.8 x 49mm 39mpixel sensor that has no vignetting at all when using 50x50mm square filters with pentax 6x7 lenses.... the rear element is smaller than the film it is designed to expose.... it all depends on the shape of the light cone and the design of the optics..... you are thinking much too simplistically in a number of points.... optics, flats, light sources, signal level in the flats,.,, not considering noise for dark signal and exposure times....

this is a highly technical endeavor to do using best known methods and obviously there are many relevant things you are not considering... give that tutorial on flats a read and we can have a follow up discussion.

[Richard Crisp]

1) I never said dew forms on filters: it forms first on windows... there's a tradeoff between how closely you can put your sensor to the window and how much you can cool it and avoid problems. QSI made a tradeoff that limits how deeply they can cool the sensor without frosting/dewing the window. They also have a poor thermal design that further limits their cooling.
2) what's the relationship between long downloads and sky flats? simple: you are in a race between getting your flats done and the sun setting.... if it takes you 35 seconds to cycle a flat and you are trying to shoot a luminance flat, you have big problems. Isn't that obvious?

[Charles Dunlop]

I already have a Lodestar which fits into my proposed setup. Think I'll just get it, and when I upgrade optical tubes to an expensive one in a year or so I'll get an FLI that fits the scope. This will become my second setup sort of speak. I'm really just excited to be off axis at this point, get rid of flexure. Just last night I was getting a solid shot, and about halfway through a 4 hour run slight elongation appeared. Pain in the rear!