9 replies to this topic

[SimonIRE]

Hi All, 

 

This question has been done before in many different ways but I am really keen to get advice on my particular situation. 

 

In a nutshell, it's a CCD vs CMOS question but with some specifics and preferences. 

 

I have a QSI 683, several ZWO cameras (the 1600mm Pro being the relevant on to this conversation) and a newly bought QHY600. Some of my gear is going to a remote site and I am looking to buy another mono camera for my home observatory. I will be shooting larger targets with an FSQ106 and smaller targets with a 152mm refractor. 

 

The issue is this:

 

I have been experimenting with the QHY600 and to put it bluntly, I just don't like the images it produces, particularly when I comparing them to the QSI (obviously taking into account the different sensor size). The QHY has Chroma filters and the QSI has Astrodons. I fully understand that this is a really subjective comment but to my eye, the QSI produces much richer looking images. It's also a lot easier to handle with its large pixels and easier, in my opinion, to process. I should add, as you may guess, I am not an experienced imager. 

 

My dilemma is twofold relating to 1) FOV and also 2) the future of the technology. I would like a larger sensor but I don't really want another full frame - I have numerous posts elsewhere on problems I am having with tilt and problems attaining a flat field with the QHY. I thought the QSI 6162 might be a good bet. However, I am reading that CCD technology in our hobby in terms of further development is dead in the water. I may be overstating this so my apologies if I am wrong. 

 

Given the above comments, my question to you is?

 

1. Are my comments relating to "richer images" with the CCD something that is well-documented or am I making this up? - some of the experienced imagers near me still claim that CCD sensors are better than CMOS for deep sky imaging and haven't made the transition. What do you think of this?

 

2. Are CCD sensors on the way out (for our hobby)?

 

3. What CCD and CMOS options would you recommend/do you think I should consider? - 

 

Lastly - I full get that some of this post is highly subjective and smacks of "inexperience" so if you could bear with me I'd be grateful. 

 

Cheers, 

 

Simon

 

 

[Bart Declercq]

Given the above comments, my question to you is?

 

1. Are my comments relating to "richer images" with the CCD something that is well-documented or am I making this up? - some of the experienced imagers near me still claim that CCD sensors are better than CMOS for deep sky imaging and haven't made the transition. What do you think of this?

 

2. Are CCD sensors on the way out (for our hobby)?

 

3. What CCD and CMOS options would you recommend/do you think I should consider? - 

 

 

1. not really sure how to comment on this - "richer" isn't really a useful description, there really is very little difference between CCD & CMOS that I could ascribe the word "ritcher" to - if you could provide examples it would be easier to comment - perhaps you're "doing it wrong" - CMOS does require a slightly different approach from CCD to really get the most out of it usually, but the QHY600 is probably one of the most "CCD-like" of the current breed of CMOS in that it's got massive well depth, very smooth sensor noise and a 16-bit A/D conversion so this is a bit of an odd observation.

2. Yes, CCD sensors are on the way out, simply because sensor manufacturers are dropping them in favour of CMOS. No (or virtually no) new CCD-sensors are coming to the market and what still is is likely to be prohibitively expensive for amateur usage.

3. The ones you mentioned you already have a pretty darn good, and if you find larger pixels easier to handle, you can simply 2x2bin the QHY600 and you'll still have a 15 Mpixel camera with even more awesome well depth and a full frame FOV.

 

Though regarding point 2 - if you're going to buy a camera now, the "future of technology" isn't really all that important - if you buy a KAF 16200 today, it's not going to stop working because of being "on the way out" so if you buy it now because it is good, it's not going to be bad in future.

 

That being said, a 2x2 binned QHY600 has almost exactly the same pixel size as the QSI6162, almost 5x the effective well-depth and a little over half the read noise at gain 0 and 3x lower read noise at gain 100 (which does lower the well depth).

 

 

 

 

[SimonIRE]

1. not really sure how to comment on this - "richer" isn't really a useful description, there really is very little difference between CCD & CMOS that I could ascribe the word "ritcher" to - if you could provide examples it would be easier to comment - perhaps you're "doing it wrong" - CMOS does require a slightly different approach from CCD to really get the most out of it usually, but the QHY600 is probably one of the most "CCD-like" of the current breed of CMOS in that it's got massive well depth, very smooth sensor noise and a 16-bit A/D conversion so this is a bit of an odd observation.

2. Yes, CCD sensors are on the way out, simply because sensor manufacturers are dropping them in favour of CMOS. No (or virtually no) new CCD-sensors are coming to the market and what still is is likely to be prohibitively expensive for amateur usage.

3. The ones you mentioned you already have a pretty darn good, and if you find larger pixels easier to handle, you can simply 2x2bin the QHY600 and you'll still have a 15 Mpixel camera with even more awesome well depth and a full frame FOV.

 

Though regarding point 2 - if you're going to buy a camera now, the "future of technology" isn't really all that important - if you buy a KAF 16200 today, it's not going to stop working because of being "on the way out" so if you buy it now because it is good, it's not going to be bad in future.

 

That being said, a 2x2 binned QHY600 has almost exactly the same pixel size as the QSI6162, almost 5x the effective well-depth and a little over half the read noise at gain 0 and 3x lower read noise at gain 100 (which does lower the well depth).

 

Very useful Bart, 

 

I might edit my use of the word "richer" as it is almost impossible to gauge. 

 

I might try binning the QHY600 and see what that looks like. Is this exactly the same as having a smaller sensor with pixels 4 times bigger?

Edited by SimonIRE, 03 March 2021 - 05:13 AM.

[Bart Declercq]

Very useful Bart, 

 

I might edit my use of the word "richer" as it is almost impossible to gauge. 

 

I might try binning the QHY600 and see what that looks like. Is this exactly the same as having a smaller sensor with pixels 4 times bigger?

Pretty much - there are even camera's (like the ASI294) where this is the default and expected behaviour. Read noise is higher than in single pixel mode (as it justs adds the read noise of the four pixels together - statistically that means 2x the read noise for 2x2 binning) but still lower than comparable CCD-camera's - and since the "signal" is increased 4 times due to binning, the proportion between read noise and signal is better in 2x2 binned mode than in 1x1 mode.
 

[leviathan]

To understand better we should know which of both cameras have you used on which scope. I have FSQ-106 and QHY600 and this is a perfect combo ! Almost all targets fit nicely to FOV (exception are really huge targets like Simeis-147 where you need to mosaic), pixel scale is 1.46", QE is very high and images are clean with readout noise of just 1.7e. Previously I had SBIG-8300 and IMX455 images of the same targets are MUCH better. While I understand that electronics of QSI could be better, but at the end it's still very much outdated old KAF-8300 sensor.

 

I suggest that you run QSI 683 on 152mm refractor where pixel scale will be better considering 5.4 micron pixels of KAF-8300.

[SimonIRE]

To understand better we should know which of both cameras have you used on which scope. I have FSQ-106 and QHY600 and this is a perfect combo ! Almost all targets fit nicely to FOV (exception are really huge targets like Simeis-147 where you need to mosaic), pixel scale is 1.46", QE is very high and images are clean with readout noise of just 1.7e. Previously I had SBIG-8300 and IMX455 images of the same targets are MUCH better. While I understand that electronics of QSI could be better, but at the end it's still very much outdated old KAF-8300 sensor.

I suggest that you run QSI 683 on 152mm refractor where pixel scale will be better considering 5.4 micron pixels of KAF-8300.

Very useful.

May ask about your imaging train with the FSQ/QHY; have you had issues with tilt or curvature?

Edited by SimonIRE, 03 March 2021 - 06:45 AM.

[FredOS]

1) I have compared a ASI6200, ASI1600 and FLI16200 in mono versions only. I didn't really find any major differences. I do believe that the raw frames from CMOS are easier to work with, less variation in background and low read noise for narrowband. However, they add complexity : (i) dealing with various gain levels (including for darks, flats, etc...), (ii) tip tilt, (iii) larger files. For tip tilt, I'm not sure you really need perfect stars in the corners as it most cases we crop the frame. It becomes mostly relevant in my view if you do mosaics as corner stars can become centre stars.

2) I think so but you can still get hold of a few FLI cameras which are the gold standard. I plan to use my FLI16200 for many years.

3) On CCD, per above the FLI16200. On CMOs, until we get larger pixel cameras I would stick with the full frame ones per your QHY600. You can always crop but there will be cases where you are happy with the option of full frame. And you can also software bin if using a longer FL scope. I would jump into a quality CMOS camera with 6micron pixels.

[leviathan]

Very useful.

May ask about your imaging train with the FSQ/QHY; have you had issues with tilt or curvature?

Curvature - no, just a bit on the very edges with 2" mounted Ha filter. 50mm unmounted is the way to go.

Tilt - yes, sometimes. I suspect this to be because of the stock Tak focuser. I already dropped CAA and using only solid threaded adapters all the way down to CFW. QHY FW + Camera weight shouldn't exceed 2kg, that is not that heavy compared to old CCD cameras. But at the same time I'm considering to switch to Nightcrawler once.

 

If you have curvature I suggest that you replace the narrowest part in your imaging train that is limiting light cone. In my case it's M98 all the way down to CFW and then M54 very close to the filter wheel. From filter wheel it's M54 threads on both adapter and camera sides and that is enough to not limit the light cone. If I would still use my old 645 QE 0.72x reducer, I would even go with M68 threads and maybe 50mm SQUARE filters not to have any curvature.

[Alex McConahay]

I disagree that this is an issue of CCD v. CMOS. 

And I heartily agree with Bart (post 2) "if you're going to buy a camera now, the "future of technology" isn't really all that important -"

 

The issue here has to be the same issue as it is every other time you have ever bought a camera, or ever will.....what best matches your equipment, offers the lowest noise, highest dynamic range, and all that. 

 

Perhaps someday CMOS chips will completely supersede CCD technology. Okay, wait till that day. But, realize that a year after you buy that new camera some other engineer will figure out a way to supersede that camera you just bought. So, which camera, TODAY (if you are buying today) can best meet your needs? 

 

No doubt CMOS is still developing....but some doubt as to whether it has surpassed the highly refined alternative technology. 

 

And, yes, I think "richness" is a quality. I would say it is a combination of dynamic range and low noise. A higher quality camera will get more of it than a lower. And so, look for the camera with those qualities however the engineer and manufacturers managed to achieve it. 

 

Alex

[Bart Declercq]

And, yes, I think "richness" is a quality. I would say it is a combination of dynamic range and low noise. A higher quality camera will get more of it than a lower. And so, look for the camera with those qualities however the engineer and manufacturers managed to achieve it.

The combination "dynamic range" and "low noise" is "effective dynamic range" (though some might argue that "dynamic range" already takes noise into account - the "DR" values in the graphs from ZWO certainly take the read noise into account which is why there's a big drop at gain 100 for the ASI6200/2600 - they switch to a low read noise readout mode at that gain)