Is CMOS (rather than CCD) the future?

[riklaunim]

You just need sensor data, not camera, to get the QE. And that for astro-cameras is given. For other consumer cameras the biggest problem is to determine what sensor is inside and then to find any datasheet for it.

[Andrea Ricciardi]

Professional observatories use both CCD and CMOS, for different instruments. For example a wavefront sensor for adaptive optics systems needs to run at high framerates, so CMOS wins, while custom wide field cameras use CCD or mosaics and are very different from commercially available ones, with very large pixels due to image scale considerations and very low noise and slow readouts.

But it's a fact that for the commercial world CMOS is largely winning for many economic reasons. The point is: which product is driving the market? Well, it looks like smartphones are leading now, which means technology will mainly advance on small sensors with small pixels, not exactly ideal for astronomy...

Anyway medium/large format digital cameras, even if a minor market, are pushing some development (and remembering some old truths, like big pixels are good for low light...) , but then you can only count on CMOS color sensors, and anyway an astronomy company would not produce the volumes needed to gain access to those sensors: typically those sensors are not available for the industrial market, this is why all the major players in the CCD camera market still mostly use Kodak (well, now ON semiconductor) CCD sensors for their large area cameras.

There are exceptions, with very good monochrome large format CMOS sensors that are accessible to astronomy companies, like CMOSIS ones (good cost, not bad QE, dark and readout noise, but only 12bit readout, so you have to saturate very quickly to have enough gray levels near sky level) or better sCMOS (much higher price but over 70% QE, 1e- readout noise, multigain parallel readout), which could well be the future for astrophotography but not for people on a budget

Andrea

[gorann]

Thanks Andrea, you got me surfing the net for monochrome CMOS and I found that there is chilled 18 Mpix 24x36-mm from Nikon sold for microscopy called DS-Qi2:

http://www.digitalversus.com/nikon-ds-qi2-first-camera-with-monochrome-24x36-mm-cmos-sensor-n36083.html

http://www.nikoninstruments.com/en_EU/Products/Cameras/Camera-Heads/DS-Qi2

Much biomedical microscopy is nowadays done with fluorescent dyes emitting at specific wavelengths and then monochrome cameras with filters are what is needed.

There was a short discussion about the Nikon camera on Cloudy Nights which ended when someone found out it was 16 000 USD. Max exposure time is also a bit short: 60s.

Apparently Sony is also developing monochrome CMOS cameras (mentioned in the first link above) and Leica is selling such a DSLR, as was pointed out here earlier.

I wish these companies would realize that there is a potential market for these sensors among amateur astrophotographers, if the price becomes reasonable.

[chrisvdberge]

There is of course always the option to have someone 'debayer' your normal DSLR sensor. I think there are some people who have mastered this reliable and actually offer this as a service and even product. (besides the regular 'modification' of dslrs)

[gorann]

The Nikon DS-Qi2 may not be as expensive as suggested by someone on Cloudy Nights. I work in biomedical science and 16 000 USD for this type of microscopy camera seems very high to me. Unfortunately, it is nearly impossible to find a price using Google, and maybe I should ask for a quote. However, I found this Indian government  site stating that the price is 235827 Indian rupi, which is about 3200 Euro - much more reasonable.

http://geosquare.in/Ds-qi2.htm

If that is correct then there may be a market for a similar product in amateur astrophotography, especially if Nikon can offer longer exposure times than 60 s.

[chrisvdberge]

Here you seem to be able to get a monochrome D800 for $6050

http://www.maxmax.com/monochrome_camera_order.asp

(he offers some Canons as well)

[gorann]

Thanks Chris for the link. Very interesting.

It is a pity that their rebuilt ones are so expensive (A Canon 5D Mark III that is 2800 USD new becomes 6650 USD if you buy it from them  debayered). On the page from your link they say "We are not converting any customer supplied cameras because of risk to sensor during conversion", while on the page called "Conversions" (http://www.maxmax.com/IRCameraConversions.htm)  they say that you can supply a camera for conversion and it may only be 450 USD. I assume that one of the pages is outdated. It may also be that monochrome conversions are particularly expensive and a high risk operation (so many end up in the bin).

There is also a company in your country, JTW Astronomy, doing similar rebuilds:

http://www.jtwastronomy.com/products/monochrome.html

They sell a monochrome Canon EOS 700D that is chilled with a delta T of -85 °C for 660 Euro. One drawback is that it is not a full size sensor and the pixels are a bit small (4.3µm), but it is certainly a very interesting offer. They also rebuild cameras that customer send in but it worries me a bit that they do it at the risk of the customer. They also state "It is important to know that this process is not perfect, and you cannot expect a perfect flat field from one of these cameras although it is close." Maybe they are just very honest about what to expect.

It would be great to hear from someone that has tried a converted monochrome DSLR. I think I will start a separate thread about it so the question gets noticed.

[Zakalwe]

I don't think that JTW are doing modified cameras any more.

[chrisvdberge]

Im not sure if JTW still does it's 'ultimate' conversions indeed. I've been inclined to go for one in the past, but the fact that you loose the advantage of not needing a laptop to control the dslr made me go for a 'regular' nikon d600 for now. 

All in all with current pricing you end up more or less in the range of the big sensor monochrome ccd's anyway, and that being a proven product seems quite the advantage. 

It would become very interesting if price can remain low for monochrome dslr's, but with removing the bayer matrix being a high risk operation I think it won't be cheap any time soon

[Zakalwe]

The Nikon DS-Qi2 may not be as expensive as suggested by someone on Cloudy Nights. I work in biomedical science and 16 000 USD for this type of microscopy camera seems very high to me. Unfortunately, it is nearly impossible to find a price using Google, and maybe I should ask for a quote. However, I found this Indian government  site stating that the price is 235827 Indian rupi, which is about 3200 Euro - much more reasonable.

http://geosquare.in/Ds-qi2.htm

If that is correct then there may be a market for a similar product in amateur astrophotography, especially if Nikon can offer longer exposure times than 60 s.

Here you seem to be able to get a monochrome D800 for $6050

http://www.maxmax.com/monochrome_camera_order.asp

(he offers some Canons as well)

I can't see where these choices would be a better bet than an Atik 4000 or a QSI 6120

http://qsimaging.com/6120-overview.html

[chrisvdberge]

I can't see where these choices would be a better bet than an Atik 4000 or a QSI 6120

http://qsimaging.com/6120-overview.html

Maybe not better, but definitely different: 12mp vs 36mp for starters. 

QE and full well will be different as well as the noise performance. Maybe difficult to compare, but definitely a whole different chip

And remember; we are talking DSLR vs CCD here as well, not just comparing the sensor. Other factors may play a very important role in which is better  

[gorann]

JTW homepage is rather contradictory regarding what the do or do not offer - probably needs updating. However, since April 2015 they seem only to sell kits and maybe they want you to do the debayering yourself - could become a very frustrating and expensive exercise.

Regarding the CCDs in comparable price range, they are not really comparable as Chris points out. They have smaller chips (13x9 mm in the 12 Mpix QSI6120) and the 15x15 mm chip in the Atik 4000 is only 4 Mpix. If I give up waiting for a reasonably priced mono DSLR,  the best CCD value now appears to be KAF-8300 (18 x 14mm) based cameras that have recently fallen in price (SBIGs version is just under 2000 USD)

[gorann]

I just got a quote from a Norwegian company selling the Nikon DS-Qi2, and it is 103,000 NOK = 11,700 Euro before VAT, so it is not for us.....

[Uranium235]

Talking about big CMOS sensors is all well and good, but ultimately pointless if you dont have the telescope to take advantage of it - as large, flat fields are not cheap either.

[Physopto]

Talking about big CMOS sensors is all well and good, but ultimately pointless if you dont have the telescope to take advantage of it - as large, flat fields are not cheap either.

Yes Taks and Vixens to name two.

Derek

[Stu_2011]

The Nikon DS-Qi2 may not be as expensive as suggested by someone on Cloudy Nights. I work in biomedical science and 16 000 USD for this type of microscopy camera seems very high to me. Unfortunately, it is nearly impossible to find a price using Google, and maybe I should ask for a quote. However, I found this Indian government site stating that the price is 235827 Indian rupi, which is about 3200 Euro - much more reasonable.

http://geosquare.in/Ds-qi2.htm

If that is correct then there may be a market for a similar product in amateur astrophotography, especially if Nikon can offer longer exposure times than 60 s.

Hi

If it helps with your cost estimates I purchased a Nikon DS Fi2 for work last year and paid just under £5k for it. This did include the optional touch screen control monitor. Not the same camera but should give an indication.

It's a fantastic camera, 5 meg colour camera.

Thanks

Sent from my iPad using Tapatalk

[gorann]

Even APS-C sensors of DSLRs (20.7×13.8 mm to 28.7×19.1 mm) are quite a bit bigger than most affordable CCDs, and still within the range of the flat field of affordable telescopes. I have had no vignetting problems with my Canon EOS 60Da (22.3 x 14.9mm = 27 mm image circle) connected to my Celestron Edge HD 8" (with and without the 0.7 reducer) or  my ES 80 Apo with field flattener, so stepping down to 13x9 0r 15 x 15 mm (= 16 - 21 mm image circle) seems a waste of photons.

It may be overkill to get a DSLR with a full size sensor for many scopes, as it has a 43 mm image circle, but even my modest scopes may handle that with some slight cropping (42 mm image circle when using T2 connectors). I would feel more in control if I do the cropping in post-processing than if the camera does it because of its small chip, and then with a 36x24 mm CMOS chip I could one day upgrade the scopes. Still, I would be happy if I could get a monochrome CMOS in APS-C format (in a DSLR or other housing) at the low price and low noise level and less need for cooling that one could expect from the CMOS technique.

However, if nothing exciting and affordable happens on the CMOS frontier within a year or so, I see myself loosing patience and going for a CCD.......

[chrisvdberge]

I wouldn't count on anything getting 'mainstream' or 'readily available' within a year. So you're either looking at modifying a regular dslr or getting that CCD  

[Alien 13]

The problem I can see for future development is that camera manufactures will end up putting more hardware on the camera chips for assisted live view focussing etc so a nice astro camera based on one will never happen.

Alan

[Physopto]

Even APS-C sensors of DSLRs (20.7×13.8 mm to 28.7×19.1 mm) are quite a bit bigger than most affordable CCDs, and still within the range of the flat field of affordable telescopes. I have had no vignetting problems with my Canon EOS 60Da (22.3 x 14.9mm = 27 mm image circle) connected to my Celestron Edge HD 8" (with and without the 0.7 reducer) or  my ES 80 Apo with field flattener, so stepping down to 13x9 0r 15 x 15 mm (= 16 - 21 mm image circle) seems a waste of photons.

It may be overkill to get a DSLR with a full size sensor for many scopes, as it has a 43 mm image circle, but even my modest scopes may handle that with some slight cropping (42 mm image circle when using T2 connectors). I would feel more in control if I do the cropping in post-processing than if the camera does it because of its small chip, and then with a 36x24 mm CMOS chip I could one day upgrade the scopes. Still, I would be happy if I could get a monochrome CMOS in APS-C format (in a DSLR or other housing) at the low price and low noise level and less need for cooling that one could expect from the CMOS technique.

However, if nothing exciting and affordable happens on the CMOS frontier within a year or so, I see myself loosing patience and going for a CCD.......

Life is too short,you don't know what awaits around the corner !

Derek

[gorann]

Yes, friends you have convinced me that life is too short, so yesterday I ordered an ATIK460EXmono from Teleskop Service. I was thinking long about getting an ATIK383 with a rather large chip but finally decided on the 460 with a Sony chip because of its reputation of being very quiet = darks hardly needed). I also ordered all the other stuff like filters that I need for going mono and a 0.8x TS reducer/flattener for my refractors to compensate a bit for the smallish chip in the 460 - should make me able to fit in M31

[kalasinman]

Yes, friends you have convinced me that life is too short, so yesterday I ordered an ATIK460EXmono from Teleskop Service. I was thinking long about getting an ATIK383 with a rather large chip but finally decided on the 460 with a Sony chip because of its reputation of being very quiet = darks hardly needed). I also ordered all the other stuff like filters that I need for going mono and a 0.8x TS reducer/flattener for my refractors to compensate a bit for the smallish chip in the 460 - should make me able to fit in M31

You'll be making mosaics. Kaf8300 perhaps a better choice. Some of the best imagers here use that chip and get very good results. Just my opinion.

[perfrej]

My money is on the 8300 as well. Never used any darks with it... Like this last one; all 8300, no darks...

[Physopto]

Well rightly or wrongly you have made the decision. I hope it works out well for you. As I said life is too short to waste, even if it is the wrong decision you will know and still get results and a learning curve is started. Mine is still going and going and going

Derek

[gorann]

Lovely picture Per!

Well friends, I read so many threads about these two sensors and I finally had to make a decision, right or wrong. A lot of 460 owners seem to love that camera, and i get the impression that it is based on a more modern technology with less noise and higher Qe. I also got the impression that the quality of the 8300 chips varies since some love them, like you guys, while some are very disappointed about noise and dead pixels. And the 460 chip is not that small. On my 80mm ED refractor (FA 480 mm) I should even be able to fit in M31, but in some cases I may try to do some mosaics. The 460 chip area (125 mm2) is slightly more than half the size of the 8300 (238 mm2) so with two frames I get about the same field of view. I still have my Canon 60Da (332 mm2 chip) that I aim to continue to use for large objects. In any case, I assume most agree that trading in the Netscape was the right decision.