Hi all
Houghy has kindly lent me his cooled , modded 350D camera, and it takes great images, but just now I really tested it by dong a staggering 45 minute dark at ISO 1600. Sensor temp was 5 degrees.
At ISO 1600, a shot even 1/2 as long would see the image covered with coloured pixels, quite bright in levels. But take a look at this. Yes its a 45 minute dark frame, only amp glow and some hot pixels are visible. Even on a full res version theres some more hot pixels but the vast majority are black!
If Cannon would make available monochrome versions of this sensor, that is with clear instead of coloured lenslets on the pixels, they would have a sensor that would eat anything SBIG and others for breakfast. If it was mounted into a custom body like a dedicated astro camera, single stage Peltie cooling, going, say 20 or so degrees below ambient wound see breathtaking low noise images. Give it 2 stage cooling and Id say youd pick up more cosmic rays than thermal noise.
If I was one of the big astro camera makers like Sbig Id negotiate with Canon to make a large run of monochrome 10 megapixel (400D) chips and the basic processing /control hardware, but with 16 bit processing then mount it in a body designed for thermoelectric cooling, usb2 support etc, then sell it for the same price as a 400D camera. Id either mount the amplifier chip away from the sensor or better yet have it turn off during imaging (its only needed for readout), eliminating the amp glow. Overnight Id corner the market, selling 10 MP astro cameras for a fraction of the price of others, with much lower thermal noise to boot!
Scott
From this information I have figured out that there is no point going below about 500 ISO with the 5DH as its dynamic range is limited by the 12 bit A to D converter.
Check out where the various CCD's place in the scheme of things! Without extreme cooling they are found wanting!
I have been spurred on by the performance of that cooled Canon DSLR and I am designing a peltier fridge for my 5DH. Even a 10 or fifteen degree centigrade drop in temperature makes a huge difference to the thermal noise.
Bert
PS. Check out which sensor wins in every figure or graph.
Indeed, I take my hat off to Canon for their CMOS engineering. If they provided pixel binning and an NABG version, then they would have a real solution to work with. I can't imagine Canon would produce a monochrome chip. Commercially its probably not viable considering their goal is to produce chips for "prosumer" cameras in bulk. Typically, users want color images.
Cooling benefits all chip types, that's a given. I'm afraid I won’t be trading-in my STL11k anytime soon. The flexibility of a monochrome chip (even with the low QE of the KAI-11002) and individual band pass filters (either broad or narrow band) is a huge benefit. I can live with noise as this can be remedied through image calibration and frame combining to improve the signal. Actually, you'll probably find that noise is introduced during your image processing routine, so some noise is to be expected.
There are also emerging technologies such as EMCCD which has the potential to combine the benefits of both realms. High sensitivity and low read out noise to state a couple. The horizon looks bright...
scott - go ahead and pitch the idea - there can be a future ahead if you make the project sound feasable. I am impressed with the cooled 350D, which was why I did what i did.
Hell thats very impressive, wonder how much to modify my 400D hmmmmm.
About $1500, go to www.centralds.net and click on prices, but i would email Yun to get a firm price, it will be different to the online stated price. That is sending your own camera over for the "fix"
I had contacted Yun a few months ago about getting my 20Da modded and cooled but he said he only does the 350D and 400D's.
Now that would have been awesome to have the 20Da cooled.
It already has the "amp off" function and you'd also have live focus to boot !
Yes it's great isn't it. The 350D already has excellent low dark noise.
Bert, try mounting a 12 V 0.5A fan near your 5D and blasting it with air. It definitely helps reduce dark noise.
I don't think SBIG/Apogee/Fli-Cam type people have too much to worry about though. If you look at the specs on some of the new Kodak sensors they are very impressive. The KAF-9000 CCD use in the Apogee U9000 for example has a 38 x 38mm square sensor area, 3k x 3k pixels, 7e- readout noise, 70% peak Quantum efficiency, 5e-/pixel/sec dark current at 25 C. That Quantum efficiency is achieved with antiblooming control!
