Is this the death knell for CCD cameras?

[Stub Mandrel]

I followed an ad on the S&T website claiming a CMOS camera taking stacked 10 90st images has better S:N ratio than a 900s CCD image. It is even better comapring 900s with 900s and the CMOS has greater dynamic range.

http://www.flicamera.com/kepler/16803vs4040.html

We are talking $16,000 though...

[Thalestris24]

Probably! As far as mass market is concerned I think only cmos sensors are being manufactured nowadays. Actual ccd's being limited to the scientific and technical markets. That's the impression I get, anyway.

Louise

[pete_l]

Well, in this comparison it gives the (unsurprising) result that the more expensive camera is better! I don't think the company would be publicising the data if it was the other way around.
The new camera benefits mainly from a better QE, USB3 (so faster readout) and a lower read noise.

I would also suggest that to some extent CCD development stopped some years ago and that most of the research effort has been moved into improving CMOS sensors. They are cheaper to produce, so it is somewhat inevitable that they will be equipment makers' preferred option.

[DaveS]

Hmm.. Anyone up for the KL6060. Mind you, I think even a Tak FSQ106 might be pushed to cover the 87mm imaging circle .

[Stub Mandrel]

1 hour ago, pete_l said:

The new camera benefits mainly from a better QE, USB3 (so faster readout) and a lower read noise.

Absolutely.

Interestingly the ASI1600MM claims QE about the same as the 16803 and read noise better than the 4040!

Some genuine side by side images would be interesting to see.

 

[Stub Mandrel]

7 minutes ago, DaveS said:

Hmm.. Anyone up for the KL6060. Mind you, I think even a Tak FSQ106 might be pushed to cover the 87mm imaging circle .

I suspect their target market is working with apertures closer to 1060mm than 106!

Not for the like of us, although if you make a canny choice of house swap, who knows!

[Jessun]

Very interesting. I remember a colleague who bought a 40" Plasma for £10.000 back in the days.

Early adopters will always pay the price. I'll stick with CCDs for a while longer, because I'm used to them and don't mind spending some more time on a target to make up for lower QE.

It's interesting though how the rise of CMOS blows life in the age old argument about sub length - short vs long, which was often simply dealt with on a theoretical level.

/Jesper

[wimvb]

On 2018-05-15 at 10:42, Stub Mandrel said:

Absolutely.

Interestingly the ASI1600MM claims QE about the same as the 16803 and read noise better than the 4040!

Some genuine side by side images would be interesting to see.

 

But the Kepler has a HUUGE full well capacity (no surprise with 9 um pixels) and a dynamic range of 86 dB, that's what the ASI would be competing against.

As you (Neil) also noted, it is marketed as a scientific camera, not intended for us mere mortals.

The KAF 8300 and its cousins are still in production, by ON Semi. CCD isn't dead just yet.

[Kevin Morefield]

I talked extensively with the FLI folks about this camera.  It is very interesting and I'm betting it is the future.  BUT, they really don't have any real-world images with it yet.  So any side-by-side comparisons with the 16803 to verify SN improvements are not yet there. 

The double read out of the sensor that produces the 16bit result presents some new choices for the imager based on the dynamic range of the target:

Choose HDR mode for a dynamic range greater than 0 – 4095 counts

Choose LDC when your exposures are sufficiently long that dark current growth uses a significant percentage of full well capacity. (Also cool sensor to lowest possible operating temp.)

- How long is sufficiently long?  Is there any camera control software that handles this choice?  Any automation software that would allow you to switch this choice during a single night's run?  What happens to bright stars if you choose LDC?  

The SN calculator they provided me showed a nice improvement over the 16803 with either more short subs or even with longer ones.  But how do these SN maths differ with the HDR and LDC options?

My bet is that the camera does end up with better SN than the 16803 but we need to see what the trade-offs are.  One certain advantage is the improvement in overhead times.  The USB3 download speeds vs 27 seconds for a 1x1 download on my 16803 would be amazing.  

Kevin

[ollypenrice]

On 15/05/2018 at 10:35, DaveS said:

Hmm.. Anyone up for the KL6060. Mind you, I think even a Tak FSQ106 might be pushed to cover the 87mm imaging circle .

According to Tak the present FSQ106 has a mm to spare!

Olly

[wimvb]

Now, where did I put that lottery ticket?