Problem with Flats

dmcdona wrote:

albertw wrote: I did it for my camera recently and the results were a little different from the numbers I've seen quoted

Significantly different? or just a tad? The write-up, if you get a chance, would be an interesting read.

Just a tad, but even a little with the low gain and noise nearly doubles my minimum exposure time to a whopping 20s with 2x2 binning, using the idea that read noise should be 5% of sky noise.

Atik don't state the gain but it's quoted in most places online as 0.267 e-/ADU . They do state the readout noise as '4 e- typical value'

My results gave a gain of 0.253 e-/ADU, & a read noise 5.34 e-.

Edit: The chip seems to be linear across the whole range. And has a full well depth of of 16646 e-. The most interesting thing I found out was that atik's claim that darks were not needed seems to be correct.

Cheers,
~Al

albertw wrote: I'm guessing Dave is on about measuring your own gain, dark current, readout noise etc. Doing it will give you an idea of where your noise is coming from, whether you need darks even, what your minimum exposure length given your skies and read noise etc.

Sorry - I meant to add (and bringing this thread totally back on topic), the main reason I did the CCD characterisation was to figure out if my flats were taken correctly. In order to do that, you need to know the Full Well capacity of your CCD (it may not necessarily be what the vendor reports) and its linearity. You don't want to take flats that are in a non-linear portion of the signal. For example, my 1001e is non-linear from 0 to 10,000 and 55,000 to full-well (64, 861)DN. So when taking flats, they need to be between 10,000 and 55,000 DN* for sure. I usually go for about 2/3 DN or 35-40K DN. But the whole project meant I knew (scientifically) what that range was.

As it happens, when I went through the process I found out that the Kodak 1001e chip came in two flavours of A/D circuitry with different gains. I quickly found out which flavour mine was.

Dave

*DN AKA ADU (Digital Number AKA Analogue to Digital Unit)

dmcdona wrote: As it happens, when I went through the process I found out that the Kodak 1001e chip came in two flavours of A/D circuitry with different gains. I quickly found out which flavour mine was.

To clarify - I was using the FLI spec sheet which stated a full well capacity of 500000. I never could understand why my flats were rubbish until I constructed the PTC/DTC curves. I found out my full well capacity was actually 147660 - less than a third of the quoted value. Why? the 1001e actually has two circuits - low gain and high gain. It seems it is the camera manufacturer's electronics that determine which CCD circuitry is used and therefore what your full well is in actual use(at least for the 1001e chip). I was a bit miffed that the spec sheet appeared to be misleading but didn't follow up since I'd completed the project and was able to take accurate flats.

So, never trust manufacturer's implicitly. In fairness, they are probably not deliberately misleading us - just hiding the fine detail that they know would melt most brains from 5 yards away. But its always good to know the truth of your equipment.

Dave

I completed a characterisation of my CCD camera awhile back. I also wrote it up explaining both the method and results. This report would be very useful to follow if you want to do the same thing and find out about everything about your own CCD camera I'm going to the Galway Astronomy Festival this evening. I'll bring a copy of this report if any body wants to use it.

Eamonn
www.kingslandobservatory.com

Here's a quick write up of my analysis. I didn't go too much into the methods as they are better explained on the stark-labs site.

www.cademuir.eu/blog/2014/01/31/analysin...timal-exposure-time/

Cheers,
~Al

My head is melting at the thought of doing this, all seems very confusing, think I need some red bull and a quiet room to contemplate doing this