Measuring CCD Linearity

mjc wrote: However, I really don't think that this Intensity value is what you want for the task at hand.

Mark,

It is PRECISELY what we desire since the intensity tells us the behaviour of the 100x100 pixels as a group (or the star if using this route). Now, double the exposure and the individual as well as the collective behaviour of the 100x100 pixels (or star) should also double etc etc etc etc.

Anthony.

Gents,

It is unfortunate that one cannot upload (attach) Excel spreadsheets or, otherwise, I would have provided you with my linearity testing and results from a few years ago and where everything including data and graphs is laid out when I sought to do the same.

Anthony.

Folks - the last week has been a fairly steep learning curve for me. I still didn't get AIP4Win, but I *did* get a lot of help on another forum.

The technique is easy enough but it can be quite time-consuming. But by generating a Photon transfer curve and a Dark transfer curve, I've gleaned some interesting data about my CCD.



The full well is about 148K e-
Gain is about 2.3 e-/DN
Read noise is about 15.9 e-
Linearity worked out pretty good - the CCD seem very linear between about 10 and 90% full well.

I have to get more data at the lower and upper ends because what I have is too sparse to give clear cut-off points, particularly for full-well.

The data collected has many more uses too - I'm going to figure out what they are and start plotting lots of curves

Cheers
Dave

Glad you're getting somewhere.

You might want to revisit your readout noise - that looks a little high.
Do you know what the spec says so that you have a benchmark to compare your results to?

I have an Atik 16IC - a budget CCD. Its spec has readout noise at 7e.
When I tested I got something less - but I'm not happy that I got the gain right so I have to recheck (I'm going to try Janesick's method - which is detailed in the HAIP). Linearity was very good with a very sharp deviation from linearity somewhere near to - but above - 52000 ADU's which is at least 79% of ADU range. I think that's okay for a chip with ABG. I was expecting deviation from linearity to start earlier - but less severe. So I'm quite pleased. However, it is still subjective and I suspect that one would need to perform some mathematical test to quantify the conformance to linearity. The term "least squares fit" seems to apply to such situations - but I think that's for another day.

It was a - what's the word - polava? - to get usable data. It took many attempts.

My dark current measurements where unbelievable - (I don't believe them) - 1/100 of what others have measured.

Thanks for your original post - made me complete what I started back in November (you can tell I don't take out the scope much...)

Mark C.

Mark - the spec from Kodak is actually 15 e- for readout noise. So looks like its within spec, allowing for the fact that my data needs to be refined a little more.

I'm hoping to scrub the PTC tonight if possible, with new data. I'll see if that makes a difference.

I also still need to analyse the data a bit more to construct more of the parameters and then compare them to the spec. But judging by what others have said, the 1001E seems to preforming pretty well.

There are still some question I have out to Finger Lakes about the settings - the 1001E has two amps on board - its unlikely they are user selectable but you can choose (from the camera manufacturer) a full well of about 200K or 500K - its probaly then hardwired when the CCD is put on the board. I got about 150K so I've asked FLI if there is any discrepency. Probably not. Besides, 150K for 16 bits seems a good value to choose (via the gain) - if it was set to 500K, there'd be a loss of resolution - you're squeezing the electron count into the same 65535 ADU's.

Do you have the Janesick book? I was thinking or getting it. (what's the HAIP? is that the Berry book?)

I hear you on linearity - I have a chart but I'd prefer some kind of statistical description. I'm sure its out there somewhere. But as you say, another day...

Glad I helped to get you back on track And I hear you on the polava comment! But now I've done it a few times, its not that hard. I'd recommend others here have a go and characterise their CCD's scientifically, rather then just go on the manufacturer's data...

Cheers
Dave

Dave

I don't have either Janesick or S.B. Howell - I've had them in mind for purchase over the past few months.

I know I'd learn a lot of good stuff from Janesick - but I'm not sure that I'd get much from S. B. Howell over an above either what I know or what I can easily find out elsewhere - but it is a classic reference. I get from looking at the table of contents (via amazon) that Janesick looks more like an engineering book rather than one of practical application. Coming from a techie background that appeals.

As you have S.B. Howell can I ask if you feel that you got value out of it?

I don't like to post a reference to stuff that I haven't properly read - but I have quickly skimmed it - and it appears to discuss using two A/D converters to extend dynamic range. I'll try to read this tomorrow (if for no other reason than for my own education). It appears relevant. Looking up the title and authors in ADS I believe it was published in 1998. (Proc. SPIE Vol. 3301, p. 75-79, Solid State Sensor Arrays: Development and Applications II). An Aussie educational site has public access to it.

A Method for Extending the Dynamic Range of CCD Instruments

In case the link doesn't work (no preview) it is
rsaa.anu.edu.au/observing/detectors/ccdl...inearity/wide_dr.pdf

Re my comment about your readout noise. This maybe related to the gain (I'm out of my depth now). Your gain - at about 2.2 e / ADU compared to my approx 0.2 e / ADU (roughly five ADUs per electron - I have a small full-well capacity!). Bottom line - your readout noise maybe really quite acceptable.

I'm on a journey of discovery...

Mark C.