Measuring CCD Linearity

Hi Mark - the software (or rather, what it reports in terms of numbers) is where the issue lies.

For example, with a star selected in an image using an aperture, the MaximDL info window reports the following:

1. Pixel
2. Maximum
3. Minimum
4. Median
5. Average
6. Std Dev
7. Magnitude
8. Intensity
9. SNR
10. Bgd Avg
11. Bgd Dev

The first 6 are given in ADUs (though no units are given for *any* of the numbers. In the case of my CCD, max is 65535 and that's the max value I ever see in these 6 parameters. So it must be ADU.

7 is related to some internal calibration data so safely ignored.
8 is wierd. No units give and it *does* exceed 65536 (it regularly hits 6 figures alright). Could be elctrons but who knows...
9, 10 and 11 are easy enough.


The help file gives nothing at all away...

So - Anthony says "flux", you say "peak value" but I'm none the wiser... Hope this makes sense to you though I do realise you (and Anto) do not use MaximDl.

Magnitude

ayiomamitis wrote: it is a trivial exercise to get the flux of the complete 100x100 "image"....

Trivial if you *have* AIP4Win...

I just received confirmation that my book (and software) will arrive within the next two weeks.

Out of interest, I *did* analyse some of my images and, I could be completely wrong given the lack of units given by Maxim, it seems I'm linear to about 60K ADU. Pretty good since max ADU is 65.5K.

But here's another troubling issue. The CCD data for the 1001E chip varies depending on where you go. This data of course matters when doing characterisation studies. For example, I have Gain values of 2.0 and 2.2; I have full-well values of 150K, 200K and even 500K.

If I assume an e/ADU of 2.0 and a full-well capacity of 150000, that gives me 75000 pixel values. But, with a 16 bit AD amplifier, I can only "see" 65535 discrete values. So I assume 9465 values are just discarded. Perhaps that is why I'm seeing a linearity of (apparently if my calcs and assumptions are right) across 90%plus of the CCD's value range.

I can't beleive a full-well capacity of 500K. That's just mad. Unless I've been looking at data sheets for different flavours of the same 1001E chip. In that case, I'll need to go back to basics and calculate e/ADU, full-well capacity etc. I know that this can be done but I fear yet another (minimum) 24 hour period of me, prone in the darkness, with a damp face cloth on my forehead...

I really apprciate you guys weighing in to help out. I know its complex and I know its not easy when we're comparing apples to oranges to lemons....

AIP4Win should level the playing field anyhow (or the flat-field if you prefer...)

Mark/Dave,

Another word for "flux" is "intensity" and this is certainly available in the information window of Maxim/DL and in the right-hand side column (by the way, I do use Maxim/DL).

What is "flux" and/or "intensity"? It is the sum of the pixel values of an image OR of a star. It is something which I study prior to beginning an exoplanet session, for I want "good" flux, low background and non-saturated pixels and which translate to good S/N.

Anthony.

Dave I just had a quick trawl and found some tutorials on the Cyanagen site
in which the following dialog box was presented.



From this I gleam that the pixel value is the value of the pixel under the cursor.
Maximum - or peak value is the maximum value of all the pixels in your aperture
(This is the one you wan in my opinion).

Minimum - the smallest value
etc etc

On this screen shot Intensity is next to SNR and are both zero here so I think Anthony is right and Intensity is a measure of flux - and I suspect is reported in photo-electrons which needs calibrating with a known value for gain somewhere. There is numeric values for background so there's no reason for the screen shot to have zero for intensity if it were intended to be in ADUs. Not sure what its missing that prevents a measurement of SNR though. Now I could be wrong but that's my interpretation of this screen shot.

I'd choose Maximum form your list.

I'm quite comfortable that that is the one to go with.

BTW that other stuff like gain and full-well capacity (expressed in electrons) can also be measured - and from the same data that you are collecting - all explained in The Handbook of Astronomical Image Processing.
I can post details of method or you can wait on your book.

Also BTW - I've been trying to order SB Howell's book since October from a bricks n Mortar store with no joy.

Mark C.

Anthony - that's good information. I guess the trouble is whether we are talking photons (or electrons) or ADU's. That's been the difficulty I've had with Maxim. I can sure tell that the pixel values are in ADU's but there is no information anywhere as to what the units for intensity are. Presumably, they are in number of electrons.

*However* - I'm not sure how Maxim arrives at "intensity". For example, I know (or at least various data sources tell me) that the gain (e/ADU) of the KAF1001E is 2.0. So, whatever number of electrons are read by the A-D converter, they are recorded as e/2. If a well contains 100,000 electrons, this is delivered as 50,000 ADU's (via the ADU and amplifier) . But, there is nowhere in Maxim that I have told it that the gain of my CCD is 2.0...... So what does that intensity number represent? Again, nothing in the MaximDL help files at all. Nothing in setting that I can see and certainly nothing in the FITS header that gives the gain of the CCD.

To be honest, I'm very new to a lot of this and may not be reading things right. But based on what I've read, I think I have the general gist. But I'm coming up with far more questions than answers.

On another forum, I have been informed that the methodology used by professionals is to contruct a Photon Transfer Curve - PTC. So my next direction is how to do that. It will give me empirical data that can be used to determine the precise gain and full-well depth of my CCD. Given the variability in manufacturer's data for gain and full well depth, seems its about the only method that doesn't rely on external data.

I'll get to the bottom of this and as ever, will share with you all.

Thanks so much for all your help - it really is appreciated.

Dave

Mark/Dave,

Please note that "flux" (aka "intensity") is in ADU's !!!!!

This is also the case for the Maxim/DL information window referred to earlier. I am 100% certain about this.

As to its measurement, it is the sum of the pixel values within the annulus. Although I am not very impressed with the measurement of linearity using a single star, I suspect it would lead you to the general direction.

Anthony.