Why are flats exposed at 30-50 percent?

[focaldepth]

Ok, just one question, why is it recommended to expose flats at 30% to 50% of maximum ADU?

I have just been through a hair pulling flats nightmare. (Sorry a character building learning experience)

I bought an astronomy EL panel a few years ago and thought it was fantastic but have always had random bad flat like variations in parts of my images, such as coloured patches that can't be DBEed out.

The panel had sheets of paper to get the speed up to 5 seconds (for my quaint shuttered CCD).

Recently the exposure times had crept upward reaching 15 seconds this week before, PHUT!, the panel (probably PSU) died.

This made me think!

1) I have always exposed to the right. (as my brain thinks it knows better than everyone else, despite me telling it otherwise)

2) I have always trusted the EL panel to be flat.

Anyway, having a heap of lights with suspect flats I had to do something, so I grabbed my LED light box (which is not very flat), stuck a pile of A3 paper on it, balanced it near the scope and fired away at 50% (1 second).

And POW most of my weird problems have gone.

I have also noticed that in photometry discussions, they place the flat light as far away from the scope as possible to keep the light rays as near parallel as the can. I allways placed it as close as possible to keep out stray light.

So now back to basics.

1) I am going to build/buy a new panel.

2) I am going to measure it's flatness.

3) I am going to expose the flats and dark-flats correctly.

4) I amd going to find the correct scope to light distance.

Hence my question. My brain says expose to the right but general consensus says below 50%, Googling the subject just adds confusion.

What is the reason for the reduced flat exposure level?

 

[thanders]

You should have linear response from your chip.

1) Not too much light to avoid saturation.

2) Not too less to have good signal to noise ratio.

/T

[focaldepth]

Thanks Thanders.

You should have linear response from your chip. (keep it below 80-90%)(CCD dependant)

1) Not too much light to avoid saturation. (Expose to the right)

2) Not too less to have good signal to noise ratio. (Expose to the right)

[skybadger]

Most chips become non- linear below 50% saturation, either because of the anti blooming gate or due to charge diffusion. Hence keep the charge around 30%.

[focaldepth]

Ok so it is for nonlinearities. (Thanks Thanders and Skybadger)

I had dismissed that because the 8300 chip I am using is reported to be very linear even as its buckets fill up. Figures of 1%, or less error in the top 10-15% of it's range.

So am I right in assuming that the 30% is a catch all for those non linear chips out there and the more linear chips can be pushed to longer exposures?

Sorry if I am asking daft questions but having been through flat issues I am anxious to get it right in the future.

I am in the process of comparing different flat exposures to see if I can detect any difference, although that will be with my makeshift flat light for now.

[ollypenrice]

Using a Starlight SXVH36 in a big reflector I found that the only way I could get flats which didn't over correct was to come right down to about 18000 counts or less. I have no explanation...

Olly

[davew]

Trouble with a lot of the answers you'll get is that most people aren't talking about full frame sensors. Now that you've mentioned you have an KAF 8300 chip then check out the QSI knowledge base and look at the 583 or 683. You didn't say what camera you were using so your gain may be different. I tend to aim at just below 35000.

http://store.qsimaging.com/kb_results.asp?ID=28

The H36 chip is interline.

For a cooling summary look at -

http://store.qsimaging.com/kb_results.asp?ID=21

Does that help more ? They've done all the hard work for us

Dave.

[focaldepth]

Thanks olypenrice and davew.

I started thinking that the 30% rule would apply to all set-ups, and it did not make sense to me. The QSI site does show that it is not as simple as I initialy thought.

I guess I should have mentioned my set-up, it is Sky Watcher Quattro 10" F4, Starlight xpress H18, Starlight Filter wheel.

I have now grabbed some flats at different exposures and will apply them to some shots of M31.

It will be interesting.

I am also wondering how important it is to ensure the flat light does not iluminate the insides of the scope and draw tube etc. Too many theories and not enough time to experiment.

Althought there is now a picture forming as there is a lot more to this than a simple rule of thumb.

.....