Flats & ADU madness with the red filter

[swag72]

This morning I came out to do my flats with the red filter that I'd been using all night. I set up as usual, using my EL panel and Artemis. I generally go for an ADU of about 23k.

I'd done red flats before and it was around the 0.13s mark. That was a little low, so I started adjusting the exposure to get to the right ADU. I was very surprised to find that I could not get to 23k, instead only 21.7k or 34k. Here's how the figures were stacking up.

0.13s - 19.5k

0.132s - 21.6k

0.133s - 34.2k

There's no way that 0.001s difference should make over a 10k difference. This was tried numerous times and still I got roughly the same ADU average.

I even went so far as 0.1325s - 21.8k and a tiny increase in shutter speed to 0.1326s gave me an ADU of 34k.

I swapped the filters over and the green behaved impeccably and the blue had worked fine the night before.

Can anyone suggest what on earth is happening here? It's got me well and truely confused and I've never had a problem like this with red flats before either.

[jnc71106]

i've had issues like this before with artemis where no matter what i did i couldnt get it high enough. i solved my problem by closing the program powering down the camera and restarting it all

[swag72]

I'll have another go when I am next out and see if the issue is still the same. I can't help thinking if it was purely a power down and restart issue I'd have found the same with the green filter that I tried straight after?

[opticalpath]

That's weird and I don't know the answer. A few thoughts occurrred to me though:

- Is the electronic shutter accurate and repeatable to 1/000 second? With exposure set to 0.132, is it possible that the actual exposure achieved each time is 0.13n, where n is uncertain and varies. Or is the camera rounding up the third digit of the exposure value set in some not-obvious way? Are your green and blue flat exposure times also expressed to 3 decimal places?

- Could the capture software have the same issue - i.e. could the program be losing a significant digit somewhere between reading the exposure value set, and sending it to the camera?

- Is there a gain setting on the camera that might be interacting with the exposure setting somehow?

- Could there be a memory effect, the sensor not flushing fully between exposures and subsequent exposures being affected by residual electrons in the wells from the previous exposure? Some cameras have a setting for how thoroughly the flushing is done after each image.

I would try taking a sequence of exposures at the same setting and see if there is any significant variation in ADU between them. Try a different way of measuring average ADU over the same sub-frame area in each image just in case something weird is happening in the measurement. Try exposures of 0.110, 0.120, 0.130, 0.140, etc and see if you get expected ADU levels when exposures are set at hundredths instead of thousandths of seconds.

Hope you get to the bottom of it!

Adrian

[davew]

After that excellent post all I can suggest is, as you've got Maxim, try that and see if the problem repeats.

Dave.

[swag72]

I had another go last evening, then uninstalled Artemis and reinstalled it. I'll see how it goes next time. That was a brilliant post Adrian,and I will definitely do a sequence of exposures next time, then probably fire an email to Atik. As far as I know there's no settings on the camera that can be changed for flushing or gain.

Will keep you posted!

[Mike_F]

I had another go last evening, then uninstalled Artemis and reinstalled it. I'll see how it goes next time. That was a brilliant post Adrian,and I will definitely do a sequence of exposures next time, then probably fire an email to Atik. As far as I know there's no settings on the camera that can be changed for flushing or gain.

Will keep you posted!

Will be interested in your findings. My 460ex and Artemis exhibit exactly the same problem. I just can't get to the ADU value I'm after with the red filter in the train. Blue and Green are both fine.

[swag72]

That's interesting Mike. Have you ever emailed Atik about it? What filters do you use? Mine are Baaders

[Mike_F]

That's interesting Mike. Have you ever emailed Atik about it? What filters do you use? Mine are Baaders

Yes - also using the Baader LRGB filters here.

No, I haven't emailed Atik about it, I'm too busy using the miserable allocation of clear nights to sort out guiding issues at the moment (turns out to be a nasty case of focuser flex in my Equinox 80 apo pro - so am currently converting to OAG)

I had planned to give it a go in Nebulosity at some point to see if it was Artemis rather than the 460 at fault - but haven't quite got around to it yet. I'll keep you posted if/when I do.

[swag72]

Could it possibly be something to do with the filters? If it only happens with the red in line and the green and blue are fine, then that suggests to me that it's not the camera or the software otherwise it would happen with all filters. Neither the 460 or Artemis knows what filters are in the train.

Perhaps I'll drop Baader a message. Anyone else concur with it potentially being a filter thing based on this?

[ollypenrice]

Apart from its curiosity value does this matter? Anything between about 13,000 and 40,000 should give you workable flats.

Olly

[swag72]

I used flats for the red channel at 21k and I have to say that they didn't seem to work very well, whereas 23k works fine with the other filters.

[RikM]

I use the Baader filters with a SXV-H9 and haven't had any issue with flats for red. I normally go for an average of 20K though, as that seems to be the magic number for me.

[Tim]

Intriguing Sara but not the end of the world, the flats at 34k would still likely have worked fine?

[Mike_F]

Apart from its curiosity value does this matter? Anything between about 13,000 and 40,000 should give you workable flats.

Mm, indeed. To be absolutely honest, that's why I hadn't given this issue much focus thus far.

My mindset when I encountered the problem was "meh.. I'll just re-expose the L,G & B and make them all ~24K to match each other once I've got some usable subs." My guiding issues definitely have my undivided attention right now - but it was very interesting to see someone else having exactly the same problem with mean ADU values in the red channel. To my mind it rules out a number of software/installation variables..

[opticalpath]

It's very strange, Sara. If the 'non-linear' behaviour only happens with the red filter, I would try a sequence with identical exposures on all three filters to see if you still only get the unexpected results with red. The average ADUs will be different of course, but it would eliminate the possibility that the non-linear effect was something odd to do with the particular exposure times you set for red.

How are you measuring average ADU? A central sub-frame measurement, rather than the whole frame average, would eliminate the possibility of uneven vignetting (filter not recentring reliably?) affecting the whole frame average ADU.

The EL panel must flicker at the high voltage supply frequency (about 1 kHz?) - I don't suppose that would cause odd effects but it would be interesting to compare results with an alternative light source.

Clutching at straws here!! My bet is still that it may be something to do with exposure times set to 3 decimal places of seconds. Try some rounded up times (0.11, 0.12, 0.13 etc) and see if the effect is still there.

Adrian

[swag72]

Regarding the 3 deciaml place exposure times, the reason I went to such decimal place madness was because the ADU was changing from about 19k at 0.13s to 35k at 0.14s. My inital thoughts were that if the change was that sudden (wholly unexpected based on my green and blue filter experience) then there would be a point at about half way, hence 0.135s where the value would be about right. That's where I discovered the massive jumps.

I've dropped an email to Atik as when I looked back over 314L+ colour flats I never had an issue, so in my mind that rules out the filter, as the red is well behaved with the other camera. The only difference between my 314L+ and 460EX flats in the red channel is the camera.

[Merlin66]

Sara,

Is it possible to "tone down" the flat screen to allow a longer (and probably more controlled expsoure)???

[ollypenrice]

It has never occurred to me to make my flats at exactly the same exposure value as each other and I've never ever done so, so all the pictures I've ever posted have been calibated by flats varying over a range of two or three thousand ADU. Next time I take flats I'll take two different values for one channel and compare their effect on the image. Given that the flats application is by division I'm near certain that it will have no effect whatever. Clearly this is not comparable with darks which are very sensitive to temperature and work by subtraction. Even a slight mismatch in ADU values due to temperature (or anything else) is a disaster. But flats? I'm betting it is totally unimportant.

Olly

[swag72]

1360053717[/url'>' post='1815415']

Sara,

Is it possible to "tone down" the flat screen to allow a longer (and probably more controlled expsoure)???

I will try that Ken and see how it goes.

My flats figure does vary Olly. I use Artemis white point figure to get an average the just run off a load at that exposure. It changes a little, but I'm not bothered by that.

[martin_h]

Sara,

Is it possible to "tone down" the flat screen to allow a longer (and probably more controlled expsoure)???

I would agree with this, those exposures seem very short, try a couple of sheets of A4 over the screen and get the exposure time up into the whole seconds.

[opticalpath]

Something here triggered an old memory. Back in the early days of the Artemis camera, forerunner of the rebranded Atik range, there were reported problems of 'bleeding' when imaging bright sources at short exposure times. There was a big discussion about it on the Y! group back then: http://tech.groups.yahoo.com/group/ArtemisCCD/message/1316

May be nothing to do with it, but I wonder if there could be something like this going on. It will be interesting to hear what Atik say.

Adrian

[opticalpath]

Sara,

Is it possible to "tone down" the flat screen to allow a longer (and probably more controlled expsoure)???

If your EL panel is the version that is powered by a 12V inverter (rather than mains powered), you can vary the brightness to some extent very easily by reducing the supply voltage; I usually run mine at 9V and it works even at 7.5V, giving a useful reduction in brightness.

Adrian

[swag72]

I can dim my panel as I have a variable current adaptor - I'll put a couple of pieces of paper in front and / or dim the panel and see where I go. Interesting link there as well.

I've not heard from Atik yet, despite emailing them yesterday morning - I hope they do get back to me.

[Martin-Devon]

Sara, when I purchased my EL panel I had it pre-fitted with two diffusion sheets to dim the intensity. With my Atik 460EX, I need 3.7 second exposures with a Red Baader filter to give an ADU of 20K. So as you suggest, I think you may get more accurate and reproducible ADU figures by dimming the light source, this way the shutter on the Atik is not pushed to the limit.