Atik 383L+mono -- Image Gradient

[niteman1946]

Hi Group,

I'm wrestling with a problem and trying to determine if it is pilot error or hardware related.  My new Atik is showing a consistant problem with the calibrated lights.  The upper left hand corner of the image shows a flare that is much brighter than the rest of the field (excluding astro objects).
1.  Lights, even when heavily stretched, do not show this feature.
2.  Calibrated lights (next step) do show this feature.  And this persists through the rest of the process.

Here's a stretched Light sub (no flare):
http://www.astrobin.com/22229/

Here's the same sub after calibration (flare in upper left corner):
http://www.astrobin.com/22231/

Here is the Flat High Pixel Rejection.  Similar to the Low rejection and correlates fairly well with the feature in the calibrated Light:
http://www.astrobin.com/22237/

DBE can remove most of this feature, but I suspect that is just a bandaid.  I also feel that DBE raises the noise level (at least with the mono images).

Bias are integrated from 30 subs (0.20sec each at 0C, 1x1)
Flats are integrated from 30 subs (approx. 3.0sec each at 0C, 1x1)
Darks are integrated from 30 subs (same time as lights at 0C, 1x1)

a)  I've processed the Flats and Darks by integrating first and then selecting "calibrate" at the Light's calibration.
b)  I've also calibrated them first (i.e. subtracting Bias), and then de-selected "calibrate" at the Light's calibration.  No difference in results.
c)  Flats are done at dusk using the sky.  I try to keep the ADU value around 35,000 by varying the exposure (from 1sec. to 6sec.)
d)  Optical train is a 12"LX200, F6.3 FF/FR, Orion Skyglow LPF, and a Hutech OAG5.
e)  Light exposures are 600sec., 0C and 1x1 binning.

Here are the stretched and unstretched version of the Flat Master.  Individual Flat subs look very similar.  To my untrained eye they don't look too bad, but something's not right:
http://www.astrobin.com/22214/
http://www.astrobin.com/22202/

I hope the collective wisdom on this group can point me in the right direction.

Thanks,

Mark
 

[Nocturnal]

Hi Mark,

unfortunately all those stretched images don't really say a whole lot. For example the stretched light is completely black where the flare appears in your calibrated image. I see two possible causes:

- your darks don't match your light. You light perhaps have amp glow and the darks do not. After subtraction you are left with residual signal in that corner.
- your flats don't match the curve in your lights. How do you take your flats? If you use an artificial light source have you taken two sets with the light source rotated and verified that divided you end up with a white image?

I recommend uploading a single light, flat, bias and dark as they come from your camera.

[Nocturnal]

Oh, since DBE is subtractive it does not alter the noise in your image. But that's a different topic for a different thread.

[cfranks]

Hi Mark,

I think your Flats might be over exposed.  My QSI 683 uses the same chip as your Atik and the specs say Full Well is ~ 25500 ADU.  You are trying to get 35000 ADU.

Charles

[niteman1946]

Thanks Sander and Charles for your responses.

Charles,
I will try your 25k ADU recommendation.  There's a lot of chatter of going higher and that's why I settled on around 35k.  But this is easy enough to try, hopefully this evening.

Sander,
Not having any luck getting fit files up into the cloud to share.  Any suggestions on how would be appreciated.

Mark

[topboxman]

I have gone as high as 35K ADU flats without issues for SXVR-M25C and Atik 460EX cameras. I typically aim for around 30K.

If you are using flat box, do what Sander said, rotate the flat box 180 degrees and see if the gradients also rotate 180 degrees. If that's the case, then your flat box may have issues.

I use T-shirt flats and never fails me. Sunlight provides the most even illumination.

Try upload your FIT files to www.sendspace.com .

Peter

[niteman1946]

Sander,
Here are the individual fit files for Bias, Dark, Flat and Light.  My apologies if they don't show up, it's my first try with Sendspace.
http://www.sendspace.com/file/ebyl4w
http://www.sendspace.com/file/i68yct
http://www.sendspace.com/file/9qvck6
http://www.sendspace.com/file/8dxful
Flats are done at dusk using the sky.  Currently there is no moon or bright area of the sky that might be causing mischief.  I appreciate you looking at these and offering any suggestions. 
And I would like to further that conversation about DBE and noise sometime.

Peter,
Thanks for the link to Sendspace (hope it works).  I had tried Dropbox earlier but could not figure out how to point to the files once they were uploaded.
As mentioned to Sander, my Flats are sky flats (no artificial lighting and no tee shirt).  I will try Charles' suggestion and go to shorter Flat exposures (i.e. lower ADU).  If that works, it would be an easy solution.
And to respond to your suggestion from the Atik forum to try another software ... I would if I had one.  I'll see if Nebulosity still has their trial download.

I'm also going to revisit my processing and see if there is anything there.  I'll post my findings when I have something.

Thanks guys,

Mark

[topboxman]

Mark,

What capture software are you using?

Peter

[niteman1946]

Images Plus v.4.50B

[topboxman]

Can your Image Plus do any processing like calibration?

Peter

[niteman1946]

Peter,

No, it's strickly image capture.


Mark

[topboxman]

Hi Mark,

I was able to download your images. According to Nebulosity, the Max ADU of your flat is 45K and Mean ADU is 35K. Your flat may be a little too high but I don't see anything wrong at four corners.

I used Nebulosity to perform calibration with flat, bias and dark. I do see some gradients at the same corner as yours. Then, I re-ran calibration but not use flats and I don't see gradients but I see four dark corners just like your light. The dark corners are likely vignetting from your optics/camera.

It seems to be something may be wrong with the flat. Try take some flats at different exposures and get Max ADU of 25K, 30K and 35K and see if it makes a difference.

Peter

[topboxman]

Your camera has a mechanical shutter and reducing exposure to lower ADU may show shutter shadow on your flats. You may need to expose a minimum of 2 seconds and maintain maximum ADU between 25K and 30K. Try use T-shirt method to take flats. You may need several layers of T-shirts to lower the ADU and above 2 seconds exposures.

I also use T-shirt method to take flats. My camera does not have mechanical shutter so I can expose as little as 0.05 second and still works fine.

Peter

[niteman1946]

Hi Peter, thanks for sticking with this.

I did some more flats last night and they provided better results.  The two things I changed were the ADU value and the length of their exposure.
1.  Exposure:  No less than 3sec, and as much as 13sec.  Previous flats were predominantly 1 and 2sec.  This was an attempt to eliminate any questions about shutter shadow.
2.  ADU:  I let some of the flat subs run around 22k ADU and some as high as 40k.  This tended to drive the average down.  The master flat was slightly darker and its stretched version featured a very dark field outside the hightlighted center.  I was pleased to see the Pixel Rejection images to not have the negative (reverse) version of the upper left flare.

I know changing two things at once is not a good way to diagnose the problem, but sky flats and somewhat longer exposures tend to run one out of time.  Need to change my technique for acquiring flats.

I suspect there is something quirky about the 383L+mono camera.  But likely one or both of the above actions can mitigate that tendency somewhat.

Here's the final result of the image on which I was working.  Got rid of most of the corner flare but feel I lost some of the faint nebulosity to that side.
http://www.astrobin.com/22358/0/?mod=none

[niteman1946]

I posted my message before seeing your follow-up.

Mark