383L+mono -- Image Gradient

[niteman1946]

Looking for your thoughts on this.  Here are four images from my CCD.  Apologies for how this is presented.

1.  Image of NGC7293 freshly calibrated, aligned and integrated - stretched, but no processing.  This is an integration of 52 subs at 10min. each and 0 degrees C.  Notice the upper left corner.  It is noticeably lighter than the other three corners.
http://www.astrobin.com/22199/

2.  Bias Master.  Made up of 30 subs, all at 0 degree C and 0.2sec (Images Plus limitation). These were shot in the dark.  Heavily stretched.  Notice no significant variation from corner to corner, although there is a strong gradient from bottom to top.
http://www.astrobin.com/22200/

3.  Dark Master.  Made up of 30 subs, all at 0 degree C and 10 min. each.  Heavily stretched.  Notice no significant variation from corner to corner.
http://www.astrobin.com/22201/

4.  Flat Master.  Made up of 30 subs, all at 0 degree C and about 3 sec. each.  Heavily stretched.  Bright spot is slightly below center, but otherwise there is no significant variation from corner to corner.
http://www.astrobin.com/22202/

The lighter upper left corner in image no.1 shows up in all three targets (NGC6946, NGC6888 and NGC7293) that I have shot so far.  Subs for no.1 image were shot towards the south at about 35 degrees elevation and no moon.
DBE in Pixinsight does a pretty good job of normalizing gradients, but (IMHO) also adds noise to the image.  So I try to keep from pushing this operation too much.

So far I'm pleased with the camera, but am curious if this is a known issue, or if I'm simply missing something obvious.

All subs were shot with the camera powered using an Atik 120vac to 12vdc converter.  All subs were shot at binning 1x1.

Thanks for your help,

Mark

[chrisjbaileyuk]

Mark

Most such illumination issues come from poorly acquired flats.

I realise they are stretched but your flats looks wrong to me. What did you use as a light source and what does the unstretched histogram look like?

Chris

[niteman1946]

Thanks for taking the time Chris.  The flats are done at dusk as sky flats.

This is the Master Flat in unstretched mode:
http://www.astrobin.com/22214/

This is an Individual Flat sub in unstretched mode.  It has an ADU value of about 29,000 in the corners and about 38,000 in the center (according to Images Plus):
http://www.astrobin.com/22215/

Here are the unstretched histograms from PixInsight.  The left is the Master and the right is the Individual sub:
http://www.astrobin.com/22216/


Mark

[niteman1946]

Chris, looks like your instincts are on target.

Here is a stretched Light sub.  No flare in the upper left corner:
http://www.astrobin.com/22229/

Here is the same Light sub after calibration.  The upper left corner flare shows up.  Calibration includes Bias, Darks and Flats:
http://www.astrobin.com/22231/

Any suggestions would be greatly appreciated.  What about the Flats looked wrong to you?

I will also take this over to the PixInsight forum to see what they think is happening - after I fiddle a bit with calibration.

Mark

[niteman1946]

Well, I've had no luck eliminating the flare feature in the upper left corner.  I've done some tweaking with the process, but have seen no improvement.
I did stumble on a couple of supporting data for the Flats.  These are PixInsight's presentations of pixel rejection resulting from the Flat integration.  Both the high and low rejection images show the same pattern.  And this matches up closely with the corner flare.
http://www.astrobin.com/22237/

Any thoughts?  Could this be an image chip problem?

Mark

[topboxman]

Interesting. Try calibrating with different software like Nebulosity, Maxim or Artemis.

I use PixInsight for stacking and post processing but use Nebulosity for calibration. I find it easier to use Nebulosity plus I like Nebulosity's Bad Pixel Mapping.

Peter

[niteman1946]

Just to close this loop, I believe that the fix is in. 

Joel Short from another post had experienced similar problems and had solved them by going to longer flat exposures (I believe to 3 seconds).

Thanks to his encouragement and Peter's (Topboxman), I lengthened my subs to 5 seconds (if a little is good, more is better).

I also used the Tee shirt technique and shot during the daylight.  This is sooooooo much easier than doing dusk shots.  Here I easily kept the center value of the flat subs around 31k ADU.

The Flare (at least for now) is gone.

Thanks again,

Mark