A problem with flats

[mackiedlm]

I had first light with my new Sharpstar 61EDPH II last night and have a problem with the integration in that there are curved lines showing over the image.

Marked for clarity

These are screengrabs of the autostreched masterlight straight out of WPBB (fully integrated in Pixinsight with flats, dark flats and a master dark)

When I inspect the flats and masterflat, it is clear that these curved marks are associated with dust bunnies but they do not seem to exactly line up with the bunnies in the lights and therefore have not been properly corrected.

Capture details Sharpstar 61 EDPHII (275mm) ASI2600MC pro, 2" L-enhance placed in a ZWO Filter holder (taped up to ensure no light leak),

Flats were taken using APT's CCD flat tool, with ADU targeted at 35000 - led flat panel in front of several layers of white fabric (used on many occasions with my 80 ED with no such issues). Flats were taken immediately after the imaging session without moving the scope or changing the optical train.

My current thinking is that the moon was very bright, I was probably too close to the moon for the L-enhance, and that some type of reflection or shading has happened to make the dust bunnies larger in the lights than in the flats. There are also wicked gradients that I cant seem to tame.

I have put the master light, master flat, master dark together with a couple of single subs, single flats and single dark flats into this folder if anyone would like to look at them to  help me diagnose why this has happened and if there is a way to resolve it (new flats taken a different way?)

https://1drv.ms/u/s!...X_qGmi?e=NDOaCH

If taking new flats will not fix this is there any to fix this in processing.

Thanks for any help.

[wimvb]

Dust bunnies, as caused by small particles, tend to be round. What you have is, I suspect, a fibre of sorts in you imaging train. Fibres are larger and move around more. You might want to check the cause and remove it. Calibrating out anything that is likely to move around, is almost impossible.

[mackiedlm]

Thanks Wim, but if you look at the lights and the flats youll see that the curves are concentric to the dust motes. Also, they dont show up as separate moving lines in sequential lights. its only the dust bunnies that show. So what seems to be happening is that most of the dust mote is being calibrated out leaving only part of the circumference of the original mote.

I appreciate your help.

 

 

[ollypenrice]

Something dark in the calibrated image will be created by something light in the flats, if the flats are the cause.  I can't download large files with our rural French internet connection but it would be good if you could post a JPEG of a stretched flat. I'm wondering whether the bright light source used for the flats has created a reflection from, say, the edge of the filter and that this is being shunted around back and forth as an internal reflection. The artifacts seem to be segments of a consistently sized circumference.

Olly

[wimvb]

40 minutes ago, mackiedlm said:

Thanks Wim, but if you look at the lights and the flats

Ah, ok. I could only see you image in the original post, as I am reading this on my mobile device.

[mackiedlm]

14 minutes ago, ollypenrice said:

Something dark in the calibrated image will be created by something light in the flats, if the flats are the cause.  I can't download large files with our rural French internet connection but it would be good if you could post a JPEG of a stretched flat. I'm wondering whether the bright light source used for the flats has created a reflection from, say, the edge of the filter and that this is being shunted around back and forth as an internal reflection. The artifacts seem to be segments of a consistently sized circumference.

Olly

Thanks Olly,

So below is a master flat, a single flat and the full integration (this is out of APP because I ran the data through that too see if it calibrated differently - it didnt!!)

I've tried to screengrab them at the same size to allow for comparison of location and you can see that the marks on the integration correlate closely to the left circumference of a dust mote. They are not exact line ups, rather the marks in the final are offset to the left of the dustmote and the calibration has done a fairly good job on the rest of the dust mote. On the left hand side where the image is darker I dont see any of the marks but that may just be because the background is darker.

single flat

Masterflat

Final

[Laurin Dave]

Looks to me like your filter has moved slightly between lights and flats or the camera has rotated slightly with respect to the filter.  Also in my experience the moon makes flat fielding trickier.  I'd make sure everything is screwed up tight before your next session.  

Nice image by the way,  to fix I'd remove the stars and sort the arcs in photoshop, then put the stars back 

Dave

[mackiedlm]

Just for completeness this is a single light but stretched further than the images above just to see whats going on.

[mackiedlm]

6 minutes ago, Laurin Dave said:

Looks to me like your filter has moved slightly between lights and flats or the camera has rotated slightly with respect to the filter.  Also in my experience the moon makes flat fielding trickier.  I'd make sure everything is screwed up tight before your next session.  

Nice image by the way,  to fix I'd remove the stars and sort the arcs in photoshop, then put the stars back 

Dave

Thanks Dave, yes I see how that could happen. But the filter is in the drawer, screwed tight and, because I worry about light leak, the drawer is then taped shut with electricians tape so I cant really see that. I'll check everything is tight but all I did from last light to first flat was park the scope - so I'm struggling to see that either.

 

And yes thats what I'm thinking of doing. I got rid of the worst gradients in APP, (PI failed miserably!) but its still not great.

Edited January 17, 2022 by mackiedlm

[Laurin Dave]

Strange then... was there a meridian flip involved?  if so maybe integrate pre and post flip subs and see if there's a difference, its possible something shifted on the flip.  Also your single sub doesn't look too bad but that I believe is because its only one and integrating however many you have would show the dust bunnies more clearly.. 

[carastro]

I have had similar arcs in images where the flat have failed to fully correct the dust bunnies.  I think you are correct in that it is related to the dust bunnies.  IN some cases I was unable to remove them because they were too overpowering for the flats to cope with and probably because I imaged in less than perfect skies i.e. too much LP and as Laurin Dave suggested, the Moon making this more difficult. 

Not sure what to suggest as it may not happen again.  

Carole

 

[carastro]

Quote

was there a meridian flip involved?  if so maybe integrate pre and post flip subs and see if there's a difference, its possible something shifted on the flip.

Was wondering that too.

Carole

Edited January 17, 2022 by carastro

[mackiedlm]

22 minutes ago, Laurin Dave said:

Strange then... was there a meridian flip involved?  if so maybe integrate pre and post flip subs and see if there's a difference, its possible something shifted on the flip.  Also your single sub doesn't look too bad but that I believe is because its only one and integrating however many you have would show the dust bunnies more clearly.. 

Thanks Dave and @carastro but no, the clouds did not lift till nearly nine so it was past meridian by that stage.

 

[Laurin Dave]

One more thought would be to look at the first and last calibrated registered subs and see if the arcs move..  this may show whether the filter is moving as the scope tracks

[mackiedlm]

14 minutes ago, Laurin Dave said:

One more thought would be to look at the first and last calibrated registered subs and see if the arcs move..  this may show whether the filter is moving as the scope tracks

Good idea. They are not so easy to see in the individual subs but they are there and are not moving around.

[ollypenrice]

Having copied and pasted the JPEG onto an equalized version of the California image I did finally concur that the bunnies often correlate with the dark arcs. However, there are some dark arcs without correlating bunnies and I did have to offset the flat a good way to the left to line up arcs with bunnies. (The equalize adjustment in Ps is really a diagnostic tool for greatly exaggerating local contrasts and made the arcs jump out.)

This offsetting to the left sets a mental alarm bell ringing when I also remember that you use a filter drawer. If the bunnies are caused by dust on the filter and the filter was offset slightly to one side when the flats were taken, then this offsetting of the bunnies between flats and lights would be exactly what we would expect.

Here's a screen grab showing the extent to which the flats had to be offset to get the bunnies aligned with the arcs.

An offset filter draw when the flats were shot seems to me to be a good fit with what we're seeing... It's not a perfect fit but these things rarely are.

Olly

[mackiedlm]

Thanks so much for doing that Olly. I can see exactly what you are getting at. As I said to Dave above, the drawer was taped in to avoid light leak but I did send the scope to park remotely (via my tablet from my bed!!) so there has to be the possibility that it moved during that skew. Although I did go down and run the flats immediately, I didnt think to check the filter drawer.

If we take the image width as being equal to the sensor width (25mm), then I'd calculate that offset as being less than 1 mm movement of the filter drawer.

The rig is still set up exactly as I left it this morning so I'll take a look when I get home. Its about the only thing thats making any sense at the moment. I'll report back.

Thanks again.

 

David.

Edited January 17, 2022 by mackiedlm

[mackiedlm]

17 hours ago, ollypenrice said:

Having copied and pasted the JPEG onto an equalized version of the California image I did finally concur that the bunnies often correlate with the dark arcs. However, there are some dark arcs without correlating bunnies and I did have to offset the flat a good way to the left to line up arcs with bunnies. (The equalize adjustment in Ps is really a diagnostic tool for greatly exaggerating local contrasts and made the arcs jump out.)

This offsetting to the left sets a mental alarm bell ringing when I also remember that you use a filter drawer. If the bunnies are caused by dust on the filter and the filter was offset slightly to one side when the flats were taken, then this offsetting of the bunnies between flats and lights would be exactly what we would expect.

Here's a screen grab showing the extent to which the flats had to be offset to get the bunnies aligned with the arcs.

An offset filter draw when the flats were shot seems to me to be a good fit with what we're seeing... It's not a perfect fit but these things rarely are.

Olly

Hi Olly and others, so I went home and checked the filter drawer and it was rock solid, I could not see anyway it had moved.

I then made a stack but without calibration frames (no flats or darks) and the dark arcs are present (infact they are darker than the bunnies) in the light subs before calibration - so are not infact related to the flats or the calibration. (Sorry, I probably should have done this earlier.)

This is the uncalibrated stack, heavily stretched and stars removed as it shows up the abberations so much better

And this is the master flat which I have adjusted to try to bring out the dust and the vignetting

I then checked in stellarium and when the lights were taken the bright full moon was to the right and just below centre of the light image. It was only about 45 degrees from the target. I think this probably accounts for the heavy dark shadow on the right - cast by the dew shield. Yes there is some vignetting in the  flat but not nearly as bad and it does not match the shadow on the image which is lighter at top/heavier at bottom.

My conclusions, but I'd really appreciate a sanity check from those of you who know better than me, are;

• These dark arcs are a result of the bright full moon, relatively close to the FOV
• This uneven bright illumination has resulted in the deep shadow (vignette) to right of image and by some type of reflection abberation, to the dark arcs which correlate with the edge of the dust bunnie furthest from the direction of illumination.
• The L-enhance flter is probably the physical location of the dust (ses this sound right according to the size of the bunnies??)
• The L-enhance filter with its highly reflective surface may be causing/adding to the effect
• The fact that no dark arcs are present in the flats is telling me that the issue is not intrinsic to the scope itself i.e the optics of the scope are not at fault. (This is really important because its a new scope and if there is a potential optics issue i should be informing the supplier right now rather than wait for later confirmation - it will likely be weeks before I get out again).

Thanks for all your help.

 

David

[Laurin Dave]

Hi...   Astronomy Tools   astronomy.tools     has a Dust donut Calculator which will confirm where the dust is,  looks like  you've posted a crop as it gave a distance of 13.5mm which is impossible given the 17.5mm backfocus of your camera. 

I'd agree it looks like the Moon was the issue, hopefully a moon free night will confirm this.  Another thought about your system, is there any way the filter can be inserted facing the wrong way, if indeed it has a right way.  Having the anti-reflective side facing the wrong way can cause issues.

[mackiedlm]

22 minutes ago, Laurin Dave said:

Hi...   Astronomy Tools   astronomy.tools     has a Dust donut Calculator which will confirm where the dust is,  looks like  you've posted a crop as it gave a distance of 13.5mm which is impossible given the 17.5mm backfocus of your camera. 

I'd agree it looks like the Moon was the issue, hopefully a moon free night will confirm this.  Another thought about your system, is there any way the filter can be inserted facing the wrong way, if indeed it has a right way.  Having the anti-reflective side facing the wrong way can cause issues.

Thanks Dave,  when I do that having measured the donut in PI I get a distance of around 11mm - which actually makes sense if the dust is on the camera window - NOT the filter. The 17.5 is to the front of the tilt plate and the window is set back some mm from the tilt plate.

No the drawer only goes in one way and there is only one thread on the filter so no way to muck it up - or else I surely would!!

 

 

[Shibby]

Having read through this, I would have to conclude the same as you - the arcs are some sort of reflection of the dust bunnies. Although it's hard to fathom exactly, it makes sense that the moon could be responsible, giving you light from a different angle than your light panel does.