Vignette Problems

[niteman1946]

Hi Guys,
I would like to get some feedback from Atik support, and also anyone from the forum.

My images suffer from vignetting, caused by light drop-off in the corners of the imaging chip.  I purchased my 383L+mono camera in August of last year and imaged through my 12”LX200 classic using a Meade f/6.3 reducer and Orion SkyGlow filter.  There was at times some vignetting (or something that looked like it), and this seemed to depend on subject and background brightness.

I received my EFW2 in December and started getting serious about both LRGB and NB in January of this year.  Here is the setup with the EFW2.  The photo shows the initial arrangement with an approximate f/6.11 focal reduction.
              http://astrob.in/30273/
I have since removed a short (7.5mm) spacer that sat between the OAG and EFW2, making the reduction closer to f/6.56 (I think).

I’m using the 1 ¼” mounted LRGB and NB Astronomik filters in the wheel.  According to Atik, 1 ¼” filters should not introduce relevant vignetting with my camera until reduction gets down to f/5.
 But for me, vignetting is very much a problem.  Here is a Line Profile comparison between a single sub shot through the Luminance filter (upper view), and one shot through an open port in the wheel (lower view). 
               http://astrob.in/51881/

Giving the Atik folks the benefit of the doubt, the f/5 value may depend on the optical train.  And mine might simply fall outside of their design intent.  Nevertheless, the line profiles (and visually the subs) do show a significant drop-off of light towards the corners of the chip.  And notwithstanding my best efforts with Bias, Darks and Flats, the only solution I’ve found for eliminating vignetting is surgery.

So, here are my thoughts,
a)   I believe that if I change out the filters (and filter wheel) for larger ones (36mm), the problem may be solved.  I’m basing this on what I see in the above line profile example where the open port allows more light path room.  However, this gets fairly expensive quickly.
b)   I am also considering shortening the component stack up further by modifying the Meade f/6.3 ff/fr so that it slips into the Feathertouch focuser.  This would save me about 1 ¼” of stack up.  I honestly do not know whether this would mitigate or aggravate the problem.

I would genuinely appreciate input this issue.

Thanks,
Mark

[chrisjbaileyuk]

Mark

On my FSQ85 with OAG/EFWII/383L at f5.6 I still get some light fall off with 36mm filters and even with 2"! I went down from 2" to 36mm to get 8 in the wheel. Its nothing that cannot be sorted with decent flats though. I do find the normal advice of exposing flats to half of full well to leave them a little hot and they dont correct well. At an ADU level of 16-18,000 they seem to correct the lights much better with the 383L.

Chris

[NickK]

I bit the bullet and went for 2" par-focal baader filters (I bought my 383L when they first came out) with the few that they would allow more flexibility with cameras/backfocus at a later date.

Only thing is 2" means only five slots in the EFW2. It would be nice to have both wide and narrow band filters in one device!

[joelshort]

Mark,
Your flat comparison shows the same results I get with my 383L and refractor at f/5.4.  There is vignetting in the corners that looks exactly like what your image shows.  However, good flats completely take care of this issue for me.  I aim for flats around 25,000adu and make sure that the flats are exposed for a minimum of 2sec to avoid shutter shadow.

To answer your question, I do think that 36mm filters would eliminate vignetting.  I'm rather surprised by Chris's statement that he has light fall off with 36mm or even 2" filters. 
joel

[niteman1946]

Thanks Chris, Nick and Joel for your replies.

Chris, I tried your suggestion of lower ADU for my flats.  Unfortunately, no gain in mitigating the concentric ring that occupies the corners of my integrated image.

Nick, before I spend the cash on a new set of filters and holder wheel, I'll use the line profile tool to revisit older images prior to the installation of the EFW2.  And see if I can justify it.

Joel, we've visited before on the minimum exposure duration for the flats.  I normally do a minimum of 4 seconds just to avoid the shutter shadow issue.

It's both discouraging and encouraging that most of you (and others on other fora) are able to "fix" the vignetting issue with flats.  So, it can be done, but I've yet to find the correct technique.

FWIW, I do my flats with t-shirts and the scope pointed to a neutral daylight sky.  Minimum of 4 second exposure (more often 5 to 10 seconds, depending on filter).  And around 25k ADU (I did try Chris' suggestion of around 16k).  All optical components are in the shade, and I did cover the back of the scope one time with a black bag to see if stray light was the problem.

I process with PI, although I did try DSS once and got the same vignetting/ringing, so my problem is not unique to PI.

Here's a recent example of the problem.  With effort, I can eventually get the background to where I want it.  But in the process, I have to give up on quite a bit of the faint nebulosity.  Alternatively, I just cut.

Mark

[joelshort]

Sorry if we've been through all this already.  I post in so many places its hard to keep track of them all!

Something is obviously not right.  Your image of NGC5566 looks to me like there is severe light pollution in the lower right corner, but it's hard to tell if that is LP or incorrectly applied flats.  Can you provide a way for me to download a raw flat and an individual sub of NGC5566? 

[niteman1946]

Joel, not a problem.

If this works, it's of the following:

1.  32k ....... Individual flat at 32k ADU (approx.) and -10C
2.  -10k....... Individual flat at 12k ADU  (approx.) and -10C (used in previous example)
3.  Image_Light_600 ..... NGC5566 single sub at 600s and -10C

https://www.dropbox.com/sh/s0rfbqwlw8uu4v9/h6aoiaUc6p

Mark

[joelshort]

Well, I hate to say it but I am at a loss with this.  Both flats you provided look "ok".  There is definitely more light drop off than I would expect at f/6.xx, but they don't look bad.  There is more light drop off than I get at f/5.4

Your individual sub also look ok to me.  There does not appear to be light pollution or gradient interfering.  It looks pretty uniform across the image.

So in my mind there are two possibilities.  Either the flats are somehow being incorrectly applied to the lights, or the flats are not showing the actual defects of the optical train.  In other words, perhaps the flats are being shot in such a way that they don't show the correct image circle, dust patterns, light fall off etc.  I've never taken t-shirt flats, but I would really encourage you to try and take flats some other way.  Put a white posterboard up somewhere and try to illuminate it as evenly as possible with ambient light sources and take the flats again, without t-shirts.    Basically, try to take your flats a different way and see what happens.

When I applied your flat to the light using DSS it came out very wonky, showing the same pattern that you see.  I don't understand why.  Both the flat and light look decent to me.

Sorry I couldn't be more help.

[niteman1946]

Thanks again Joel.  You're quick!

1. Flats incorrectly applied (i.e. processing software issue):  your test with DSS obviates this, no?

2.  Fundamental flaw with Flats (due to the way they were taken):  i'll try some variation of your posterboard suggestion.

Will let you know.

Mark

[joelshort]

The flats could be incorrectly applied if somehow they were rotated or mirrored before being applied to the lights.  But yes, my test with DSS would almost certainly eliminate the possibility of incorrectly applying the flats.

Take some flats using some other method and see what kind of results you get.  Good luck!

[chrisjbaileyuk]

Mark

I have just had a look at your flats and compared them to mine.

Yours have a level of ~35,000 ADU in the middle and 15,000 in the corners.

My L master flat that works very well has a level of ~20,000 ADU in the middle and 18,000 in the corners.

I had a look at your setup photo. How far is the image plane from the back of the FR? The problem with the FR might be that it also reduces the image circle. Have you tried it without the FR in the chain?

Chris

[niteman1946]

Joel,

I hung a white vellum sheet inside my dome and mounted the dew shield on the scope.  The open end of the dew shield was about 6" away from the vellum (to cut down on stray light).  The dome is white and somewhat translucent, so ambient light did illuminate both the dome side and scope side of the vellum.  I shot the Flat subs at dusk and got about 25k ADU at 3sec. exposure.
I shot both Flats using the Luminance filter, and through an open port in the EFW2 (i.e. "No Filter").  So this was a different method than my Tee shirt approach.

Chris,

I looked at one of the "No Filter" flats, and Images Plus reported about 21k in the center and 16k at the corners.  So this was an improvement over the Luminance flats.

Here are four sets of images and line profiles:
1.  Luminance Flat Sub and Master Flat with Line Profile:


2.  Luminance Image Sub and Calibrated Image Sub with Line Profile:


3.  No Filter (open port) Flat Sub and Master Flat with Line Profile:


4.  No Filter (open port) Image Sub and Calibrated Image Sub Line Profile:


There were clouds last night, so item #4 got short shrift.  I just pointed to a more clear area and did not bother to guide.  However, as ugly as it looks, I think it proved the point.

With my current setup, it looks like the 1 1/4" filters (at least the Luminance) cause severe vignetting, while the open port does not appear to.

Next step is to tear the setup down and shoot without the f/6.3 fr/ff.

Mark

[joelshort]

In looking at your image train, I agree with Chris that the reducer seems to have really reduced the image circle.  It seems like there is an awful lot of distance between the reducer and the focal plane.  I might suggest trying to shorten the distance by removing the OAG or something and see how that compares.  But completely removing the reducer is a good test too. 

I still don't get why the calibrated image shows a bright left side and dark right side.  If the flat is being used correctly it should be even or at least closer. 
joel

[niteman1946]

Joel, Chris,

The f/6.3 fr/ff is set up at 101mm from rear shoulder to camera focal plane.  Depending on the source the 101mm should return a reduction of around 6.56.  But it could be more or less.

I've run out of space to shorten this setup any further.  The reason for the fr/ff is to tame the stars away from the center.  The old classic looks pretty bad without the reducer.  And I won't give up my OAG.  So my taking the reducer off is for test purposes only. We will see.

I was so glad to see some improvement with the "No Filter" approach.  And I agree, the resulting calibrated image through the Luminance is bizzarre.  For whatever reason, the lights and flats do not play nice.

In my initial post I mentioned that I am considering modifying the reducer's housing so that it can slide into the focuser.  This will NOT change the 101mm distance between the reducer and camera, but it will shorten the overall backfocus (?) and give me an 1 1/4" of more clearance for the camera.

Welcome any thoughts on this one.
Mark

[chrisjbaileyuk]

Mark

101mm is a lot more than the designed back distance of 87mm and I think the FR is going to be a major factor in this so will be interesting to see the results of it not being in the mix.

Chris