Flat frames: do you take them with dew shield on or off?

[opticalpath]

I normally take flats with an EL panel resting on the end of the dew shield - which is in place as it would be for imaging - but I'm starting to wonder if this is a good idea.  The interior surface of the dew shield is certainly not 100% non-reflective and the EL panel is relatively bright and close, so I could imagine flats being somewhat affected by reflected light.  Would it be better to remove the dew shield and place the panel directly on the OTA to eliminate possible reflectionst?  Or would it be better actually to get the EL panel even further away from the scope objective so the illumination is coming in more nearly on-axis?

My flat correction is not as good as I'd like it to be and at some point I'm going to do some experimenting, but meantime I just wondered what others' experience and usual practice is.

Adrian

[sharkmelley]

Hi Adrian,

What a great question! 

I have had lots of problems with flats partly because of using a large chipped camera (a DSLR) on relatively complicated optics such as an SCT and Hyperbolic Newtonian scopes and the fact that I tend to apply significant stretches to the image during processing so any mismatch in the flats and the lights gets exaggerated.

The short answer is that my best flats have been twilight sky flats taken after sunset with the dewshield in the same position as for imaging.  A EL panel on the dewshield is next best.

The longer answer is as follows:

You have correctly identified that stray light is a problem but I think these "reflections" within the dewshield are as much of a problem for twilight sky flats as for EL panel flats.  And it certainly also happens during imaging.

I concluded some time ago that the problem with flats is that you want to record characteristics of the direct optical train i.e. vignetting, dust spots etc.  However, on top of this is recorded all the stray light that reaches the sensor: internal reflections because the scope interior is not perfect black, smudges, dust and dew on the lens, mirrorand filter surfaces that act as secondary sources of light etc.  However, similar stray light reaches the sensor during the course of normal imaging.  Hopwever, it's not at all clear to me that the flat is able to correct this.  In an ideal world, the flat containing the optical train characteristics should be separated from the flat that contains the stray light characteristics.  But I've yet to think of a way to achieve this - indeed I am sure it is impossible for practical purposes.

Is stray light during imaging really a problem?  Think about it this way: suppose internal blacking and baffling of your scope prevents 99.99% of stray light reaching the sensor - this is just a figure I have plucked out of the air but it sounds pretty effective doesn't it?  But now consider that if you are imaging a 2x2 degree patch of sky then this represents only 1/5000 of the total area of sky in front of your scope.  All the rest of that sky is pouring light through the front of your scope - all the stars, all the light pollution, maybe some distant street lights etc. And if 99.99% of this stray light is prevented from reaching the sensor then it means that 0.01% does reach the sensor - which is a significant proportion of the light coming from the 2x2 patch of sky you are actually imaging!  Hopefully, just hopefully, all this stray light is distributed over the sensor in exactly the same proportions as when you take a flat but I strongly suspect this is not the case.  Especially since the sky tends not to be uniformly lit, it could also be partially blocked by darker trees and buildings and observatory walls or there maybe additional brighter sources of stray light such as other buildings and street lights etc.

This might explain why sometimes my flats work almost perfectly and at other times they leave significant residual gradients. I'd love to hear from an expert who has the definitive answer to this conumdrum.

Mark

[Gina]

So would I - this is a very interesting point.

[RikM]

I always take the dewshield off my Newts to take flats, because they are fairly floppy camping mat foam. I have only done a couple of refractor images but here I left the dewshield on because it is stiff aluminium or whatever.

[pubquiz]

Yes interesting question as I will possibly be getting an A3 size E L panel soon....I have an A4 sized one for the refractor and love it but always used a tee shirt stretched in a frame I made and pointed at a lit wall on my C8 but the E L panel is nicer to use so want one for my RC 8

I would be tempted to remove the dew shield and place it directly on the end of the scope.....like you say if there is dew in the shield I suppose it could cause unwanted reflections due to the brighter light of the panel.....reflections that wouldn't be there from dim starlight but would be interested to hear the views of others

Tom

L

L

[sharkmelley]

like you say if there is dew in the shield I suppose it could cause unwanted reflections due to the brighter light of the panel.....reflections that wouldn't be there from dim starlight but would be interested to hear the views of others

The brightness of the panel can't make any difference at all to whether the dew shield reflects or not.  Similar reflections wilI occur in dim starlight - the camera will certainly see them even if your eye cannot.

Mark

[ollypenrice]

SBIG published a paper on internal reflections affecting flats. It turns out that some apparently mat black finishes are quite reflective in UV, to which cameras are sensitive. The offending paints are ones using dyes rather than pigments. Barbecue paints, however, use pigments to withstand the heat and tend to be a better bet.

In the short term I guess the upshot would be to minimise the possibility by not using the dewshield for flats. Then paint your dewsheild with barbecue paint inside.

I'm not persuaded that a panel on top of a dewsheild will exactly replicate any night sky stray light or internal reflections. It isn't the increased intensity which makes me say this but the proximity of the source, which I can imagine creating new shadows or reflections in any nooks and crannies otf the tube. I've no idea whether this is really the case but I thought about it during a thread on using a panel with a Hyperstar, where you'd be illuminating the camera and cables from close-to, creating who knows what effects? Also, the spectrum of the panel might play here. If it were strong in the UV and the paint issue were operating then you'd have reflections from the panel which were not present while imaging.

Olly

[opticalpath]

Ah, I remember that article now you mention it, Ollie.  I think it was IR reflection that was the issue, but same difference.  So that's probably another reason to remove the dew shield for flats. I use one of those felt-lined Astrozap things.  Who knows if it reflects a load of IR .... but painting it with BBQ black isn't an option.  Proper baffling of an oversize dew shield is probably worthwhile to reduce these problems. Time for some experiments.  Bring back T-shirt flats!

 

Adrian

[sharkmelley]

Olly, that's very interesting what you say about reflectiveness of paints.  Especially that the EL panel has a different spectrum to the night sky and so reflections may occur in one case but not the other.

I totally agree that an EL panel on top of a dewshield is unlikely to exactly replicate night sky stray light or internal reflections but I think it is better than the alternative of pointing the EL panel directly on the front of the scope.

My reasoning is as follows: during night time imaging the dewshield acts as a light shield and provides valuable shielding from stray light - i.e some stray light that would otherwise directly enter the front of the scope can now only reach the scope by (hopefully) attenuated reflections off the dewshield.  However, by removing the dewshield during flat frame acquisition and putting the EL panel directly on the front of the tube you are introducing stray light paths diectly into the front of the scope that could not have been present during imaging.

In the end, it's really easy to take two sets of flats and see which work best for your setup.  I did this and found the dewshield flats to be the better match, at least on my own equipment (DSLR on Hyperbolic Newtonian).  I also found that twilight sky flats were better than EL panel flats.  I am guessing that the twilight sky better replicates night sky stray light and internal reflections than does an EL panel - for whatever reason.

I've also recently acquired a C11 with Hyperstar so I also have that flat frame pain to come!  But I've removed the mirrors from my DSLRs and that already makes a huge difference to the shadow area on the sensor.

Mark

[ollypenrice]

Sorry if I've confused IR with UV. I agree with Mark, though, that the pragmatic solution of trying both is always going to be best. In assorted discussions I've formed the opinion that reflectors may prefer sky flats. Only after a lot of experimantation did we get panel flats to work for the ODK14, though now we seem to have it sussed. I find panel flats generally good with the reflectors though just ocassionally, and for no reason which I can fathom, they give gradients. (As do sky flats when they feel like it.)

What a caper.

Olly

[sharkmelley]

Agreed - a right caper!

What I would really love to know is just how much unfocused stray/scattered light typically reaches the sensor, though obviously it varies from scope type to scope type and depends on many factors such as internal baffling/blacking, cleanliness of optics/filters, condensation build up on optical surfaces etc.

I am minded to do some experiments to find out, because this question has always bugged me.

Mark

[sharkmelley]

I think I've found the article you referred to: "SBIG - Flat Fields and Stray Light in Amateur Telescopes"

https://www.sbig.com/about-us/blog/flat-fields-the-ugly-truth/

[RikM]

I have always used Newtonians and used to routinely get problems with gradients and flats not applying properly when shooting T-shirt flats. Since moving to an EL panel on the end of the scope, I have had a much easier time of it. Flats have been much more consistent.

[ollypenrice]

I have always used Newtonians and used to routinely get problems with gradients and flats not applying properly when shooting T-shirt flats. Since moving to an EL panel on the end of the scope, I have had a much easier time of it. Flats have been much more consistent.

Bing

I think I've found the article you referred to: "SBIG - Flat Fields and Stray Light in Amateur Telescopes"

https://www.sbig.com/about-us/blog/flat-fields-the-ugly-truth/

Yes, that's the fella and it was indeed IR. Thanks, I'll save it this time.

Olly

 goes my theory, then!!

[dph1nm]

Professionals have scattered light issues with their flats too. There is no real answer to it other than better baffling.

http://casu.ast.cam.ac.uk/surveys-projects/vst/technical/scattered-light-and-illumination-correction

NigelM

[ollypenrice]

Sorry, my cryptic comment 'Bing' above was meant to say, 'Bing goes my theory...'

Olly

[sharkmelley]

Here's a very interesting explanatory article on flat fields.

http://www.astroinformatics.de/index.php?option=com_content&view=article&id=139

It shows an example of where stray light during imaging (moonlight) caused the flat field to perform suboptimally on the image.

They also say that the flat field itself suffers from stray light so,

"This means, the recorded flat field is a sum of a multiplicative term and a certain additive background illumination from scattered light. "

It also describes a method of taking an image of the stray light effects within the telescope, essentially using a pinhole camera.

It doesn't answer the original question though - should the dew shield be on or off for flats!

[PhotoGav]

Hello all,

I was about to post a new thread asking exactly this question - should EL Panel flats be taken with the dew shield on the end of my 8" SCT or with dew shield off? I did a quick search first and found this thread... Very interesting.

However, there is not really a definitive answer in the thread! The suggestion is that it's probably better with the dew shield on, but the best thing to do is try both and see which works best. That I will do, but just wondering if anyone has any results of experiments that they would like to share and hopefully get closer to a guideline answer?!

Thank you.

[sharkmelley]

If I'm using a dew shield when imaging then I take flats with the dew shield on because it then matches the imaging conditions and my own experience on my scopes (SCT and Newtonian) is that I get better flats like that.  On other scopes (Refractors) it may not matter so much but I don't know.

But using a EL Panel with a dewshield on may not be easy so practicality plays a part.

Mark

[PhotoGav]

Excellent - thank you for that. I will give it a try with dew shield on first and see how I get on. I suppose that is effectively what I do with the ED80 and it's 'built in' dew shield and it does match the lights shooting conditions, so in theory should be best. Fingers crossed!