Optical shadow- can you help determing the cause

[pyrasanth]

I took a number of eight minute RGB subs using a Celestron C11 edge, 0.7 focal reducer, Innovation foresight ONAG & Optec secondary mirror focusing system. All the subs are affected with the strange donut object at the top. The picture attached is a life size crop. I've been communicating with Olly & he seems to think it might be the secondary mirror I'm seeing but asked me to open the discussion to the forum to get your views. Any advice would be appreciated as obviously I need to remove or minimise the defect & I will add more information tonight.

This is a single sub but when stacked (I will post a sample tonight) the defect is very nasty.

[Davey-T]

No idea, but please keep us posted on the ONAGs performance as I'm still tempted by one.

Perhaps it's the great doughnut nebula

Dave

[pyrasanth]

7 minutes ago, Davey-T said:

No idea, but please keep us posted on the ONAGs performance as I'm still tempted by one.

Perhaps it's the great doughnut nebula

Dave

The ONAG is great but I reserve judgement until I've eliminated it from this investigation. I wish I had captured a n new object but alas......we digress.

[Peter Drew]

My first thought would have been a "dust bunny", I would have expected a secondary mirror shadow to be more centrally placed, larger in area and less obvious. But what do I know about AP! 

[pyrasanth]

3 minutes ago, Peter Drew said:

My first thought would have been a "dust bunny", I would have expected a secondary mirror shadow to be more centrally placed, larger in area and less obvious. But what do I know about AP! 

The difficulty is that this "bunny" shows up after the flat field correction has done its stuff & the artifact is not on the calibration flats. I think a good start will be to clean the system & retest.

[Peter Drew]

I refer you to my last sentence. 

[derrickf]

It's difficult to be sure from the JPEG image but I think I can see other, fainter artefacts that look similar; this would indicate badly matched flats (they seem to be over correcting in this image).

 

 

[almcl]

Something of a long shot this (and here's my complete lack of knowledge disclaimer)  but how well shielded is the scope from extraneous light sources?

I saw somthing similar where  a light behind the camera was causing internal reflections.

[andrew s]

I think it is a residual dust bunny. The application of a flat requires the system is linear both in the response of the CCD and in the illumination. The use of the ONAG  makes me suspect that the between the illumination of the flat and sky differ enough that the visual/IR reflection/transmission is shifted enough to make the system slightly non linear in the cross-over region where CCDs can be very sensitive.

However, this maybe completely wrong but I don't have enough detail / data to tell so please take this as a possible cause.

As others have said cleaning might help. I also seem to recall you can locate where the dust is by the size of the bunny assuming it is a dust bunny.

 

Regards Andrew

 

PS You could look to see if there is a similar artifact in the guide image. It does not need a flat. If it has then the cause is ahead or on the ONAG if not after.

[michael8554]

Too small to be a reflection of the guide camera nose onto the diagonal ONAG mirror?

Michael

[ollypenrice]

The over correcting flats idea is interesting. Note that the outer circle is bright at the top and dim at the bottom but the inner circle is the opposite. While I notice this I don't know what to make of it.

Olly

[MartinB]

I think that's a bas relief flat artefact.  It's produced when you have a bit of flex in your image train, typically when it is extended by the need to accomodate a focal reducer.  If you then take your flat with your scope at a different angle the tilt of the chip is slightly different for your lights and your flats.  It tends to be most noticeable when you are imaging at low declinations.  That's my guess anyway

[pyrasanth]

I have cleaned all the optics however take a look at the stack of images below. 6 frames before a meridian flip & eight after the flip. If you stack the frames before the flip you get the donut at the top. If you stack after the flip you get the donut at the bottom & if you integrate all the frames you get 2 donuts but it shows the magnitude of the problem- is it too big for just dust?- I will find out when I retest the newly cleaned system.

If you look at the bottom image it looks like a shadow has been cast- could this be the new SMFS assembly? It carries the secondary mirror which is moved on a threaded rod. It looks like one of the mirrors to me.

 

[andrew s]

The 180 rotation symmetry seems to me to indicate this is due to something in the optical train i.e it stays at the same location on the CCD. Rather than continue speculation I would propose you look at the raw images, flats and any other calibration frames you are using and seeing if they show anything at the location of the artifact. As you have an ONAG you can also look at the guide images if you still have them again looking for artifacts. This should help you pin down the cause.

 

Regards Andrew

[pyrasanth]

10 minutes ago, andrew s said:

The 180 rotation symmetry seems to me to indicate this is due to something in the optical train i.e it stays at the same location on the CCD. Rather than continue speculation I would propose you look at the raw images, flats and any other calibration frames you are using and seeing if they show anything at the location of the artifact. As you have an ONAG you can also look at the guide images if you still have them again looking for artifacts. This should help you pin down the cause.

 

Regards Andrew

Very useful. I went back to the lights & stacked the red without applying a calibration & lo & behold 2 dust bunnies in the place where the donuts sit after the calibration. The calibration seems to be amplifying the donut & not removing them at least in these 2 places- any thoughts on this?

See the attached frame which is the raw uncalibrated stack with frames including the flip.

[MartinB]

9 hours ago, MartinB said:

I think that's a bas relief flat artefact.  It's produced when you have a bit of flex in your image train, typically when it is extended by the need to accomodate a focal reducer.  If you then take your flat with your scope at a different angle the tilt of the chip is slightly different for your lights and your flats.  It tends to be most noticeable when you are imaging at low declinations.  That's my guess anyway

You haven't commented on my post but the change after the meridian flip is entirely consistent with the above

[andrew s]

8 hours ago, MartinB said:

You haven't commented on my post but the change after the meridian flip is entirely consistent with the above

Interesting Martin, having looked up "Bas Relief" I have a better idea of what you are proposing. If I understand correctly you are saying, for example, a dust bunny will be in one place on the flat and placed differently on the target image. This means the flat will not remove the bunny completely and leave a bas relief. Please correct me if I am wrong. In addition as the flip will invert the shift you might expect different bas relief patterns as the image shows.

One way to test this would be to see if the raw dust bunnies are in different place on the CCD before and after the flip and also how this relates to their position on the flat. A simple way to do this might be to subtract a before flip and a 180 rotated after flip image and look at the dust bunny residual.

 

Regards Andrew

[MartinB]

That might give an indication.  Are you using an EL panel?  If so you could try a few flats actually holding the panel in place with the scope in a similar orientation to when taking lights and try applying those to the subs taken on that side of the meridian.  You only need 1 flat for testing so shouldn't be too tricky.

I have had exactly this problem when using a long imaging train

[Stub Mandrel]

Try turning the camera through an arbitrary angle (say 60 degrees) without moving the scope. That way you will know if its scope or camera.

P.S. forget duct bunnies, I think that's a dust leopard.

[andrew s]

1 hour ago, MartinB said:

That might give an indication.

43 minutes ago, Stub Mandrel said:

Try turning the camera through an arbitrary angle (say 60 degrees) 

[andrew s]

Having another go without quotes. As you cleaned your optics new flats will need to be used with new images. I have got out my trusty "The Handbook of Astronomical Image processing" by Berry & Burnell. It has an example image where a "Flat" has moved with respect to the image and it is full of artifacts like yours. The movement can be due to many things - filters not always returning to the same place, focuser shifts when focusing and image train droop etc.

You can locate where the dust is/was using D = Pdf where D is the distance in mm P is the width of the dust donut in pixels d is the pixel size in mm and f the focal ratio of the system (Equ 3.14 from the above book).

While some of it is a bit dated this is a very good book to learn about CCD imaging and full of insight.

 

Regards Andrew

[ollypenrice]

I think that, of all the ideas posted, I'd back Martin's.

I'm sure I've seen this come up before but can't remember the outcome.

Olly

[andrew s]

Two last thoughts on this. The ONAG unit puts the imaging camera off-axis and this applies a torque that needs very secure fixings to resist.

I noticed you use the Optec secondary focusing system. Unless this is engineered to exceptional standards it will introduce some shift that could account for the problem if the focus changed between the flats and image taking.

In any case a set of carefully planned experiments should allow the root cause to be found.

Good luck Andrew

[sharkmelley]

Would you post an image of the whole frame after flat frame correction.  The pattern of correction (or not) across the whole frame will provide a lot more information for diagnosis than image crops.

Mark

[Baudat]

On 2/17/2016 at 06:57, pyrasanth said:

I took a number of eight minute RGB subs using a Celestron C11 edge, 0.7 focal reducer, Innovation foresight ONAG & Optec secondary mirror focusing system. All the subs are affected with the strange donut object at the top. The picture attached is a life size crop. I've been communicating with Olly & he seems to think it might be the secondary mirror I'm seeing but asked me to open the discussion to the forum to get your views. Any advice would be appreciated as obviously I need to remove or minimise the defect & I will add more information tonight.

This is a single sub but when stacked (I will post a sample tonight) the defect is very nasty.

I am not sure which camera you are using, I'll assume the ATIK460 (if it is the ATIK11000, just double the numbers below).

The pixels are 4.54 microns (square). It seems, based to the picture, (assuming no binning or compression effect) that artifact is about 100 pixels across, or about 0.5mm (on the sensor plane).

If it is some dust particle creating this out off focus shadow, at f/7 it should be, roughly estimated, located around 7*0.5 = 3.5mm away from the sensor. The ATIK460 has a 13.5mm optical back focus, assuming a 3mm thick window (1mm optical back focus effect)  and a 3mm deep T2 thread (based on ATIK drawing) this brings the window around 9mm form the chip. I would guess that the particle (if any) is either on the window (either sides) or somewhere inside the camera sensor chamber.
Dust particles do not need to be big to create such effect when close to the sensor plane.
Either ways a proper flat calibration should have fully removed it.

It may still be some stray light issue too, but it looks very much like a dust particle (with this typical donuts shape due to the secondary central obstruction) not fully compensated by the flat processing due to some offset/difference between the flat and the light frames for some reason.

Maybe the dust particle has moved a bit after the flat frames were taken, or the flat frame light pattern is somehow different from the sky background. This may happen when using a light box close to the OTA, it adds more stray light effect (rays coming at large angles inside the tube are more likely)  and may impact the flats in the different way that the sky background would.
One option would be to take flats at dusk/dawn using the sky itself to compare.

The ONAG XM has a very rigid body, the connection between the imaging camera and the scope (dovetail system) is made of a single aluminum alloy profile about 10mm thick. Even under the most heavy camera/FW load (FLI Proline for instance) there is no flexure. Also having the imaging camera weight closer to the scope visual back (on top of the ONAG unit, see attached picture) reduces the torque applied to the scope connection (smaller arm) versus the on axis classical configuration for which a longer distance (arm) is required for the same optical back focus distance.
Finally the ONAG dichroic mirror is acting exactly as any interference filters used with monochrome cameras and filter wheels, it is fundamentally a star diagonal. Since this is a new kind of configuration I have attached, for information, a picture of a setup using the ONGA XM and a SBIG STL11000 on an iDK scope to give an idea of the typical setup with an ONAG.

Maybe an adapter/spacer is loose.

Dr. Gaston Baudat

Innovations Foresight president