Flats when using rotators

[teoria_del_big_bang]

In theory I believe flats should be taken after each session, they are quick to take so not a big issue.
I do hear that many do successfully reuse flats, so long as the image train remains in place and camera etc are not removed for storage.

But does this mean (again in theory) that when using a rotator controlled by the sequencing software that I should take flats before moving the rotator ?
I have not long had a rotator and in fact only just got it working reliably in EKOS when plate-solving and now seems to work well and will orientate camera to match a previously taken image.
But in a recent session I noticed that this now means that after a flip the rotator now moves 180 degrees to keep the orientation of the image, whereas previously this had to be done in the processing software.

So I was thinking that I would need to take flats twice 180 apart as the camera and filter wheel had moved with respect to the other optics.
So to be safe I did do this but on close inspection I am struggling to see any differences between the two master flats so assume nobody does this normally and also assume the vast majority of dust motes etc that affect images are just on the CCD cover glass and filters and that rotation of camera does not affect vignetting.

Any thoughts on this ?

Steve

[newbie alert]

I'd assume that the rotator rotates the whole imaging train not just the camera.. so the camera relevant to a filter is the same.. only differences are the optics, as a rotator is rotating the imaging train around the optics 

That's how imagine it ...

[The Lazy Astronomer]

Have you tried calibrating one of the master flats using the other? If all the dust shadows and vignetting are completely removed then you'll know there's no appreciable difference between them. 

Out of curiosity, what rotator are you using? Sensor orientation is the last thing I have to do manually, using my hands like some kind of common visual astronomer (I kid! 😜) - I'd love to be able to computerise it, but the rotators I've seen are just too expensive for my liking. 

[discardedastro]

Practically, for ease of calibration, I'd want to have flats for each rotation position (and maybe stick to a few common rotation positions if possible if you're able to and don't want to be taking flats all the time). I know that my flats generally correct more than just what's in the optical train and that tilt of the train isn't going to be exactly nil (e.g. vignetting etc won't be perfectly symmetrical). Better to be safe than throw out a bunch of exposure time, I think.

I also don't think you'd want to do synthetic flats e.g. rotate your existing flats to calibrate the flipped ones. This has the risk of correcting correctly all the optical train artefacts but incorrectly correcting for the scope.

After flips you don't strictly need to rotate the camera - there's no downside to rotating the image in software later, and then all your calibration data will match with only one set of flats. Obviously having the rotator is really helpful between subjects, though.

And while lots of people do suggest flats after every session I practically do just fine with everything left in place for months at a time with occasional imaging. I redo flats on occasion just to update for any new dust or debris!

[vlaiv]

3 minutes ago, The Lazy Astronomer said:

Have you tried calibrating one of the master flats using the other? If all the dust shadows and vignetting are completely removed then you'll know there's no appreciable difference between them. 

This is excellent piece of advice to determine if flats are impacted by rotator.

@teoria_del_big_bang

It really depends on scope. We could say that there are two main components to flats - one is dust shadows and other is vignetting.

Depending where rotator sits in optical train - dust that is responsible for shadows might rotate with camera and it won't change position.

In scopes that are symmetric if sensor is placed dead center on optical axis - vignetting will be symmetric as well and rotation won't matter. Refractors are prime candidates for this, but do be careful - camera can be offset, or rather sensor.

For example - look at latest ASI485 - sensor offset is very clear with that camera:

They did their best to make it centered (even if silicone is offset) - but I don't think it's centered properly. Usually this does not matter for imaging, but it can be problem with flats.

Another thing that can create asymmetric flats is type of scope. Newtonian scopes and especially fast newtonians have asymmetric flats (there is some secondary offset that I don't really understand - never bothered to properly grasp it since I never owned fast newtonian)

[teoria_del_big_bang]

1 hour ago, newbie alert said:

I'd assume that the rotator rotates the whole imaging train not just the camera.. so the camera relevant to a filter is the same.. only differences are the optics, as a rotator is rotating the imaging train around the optics 

That's how imagine it ...

Yes thats correct

Steve

[newbie alert]

46 minutes ago, vlaiv said:

 

@teoria_del_big_bang

This looks like a focuser that's offset to me

[teoria_del_big_bang]

54 minutes ago, The Lazy Astronomer said:

Have you tried calibrating one of the master flats using the other? If all the dust shadows and vignetting are completely removed then you'll know there's no appreciable difference between them. 

Out of curiosity, what rotator are you using? Sensor orientation is the last thing I have to do manually, using my hands like some kind of common visual astronomer (I kid! 😜) - I'd love to be able to computerise it, but the rotators I've seen are just too expensive for my liking. 

Good idea I can try that.
 

I am using the Deep Sky Dad rotator, reasonable price, even after import duty and seems very rigid, some videos somewhere showing them adding big weights to it and rotating it with no deflection.
Works well in EKOS and assume it will also do so if using ASCOM.

I did originally want the Pegasus Falcon rotator and did originally get one but the adaptors were a mm or two wider than stated on their website and I just couldn't get it in the back focus distance required, despite having a thin filter wheel (18mm) so thankfully RVO allowed me to return it and I then got the DSD one. I think the thickness is about the same as the Pegasus but did not need any adaptors and so just managed to get it in the 55.7 mm backfocus after the flattener.

Steve
 

[teoria_del_big_bang]

49 minutes ago, vlaiv said:

This is excellent piece of advice to determine if flats are impacted by rotator.

@teoria_del_big_bang

It really depends on scope. We could say that there are two main components to flats - one is dust shadows and other is vignetting.

Depending where rotator sits in optical train - dust that is responsible for shadows might rotate with camera and it won't change position.

In scopes that are symmetric if sensor is placed dead center on optical axis - vignetting will be symmetric as well and rotation won't matter. Refractors are prime candidates for this, but do be careful - camera can be offset, or rather sensor.

For example - look at latest ASI485 - sensor offset is very clear with that camera:

They did their best to make it centered (even if silicone is offset) - but I don't think it's centered properly. Usually this does not matter for imaging, but it can be problem with flats.

Another thing that can create asymmetric flats is type of scope. Newtonian scopes and especially fast newtonians have asymmetric flats (there is some secondary offset that I don't really understand - never bothered to properly grasp it since I never owned fast newtonian)

Interesting and good info, I need to look into this further.
I will check the flats as suggested
I do have some offset in the vignetting so also need to check further but when I initially looked at the two masters 180 degrees apart they looed the same by eye but need to do the calibration suggested.
My only other option without having to do individual flats, which would be a faff each session, would be to not use rotator after a flip but not sure I have the option to do that in EKOS. That way I might need different flats for different targets but not if on the same target.

Steve
 

[vlaiv]

1 hour ago, newbie alert said:

This looks like a focuser that's offset to me

That can easily happen with any system that can be easily collimated - like mirrored systems.

Collimation aligns optical axis of the mirrors, but both mirrors can end up being lightly tilted with respect to focuser or baffle system.

If one has focuser tilt - there is tilt plate on camera or separate tilt system to compensate and in the end everything is square except baffle system - and that results in vignetting that is not completely symmetric.

[ONIKKINEN]

12 hours ago, vlaiv said:

Another thing that can create asymmetric flats is type of scope. Newtonian scopes and especially fast newtonians have asymmetric flats (there is some secondary offset that I don't really understand - never bothered to properly grasp it since I never owned fast newtonian)

This looks like the secondary mirror was not perfectly centered under the focuser, which in my experience causes asymmetric flats (and other issues).

Take a look at some examples from my previous flats. These screenshots have been taken with the false color and histogram preview modes in Siril, very quick to see differences. Also there is a change of camera, hence the very different look to them.

One of the first flats i ever took, in almost stock collimation with my VX8 which after inspection with proper tools i found to be not well centered. Not a huge change in collimation, but a huge effect. On the bottom we see shutter shadow from a DSLR not quite being fast enough for flats. On the left side we see the same thing as in your example, an asymmetric brightess change.

One of my more recent flats with care taken to achieve as good a collimation as i can, but you can still see the red bright regions are not perfectly centered and probably some tilt remains. But the asymmetry is mostly gone because i centered the secondary with a concenter eyepiece. Also, this process includes adjusting the focuser and not just the secondary so yes one could say the secondary and focuser were out of collimation.

[teoria_del_big_bang]

22 hours ago, discardedastro said:

Practically, for ease of calibration, I'd want to have flats for each rotation position (and maybe stick to a few common rotation positions if possible if you're able to and don't want to be taking flats all the time). I know that my flats generally correct more than just what's in the optical train and that tilt of the train isn't going to be exactly nil (e.g. vignetting etc won't be perfectly symmetrical). Better to be safe than throw out a bunch of exposure time, I think.

I also don't think you'd want to do synthetic flats e.g. rotate your existing flats to calibrate the flipped ones. This has the risk of correcting correctly all the optical train artefacts but incorrectly correcting for the scope.

After flips you don't strictly need to rotate the camera - there's no downside to rotating the image in software later, and then all your calibration data will match with only one set of flats. Obviously having the rotator is really helpful between subjects, though.

And while lots of people do suggest flats after every session I practically do just fine with everything left in place for months at a time with occasional imaging. I redo flats on occasion just to update for any new dust or debris!

I have only had the rotator a few months and its taken till now to get it working automatically and maybe that's the issue.
In Ekos as far as I see I have no real control over it after a flip. Ekos knows it has a rotator and just does it as part of the plate solving, which I need to do otherwise the whole image could be completely out. Personally if there is a way I am perfectly happy to let the processing software flip the images as happened before I had the rotator.

I agree about not rotating taking one set of flats and rotating them for the images before the flip, that won't work.

I need to look into the two sets of flats I took properly, on first viewing they looked the same for both angles so to get on with processing my images I just made one master flat with all the frames, technically not correct I know but it does seemed to have worked to an extent, but when I have processed this image I will revisit it and look again at the flats and maybe re-do it, I am sure the North of UK will provide me with plenty of cloudy nights to do all this 🙂 

Steve

[ollypenrice]

My feeling, like Vlaiv's, is that everything causing bunnies will rotate with the camera and so be unaffected. If your objective's beam is on-axis then vignetting will be so nearly symmetrical as to be unaffected by rotation as well.

However, my other feeling is that it is counter-productive to rotate after a flip. You are losing an opportunity to redistribute the residual camera noise and so reduce it.  Or so I suppose.

You could put one flat as a layer on top of another in Ps and set the blend mode to 'difference' to see how different they are.

I use flats for months at a time. Never a problem.

Olly

[vlaiv]

4 minutes ago, ollypenrice said:

You are losing an opportunity to redistribute the residual camera noise and so reduce it.  Or so I suppose.

Quite true. It's like dithering but better

[teoria_del_big_bang]

1 hour ago, vlaiv said:

Quite true. It's like dithering but better

I agree just not sure if I can stop EKOS doing it 😞 
I don't want to do it, but I still need EKOS to platesolve after the flip to ensure the FOV is as near as damn it the same but if EKOS knows there is a rotator then it just does it as part of the plkatesolve.
I will have a look at EKOS and see if any way to stop this (you would think there is wouldn't you as there is just no reason to move rotator after a flip).

Steve