This is coma - not rotation

[MarkS]

Mac,

After our conversation at the last meeting I promised to post an image showing my problem.

Here is a good example:
http://gallery.orpington-astronomy.org.uk/albums/userpics/10046/notrotation.jpg - it is a single sub with flats applied

If you look in the top right and bottom right you will see "rotation" that is similar to what you are getting.  Coincidently there was a post on UKAstroImaging this week from someone with the same problem:
http://ukastroimaging.co.uk/forums/index.php?topic=41010.0

I'm 95% certain this is due to field curvature i.e. we are seeing a type of coma.  Examine the common factors between, yours, mine and Euan's from the thread above.
1) Each is using an SCT
2) Each is using a reducer/flattener
3) Each is using a DSLR

Problem is that the DSLR has a chip much larger than the properly illuminated Field Of View.  The coma appears towards the corners - this is outside the stated FOV of the scope and so I think it is also outside the area that is properly corrected for by the flattener. So there is strong field curvature and therefore coma.

In my image above, the coma effect is quite different on the LHS of the image compared with the RHS.  This is simply because something, somewhere is very slightly out of alignment - it _might_ also be due to collimation.

Mark

P.S. Can anyone work out what I was trying to image?

[Mike]

P.S. Can anyone work out what I was trying to image?

Is it the Tulip?

[MarkS]

Yes, the Tulip.

[Mac]

Very interesting read,

I can see clearly in your photo, where the flatness ends, and the coma, begins.
As for my photo, it was only on one side,
So i am assuming that the edge of the chip on the DSLR must be slightly off on one side, so it was
only appearing on one side, and not the other.

Any way, hopefully i wont have to suffer that problem any more
as my H9 arrives tomorrow.
So apoligies to everyone for the forthcoming bad weather.

As for the image shift problem, still haven worked that one out,

I'll see if the problems carries over to the H9, but im hoping it was the canon adapter.

Everything was tight.

I can understand if my polar alignmet was off, but i would have star trails as well.
Each image seems as if the telescope has moved, you can see by the dy dx values compared to the middle one.

Here is a screengrab for DSS for the offsets.



Nothing was moved in between the shots, The mirror was locked down.

Is this just the perodic error of the mount?

if so would this not also induce field rotation? albeit slightly?

The photos were sequential.

Here is an overlay in photoshop, with the starts rubbed through to show the difference between each frame.



any ideas of this one?

Also my colomation is off quite a bit. I used CCDInspector to check.

will sort that out later.

Mac.

[RobertM]

I had a similar (but not the same) thing happen but the images were in a straight line and it was due to Polar alignment being way out.  It was all over an image taken with a 450d and the C9.25 at F10 (unguided) not just at the edge of field.

If you are guiding in the central part of an image then it would be reasonable to expect that field rotation to be worse at the edges of the image and that it would get worse nearer the pole.

[Mac]

If you are guiding in the central part of an image then it would be reasonable to expect that field rotation to be worse at the edges of the image and that it would get worse nearer the pole

Thats just the problem,

The field rotation which is visible is only at one edge.
Which would sort of rule out field rotation.
as every star should have rotation about the central point.

and the stars on the left hand side show no signs of rotation, only the ones on the right.

When i polar aligned the pier i used the iteritive method,

Align pole star.
Select another star,
goto star,
center star using handset
sync on star
return to pole star
adjust scope to remove half of the error,

Repeat untill satisfied,

I done this for about 15-20 mins, halving the error each time,
untill the error was indistinct.

With a 9mm eyepiece centered on a star, i left the scope for about an hour,
after an hour, the star had drifted about 1/3 of the eyepiece view,

Now this could have been perodic error, as i havent trained the scope yet, to remove it.

but each of the subs were only 60 sec long.

Field of view in eyepiece = 12'
Drift in 1 hour ~1/3 view = 4'
Drift in 1 min 4'/60" = 4"

Quick calculation.

Field of view on canon 350D on  my scope
x = 30'
y = 20'

which gives 0.53" per pixel.
which should equate to about 8 pixels, over the 1 min,
which thinking about it is quite a lot.

but it still dosent tally with the dx dy values, as these seem to be sort of random values between them.

its the whole image that has shifted on the camera chip, either up, down, left, or right,

but each image is the same, but just shifted on the chip. The stars on the right hand side show rotation,
the ones on the left dont!

the star pic, was to show how much each star has moved, with respect to the chip.
Dont forget each picture was 60 seconds long, and the difference between each photo was 5 secs ish

I'll have another look at my polar alignment at some point, and train out the PEC and see if that makes a differece.

But i might just see if i  get the same results with the H9, and hope that it was just the camera and fitting.

Mac.

Just had a thought, dumb one at that.
I'm just going to recalculate the images that were taken sequentally, from
02:48 through to 03:11 and see if the dx & dy values increase or if they are totally random ish.
looks like my polar alignment might be out.
the DY values go from 0, reference frame to -25. over a period of 20 mins. roughly sequentially
but the DX values. look like PEC.
Think i'll sit in the corner for a while, quietly.

[RobertM]

The field rotation which is visible is only at one edge.
Which would sort of rule out field rotation.
as every star should have rotation about the central point.

No it will rotate about the guide star as that's the point that's fixed.

Have a beer

[Mac]

No it will rotate about the guide star as that's the point that's fixed.

Thats what im saying,
the central star would be fixed, whilst the other stars show trailing,

Similar to a long expoure on the pole star,

But thats not the effect im getting,

Its only the stars on the right hand side of the central star that show the rotation.
The left hand side the stars were pin sharp.

Mac.

[RobertM]

Are you sure that the guide star was in the center ?

It's probably best if you start from first principles.  Get the scope in collimation, make sure it's polar aligned i.e. there is virtually no drift in a 5 minute sub, make sure the guide scope has ZERO flex then take an image where you know there will be little rotation - on the celestial equator.  If you still get this odd effect then it is indeed bazarre.

What scope is this ?

[Mac]

Its my LX200 10"
I had the same problems with the Nikon as well as the Canon,

I'll probably give it a go tomorrow nihgt as im not working.

Its one of those problems, that when its all sorted, i'll be able to sit back and go

of course it was.

Mac.

[Mac]

Well after about 5 hours last night,
tweeking, adjusting, banging, cursing, kicking the dog, and the pier, i've found out the following.

Before i started, my polar alignment was only very slightly off, very slight drift, in a 4mm eyepiece over 5 mins.
both in an eastly view and southerly view.

So no issues there.

but i did tweek it a little to improve it.

When i put my newly aquired H9 on the scope, with filter wheel and started taking photos,
I had exactly the same results, one photo would be fine
the next one it would have moved 20ish pixels,
some would show signs of star trailing.

Anyway, i realised at the end after i had packed away, that i've forgotten to put
the balance weights, on,

I also forgot previously to do the same, and the weight of the kit was causing the scope to slip.

that aside, an absolute nightmare trying to get anything in focus.

I recolimated the scope with an eyepiece and spent about 20mins getting it spot on.
then using CCD inspector, it said it was miles out?
Adjusted it using CCD inspector, and ended up with cones for stars.
Removed everything and recoliminated again with eyepiece

Checked and again it said the colimation was miles out.

Im just wondering if the light path to the H9 was twisted, or the chip is not square to the field.

at the moment it was set up as follows.

SCT ¦ 6.3FR ¦ Filterwheel ¦ H9.

Using an eyepoiece at parfocus, the stars are pin sharp.


Swapped owver to the H9, and using a hartman mask the stars are a single points indicating in focus.
remove the mask and the stars appear as cones!!

i can only conclude that the chip is not square to the scope.

Any ideas of how i can check if the ccd chip is square on to the scope?

at which point, i put everything away and took about 10 1min widefield subs of M45.

Mr Frustrated.

I know how you used to feel Fay.

Mac.

[MarkS]

Mac,

I share some of your problems. I discovered last night (after shooting taking hours of data) that the CCD of my Canon was not square to the raypath.  So either the raypath is wrong (i.e. collimation is out) or the camera/CCD is not square with the telescope.  Does this happen all the time or was it a one off - I have yet to find out.

Anyway, the effect of this is that I cannot bring all stars simultaneously to focus across the whole CCD - here I'm talking about the main illuminated part of the CCD not the bit beyond that, which is always a problem. Let me clarify:
1)  Outside the main Field of View (of the flattener) I have strong field curvature - these stars will always show coma.
2)  Inside the main FOV the CCD is not square so only some stars can be in focus - the rest will exhibit varying amounts of defocus and slight coma.

You mentioned the Hartmann mask.  I guess it will disguise coma.  Coma is a result of raypaths coming from the whole surface of the lens but the Hartmann mask cuts most of these raypaths out.  So something that looks good using the mask may show coma once the mask is removed.

I think we're both on a learning curve here!

I'm not familiar with CCD inspector.  What does it do?

Mark

[Mac]

I'm not familiar with CCD inspector.  What does it do

amonst other things, it measures your colimation. as well as checking focus in real time.

http://www.ccdware.com/products/ccdinspector/

so long at you have a lot of stars in the field of view, it then mathmatically defocuses them and
works out if your scope is or isn't in colimation,
it can show you a 3D view of the curvature of your chip,

so i might have a look at using this to see if the ccd is square on.

I'll probably do my next set up in daylight and check it against some small type on a sheet of A4 against the house.
That way i can check if the chip is square, by first checking with the eyepiece
and then the CCD.
As i know the type on the A4 is in focuswith the eyepiece, then the out of focus bits must be due to the CCD not being square.

I might also have a play with where i put the filter wheel as well.
Might try putting it before the focuser, so im only moving the CCD camera and not the filter as well.

SCT ¦ F6.3 ¦ Filter wheel ¦ focuser ¦ CCD.

Mike, how is your set up, with respect to filter wheels, focuser & ccd, or anyone else for that matter.

Cheers

Mac.

[RobertM]

Mac, I now ALLWAYS use reducer then filter wheel then ccd in the light path, the further the FR is away from the CCD the more vignetting you will experience (yes you will even get it on the H9!)   These three components are screwed together as a unit and whenever I move from scope to scope they move together.  I used to have Scope -> reducer (sct thread) -> focuser (JMI NGF-S) -> filter wheel -> CCD.  I didn't like this arrangement because of the weight on the coupling between the focuser and the focal reducer - it wasn't rigid enough and I thought it may have been the cause of some of my image shift.

When I attach any camera I always check that it's fully seated, then check and recheck.  If the CCD units is rotated then I do the same again.  The depth of focus on the Sky 90 at F4.5 is a few thousanths of an inch so the chip has to be dead square otherwise I get flaring on one side of images.  With a Cat the depth of focus even at F6.3 is much more because the light cone is shallower so I would only expect a problem if something in your imaging train is not square.  That could be anything attached to the back of your scope.  Do you get the same problem with a DSLR ?  It's got a much bigger chip so I would expect the probem to be much worse, if it isn't then it could be the H9 - if you have a guide cam then try that as well.

Hope that helps.
Robert

[MarkS]

Mac,

Thanks for the tip on CCDInspector - I've just analysed some images I've taken over the past few weeks and the collimation is a long way out: 15-20arcsec according to its analysis.

This may well explain the issue I noted in my "Lack of Orthogonality" thread: http://forum.orpington-astronomy.org.uk/index.php?topic=3936.0

So, the next clear night will be spent collimating.  It'll be interesting to see if I notice it (lack of collimation) in the eyepiece i.e. lack of concentricity when defocused.

Mark