Tilt, Collimation or Something else?

[rubecula]

This is a composit image of the four corners of a frame I took last night,  There are clearly misshapen stars a-plenty.

 

X6gas (Ian) kindly ran this through CCDInspector and the results are thus:

This is my set-up, Focal reducer - spacer - OAG - Filter Wheel - Camera:

 

The whole assembly feels very secure with no sign of any play.  The mount was recently polar aligned. 

Any suggestions as to what might be causing this?

Thanks

[steppenwolf]

I hate to be a pain but a centre image would also help for reference - if you want to send me the full image, I can generate a complete analysis set for you.

[x6gas]

I'm sure Robin won't mind so here is the single sub used to generate the curvature and 3D plots above.

L 21665.fits

Apparently nothing obvious has changed since he was successfully using the rig and getting nice round stars.  Visual inspection of the sub looks like tilt to me, but that's not what CCD Inspector is suggesting and Robin reports secure connections...

I'll be watching this thread with great interest!

Ian

[steppenwolf]

22 minutes ago, x6gas said:

I'm sure Robin won't mind so here is the single sub used to generate the curvature and 3D plots above.

Thanks Ian. I too thought tilt as soon as I saw the image but I agree that CCD Inspector doesn't confirm this. However, to perhaps give a better overall view of the issue, here is the visual analysis mosaic from the data you kindly provided:-

[rubecula]

Gentlemen, thank you for coming to my assistance. To be quite frank I've not heard of a visual analysis mosaic and do not understand what it's telling me.

[steppenwolf]

Fancy name (my fault) for a more detailed version of what you submitted in your first post! It shows the centre, corners, top/bottom and left/right centres which makes it easier to visually analyse an image.

To me, the image is not showing classic field curvature for example as both top right and left corners trail in the same direction, as do the left and right centres. However, the centre tile also shows the same trail direction indicating that this is partly due to a tracking issue, which is also visible in the bottom and right centre tiles.

That the trailing is much less in the bottom left and top centre is puzzling unless it is tilt negating the small tracking error.

[steppenwolf]

I have done a quick analysis to try to determine the dynamic here by superimposing the mosaic image on top of what I believe to be a correctly orientated facsimile of your sensor's orientation during the capture of this image (the Red rectangle) normalized to North/South to see what the effect of gravity would be on your image train. This doesn't indicate that the tilt (if that is what it is!) is slop in the focuser (or more accurately, the visual back and extensions on your OTA) but I present the image here to perhaps help the discussion!

[rubecula]

Well, I'm out of my depth here.  The visual analysis mosaic is indeed useful in being able to concentrate on different areas of the image.  I can see that the egginess varies from square to square but in my ignorance am unable to see what is causing this varience.  The image was taken about 30 minutes before a meridian flip and the guiding was good.  The image train feels very solid and I've never knocked into it, but could it sag over time even though it feels solid?  I have knocked into the weight bar but not particularly violently.

I'm stumped.

Thanks

[steppenwolf]

29 minutes ago, rubecula said:

The image train feels very solid and I've never knocked into it, but could it sag over time even though it feels solid?  I have knocked into the weight bar but not particularly violently.

Looking at the photograph of your telescope/camera, it looks like the Baader branded extension tube is a push-fit connector to a nosepiece on the Off Axis Guider. If this is the case, you could try the following:-

1. slacken off the Baader tube's grip on the nosepiece

2. press the camera/filter wheel/OAG onto the throat of the Baader tube and rock it slightly to feel the point at which the two mating surfaces (rear of the Baader tube and flange on the nosepiece) are absolutely flat against one another

3. while pressing the two mating surfaces together, tighten the Baader tube's grip

This gives you the best chance of a sag-free connection.

[x6gas]

Well I'm stumped too!

It seems there is something very odd about this.  If I look at the whole frame, the stars in the bottom left in @steppenwolf's mosaic seem OK if a little soft.  I *think* I am correct in saying that if this was tilt, then the coma would be pointing away from that corner of the frame but in actual fact the trailing is perpendicular to it.

I've analysed the corners and centre of the frame using PSFimage (a very useful PI script) and the results more or less confirm what the eye tells you... except it seems that the aberration is worst across the diagonal from top left to bottom right.  I'd expect it to be worst in the top right if the star shapes are best in the lower left.  Here are point spread functions arranged as they have been sampled from the image:

I did wonder if this had something to do with a problem with your guide cam not being orthogonal to the camera, but (a) I've successfully guided with all manner of aberrations on the stars produced by my OAG and (b) your camera sensor is parallel to the pick off prism of your OAG so if that were the problem I'd expect the best stars to be in the centre of one side of the frame and not in one corner and (c) guiding problems would give the same problem across the whole frame (unless there is a second problem confounding things).

Upshot: even though CCD inspector seems happy with the collimation, that's all that I can think of that would giving such odd results.  I'd suggest a good old fashioned star test - look at a defocused star and see if the shadow of the secondary is bang in the middle.  This is something that can easily be checked with an artificial star so at least you don't have to waste a clear night on it.

All very frustrating, I'm sure!

 

[rubecula]

@steppenwolf and @x6gas thank you so much for all the trouble you have taken to help me.  I've re-seated the Baader extension tube / OAG connection and am hoping that the clear forecast for tomorrow night holds so that I can see if that was the issue.

[rubecula]

Ha ha...glorious round stars, corner to corner! Thank you again @steppenwolf and @x6gas

[steppenwolf]

10 hours ago, rubecula said:

Ha ha...glorious round stars, corner to corner! Thank you again @steppenwolf and @x6gas

Excellent news - very pleased for you!

[Craney]

One day I will re-read all this and it will make sense.....

but what really is crystal clear,  is the wonderful community of help that  these pages offer when things go egg-shaped.  Well done everyone.  

[x6gas]

14 hours ago, rubecula said:

Ha ha...glorious round stars, corner to corner! Thank you again @steppenwolf and @x6gas

Ummmm.  I think that @steppenwolf deserves all the credit.  If I am honest I don't understand - at all - how on Earth a problem with the OAG caused your problem.  If the OAG and guide camera are securely anchored I can't see how it was causing this trailing.  But then, that's why Steve is the Scope Doctor 🌙🪐⭐🔭 👨‍⚕️

[steppenwolf]

17 hours ago, x6gas said:

Ummmm.  I think that @steppenwolf deserves all the credit.  If I am honest I don't understand - at all - how on Earth a problem with the OAG caused your problem.  If the OAG and guide camera are securely anchored I can't see how it was causing this trailing.  But then, that's why Steve is the Scope Doctor 🌙🪐⭐🔭 

I don't think so, your idea of re-purposing PSFimage for star shape detection was inspired!