Help and advice please - eggy corner on subs

[Owmuchonomy]

Ey up.  So last night I set up to test my guiding rig/software after some issues, which is the content of another thread.  The PA and guiding went very well indeed according to the PHD readouts so happy on that front.  However, and despite the moon, I thought I would take some photos for a practice.  For some reason, after previous successful imaging runs in the Spring with exactly the same set up, I now have eggy stars in my bottom left corners.  Here are some crops (low JPEGs) of one sub (they are all identical) taken in the NGC 896 region. The main issue is in the bottom left corner.

I have also put a full frame in of the last run I did in the M81 and M82 area for reference where all the subs had round stars to the edges.  Could you advise me on the likely cause(s) of this scenario and any possible remedies please.

All subs are around 120secs guided.

Equipment:

HEQ5 Pro Synscan

AA 102 EDT APO

Canon 60D modded

Dedicated AA x1 field flattener and Canon M48 adaptor

Guiding: SLX Lodestar, AA f4.1 guidescope, PHD2

Centre

Bottom left (main issue(

Top left

Bottom right

Top right

Springtime sub on M81/M82 (no issues)

Any help and advice much appreciated.

[IanL]

All the corner crops show some elongation in to the relevant corner, though bottom left is worst.  I'd say most likely you have a slight spacing issue between the flattener and the camera chip; if the optical axis is off centre from the chip centre then you'd see more elongation in the corner furthest out from the optical axis - I certainly had a similar uneven elongation effect when I had the wrong spacing with my reducer/flatter and DSLR.

[FunkyKoval35]

Hmm... I see that not only bottom left corner has eggy corners - all corners has, more or less eggy stars. The first thing you need to do is check if your camera is in correct position - distance from flatterer and angle(the perpendicularity to the optical axis).

[Owmuchonomy]

Regarding the above comments I'm not fully convinced. This problem has only just arisen but my set up i.e., spacing has not changed on my equipment. Can you advise how to check the spacing? Many thanks.

[Earl]

i would take the adapters apart give them a clean a very thin gentle wipe of boot polish on the threads and put it back together and see what you get.

[Owmuchonomy]

i would take the adapters apart give them a clean a very thin gentle wipe of boot polish on the threads and put it back together and see what you get.

Hi Earl, a very neat remedy well worth a punt. I'll give it a go. It seems so strange to have this issue now as in the past I've had nice flat fields and round stars.

[Earl]

yes it is odd, i wonder if the threads have shifted over time.

[Owmuchonomy]

All the corner crops show some elongation in to the relevant corner, though bottom left is worst.  I'd say most likely you have a slight spacing issue between the flattener and the camera chip; if the optical axis is off centre from the chip centre then you'd see more elongation in the corner furthest out from the optical axis - I certainly had a similar uneven elongation effect when I had the wrong spacing with my reducer/flatter and DSLR.

Hmm... I see that not only bottom left corner has eggy corners - all corners has, more or less eggy stars. The first thing you need to do is check if your camera is in correct position - distance from flatterer and angle(the perpendicularity to the optical axis).

Thanks both. I'm becoming more concerned about your conclusion.  So last night I tried to investigate this issue further, within my abilities anyway.  I did the following;

1) I took subs of the same sky at 30secs, 2 mins, 4 mins and 8 mins.  The frame corner 'egginess' (direction and scale) does not seem change.  Therefore, am I excluding the possibility of field rotation by this method?

2) The egginess is not uniform across the frame so I am also excluding differential flexure?

3) I swapped my 60D modded for my 600D.  The frames looked pretty much the same.

4) I stopped using the EOS software and did a cable release frame.  Still the same issue, so not a software issue.

There is a very slight egginess at the centre of frames when you zoom in a lot but not anywhere close to that seen in corners.

I did all this after cleaning threads on the image train etc.  So now I'm stuck to know what to do next before I approach the retailer.

I'm confident in my polar alignment using the synscan handset routine (I always get to < 30 are secs of Mel/Maz error before I start guiding).

I don't think looking at the PHD graph helps much but here is a screen shot from last night.

Any further advice much appreciated.

[IanL]

Thanks both. I'm becoming more concerned about your conclusion.  So last night I tried to investigate this issue further, within my abilities anyway.  I did the following;

1) I took subs of the same sky at 30secs, 2 mins, 4 mins and 8 mins.  The frame corner 'egginess' (direction and scale) does not seem change.  Therefore, am I excluding the possibility of field rotation by this method?

2) The egginess is not uniform across the frame so I am also excluding differential flexure?

3) I swapped my 60D modded for my 600D.  The frames looked pretty much the same.

4) I stopped using the EOS software and did a cable release frame.  Still the same issue, so not a software issue.

There is a very slight egginess at the centre of frames when you zoom in a lot but not anywhere close to that seen in corners.

I did all this after cleaning threads on the image train etc.  So now I'm stuck to know what to do next before I approach the retailer.

I'm confident in my polar alignment using the synscan handset routine (I always get to < 30 are secs of Mel/Maz error before I start guiding).

I don't think looking at the PHD graph helps much but here is a screen shot from last night.

Screen Shot 2014-09-10 at 00.02.10.png

Any further advice much appreciated.

1. If it was field rotation you'd expect it to get worse with longer subs - also it would be a circular trail (like long exposure star trails) with all the trails concentric around the centre of rotation (which might be somewhere in the frame or somewhere outside it).

2. Flexure could be taken to mean two things.  Differential flexure would typically produce parallel trails in one direction as the guiding image and the main image shift relative to each other.  I think you mean focuser (or other) droop though, whereby the sensor or (flattener in this case) and the OTA focal plane are not parallel.  I 'd have expected that to produce more of a parallel pattern that a radial one but it might be worth checking.

What you could do is take an exposure, rotate the camera and flattener together (if possible) 45 degrees relative to the focus tube and take another one, then one at 90 degrees to the original. Compare the elongation in the corners between the three images.  If the pattern is more or less the same then that rules out the focuser being out of true with the optical axis (i.e. droop).  If the elongation is always worse in one (physical) direction (usually up/down due to gravity) then you need to fix your focuser.

3. See below on cameras.

4. Can't see any way that software (other than guiding) would have an effect on star shapes.

The PHD graph looks good, but nonetheless guiding or polar alignment problems would not produce radial elongation like this.

You have elongation out from the centre which is diagnostic of a non-flat field.  Scopes/reducers will only produce a flat field of a given diameter, and once you go outside that diameter you will see this kind of problem.  If you were successfully getting round stars right in to the corners previously, that ought to rule out the sensor being too big for the flattener.  Double check a few old and new subs in detail just to be sure though.  

So then you're just left with:

a )  Sensor tilt - if the tilt were sufficient you could end up with the wrong spacing across the chip (one corner or edge too close, one too far away) and that would produce elongation of some sort.  More likely a parallel pattern but I guess anything is possible.  Since you have an unmodded and a modded camera and both have the problem, it isn't inside the camera, so the only place it could be is in the mechanical connection between the flattener and the camera or the flattener and the focuser.  Check all the threads and make sure they are OK and screwed together right.  Also check your M48 adaptor - some of them are made in multiple pieces with screws or similar holding the front and back together - if a grub screw is loose or it wasn't put together right then you could tilt the camera relative to the flattener.

b ) Your spacing between the camera and flattener is wrong.  You should be able to find the optimal spacing range between the flattener and the camera from the manual or check with the supplier.  Bear in mind that you'll need to check where to measure the spacing.  Canons usually have a small 'plimsoll' line (a circle with a line through it) on the body showing where the sensor plane is - mine is embossed next to the flash hot shoe but it varies by model.  Check the manual or supplier for where to measure the spacing from the flattener - it is often the centre of the back of the last glass element but it might be from the rear face at the bottom of the threads or somewhere else.

c ) Check the flattener is ok - are all the elements secure and flat - depends on how it is made but some have big rings that screw in to hold the optics in place with a couple of small slots on each side.  I once had tilt due to my LP filter not having this screwed in properly at the factory.  Easy to fix but hard to figure out unless the glass elements are actually rattling around.

One tool I find invaluable is a set of digital calipers - I got them from Maplin but you can get them on eBay, etc.  You can measure the distances all the way around between difference faces (e.g. between the camera body and the reducer) to assure yourself everything is square, plus you can measure spacings very accurately and they're also handy for measuring focus positions and finding them again in the future without 10 mins of fumbling around trying to get close to focus first.

[Owmuchonomy]

1. If it was field rotation you'd expect it to get worse with longer subs - also it would be a circular trail (like long exposure star trails) with all the trails concentric around the centre of rotation (which might be somewhere in the frame or somewhere outside it).

2. Flexure could be taken to mean two things.  Differential flexure would typically produce parallel trails in one direction as the guiding image and the main image shift relative to each other.  I think you mean focuser (or other) droop though, whereby the sensor or (flattener in this case) and the OTA focal plane are not parallel.  I 'd have expected that to produce more of a parallel pattern that a radial one but it might be worth checking.

What you could do is take an exposure, rotate the camera and flattener together (if possible) 45 degrees relative to the focus tube and take another one, then one at 90 degrees to the original. Compare the elongation in the corners between the three images.  If the pattern is more or less the same then that rules out the focuser being out of true with the optical axis (i.e. droop).  If the elongation is always worse in one (physical) direction (usually up/down due to gravity) then you need to fix your focuser.

3. See below on cameras.

4. Can't see any way that software (other than guiding) would have an effect on star shapes.

The PHD graph looks good, but nonetheless guiding or polar alignment problems would not produce radial elongation like this.

You have elongation out from the centre which is diagnostic of a non-flat field.  Scopes/reducers will only produce a flat field of a given diameter, and once you go outside that diameter you will see this kind of problem.  If you were successfully getting round stars right in to the corners previously, that ought to rule out the sensor being too big for the flattener.  Double check a few old and new subs in detail just to be sure though.  

So then you're just left with:

a )  Sensor tilt - if the tilt were sufficient you could end up with the wrong spacing across the chip (one corner or edge too close, one too far away) and that would produce elongation of some sort.  More likely a parallel pattern but I guess anything is possible.  Since you have an unmodded and a modded camera and both have the problem, it isn't inside the camera, so the only place it could be is in the mechanical connection between the flattener and the camera or the flattener and the focuser.  Check all the threads and make sure they are OK and screwed together right.  Also check your M48 adaptor - some of them are made in multiple pieces with screws or similar holding the front and back together - if a grub screw is loose or it wasn't put together right then you could tilt the camera relative to the flattener.

b ) Your spacing between the camera and flattener is wrong.  You should be able to find the optimal spacing range between the flattener and the camera from the manual or check with the supplier.  Bear in mind that you'll need to check where to measure the spacing.  Canons usually have a small 'plimsoll' line (a circle with a line through it) on the body showing where the sensor plane is - mine is embossed next to the flash hot shoe but it varies by model.  Check the manual or supplier for where to measure the spacing from the flattener - it is often the centre of the back of the last glass element but it might be from the rear face at the bottom of the threads or somewhere else.

c ) Check the flattener is ok - are all the elements secure and flat - depends on how it is made but some have big rings that screw in to hold the optics in place with a couple of small slots on each side.  I once had tilt due to my LP filter not having this screwed in properly at the factory.  Easy to fix but hard to figure out unless the glass elements are actually rattling around.

One tool I find invaluable is a set of digital calipers - I got them from Maplin but you can get them on eBay, etc.  You can measure the distances all the way around between difference faces (e.g. between the camera body and the reducer) to assure yourself everything is square, plus you can measure spacings very accurately and they're also handy for measuring focus positions and finding them again in the future without 10 mins of fumbling around trying to get close to focus first.

Hi Ian

This is very helpful.  Here is my feedback.

1) Agreed and I also lined up the guide scope with the main scope so that the guide star was pretty much central.

2) I was thinking flexure between guidescope and main scope. As you say it would be parallel.  I can't rotate camera and flattener without rotating the focus tube (without loosening threads).  To be honest this scope is made like a battleship so i can't see any 'droop' occurring but who knows.

a) yes, checked this.  The M48 adaptor needs to be put together correctly otherwise the grub screws 'tilt' the flange ring.  I checked with my callipers and its ok.

b Probably the most likely.  The AA website says optimum spacing between M48 back and sensor plane is 51.4mm, tolerance 5mm or up to 10mm with smaller sensors.  Mine is sitting at 55mm.  Notably the test photos shown on the site don't look great with a full frame sensor but with a crop sensor like mine there should be a nice flat frame.

c) Just checked this and refitted it.

While i was on the AA website I notice they now offer a collimator tilt ring for £55.  Maybe they know something i don't.  I will have one more session at fixing it myself then I will contact the retailer or sell all my kit!  Its a shame because the frac is really nice to deal with.

Thanks so much for your wise words.

[Owmuchonomy]

i would take the adapters apart give them a clean a very thin gentle wipe of boot polish on the threads and put it back together and see what you get.

  

All the corner crops show some elongation in to the relevant corner, though bottom left is worst.  I'd say most likely you have a slight spacing issue between the flattener and the camera chip; if the optical axis is off centre from the chip centre then you'd see more elongation in the corner furthest out from the optical axis - I certainly had a similar uneven elongation effect when I had the wrong spacing with my reducer/flatter and DSLR.

So folks, the UK distributor is sending a new flattener, no quibble. Let's take it from there.

[ollypenrice]

While the scope is generously built there are a few mechanical areas I'd check, since I'd bet on tilt as the culprit. Firstly the rotator is a pretty simple affair and might have settled into a droop. I'd loosen it, rotate it and tighten it up again. When tightening it I'd support the back of the scope to help the lockscrew pull the rotated components nicely up against the back of the tube.

I'd also look at the drawtube for any ability to move within its bearings. These are adjustable.

Finally the push-pull collimation screws for the objective might have slackened but these will rapidly decollimate the cell so all I'd do initially is see that none of them is obviously loose.

Olly

[Owmuchonomy]

While the scope is generously built there are a few mechanical areas I'd check, since I'd bet on tilt as the culprit. Firstly the rotator is a pretty simple affair and might have settled into a droop. I'd loosen it, rotate it and tighten it up again. When tightening it I'd support the back of the scope to help the lockscrew pull the rotated components nicely up against the back of the tube.

I'd also look at the drawtube for any ability to move within its bearings. These are adjustable.

Finally the push-pull collimation screws for the objective might have slackened but these will rapidly decollimate the cell so all I'd do initially is see that none of them is obviously loose.

Olly

Thanks Olly. I have already checked all except your last point. So since Harrogate has been and remains blanketed in grey cloud I will trot off and do as you suggest. Thank goodness it's Thursday, that means a spin on the bike with the lads.

[Owmuchonomy]

Ok, so I examined the scope and decided as it is only a few months old and that there are no obvious collimation screws on the objective holding ring that I would not attempt that right now. The objective holder alone has 12 grub screws in it!