PHD and getting perfectly round stars

[Flea77]

In my effort to improve my astrophotos it has come time to address a few things including trying to get perfectly round stars. From what I have read this seems to be a function of:

1: Getting it polar aligned correctly. I use a compass to get the mount in the correct general direction, then a dual level to get it leveled. I have a polar scope with the offset (put polaris in the little circle off to the side of the crosshairs type).

2: Guiding. I am using PHD with the Orion Awesome Autoguider package (SSAG, ST80).

The results are good, but not fantastic. What is more interesting is that sometimes it can change throughout the night (yes, everything is tightened as tight as I dare, no pliers are used and nothing cracks). Here is a sample from last night, the last run of the evening (or should I say morning, about 3am). The target is M101, ISO800, 10 minute exposure (you can not see any part of M101 in this image):

The image above is a crop from the original 4991 pixels wide down to 300 pixels wide, then enlarged to 600 pixels to show extreme closeup. The crops come from near the center of the image, and the far corners may show slightly more of a problem but that has been mostly corrected by my new field flattener.

I fear it may be because I am using all the default settings for PHD but really am unsure if it is PHD or my ineptitude since I am still very new to all of this so I am attaching a couple of graphs from PHD below:

Anyone know from looking at all of this what is going on, or what direction I should proceed to solve this?

Thanks in advance.

Allan

[rfdesigner]

Can you point to North in your image?

As you have your DEC on Auto, it may be switching directions, which I found to be a problem, in which case it is better to choose the appropriate direction and stick to it.

i.e. try south, if the DEC error just builds and builds then choose north.

I was playing with my settings two nights ago and things got much better by ditching the Auto setting.

Derek

[Flea77]

Derek,

The target was at about 42 degrees from north, about 34 degrees high. This would mean that polaris would be towards the left edge of the image.

Is that what you wanted?

Allan

[rfdesigner]

Ah.. once again I didn't explain myself properly, sorry:o.

I was wondering if your stars were elongated in RA or DEC or a combination of..

If all in DEC then that's the problem Axis...

It should be possible to add a little North Arrow to your image, I'm guessing it would point to the top right or top left of your image?

Derek

[Flea77]

Here is where things were when I imaged this, roughly.

I hope this helps.

Allan

[George]

Looks like coma to me rather than tracking.

[Flea77]

Looks like coma to me rather than tracking.

I thought coma was only an issue with reflectors?

Allan

[Stephen]

Chromatic aberration? There appears to be blue above and red below.

Could we see a full frame?

[rfdesigner]

Here is where things were when I imaged this, roughly.

Allan

No.

I'm not talking about where this image is in the sky.

I'm talking about the orientation of the image in the sky.

Take a look at the little M51 image next to my name in this message, North is more or less directly right. (at 3 O'clock). If all the stars were streached in DEC then I would expect them to be streached left to right in my image.

Make sense?

(looking at your Phd plots I can see what looks like the DEC line reversing a couple of times, which could cause star elongation)

Derek

[Flea77]

No.

I'm not talking about where this image is in the sky.

I'm talking about the orientation of the image in the sky.

OK, sorry, I just figured that if you saw where it was in the sky at the time I imaged it, and saw my sig line to see I am using an EQ mount, it would have been fairly obvious where north was in relation to the image, my bad. I guess some people could image with their camera mounted sideways or upside down.

If you mean north as in the direction from the center of the image to Polaris, north would be roughly the lower left corner of the image.

Allan

[Flea77]

Chromatic aberration? There appears to be blue above and red below.

Could we see a full frame?

Here you go....

Allan

[dave_galera]

Its not coma as the pulling of the stars are all in the same direction and not pointing to the centre. Looks like chromatic aberration to me as well...maybe, as I wouldn't have thought that an RMS value of 0.22 would give that bad a shape of start.

Might be worth drift aligning the scope just to make sure its not polar alingment problem.

Can we see an image with greater resolution?

[kookoo_gr]

when you image you should follow a flow in order to find what's wrong with the guiding.

in your shoes first of all i would check the balance of the scope. some prefer to have the balance of the scope a bit counterweight heavy, others like myself prefer to have perfect balance at the whole setup.

also make sure that no cables are tangled at the setup.

drift align the mount so there will be less corrections from the phd, there are topics on how to drift align your mount. Also since you image with a refractor i suggest to polar align with the whole setup at the mount since there is a difference at polar alignment with and without the equipment.

Check the focus of your scope before imaging, a bahtinov mask is the right tool for the job, also after focus lock the focuser drawtube and check regularly to see if it has any slop that may cause the image to get out of focus . i found that if you get out of focus then the stars can get elongated which may lead to hair pulling situations as you will think it is flexure.

A good trick (as others mentioned) to locate where the scope drifts is to orient the camera according to the mount's movements. you can place it in way so that the horizontal axis corresponds to the the RA movements of the mount and the vertical axis corresponds to the DEC movements (in a few words place the camera at the focuser horizontally or vertically)

If the drift is at the RA axis then you can play with the RA aggression and HYS values. others mentioned that you need to make small adjustment at the MAX RA value. also try changing the min motion value. If you make any changes at the software i suggest to start with the min motion value

Use the ASCOM pulse guiding option (via the usb port) for guiding since you adjust the pulses better from the EQMod software. If you use ST4 guide pulses then you can have guide errors since the pulses are fixed at this kind of guiding.

[rfdesigner]

OK, sorry, I just figured that if you saw where it was in the sky at the time I imaged it, and saw my sig line to see I am using an EQ mount, it would have been fairly obvious where north was in relation to the image, my bad. I guess some people could image with their camera mounted sideways or upside down.

If you mean north as in the direction from the center of the image to Polaris, north would be roughly the lower left corner of the image.

Allan

No problem.

That means you're seeing the image blur in right ascention, not declination.

With the guiding graph you've shown that strongly suggests it isn't a guiding error as, if anything, the graph suggests more problems in DEC than RA.

It could as others here say be a chromatic abberation.. not sure how you manage to do this.

Ohhhhhhhhh!

I just tried splitting your magnified image into R,G & B images. The stars all look pretty round in the single colour images.

I *THINK* you may have slightly different image scales or perhaps a debayering issue.

Whatever, you might be able to correct it by splitting into RGB, then realigning and recombining. If that works across the entire field then you could have a solution.

see the result of aligning:

Derek

[26Left]

I recall reading somewhere (can't remember where) that the distance of the guide star from the target comes into play. There is even a formula for calculating the errors that would be introduced with a given distance between the two.

My own experiences seem to bear this out. With galaxies, it can sometimes be difficult to find a guidestar close by, and this shows as a slight elongation in the stars. I took an image of the flame/horsehead region a couple of weeks ago using Alnitak as a guide star, and the stars could not have been rounder.

Tom.

[Flea77]

Its not coma as the pulling of the stars are all in the same direction and not pointing to the centre. Looks like chromatic aberration to me as well...maybe, as I wouldn't have thought that an RMS value of 0.22 would give that bad a shape of start.

Might be worth drift aligning the scope just to make sure its not polar alingment problem.

Can we see an image with greater resolution?

Here you go!

http://www.allans-stuff.com/images/122911-068-2048.jpg

in your shoes first of all i would check the balance of the scope

Balance should be pretty dang close, I check it before each imaging session and although I can not tell you which side is a couple of grams heavier than the other, it is balanced as well as I can get it.

Also since you image with a refractor i suggest to polar align with the whole setup at the mount since there is a difference at polar alignment with and without the equipment.

I polar align using the polar scope built into my Sirius mount, and I do it with it completely assembled ready to image, immediately before I do the scope's three star alignment process.

Check the focus of your scope before imaging, a bahtinov mask is the right tool for the job, also after focus lock the focuser drawtube and check regularly to see if it has any slop that may cause the image to get out of focus . i found that if you get out of focus then the stars can get elongated which may lead to hair pulling situations as you will think it is flexure.

Focus is checked with a bahtinov mask immediately after the three star alignment process, right before getting on my first target of the evening. The focus lock is a nice heavy screw lock under the focuser and is always very tight. Focus is redone every time I change filters to ensure focus stays exact.

Use the ASCOM pulse guiding option (via the usb port) for guiding since you adjust the pulses better from the EQMod software. If you use ST4 guide pulses then you can have guide errors since the pulses are fixed at this kind of guiding.

Now this is one I have not heard of, will have to read up on this and see what it may do, thanks for the suggestion.

Ohhhhhhhhh!

I just tried splitting your magnified image into R,G & B images. The stars all look pretty round in the single colour images.

I *THINK* you may have slightly different image scales or perhaps a debayering issue.

Whatever, you might be able to correct it by splitting into RGB, then realigning and recombining. If that works across the entire field then you could have a solution.

Now this is weird. The images you are looking at were taken with a Nikon D7000 in RAW mode, opened with Photoshop CS5 using the latest version of Adobe Camera Raw, then saved as a JPG using PS's "Save For Web & Devices" function. So while I am certainly not arguing the point with you, I can not see how in the world that could be possible.

Allan

[Flea77]

I recall reading somewhere (can't remember where) that the distance of the guide star from the target comes into play. There is even a formula for calculating the errors that would be introduced with a given distance between the two.

My own experiences seem to bear this out. With galaxies, it can sometimes be difficult to find a guidestar close by, and this shows as a slight elongation in the stars. I took an image of the flame/horsehead region a couple of weeks ago using Alnitak as a guide star, and the stars could not have been rounder.

Tom.

Guiding with this target would have been done with that bright star just to the left of M101 in the image in post #11 above. Is that not close enough? Should I have used one somewhere else in the frame?

Allan

[rfdesigner]

Now this is weird. The images you are looking at were taken with a Nikon D7000 in RAW mode, opened with Photoshop CS5 using the latest version of Adobe Camera Raw, then saved as a JPG using PS's "Save For Web & Devices" function. So while I am certainly not arguing the point with you, I can not see how in the world that could be possible.

Allan

It could just be that your scope is producing images of different wavelengths slightly differently, I would normally anticipate the RG&B layers to be fractionally different scales with a refractor, and it would be possible to have them slightly misaligned if the refractive index in the various lenses weren't perfectly even.

Derek

[FraserClarke]

It could just be that your scope is producing images of different wavelengths slightly differently, I would normally anticipate the RG&B layers to be fractionally different scales with a refractor, and it would be possible to have them slightly misaligned if the refractive index in the various lenses weren't perfectly even.

You might just be seeing atmospheric dispersion. You're observing quite low down (34 degrees above the horizon), and this altitude the atmosphere itself acts like a (bad) lens and smears out the image by ~3 arcseconds, putting blue "on-top" and red "on the bottom" (I think it's that way round). Nothing you can do about it apart from build some very complex moving optics to correct it (trust me -- you *don't* want to do that! ). Best thing is just to realign the RGB channels after the fact as Derek suggested.

What's the pixel scale of the image roughly? (how many arcseconds per pixel??). If you work out the separation between the red and blue images in arcseconds, it should be ~3 arcseconds from atmospheric dispersion.

If you are seeing the atmospheric dispersion, I'd say your guiding is pretty darned impressive actually!!

[26Left]

Guiding with this target would have been done with that bright star just to the left of M101 in the image in post #11 above. Is that not close enough? Should I have used one somewhere else in the frame?

Allan

I that case, I'd guess that it's not an issue. It's only become an issue with me when the guide star is a few degrees away.

[Beyond_Vision]

What method of averaging did you use when stacking and have you checked each frame to see if any have this elongation?

[Flea77]

You might just be seeing atmospheric dispersion. You're observing quite low down (34 degrees above the horizon), and this altitude the atmosphere itself acts like a (bad) lens and smears out the image by ~3 arcseconds, putting blue "on-top" and red "on the bottom" (I think it's that way round). Nothing you can do about it apart from build some very complex moving optics to correct it (trust me -- you *don't* want to do that! ). Best thing is just to realign the RGB channels after the fact as Derek suggested.

What's the pixel scale of the image roughly? (how many arcseconds per pixel??). If you work out the separation between the red and blue images in arcseconds, it should be ~3 arcseconds from atmospheric dispersion.

Now that's depressing, something I can do nothing about. Interestingly enough though, I see the same thing roughly with targets directly overhead.

I have no idea what my pixel scale is, I am shooting with 770mm fl and an 18mp cropped sensor if that helps.

What method of averaging did you use when stacking and have you checked each frame to see if any have this elongation?

It is in every light frame. I am using Auto Adapted Weighted Average with 5 itterations.

I should also point out that I purchased a field flattener and that did not seem to change much. It may have helped a little but I can't really tell because older targets are not as well in focus (I bought a mask for focusing) so everthing looks worse.

Allan

[dave_galera]

On the full size image, the RGB channels are not aligned over the entire field..........still playing with it lol

[Flea77]

I am begining to think it is not PHD but that my field flattener is not working as it should. Here are five images of M101 I took the other night...

Top left:

Top right:

Bottom left:

Bottom right:

and center:

All of these are the exact same size crop. Maybe I need a spacer?

Allan

[Flea77]

Here is a set from an image before I started using the field flattener...

Top left:

Top right:

Bottom left:

Bottom right:

Slightly to the right of center:

It seems much improved, but not quite enough. Would a spacer (they sell kits just for this I think) help it more?

Allan