Do people re-collimate their Newtonians that often?

f/5 is far more forgiving then a f/4 but even then I don't recollimate that often. With the 12" maybe once a season, mostly checks and tweaks. The f/6.3 possibly once every 1-2 years, when I used to use it and then small tweaks. The 6" f/4 have been recollimated quite recently, to make sure it was all set correctly after having been left alone for few months. I had to tweak the secondary and the primary as the image was showing signs of off-axis coma on the upper left quadrant.

Fact is, once you have a good (in the sense insensitive to small tilt/despace variations) coma corrector precise collimation is less of a concern as the image is corrected off-axis as on-axis. Obviously the faster the system is the tighter are the tolerances.

Jean-Baptiste:
I took a look at your crescent nebula, does this show that the collimation needs to be checked ? For me the answer is yes, this is not a big deal but I think you could get better results after this step.

Thank you!  To enable me to learn and improve, what was it in the Crescent nebula image that pointed out the need for better collimation? I really am trying to learn here, so please be as brutal as you like with the feedback.

Look at the stars shape, take a lot for example at the open cluster of CCDMike , the stars have  a regular round shape 
sorry I am in the transport with am iPad, not easy to bring out the details with copy paste of images :-)

Clear skies

I think it depends on two things - how well the scope is set up "factory side", and your experience. And generally the built quality, obviously. Some scopes hold their collimation better than others. 

Took me quite a bit of time until I had it figured out, until then ..yes, I was fiddleing with the scope all the time. Simply because I was unable to determine what the underlying causes of certain problems were.

In my admitedly limited experience, I have found that my f/4 Newtonian holds collimation very well between sessions. 

The most useful procedure/reading I found is this one; a bit long, but very well worth the read to understand exactly what is happening and some of the pitfalls:

https://www.cloudynights.com/topic/251778-concise-thread-about-autocollimatorsimprovements/#entry3189737

I use a barlowed laser to check for primary axial error and an autocollimator to check for focuser axial error. When both instruments simultaneously show good collimation, I am confident that my scope is collimated. The thread above goes into why just a single tool is sometimes insufficient. The first time took a few back and forth iterations. After that, I generally find that very minor secondary adjustments are needed.

Andy Wray:

Jean-Baptiste:
Look at the stars shape, take a lot for example at the open cluster of CCDMike , the stars have  a regular round shape 
sorry I am in the transport with am iPad, not easy to bring out the details with copy paste of images :-)

Here's my latest image (corners and centre) .. What do you see?  I do see some tracking issues, but there may be some collimation issues also?

In my opinion, things are quite simple if you take a few subs  and look at the stars, comparing how they look. 

- Collimation/alignment issues result in non-uniform stars. E.g. stars will be wider at the bottom right of the sub. There is a change if you rotate the camera (the place where stars get wider will follow the rotation).
- Tracking/guiding issues result in elongated stars, along the same axis: all your stars will be 0-shaped. They also get worse with longer exposures.
- Correction issue tend to be radially symmetric. If stars are smaller at the center and wider as you go towards the edges of the field (all edges), then you have a correction issue. Wrong distancing, perhaps the sensor is too wide for the corrector, perhaps the corrector just isn't good.
- If the stars have strange shapes but consistently so (let's say every star is shaped like a triangle, all triangles pointing the same way, or let's say the color blue is not aligned to the other colors) then you have a build/optics problem. 
- Anything else is usually an accident. E.g. extra diffraction spikes typically mean some cable got in the way. 


In the picture that you show you have a clear (but mild) tracking issue because all stars (including central ones) are slightly elongated along the same axis (like a / ). There is perhaps a very small alignment issue because to me at least the bottom-right corner looks slightly better than the others. But it is even milder, too mild to bother in my opinion. Mind you, it does not have to be collimation. It could be e.g. a tiny tilt of the focuser due to weight or even of the camera sensor from its construction. I would not worry about it. Best to focus on the tracking/guiding if you want to see a visible impact on your images (again: personally I wouldn't worry about that either, I think it's too mild to bother.  And it is not necessary to change anything physically. Perhaps just take slightly shorter subs or reject a little more aggressively).

Regarding the original question, I must note that even though my F/5 holds collimation excellently, aligning the secondary was quite a nightmare and I had to try it three times (several hours every time) to get right. What helped me was the concentric circles collimation eyepiece.

Cheers,
Dimitris

Hello @Andy Wray 
a quick feeed back, much easier

I had a look at your crescent pictures, for me, what I see is linked to the primary axis alignment, even if the resolution is low,





It is somtehing that you see on all the sensor

Other candidate may be pinched optics, but I don't see this as being the guiding (I am talking of the crescent nebula picture !)

I don't see that in the latest pictures : Horse head and M1

This is not a big deal, I am pointing that because you asked where to look for if you want to get better results

My 2 cents advice would be :
- learn the star test with your sensor in place, it should be your final judge with your own eyes, not someone else' eyes ;-) (The guide referenced here above that Tommy Nawratil from Lacerta wrote is a must) 
- another point that you could try is a "primary ring" that blocks the effects coming from the edge of the primary, this greatly reduces some stars halo :


If I take your Alnithak, an extreme case you have :


I show an example that I took with my primary mask :



Clear skies :-)