16-35mm f/2.8L II possible lens blur problem

hiki08 wrote:

Thanks Mitch. Wonder if you may clarify the reason you say Canon might not be able to do a lot in the case of a collimation?

At F/2.8 the lateral  <x,y>  accuracy is on the order of 0.001" and the <z> number is on the order of 0.003" for every set of lens groups in the whole optical train. A good (i.e., bad) drop can distort the lens frames to the point where the lens needs to be completely taken apart and rebuilt from the lenses--not something Canon can do in a repair shop.

Higher end Telescope manufacturers build in adjustments so that we users can collimate the field perfectly. A Ritchey Chretien telescope has some very tight collimation tolerances, and some of the larger professional telescopes take a whole month to collimate down to the 0.001" level for an 8 meter optical system.

Not sure if it was mentioned above but the lens was previously dropped on concrete once, and I was told by Canon that the lens elements misaligned after. Their first fix apparently didn't fix the problem (would this be what you speak of? i.e. that it's very difficult to completely align the lens elements out of the original factory so there will always be collimation once it happens?).

I'm not particularly seasoned at astrophotography (I just like looking at the stars and take the occasional photos of them while on holiday), but I've also only ever heard of people suggesting to use the largest aperture when taking photos of the night sky. Wonder if your suggestion of a smaller aperture is only to see if it is a lens problem or an inherent lens characteristic?

Telescopes are designed to be used ONLY at maximum aperture, and the entire optical train is optimized at wide open--sometimes to the point where is you stop down you will reintroduce spherochromatism that was nulled wide open that is no longer nulled stopped down !?!

Take a minute to look up the MTF graph of your lens at several apertures and use an aperture where the MTF graph is as high as possible for as wide as possible, and time the exposure appropriately.

Thanks for the tip on star trail movement. I usually go for around 15 seconds, at 16mm, that should be a lot less movement?

16mm gives you 50% more time before blur sets in. Higher resolution pixels give you less.

I might take a few photos with a simple patterned subject over the weekend to see if there is noticeable uneven blur.

Got to take these wide open to see the collimation effects.