Problem with a zoom lens on Canon AF camera?

[Ari]

Hi all,
It's been 20 years since I used a 35mm, and back then it was a manual-focus Contax.
I just bought a Canon EOS-1 with a Sigma 17-35mm zoom lens and I noticed something strange.
If I focus with the lens at 17mm, the AF works quickly and accurately, or so it seems.
When I zoom to 35mm and re-focus (by releasing and re-pressing the shutter button halfway), I see that when it focused at 17mm, it was slightly off. At 35mm, it seems to be spot-on.
Is this merely a question of using a wide angle lens with too much to focus on, or is there something really wrong with this lens/camera or both?
The viewfinder is not standard, it was replaced with a grid screen by the last owner, so I'm unsure exactly where the AF point is, which could also be the cause.
Thanks for your help.

 

[wiltw]

Unlike what was found in better zoom lenses for manual focus cameras, which were more commonly 'parfocal' and holding focus as you zoomed, few modern zoom AF are 'parfocal' and unchaning of focus point as you zoom.
The modern AF zoom lenses are most commonly 'varifocal'...they do not hold focus as you change FL!
So the old method of zooming to tele FL to focus, and then widen FL for the actual shot does not work very well today.

 

[Ari]

I don't think it's a question of focus changing as I zoom. AF seems to be inaccurate at 17mm, and better at 35mm.
I'd expect AF to work equally well at both ends of the zoom.

 

[sfphoto]

Even current lenses sold new can have this problem. I just read a review yesterday of a Panny M4/3 lens which said it had to re-focus when zoomed.
At 17 (or 24 and wider) you should consider using the hyperfocal distance.

 

[Ari]

Hyperfocal, eh? I bought the lens to use it wide open with IR film, so HF could be problematic.
I could zoom in and out quickly, too, but that would be a minor annoyance.

 

[MattKing]

I bought the lens to use it wide open with IR film

Will the AF even work behind an IR filter?
And depending on the film, are you planning to account for the IR focus shift?

 

[Ari]

Will the AF even work behind an IR filter?
And depending on the film, are you planning to account for the IR focus shift?

I bought a #29 filter, not opaque but dark. I hope it'll work, but it hasn't arrived yet.
Modern IR films are pretend IR, so they don't need to be re-focused.
I've been through a lot of Rollei and Maco 120 film and I just focused as I would normally.

 

[Ari]

Ok, I took the camera and lens outside for a long walk-around and it focused much better.
These older AF systems aren't like what we've got now with digital cameras, they need discernible patterns and well-lit objects in order to work properly.
Problem solved, thanks for your input!

 

[Ari]

Will the AF even work behind an IR filter?

I'm happy to report that the EOS-1 and Sigma lens work very well with the Haida IR720 filter in front of the lens. This a nearly-opaque filter.
The #29 filter should work just as well.

 

[wiltw]

Even current lenses sold new can have this problem. I just read a review yesterday of a Panny M4/3 lens which said it had to re-focus when zoomed.
At 17 (or 24 and wider) you should consider using the hyperfocal distance.

The issue with the hyperfocal concept is that old manual focus distance scales were much larger rotationally, and permitted many many more specific distances on the scale, so it was much easiler to esitmate distances.
The AF of today is rotationally compressed, to preserve battery life, and there are few distances on the scale.
Looking at my Canion 17-55mm, the distances are 1.2', 2', 3', 5' and infinity...so where would I put the lens scale when my Hyperfocal distance is 35' for 17mm f/4,, or even 8' Hyperocal for 20mm f/8?!

 

[wiltw]

Ok, I took the camera and lens outside for a long walk-around and it focused much better.
These older AF systems aren't like what we've got now with digital cameras, they need discernible patterns and well-lit objects in order to work properly.
Problem solved, thanks for your input!

I was going to raise that exact point, that AF of 30 years ago was crude in comparison! that, and the possibility that the AF sensor or auxiliary reflex mirror had jarred out of position in 30 years of life.

 

[sfphoto]

The issue with the hyperfocal concept is that old manual focus distance scales were much larger rotationally, and permitted many many more specific distances on the scale, so it was much easiler to esitmate distances.
The AF of today is rotationally compressed, to preserve battery life, and there are few distances on the scale.
Looking at my Canion 17-55mm, the distances are 1.2', 2', 3', 5' and infinity...so where would I put the lens scale when my Hyperfocal distance is 35' for 17mm f/4,, or even 8' Hyperocal for 20mm f/8?!

For a wide zoom as described I use a DOF chart like this:
https://www.photopills.com/calculators/hyperfocal-table

And I typed up my own chart and printed on a mailing label hen placed on the inside of the lens cap.
FL - f4 - f8 - f11
17 - 8' - 4' - 3'
20 - 11' - 5.5' - 4't
etc.

I suspect now-a-days many would use an interactive chart on their smart phone.

 

[sfphoto]

I was going to raise that exact point, that AF of 30 years ago was crude in comparison! that, and the possibility that the AF sensor or auxiliary reflex mirror had jarred out of position in 30 years of life.

The lens review I recently read [mentioned above] stated that the (current production) lens they were reviewing (on a current camera) had the issue the OP here mentioned and that they had seen it in other lenses.

 

[wiltw]

The lens review I recently read [mentioned above] stated that the (current production) lens they were reviewing (on a current camera) had the issue the OP here mentioned and that they had seen it in other lenses.

Interesting to learn that.