Focus shift.

Hello all, Well I got my Siggy 50mm 1.4 back from Sigma repair Friday fixing SF issue. (only a 9 day turn around BTW, from my door, fixed and back to my door) and I encountered an issue that I haven't seen before and I didn't even know what to call it until I recently read someone else's post with same type of problem on a Canon lens.

from 1.4 thought to 3.5 the focus shifts. Most notably between these three stops 2.5,2.8,3.2. The focus moves like the whole distance of the center AF point. If you made a .gif out of it, it would look like three different pictures took with thee different focus point of the same subject.
The focus goes further away from the lens ie, the 2.5 is closer focus, the 2.8 is further, and 3.2 is further out yet, after that the aperture opens up and it's hard to tell if it shift that much.

The tests where done with shutter release, 2 second timer, very tightly attached to weighted tripod, 45 degree angle, center point, single shot, iso 100.

I never heard of this, is it normal, can it be fixed? Should I care, should Sigma care?

BTW, Soon as I got the lens it was so very soft I returned it straight away so this is my first real tests with the lens or I would have brought that to there attention when it went out two weeks ago.

No one?

I've tested two different SIgma 50s a while ago, and both had a bit of focus shift. I found it started at around F/1.6, and was gone by about F/2.8, with the maximum shift occurring at F/2. The amount of shift was about 1 cm at a focusing distance of 50 cm, so well outside DoF. These two specimens didn't focus very well at distances of 4 to 5 m and beyond. I also reported this a few times here and there on this forum. I still think that the Sigma to me is the second best AF 50 mm out there currently, but you may have to have a lot of patience getting a good one or have it calibrated properly.

Regarding the front focus you report: that is really weird, because focus shift is by definition back focus. So either this is a calibration thing, which it could be, or something else is causing this.

Do note that AF, just like MF, is never exact. It always is approximate, and varies with every attempt at focusing. In order to make sure, you need to do at least 10 tests or so for a single aperture, and with a Canon body you have to focus at a vertical contrast change, rather than a horizontal one.

Furthermore, I also suggest you don't do a 2 second timer stop between MLU and picture, but a 10 second one, or even better, use a remote switch. With timer operation you always have to touch the camera, and this introduces unknowns into any test. A time of less than 3-4 seconds between MLU and actual shot is not a good idea, because the camera may still show some residual vibrations. This is why I have changed my technique quite a while back to allow for at least 10 seconds between MLU and actual shot - I noticed a difference, even indoors.

With the 40D you need to shoot in proper daylight, or flash with assist light, for proper focus.

BTW. I described here how I propose to do these tests, maybe this helps:
https://photography-on-the.net …hp?p=8307750&po​stcount=16

A final comment wrt AF variation: I read an article somewhere, once where it was postulated (and tested) that AF easily varies plus or minus 5 mm between shots, with lenses with accurate AF motors, at distances of approximately 50 X FL. Unfortunately I can't find that article anymore. The best way therefore to test for focus shift is focus at maximum aperture in LiveView mode, at maximum aperture, and have the camera ideally tethered to a computer for control of the camera.

Anyway, HTH, kind regards, Wim