Acceptable AFMA difference for zoom between W and T ends?

[entlassen]

I've got two new copies of the 24-70mm f/2.8 II (playing the zoom-tick lottery). I'm wondering, what would be an acceptable AFMA difference between the Wide and Telephoto ends?

1st copy: -1 at 24mm and +1 at 70mm (so a difference of 2).
2nd copy: +2 at 24mm and +7 at 70mm (so a difference of 5).

I'm pretty happy with the first copy, but I'm not sure about the second copy. Practically, I will input the two AFMA values into my 5DM3 and it will interpolate the AFMA for focal lengths between the two extremes. My question is more general, which is "what is an acceptable difference before you should consider returning the lens to the store?"

The main reason I'm asking this question is because I read Roger Ciala's old article about how to test a lens, and he suggested that +/-3 between the two ranges for a 4x zoom is acceptable. Since my second copy has a difference of 5, I was wondering if I should be concerned. I'm also curious what you guys think the biggest difference is before you'd return a lens back to the store.

http://www.lensrentals.com/blog/2010/11/how-to-test-a-lens

My absolute requirements for an acceptable lens are:
...
3. Front / backfocus is similar throughout zoom range and at near and far focusing. How similar is a judgement call, but if it changes less than +/- 3 for a 4x zoom and +/- 5 for a 10x zoom throughout the zoom range, then I’m pretty happy.

P.S. I'm not certain about the meaning of his terminology of "4x zoom" and "10x zoom", but I'm assuming that it means it's the longer focal length divided by the shorter focal length? So 24-70mm and a 70-200 would be <4x zoom, and a 18-200mm would be >10x zoom. Please correct me if I'm wrong.

 

[Loswr]

My 24-70/2.8L II is W = 0, T = +5, which is the largest spread of any of my current zoom lenses on my 1D X. I'd say the range isn't as important as the linearity. I use FoCal, and I all lenses at 25x and 50x the focal length, and test zooms at multiple focal lengths, the ends plus intermediates - e.g., 2x zooms at one intermediate (16-24-35), 3x zooms at two (24-35-50-70), etc. (within reason - I didn't test my 28-300L at ten FLs!).

My 24-70 at the intermediate settings follows the linear regression based on the ends of the range (+1 at 35mm, +3 at 50mm).

 

[Mt Spokane Photography]

I've got two new copies of the 24-70mm f/2.8 II (playing the zoom-tick lottery). I'm wondering, what would be an acceptable AFMA difference between the Wide and Telephoto ends?

1st copy: -1 at 24mm and +1 at 70mm (so a difference of 2).
2nd copy: +2 at 24mm and +7 at 70mm (so a difference of 5).

I'm pretty happy with the first copy, but I'm not sure about the second copy. Practically, I will input the two AFMA values into my 5DM3 and it will interpolate the AFMA for focal lengths between the two extremes. My question is more general, which is "what is an acceptable difference before you should consider returning the lens to the store?"

The main reason I'm asking this question is because I read Roger Ciala's old article about how to test a lens, and he suggested that +/-3 between the two ranges for a 4x zoom is acceptable. Since my second copy has a difference of 5, I was wondering if I should be concerned. I'm also curious what you guys think the biggest difference is before you'd return a lens back to the store.

http://www.lensrentals.com/blog/2010/11/how-to-test-a-lens

My absolute requirements for an acceptable lens are:
...
3. Front / backfocus is similar throughout zoom range and at near and far focusing. How similar is a judgement call, but if it changes less than +/- 3 for a 4x zoom and +/- 5 for a 10x zoom throughout the zoom range, then I’m pretty happy.

P.S. I'm not certain about the meaning of his terminology of "4x zoom" and "10x zoom", but I'm assuming that it means it's the longer focal length divided by the shorter focal length? So 24-70mm and a 70-200 would be <4x zoom, and a 18-200mm would be >10x zoom. Please correct me if I'm wrong.

The old article refers to a camera that only has one AFMA setting. The 5D MK III has a wide and a telephoto, and it varies the AFMA linearly at points in between. Either lens should be fine, the old article you are quoting doesn't apply to cameras with two AFMA points.
The real question as Neuro noted, is how does the lens do at the intermediate focal lengths. If the AFMA falls in-between, the extremes in a linear fashion, its going to be fine.

My 24-70 is +1 wide and -5 Tele, and its just fine.

 

[entlassen]

I just sent an email to Roger Ciala and he replied back saying he would still "like to be within 3 AFMA values throughout the range of the zoom. " I don't know if he's being a little tight on his requirements, but numerous lenses have passed through his company and that's the preference he seems to have reached.

 

[AlanF]

I just sent an email to Roger Ciala and he replied back saying he would still "like to be within 3 AFMA values throughout the range of the zoom. " I don't know if he's being a little tight on his requirements, but numerous lenses have passed through his company and that's the preference he seems to have reached.

If the change in AFMA is linear from wide to telephoto then the two point AFMA calibration should be adequate, irrespective of Roger Ciala's reactions. Just do what neuro did and test your lens at intermediate zoom lengths.

 

[Canon1]

I personally would be ok with a range of 40 if indeed the optimal AFMA settings were -20 and + 20 for W/T. (This assumes that the AFMA point changes in a linear fashion throughout the zoom range.

Also, what kind of shooting do you do? If you primarily shoot landscapes (like me) then then it would be less important to have a "perfect lens in regards to AFMA" as the smaller apertures of f8 or f11 (And manual focus much of the time) and the greater distance to subject will render AFMA settings irrelevant.

If you shoot a lot of subjects near the min focus distance with wide apertures then it would be far more critical.

+/- 5 is really not a big deal IMO. I would keep the lens which has the higher IQ value.