Zoom Lenses MFA variance

For those who have a good handle on MFA, how much variance are you seeing in your zoom lenses?

Using FoCal I have found that my 80-200/2.8 L requires -10 on the wide end and -2 on the long end. I thought this was a big variance and attributed it to the age of the lens.

While browsing at Future Shop yesterday I found a good deal on an open box 70-200/4 L (non-IS) so I was playing with it today and found basically the same variance (wide -11, long -2)!

Is this normal or did I find a possible reason this particular lens was returned to the store?

I find a lot of variance with standard zooms on my 50D: Tamron 17-50 and two copies of Sigma 17-50 had a substantial range in MFA. (Tamron was the worst, with something like -10 ... +7 at different ends). My only experience with telephoto zoom - used 70-200 f4 - seems to be behaving much better: something like -4...-2 using FoCal and halogen lamp. (But seems more like -6...-4 using daylight - needs more testing.)

Do you have any fast primes - what MFA do they need? I have a hypothesis that if the camera is mechanically off (which should show in fast primes having comparable non-zero MFA), zooms should show more variance when going from one end to another. All my primes cluster around -3...-4, which suggests the camera is a bit off. I wonder if my standard zoom would behave better if the camera was mechanically calibrated.

I would also check AFMA variances among different AF points. Enable off-centre AF points in FoCal. I had a situation where I had a difference of around 9 AFMA values between the center point and one of the outer cross points, which would make the setup utterly unusable. I had initially thought there was a problem when I had a variance of 3 AFMA values, sent in to Canon only to discover that the "fix" made it worse. Had them restore to factory settings.

joeblack2022 wrote in post #15435716
Using FoCal I have found that my 80-200/2.8 L requires -10 on the wide end and -2 on the long end. I thought this was a big variance and attributed it to the age of the lens.

Did you just think it was a big variance, or did you actually check to see how big a difference there was?

If you have FoCal Pro then you can look at the graphs to see how much difference there is at different MFA values. My experience is that, most of the time, there's very little difference over a wide range of values at the top of the graph.

I'd be surprised if there was a large difference between -10 and -6 on the wide end and between -6 and -2 on the long end - certainly nothing big enough without pixel-peeking at 100%. So it's very likely that a value of -6 will be perfectly good.

pulsar123 wrote in post #15437258
My only experience with telephoto zoom - used 70-200 f4 - seems to be behaving much better: something like -4...-2 using FoCal and halogen lamp. (But seems more like -6...-4 using daylight - needs more testing.)

Do you have any fast primes - what MFA do they need?

My testing has been indoors with both halogen and LED light sources and the light level has been around EV 8 or 9 (I think FoCal documentation says EV 10 somewhere?), so perhaps my setup is not ideal.

I tested an 85/1.8 and FoCal came back with +4 required.

s2kdriver80 wrote in post #15437592
I would also check AFMA variances among different AF points. Enable off-centre AF points in FoCal.

I'm currently using Pro so I don't have this option, but may look at it further down the road. Looking at the description on Reikan's page I can just imagine the can of worms it might open!

hollis_f wrote in post #15438313
Did you just think it was a big variance, or did you actually check to see how big a difference there was?

If you have FoCal Pro then you can look at the graphs to see how much difference there is at different MFA values. My experience is that, most of the time, there's very little difference over a wide range of values at the top of the graph.

I'd be surprised if there was a large difference between -10 and -6 on the wide end and between -6 and -2 on the long end - certainly nothing big enough without pixel-peeking at 100%. So it's very likely that a value of -6 will be perfectly good.

Being an MFA rookie, I made an uneducated assumption, hence this thread. I want to get a feel of what is "normal" zoom lens variance and ultimately not be OCD about the whole process.

The suggestion to MFA for a compromise between wide and long ends sounds logical to me. How has that worked out in practice? I have read somewhere that you should MFA based on the long end since DOF increases when shooting wider and would then mask any focusing inaccuracy.

joeblack2022 wrote in post #15438905
I'm currently using Pro so I don't have this option, but may look at it further down the road. Looking at the description on Reikan's page I can just imagine the can of worms it might open!

Pro is the highest level right? That's what I have. Don't do the multipoint focus test. Do the normal fully automatic calibration and just choose an off-center point. After doing the center point, remember not to just simply rotate the ballhead so that the off-center point is over the target, rather move the whole tripod sideways so that the camera is still perpendicular to the target when the off-center point is over the target. If you don't use off-center points often, then it's not too much of a concern.

I use a work lamp that takes a 500-watt bright halogen rod type bulb, set the FoCal temp setting to 2900 K.

If you need some confidence, take a reading twice back to back to see if you get identical numbers.

s2kdriver80 wrote in post #15439315
Pro is the highest level right? That's what I have. Don't do the multipoint focus test. Do the normal fully automatic calibration and just choose an off-center point. After doing the center point, remember not to just simply rotate the ballhead so that the off-center point is over the target, rather move the whole tripod sideways so that the camera is still perpendicular to the target when the off-center point is over the target. If you don't use off-center points often, then it's not too much of a concern.

I use a work lamp that takes a 500-watt bright halogen rod type bulb, set the FoCal temp setting to 2900 K.

If you need some confidence, take a reading twice back to back to see if you get identical numbers.

Sorry meant to say Plus, shouldn't post before coffee in the morning...

I've been running the tests 2-3 times but haven't tried with daylight just yet.

Here's a quick report on what I found using zoom lenses I have access to, interesting that the cheapest lens of the lot has the smallest variance..?

Canon 80-200mm f/2.8 L @ 80mm = -10
Canon 80-200mm f/2.8 L @ 200mm = -2

Canon 100-300mm f/3.5-5.6 @ 100m = -5
Canon 100-300mm f/3.5-5.6 @ 300m = -1

Canon 70-200mm f/4 L @ 70mm = -11
Canon 70-200mm f/4 L @ 200mm = -2

Sigma 17-50mm f2.8 OS @ 17mm = 0
Sigma 17-50mm f2.8 OS @ 50mm = -13

Canon 18-200mm f/3.5-5.6 @ 18mm = -6
Canon 18-200mm f/3.5-5.6 @ 200mm = +5