Archibald wrote in post #18543188
Well, I think DP is on to something with your lens.
Just as a trial, you could try a simple focus test. Set up three cereal boxes or similar items (flat things with fine detail on them), staggering their distance from the camera. Put one 2 or 3 inches forward, and the other a bit behind. Then in good light, take a pic of them with the lens wide open, focusing on the middle one. Which box is sharpest?
Now change the MFA setting. Make a big change - like from 0 to +10 or +20 - and repeat the test.
You should soon discover if the needed MFA setting is within the range of the camera. If it is, then keep repeating, but with smaller increments, until you get it real close. This is way more time-consuming than doing Dot-tune, but has a chance of rescuing the lens when dot-tune is not working well. Even if it is out of the camera's range, setting it to +20 might make the lens very usable.
The technique described does get you to the adjustment value...the technique I would refine with this comment...
Each time you adjust, set HALF the value between the best pair, in order to arrive at the optimal setting faster (that is, fewer test iterations to perform).
Let us assume you fundamentally need + value adjustment.
- So your highest choice is +20...set that and see how it compares to +0
- Next, set +10, and compare that result to +0 and +20 tests
Assuming that it seems the optimum is between +0 and +10, set +5 and test and compare that to the +0 and +10 test results
Or if it seems the optimum is between +10 and +20, set +15 and test and compare that to the +10 and +20 test results
- Again (depending upon the 'better' of the comparison, again set HALF (for example, half way between +5 and +10 is a setting of about +7, and half way between +7 and +10 is about +8.
So you see it takes a maximum of 4 iterations of tests to arrive at the 'optimum' value for MFA, because with each iteration you eliminate half of the possible MFA values. For example
Test Iteration 1: 0 to 20 (middle value 10)
Test Iteration 2: 10 to 20 is best pair (middle value 15)
Test Iteration 3: 10 to 15 is best pair (middle value 12.5)
Test Iteration 4: 10 to 12.5 is best pair (middle value 11)
...and it should be possible to see the best of pair (10 or 11), simply choose that as your MFA setting
Not sure about current recommendations, but at one time Canon suggested test distance for MFA was 50 * FL
And to ensure that your camera is perfectly parallel to your test focus plane, set your camera so that when you look thru the viewfinder you find that your camera is aimed at its own reflection in the mirror, then afix your focus target to that mirror!