Fine Tuning AF-S 35mm f1.8G "Problem"

[bovk]

I would appreciate help. The lens was repeatedly showing 0 correction at f1.8, one time I got -1 and I tried over 7-8 times. When I stop down to f2 it is the same. Stopping further to 2.8 gives me +4, at f4 and higher I get +14 to +19 with mid +17. I read that the Life View focuses at dialed f-stop and phase detection uses f-stop 1.8 (full open) so this would indicate some focus shift.
- I would leave the Fine Tuning of this lens at 0
- I think that +17 at f4 and above does not matter since depth of field will take care of it
- With 35mm lens and subject distance 1 meter (30xfocal) the focus ring is so close of infinity -- this whole concept of fine tuning of wide angle lens looks kind of fishy.
Could you comment on my highlighted lines of thought and add your experience with tuning 35mm lenses?

BTW I was amazed how sensitive the Fine Tuning is to conditions: I changed to fresh battery, shade covered my test chart, small change in distance - all this had an effect on the results. That is why I want to leave the correction out till I can do some real life portraits at different apertures to see the results.

[Akira]

According to Erik (Erik Lund a.k.a. Lens Doctor here), the focus shift correction function is implemented in the newer camera bodies at some point.

I was also frustrated with the moody AF fine tuning function.  The right correction value was also affected by the focusing distances, even with the relatively short lenses.

However, I was told at the Nikon service here in Tokyo that the moodiness is caused by the fact that the neutral position of the SW motor will be off over time, and that you cannot correct it with the in-camera AF fine tuning.  You need to have the lens re-calibrated machanically at Nikon service.

I've had the same problem with my AF-S 50/1.8G and AF-S 24/1.8G, but after the calibration at Nikon service along with my camera body (D750 in my case), the problem was totally gone.  I haven't even been in need of switching the AF fine tune function on.

[Erik Lund]

Make sure your using a good valid AF target for AF Fine tuning as well as appropriate neutral strong lightning and AFS-C, AF-S, AF-On tripod etc. by your prefered choice


Or send in the lens for calibration

[Ilkka Nissilä]

According to Erik (Erik Lund a.k.a. Lens Doctor here), the focus shift correction function is implemented in the newer camera bodies at some point.

According to Jim Kasson's testing, this is not correct, at least for the D850:

https://blog.kasson.com/d850/d850-pdaf-tuning-and-focus-shift/

However, I was told at the Nikon service here in Tokyo that the moodiness is caused by the fact that the neutral position of the SW motor will be off over time, and that you cannot correct it with the in-camera AF fine tuning.  You need to have the lens re-calibrated machanically at Nikon service.

I don't understand how the motor could cause a systematic focus bias. The camera measures the phase difference and gives instructions to the focus motor to correct it. If in AF-C, then it will continually adjust the focus around the zero point. I can see that if the motor is not working correctly then there might be extra jitter.

I've had the same problem with my AF-S 50/1.8G and AF-S 24/1.8G, but after the calibration at Nikon service along with my camera body (D750 in my case), the problem was totally gone. 

I tried having some of my lenses adjusted in service for AF and they said they tested them and were as good as they can be. No change distance dependent focus error. But, newer cameras have this much better handled in my experience; in the D800 I had large problems with distance dependent focus error, but D810 and D850 are ok in this respect. There may still be a small effect (like 1 point when going from 1 m to 30m distance) but I don't consider this to be a significant annoyance any more. With my D800 the 105DC required -20 at infinity and 0 up close. The D810 handled that lens fine at -12 across distances.

With the D810 and D5 I've followed the focus behavior of my lenses over the years, and once the fine tuning has been set correctly for a lens, I've never seen it drift over time. There may be a small adjustment needed over a range of distances or a small shift needed in changed lighting conditions (colour of artificial light) but over time, I've noticed no change.  I realize that if the equipment is bumped then there may (and likely are) changes in the required fine tune settings but in normal use I don't find the values change over time.

[Akira]

I don't understand how the motor could cause a systematic focus bias. The camera measures the phase difference and gives instructions to the focus motor to correct it. If in AF-C, then it will continually adjust the focus around the zero point. I can see that if the motor is not working correctly then there might be extra jitter.

I tried having some of my lenses adjusted in service for AF and they said they tested them and were as good as they can be. No change distance dependent focus error. But, newer cameras have this much better handled in my experience; in the D800 I had large problems with distance dependent focus error, but D810 and D850 are ok in this respect. There may still be a small effect (like 1 point when going from 1 m to 30m distance) but I don't consider this to be a significant annoyance any more. With my D800 the 105DC required -20 at infinity and 0 up close. The D810 handled that lens fine at -12 across distances.

With the D810 and D5 I've followed the focus behavior of my lenses over the years, and once the fine tuning has been set correctly for a lens, I've never seen it drift over time. There may be a small adjustment needed over a range of distances or a small shift needed in changed lighting conditions (colour of artificial light) but over time, I've noticed no change.  I realize that if the equipment is bumped then there may (and likely are) changes in the required fine tune settings but in normal use I don't find the values change over time.

I haven't asked the reason, but, that was how they explained.  After the calibration of the lenses mentioned above, the problem was gone, that's for sure.  I once dropped the calibrated pair of 50/1.8G and D750 from a 40cm-high sofa onto the floor coated with thik vinyl, the AF accuracy was not at all affected.

Both D810 and D850 are high-res cameras without AA filters and thus they should be more unforgiving than D750 with AA filter.

[Ilkka Nissilä]

I dropped the D800 with 24-70 from some height some years ago (it slipped from the bag I had assumed to be closed); in this case it was the camera that was damaged (AF was thrown off, Nikon paid for repair and adjustment). The AF error was mostly noticeable with long focal lengths, which is curious as I would've thought wide angles to be more sensitive to such things. With the 24-70 the AF seemed to work normally. The lens got a bit stiffer in the zoom ring.  Not recommended to drop either camera or lens. 

[Akira]

Dropping cameras or lenses are not recommended also for mental reasons...

[bovk]

Thank you for interesting comments and the referral to Jim Kasson's site. The summary of his D850 testing https://blog.kasson.com/d850/nikon-d850-autofocus-accuracy-summary/ is great. First time somebody put a number on Fine Tune step (4 micrometers).