Autofocus Errors

Modern AF systems are still not perfect. The F100 is probably as good as it gets.

An AF error is when the camera focuses at a point a little closer or farther away than it is supposed to.

It is most obvious at wide open apertures.

Why this happens is too complex to explain here. Others have seen this, I'm not crazy, and it does vary from sample to sample of camera and lens.

It is a very subtle effect. Most people will never see it even if they have it. It sometimes can be seen through the viewfinder: if you can get better focus with a small manual adjustment of the lens then there is some error. Of course this might be your viewfinder being out of adjustment, too. This all gets very complex very fast.

This is not a malfunction, but a limitation or defect in design.

This simply is the limit of today's technology. It's not something that can be repaired; you need to use a different sample of lens (or less likely, camera) to change it.

Contax offers AF bracketing in an attempt to work around this, just as some people exposure bracket in an attempt to reach perfection

There are still some minor focus errors with some lenses that may be apparent if you often shoot at full aperture as I do. These are something a very careful eye may see at full aperture, but also subtle enough that most people will probably never notice them in real photography unless they look for them. It all comes down to limitations of modern day AF lenses and systems.

These limitations are understood by Contax, who now offers an AF camera that provides AF bracketing to attempt to work around the fact that AF systems often do make slight errors.

Different samples of lenses display different propensities for this effect, too. Stopped down these effects also are invisible. They become important if you shoot wide open as I do on 50 speed Velvia.

There are several factors at work in minor auto focus errors common to most AF cameras:

1.) Focus may be different on the focus screen as compared to film, depending on the alignment of the camera. For instance, my F100 focusses dead-on on film, however the image on the focussing screen looks like the camera is focussing too close. My screen or mirror could use a slight adjustment because it's misrepresenting where the on-film focus really is. There are off-center headed screws in the mirror box that adjust mirror rest position to alter this adjustment. I tried to have Nikon adjust this and they wound up giving me a new camera.

2.) Some lenses exhibit focus shift due to spherical aberration. This problem causes the lens to focus at slightly different points depending on its set aperture. If you have a lens with this problem you need to focus it at the taking aperture. (Read Ansel Adams, "The Camera" if you aren't familiar with this problem.)

With 35mm AF cameras the camera always focusses through certain points of the lens, typically through spots equivalent to about the f/8 aperture. If your lens focusses a little bit differently at the wider apertures, then your focus will be off. It may be hard to understand that even thought your lens is wide open while you are focussing, that the AF optics are really looking through the equivalent of a smaller aperture of your lens. If this is the case, so long as you have a well adjusted camera as mentioned in 1.) above, focus manually.

2a.) Beware, however, that the extra-bright laser cut screens that replaced simple frosted screens in the 1980s (including of course the F100) also only accept light from about the f/2.2 aperture in the case of the F100. This explains why the finder gets no darker in the F100 when you use the aperture preview control to stop down from f/1.4 to f/2.0, and also means you, as always for the careful artist, need to test how your system works for focussing at f/1.4 since the screen isn't showing you what happens at f/1.4. On older camera like the F2 the ground glass really did show you what went on at f/1.4, and yes indeedy, an f/1.4 lens looks brighter than anything else on those cameras. More recent cameras are optimized for the pig-slow zoom lenses everyone uses to day.

2b.) The microprism and split image RF on manual focus screens also looked through the lens at about the equivalent of f/8 and f/5.6, respectively, and could be subject to the same focus shift issues.

3.) Zoom lenses that exhibit focus shift will show different performance at different focal length settings.

4.) Different samples of lenses can show different focus shifts due to manufacturing tolerances.

5.) AF systems do not focus exactly the same every shot. Go for 2 out of 3 if you are running tests. Again, AF systems are not 100% repeatable. They are designed for speed, not absolute lab precision. Remember that their CCDs are of limited resolution if you are scoping the results under a microscope.

6.) Nikkor lenses that have exhibited focus shift problems for me (and therefore I won't use most of them on my F100) are:

28-105/3.5-4.5 AF-D (problem is at 28mm end)
70-300/4-5.6 AF-D (problem at 70mm end, focusses too close in one of two samples)
70-210/4.0 AF, first AF 70-210 from 1986, constant f/4.0 aperture (problem at 210mm end)
180mm f/2.8 AF (OK beyond 50 feet, focusses too close at closer than 25 feet)
17-35 AF-S (only one sample was off at at 28mm setting, otherwise fine and I use this lens all the time)
28-70mm f/3.5-4.5 (at 70mm)
The 85/1.8 AF-D also showed a small propensity to focus a hair close at f/1.8.

Different samples may be different.

I also saw some errors with my 28/1.4 AF-D and my first F100 from 1999, however a new F100 received in May 2001 seems to be OK with it.

To be honest, I tried several 1999 F100s with my 28mm f/1.4D AF and all focussed a little too close. My 2001 F100 seems just fine with the 28mm f/1.4D. If the 2001 F100 has an improved AF system that can focus the above lenses that would be very cool.

These focus just fine:
80-200 AF-S
400/2.8 AF-I
105/2.8 AF micro
28-85 AF-n f/3.5-4.5 (from 1988)
17-35 AF-S (at any setting other than 28mm on one sample of lens)
20AF f/2.8
70-210 AF-D f/4-5.6

7.) If you're picky enough to have read this far, you may see different performance with each of the five AF sensors. This is because of differing camera alignment as well as the fact that the effective spherical aberration in a production lens can differ at each of the five sensor points.

95% of focussing issues are your technique. I only can see these AF errors under a microscope under carefully devised test conditions. For many people the problem they will see more often is setting the AF incorrectly on the camera; if you choose or allow the camera to default to closest subject priority you'll see much worse effects!

I hope all the film I've burned experimenting with this helps illuminate this issue. I hope that my positive experience with a 2001 F100 shows that maybe this problem has gone away.

Manual Focus Errors

My F100 came back from factory service in August 2000 with the focus screen or mirror misaligned ever so slightly.

Now the ground glass of the viewfinder screen cannot be used for precise manual focussing.

Specifically, the point of best focus as seen on the screen itself is incorrect. One needs to use the little electronic rangefinder or the AF system to focus properly.

The prime disadvantage of this is that I no longer can focus the old fashioned way, a bit of a bummer since that's the main reason to buy the AF-S lenses.

This wasn't enough of a bother for me to have bothered returning this to service even almost a year later.

I did ask Nikon to adjust this when I had to go in when my brand-new 17-35 fell apart.

For some odd reason this tiny adjustment caused Nikon to give me a brand new F100 in May, 2001. That sure fixed the problem, but gosh, this seemingly minor adjustment would seem to be the last thing to cause Nikon to give up. In any case, the new F100 I got in May 2001 focusses perfectly.

NEXT: NO MIRROR LOCK-UP > >