A while back I added a split prism focusing screen from
Focusing Screen.com to my k-x to assist in manual focusing. It was a big improvement over the stock screen and I was quite happy with it. However, when trying to do some isolations of flowers with my 50mm f/1.4 wide open I noticed a distinct front focus. Stopped down I could get reasonable pictures but at f/2 or wider the focus was noticeably off. I verified this by using live view and zooming in. Using manual focus and live view I could get a good focus but not with the view finder.
When I installed the new screen I had ignored the two plastic shims supplied and re-installed the metal Pentax stock shim. So now I pulled out the screen again and replaced the stock shim with the two plastic shims supplied and re-tested. Still front focusing. I next tried only one of the supplied plastic shims. Now it is clearly a slight back focus. So the actual required shim is between the thickness of the two supplied shims. So how to find a shim that measures between what I have on hand?
After a little testing with a caliper I determined that:
Pentax stock shim = .33mm
1 plastic shim = .13mm
2 plastic shims = .26mm
post-it note = .10mm
scotch tape = .05mm
Since .13mm was back focusing and .26 was front focusing I needed something in the middle. I tried one plastic shim plus a post-it note and got .24mm. When tested this was better but still slightly front focusing. I then tried a plastic shim stuck on scotch tape which measured .19mm. This seems to have the focus point as good as I can tell with my procedure.
Testing Procedure:
- Set up a ruler at an angle, I used an 'A' clamp to hold it up on edge so the graduation markings are vertical
- Camera on a tripod, leveled side to side and front to back, height adjusted so split prism transects the graduation marks
- Use two second timer delay for mirror lock up
- Lens wide open
- Using the split prism, focus on a single vertical graduation, I used '22'
-When satisfied with the focus, turn on live view and zoom in. The '22' should be in the same range of focus as it appeared in the view finder.
In the picture the focus point is on the '22' but '23' is actually sharper.
Note that my caliper only has 3 decimal points and I think the numbers do not add up quite right because the actual measurements need 4 decimals
I repeated each step of the test using an SMC A 50mm f/1.4, an S-M-C Takumar 50mm f/1.4 and an S-M-C Takumar 85mm f/1.8. In each case calibrating the focus screen as described resulted in a clearly more accurate focus point.
Please note that unless you have lenses with very wide apertures this process is likely unnecessary and would not result in any noticeable improvement.
Also note that this has nothing to do with auto focus or any issues you may have with the auto focus system. The focusing screen is not used with the auto focus system, which I assume is why modern DSLRs have such lousy focusing screens.
Similar Threads 
