In Wag the dog, part 3, I pronounced the Novoflex adapter’s flange distance perfect. That seemed to be true for the Tri-Elmar, but further experimentation with other lenses shows that some focus beyond infinity to a greater or lesser degree. The errors are small and would likely not be significant in normal photography, but the NEX-5’s magnified live view is such a precise focusing tool that it makes the small errors obvious.
I had noticed small errors in focusing with the Leica lenses during lens testing using the LensAlign Pro. That kind of testing conflates errors due to the rangefinder focusing cam, my ability to discriminate small differences in the rangefinder, and errors due to the lens itself. Using live view to focus eliminates the first two sets of errors, leaving only the lens, and, in the case of the NEX-5, the lens adapter.
I now realize that there are small, but discernible, errors in focusing distance in some of the Leica lenses. This makes me more sympathetic to finding an NEX-5 lens adapters that’s too short. If you were making a lens adapter, and knew that the lenses that might be attached to your adapter had some variation, you’d want to make sure that any lens focused to infinity when placed on your adapter. Thus you have to make it a little short. It’s far better for the focusing indicator to point closer than infinity when the lens is focused on something a mile away then it is to not be able to bring that object into focus because of the infinity stop on the lens focusing mechanism.
Now that I have, in the NEX-5, a great tool for evaluating focus accuracy, I can unravel the mystery of the way-too-close-focusing 35mm Zeiss f/2 Biogon that I noticed in this post. When installed on the NEX-5 via the Novaflex adapter, the Zeiss lens focuses to infinity when the focusing ring says it’s focused to infinity. Therefore the focusing error when the Zeiss lenses is installed on the M9 has to be due to the lens’s connection to the M9 rangefinder.