Nikkor 10.5 mm f/2.8 fisheye:

Correction of the flawed thickness of an adapter to the Canon EOS

Nikon lenses can be easily adapted to the EOS Canon camera.

Adapter rings can be easily found and bought on the web for instance. But quality can be very different from one to another brand.

The most frequent quality problem is a "thickness" lack of precision. Or more exactly stated: the distance between the "Canon flange" and the "Nikon flange" of the adapter does not fit with the requirement that is to exactly compensate the register difference between Canon and Nikon mounts.

Check of a possible adapter thickness anomaly

A quick test for any Nikon lens adapted on an EOS camera:

The Canon register is 44.0 mm and the Nikon register is 46.50 mm.

To make sure of the correctness of the thickness of the EOS-Nikon adapter, the best way is first to take a shot at full lens aperture setting of an object at a medium distance (e.g.less than two feet ) away from the focal plane of the camera.

If the same distance had been set on the focusing scale on the barrel of the lens, the object should be "sharp" on the image. Then check if some closer object are also sharp as well as some that are located just a littls away.

If successful, this quick test gives a strong signal of a correct thickness of the adapter.

If the infinity is also correctly sharp while the lens was set at infinity on its focusing scale, then one can conclude that probably nothing is wrong.

If the infinity is blurred for this setting and if the infinity is sharp for a shorter focusing distance on the scale, the adapter is TOO THIN.

If infinity cannot be focused sharp for any distance that is set on the scale and at the same time if very close object can be sharply defined on the photograph, the adapter is then TOO THICK.

A complete test for a 10,5 mm Nikkor lens adapted on an EOS camera:

The following table was provided by Nikon and is applicable for for the 10.5 mm fisheye lens.

Foreword: Readers of this article who are more familiar with feet and inches may adapt the following process by using the above table.

Put a rule near and above the camera as shown here:

The 20 cm mark is here located 50 cm above the focal plane (BTW that is also where the threaded hole is located). Let us now shot some photos at Full Aperture (f/2.8) successively with the focus scale mark set at the distances shown on the left column of the above table i.e. 0.14 m, 0.17 m, 0.20 m, etc.

Now look at these images closely on a monitor screen. See for instance the image shot on the 50 cm mark on the focus distance scale:

The depth of field here is exactly as it was expected when refering to the Nikon's table.

If the actual depth of field of (on all shots) is found to be closer to the focal plane than it is specified on the table, the adapter is not thick enough.

If the actual field of view (on all shots) is found to be too much farther from the focal plane than it is specified on the table, the adapter is then too thick.

Estimation of the thickness anomaly value

The estimation of the thickness anomaly value is probably the most delicate part of the experiment.

If the previous step led to the conclusion of a too thick adapter, you are in a bad shape, as there is no way to correct this flaw but machining on a lathe to remove the material in excess. It's not an easy task to perform for the vast majority of us....

Fortunately, most of the defective adapter are rather not thick enough. The reason for this is that the adapter provider does generally guaranty that infinity can be shot as sharp at any aperture setting. In reducing slightly but systematically the thickness of the average adapter, the possible machining errors are thus always covered on the "safe" side. This will be of course most often at the expense of the excursion range on the manual focusing scale of the lens. In effect it shall be possible to set the lens to focus at a distance "beyond infinity". This shall be considered by the provider as a "minor" side effect but it is really annoying for the photographer!

Let us put again the camera lying on its back. Let's now put carefully the lens on the camera mount. Do not screw the bayonet. We shall then put successively the lens at different longitudinal distances from the normal longitudinal position by inserting calibrated blades pair in between the two flanges:

These thickness calibrated blades have been simply detatched from a set of spark plug gauge (buy one at the nearest harware store or borrow it!). It is always possible to insert a pair of gauge blades(1) and (2) of the same thickness under both sides of the lens. For instance 0.020 mm on the left and (0.015 + 0.005 = 0.020 mm) on the right.

Now let's shot photos with the same set-up as previously (the canted rule above the camera lying on its back). The gap between the camera side of the mount and the rear lens mount surface shall be increased from zero to1.00 mm in five steps of 0.20 mm each for example.

If the photograph of a given step shows that the gap is too small while in the next step the gap is found to be too large (by the same method as explained in the previous paragraph), the correct gap value is obviously then in between the two. Narrowing down to the value of the defect can then be done by using much thinner gauge blades. Steps of 0.005 mm shall in fact indicate the value of the errors to be corrected within about 0,002 mm accuracy. A visual interpolation allows such a precision. This is largely a sufficient accuracy in order to correct the flaw (next paragraph).

Correction of the thickness anomaly: a personnal case study

My latest adapter was 0.28 mm (i.e. 11/1000 of an inch) too slim. This value had been found by the above described method. It was then possible have the focusing distance well beyond the infinity in real when setting at the infinity mark on the engraved scale of the lens. Infinity was reached when setting just sllightly after the 0,50 meters mark on the scale!

Most of the EOS/Nikon adapters are composed of two main parts: one ring on the "Nikon lens" side simulating a Nikon camera mount and one ring on the "Canon camera" side simulating a Canon lens mount. These two component are screwed together with some (generally four) tiny screws.

On my flawed adapter an elegant spring ring is pressed between these two main rings. This is a very common design feature that is used to build lens adapters.

I have simply cut a hard cardboard shim ring (0.28 mm thick) and inserted it below the spring ring before putting back every piece at its respective place and screw them together.


After modification, the adapter/lens combination is now perfectly registered and manual focusing is at last possible by refering exactly to the focusing numbered scale on the lens barrel. The usefull excursion range to set a focusing distance from 0,14 m to infinity is also now normally extended. Furthermore, the possibilty to unwillingly set the focus beyond infinity is now remote and on par with any standard Nikon lens.

Well, it is not quite easy to manually set a precise distance on the 10.5 mm fisheye. Nikon was not very good at designing the ring: it is one of the worst and less friendly that I know of: the angular excursion from 0.5 meters to infinity is only 3.5 mm on the scale!

Furthermore, Nikon has foregone any aperture ring on this "specialty" lens. You might also find a way to recover external (but manual) control of the Aperture Stop Setting by reading this other page.

Michel Thoby

8 March 2006