[BigMackCam]

Amongst my collection of Soviet glass, I have a very nice ZM-5SA ("3М-5СА") 500mm f/8 catadioptric / mirror lens in M42 mount. It works a treat on my Sony A7 MkII, and produces beautifully sharp and detailed photos, considering the poor reputation of this type of lens.

I also have two different Soviet 2x tele-converters... a TK-2M and K-1.

Today, I tried out the ZM-5SA with each of the tele-converters, and immediately noticed extreme vignetting... not just strong light fall-off, but the type where the image circle isn't covering the whole of the sensor. The same thing happens with both converters.

I have SteadyShot turned off on the camera, as I know this can have an effect, but in any case this vignetting is too strong to be caused by that.

I'm thinking it might be caused by the M42 adapter, but the type I'm using has no obvious internal pieces that would restrict the optical path.

Does anyone else have experience with this lens and / or these teleconverters on full frame? It would be nice to rule out whether this is due to the lens + TC combination, or if the problem could be with the adapter and/or camera.

The first photo below shows a white wall capture without the TC, and the second one, with...

Thanks in advance

Attached Images

   

[photoptimist]

These kinds of issues occur if the elements of the TC are too small relative to the cones of projected rays of light coming from the main lens. There's two scenarios for this:

First, if the front element of the TC is smaller than the back element of the lens, there's a good chance of some kind of vignetting and light loss. In most of these cases (e.g., putting an f/1.4 lens on a TC), the vignetting is subtle in that light from the outer part of the aperture is being clipped by the TC so instead of getting an f/1.4 x 2X = f/2.8 equivalent lens, the TC+lens might act more like an f/4 combination. The light from the aperture is like a converging cone that ends in a point on the sensor. In the case of this lens, the cone starts at 500/8 = 62.5 mm in diameter and gets smaller in the journey to the sensor.

Second, if the rear nodal point of the main lens is far from the focal plane, then there's a higher chance that the TC will clip the corners of the diverging rectangular cone of the projected image. For a 2X TC, the rectangle that matters is exactly 1/2 the size of the sensor or about 21.6 mm across the diagonal.

The two effects actually combine in that the spray of light from the lens is both a diverging rectangular cone defined by the TC-modulated effective sensor size and a converging circular cone defined by the main lens aperture. The required minimum diameter for the TC glass elements with this 500 mm lens will be something like c*(62.5) + (1-c)*21.6 where c is approximately the ratio of the TC-to-sensor and the lens' rear-node-to-focus distances. It's hard to know the true value of c for this lens, but I'd guess it's something like 0.25 to 0.33 which means that you need a TC with a front element of at least 32 to 35 mm in diameter.

This issue is especially severe with a mirror lens because the aperture of a mirror lens is a donut. Clipping the rays coming from the edge of the aperture will cause much more severe vignetting than might be seen with a non-mirror lens.

[BigMackCam]

These kinds of issues occur if the elements of the TC are too small relative to the cones of projected rays of light coming from the main lens. There's two scenarios for this:

First, if the front element of the TC is smaller than the back element of the lens, there's a good chance of some kind of vignetting and light loss.
...
Second, if the rear nodal point of the main lens is far from the focal plane, then there's a higher chance that the TC will clip the corners of the diverging rectangular cone of the projected image. For a 2X TC, the rectangle that matters is exactly 1/2 the size of the sensor or about 21.6 mm across the diagonal.
...
The two effects actually combine in that the spray of light from the lens is both a diverging rectangular cone defined by the TC-modulated effective sensor size and a converging circular cone defined by the main lens aperture.
...
It's hard to know the true value of c for this lens, but I'd guess it's something like 0.25 to 0.33 which means that you need a TC with a front element of at least 32 to 35 mm in diameter.

This issue is especially severe with a mirror lens because the aperture of a mirror lens is a donut. Clipping the rays coming from the edge of the aperture will cause much more severe vignetting than might be seen with a non-mirror lens.

Thanks so much for this information, and for taking the time to explain it fully and clearly!

The rear-most element of the lens is approximately 27mm in diameter, while the TK-2M converter's front element is roughly 21mm. Even allowing for some variation in "c", that's way too small for the intended application

Once again, thank you. This is extremely helpful!

[UncleVanya]

These kinds of issues occur if the elements of the TC are too small relative to the cones of projected rays of light coming from the main lens. There's two scenarios for this:

First, if the front element of the TC is smaller than the back element of the lens...

Second, if the rear nodal point of the main lens is far from the focal plane...

Is this why the 1.4a-L and 2.0a-L converters exist? They have recessed elements that go up inside the long lenses that they fit and that would put their front element closer to the cone and rear element of the lens.

Very interesting and informative info. You should put that in an article for the site.