Zoom for Mitutoyos

[Lou Jost]

As we all know, zooms used as tube lenses tend to vignette when they are zoomed to lower focal lengths. For MFT sensors, a very successful solution used by one member of this forum is to use a teleconverter, so that only the central part of the zoom's image is used; there is still vignetting eventually but this extends the non-vignetting range of the zoom. The cost is lower NA. Many of us would like to find a zoom which would give a wide range of non-vignetting focal lengths without the need for a teleconverter. Preferably it should have a constant length when zoomed or focused.


[Edit and digression Apr 25: I think I am wrong about the teleconverter lowering the NA, if we keep the field of view constant. A zoom tube lens (one with a fixed maximum aperture) at 143mm with a 1.4 teleconverter will have exactly the same FOV and effective aperture as the zoom lens set at 200mm with no teleconverter. For example, suppose we mount a 10x f/2 objective on the zoom tube lens set at 200mm. EA= m*f = 10*2 = 20. Now we set the zoom to 143mm. The magnification is 143/200 * 10 = 7.15x. So the EA of that system is 7.15*2=14.3. Now put a 1.4x teleconverter on it; this increases the EA by a factor of 1.4 (= 1 stop) so we get 1.4*14.3= 20. So the effective aperture only depends on the final magnification. Rik, is that right?]

On eBay I found a a Nikon 80-200mm f/2.8 ED IF two-ring zoom with constant length for a ridiculously low price because its auto-focus didn't work. After my tests yesterday pushing down Mitu objectives, I got excited about trying them on this lens on my MFT camera. It worked well with the 7.5x Mitu! (Haven't tried the others yet.) I could get non-vignetting images with good corner quality from 200mm down to about 110mm.

Here is the whole 110mm image, PMax from Zerene, corrected only for exposure:


Here is the image at 200mm:


Exactly the same range of non-vignetting FOVs could be obtained on APS sensors by adding a TC14 teleconverter.

There is a small hit in image quality versus a prime tube lens. I can't quantify this very well with my butterfly scales because they are moving when I blast them with flash, and road construction today is making continuous light impossible. Testing this properly would require something more systematic and controlled, like Robert O'Toole's tests on wafer targets. But I thought it might be worthwhile to present this initial result.

[mjkzz]

Does NA change at all for an objective? My understanding is that it does not change regardless what tube lens you put behind it, well, according to its definition NA = n*sin(theta) where n is refraction index and theta is half the angle of cone, none of these two parameters seem to have anything to do with tube lens.

Teleconverter is just a magnifying glass that takes the center portion of beam from the lens in front of it and project it on to sensor, effectively, increases the focal length of the lens in front of it. So adding an TC is effectively increasing the focal length of tube lens, thus increasing magnification, but I do not think it could increase resolving power of the objective, which mostly determined by or related to the NA value.

[mawyatt]

Lou,,

Doesn't Effective Aperture = Lens Aperture(1+M).

Best,

Mike

[ChrisR]

I think of NA as subject side (fixed) and EA as image side (alterable by changing M).

The converter has the potential to let you see resolvable details which your pixels were too big to discern, without it.

[Lou Jost]

Right, the teleconverter doesn't increase the resolution as measured on the subject side but if the prime lens out-resolves the sensor, the teleconverter can reveal details that were not visible with just the prime lens.

We wouldn't expect to see this effect for high-m objectives, because diffraction keeps them from out-resolving the sensor. But for lower magnifications, teleconverters are worthwhile.

Mike, the formula you give for EA is for the case of magnification by extension. For infinity-corrected objectives on extension tubes, the formula is the one I gave above.

[mawyatt]

Lou,

Thanks for educating me on the EA


Doesn't the teleconverter enhanced resolution as seen at the sensor imply it's optics don't effect the IQ in the overall optical path? When I used teleconverters with the PN105 F2.8A I noted the IQ didn't appear to be degraded by the Nikon 1.4X TC much, a little more with the Nikon 1.7TC and quite a bit by the lower quality Tamron 2X TC. I think this apparent IQ degradation was a combination of increased magnification and in the 2X case the poorer quality of the Tamron 2X. I'm not trying to bash Tamron (I have a couple of their superb "Di" lenses), but having to hold up the IQ from a PN105 around 1~2X is a tall order I would think.

Best,

[Lou Jost]

Yes, of course, the optics in the TC have to be very good for this to work. In my experience Nikon teleconverters (especially in the old days when they had a special TC for longer lenses) are really good. Many independent makers produce poor teleconverters.

Mike, the difference between those two formulas for EA is why I am so hung up on looking for stacked lens combinations (or infinity objectives) rather than trying to get magnification via extension. For low m there is a significant penalty in EA for using extension rather than coupled lenses. For m= 1.4 for example, if an f/2 lens is extended to get m=1.4, the EA is 2(1+1.4) = 4.8 while if the lens is reversed on an appropriate lens, the EA is (2*1.4) = 2.8 which is quite a difference. Unfortunately, apart from infinity microscope objectives, most lenses are not designed to be used this way, so they often suffer from aberrations. But it depends on the combo, and some work really well. Theoretically, at least, they could greatly outperform an equally-fast lens on extension.

[mjkzz]

Thanks ChrisR and Lou.

OK, we have established that the NA and resolving power do not change as they are integral characteristics of an objective once it is made.

To be able to capture resolvable details, the solution is to increase magnification if image device is not able to. To do that, with infinite objective, is to increase focal length of tube lens. What a TC does is exactly that as if a longer focal length lens were used.

Maybe we can introduce a new term, effective focal length of tube lens, (or maybe this term exists already), so instead of saying TC helps capturing more (resolvable) details, we should say it is the longer effective focal length of tube lens, thus higher magnification, that does the trick Why? Because a tube lens with same focal length as effective one but without TC does the same.

[Lou Jost]

I suppose that's right as long as the tube lens + teleconverter are still faster than the objective. Remember that we get into trouble if we close down the tube lens too much.

[rjlittlefield]

I see you folks have sorted this out already, but just to confirm...

It is always true that NA = 1/(2*Feff), when NA and Feff are measured at the same point in the optical system.

It is also always true that NA_subject = NA_sensor * magnification. By implication, Feff_sensor = Feff_subject * magnification.

NA_subject is not altered by tube lens focal length, as long as the tube lens stays focused at infinity. If the tube lens is not focused at infinity, then the distance from subject to objective is altered, and that does change the NA as well as introducing aberrations.

So, if you mount a 20mm f/2 objective on 200 mm tube optics, you get 10X magnification, effective f/20 at the sensor, and effective f/2 at the subject. It does not matter whether those tube optics are packaged in one module or two, that is, long lens by itself or shorter lens plus teleconverter.

The formula that "Effective Aperture = Lens Aperture(1+M)" is true only in very limited circumstances, typically a lens with fixed focal length, manual aperture, and pupil factor 1, mounted on empty extension. Even in that case, the formulas given earlier remain true. If you slog through the formulas for that case, it happens that Feff_subject = Lens Aperture*((M+1)/M), the factor of ((M+1)/M) being caused because the lens ends up that much farther away from the subject, compared to the lens focal length. For example a 20mm f/2 lens simply extended to 10X will end up 22mm away from the subject, and 220 mm away from the sensor, with corresponding effective apertures of f/2.2 on the subject side and f/22 on the sensor side.

--Rik

[rjlittlefield]

Maybe we can introduce a new term, effective focal length of tube lens, (or maybe this term exists already), so instead of saying TC helps capturing more (resolvable) details, we should say it is the longer effective focal length of tube lens, thus higher magnification, that does the trick Why? Because a tube lens with same focal length as effective one but without TC does the same.

Yes, that is a good way to think about what's happening. As for "effective focal length of tube lens", that's an old term but perhaps it should be more frequently used in the current discussions.

--Rik

[mjkzz]

I suppose that's right as long as the tube lens + teleconverter are still faster than the objective. Remember that we get into trouble if we close down the tube lens too much.

adding a TC will cost you a stops or two, depending on its conversion rate. I think with 1.4TC, you probably lose one stop. I think I lose about 2.5 stops with my Viltrox 2X TC as its glasses are not very good.

[mjkzz]

Maybe we can introduce a new term, effective focal length of tube lens, (or maybe this term exists already), so instead of saying TC helps capturing more (resolvable) details, we should say it is the longer effective focal length of tube lens, thus higher magnification, that does the trick Why? Because a tube lens with same focal length as effective one but without TC does the same.

Yes, that is a good way to think about what's happening. As for "effective focal length of tube lens", that's an old term but perhaps it should be more frequently used in the current discussions.

--Rik

Thanks Rik and pardon my ignorance of that old term as I have not read (or forgot) about it in this forum :-)

[rjlittlefield]

I suppose that's right as long as the tube lens + teleconverter are still faster than the objective. Remember that we get into trouble if we close down the tube lens too much.

adding a TC will cost you a stops or two, depending on its conversion rate. I think with 1.4TC, you probably lose one stop. I think I lose about 2.5 stops with my Viltrox 2X TC as its glasses are not very good.

I'm not sure how you two are thinking about this, but I suspect you're worrying about a non-problem.

We get into trouble when the entrance pupil of the tube optics is too small and/or too far back. But it would take a quite inappropriate teleconverter to change the pupil. Teleconverters cause a larger f-number because they increase the focal length, not because they change the entrance pupil.

There are some cases where teleconverters can cause vignetting, especially on full frame and with a tube lens that is merely long, not telephoto. As an example, my Kenko 1.4X teleconverter has a front element that is less than 25 mm in diameter. Stick that on full frame with say a Raynox DCR-150 way out in front of it, and I'm pretty sure the combo will vignette.

--Rik

[Lou Jost]

Rik, that's not what I was referring to. I'm referring to the problem you have often pointed out, that the aperture of the system should not be in the tube lens. If the tube lens is too slow, it's aperture becomes the controlling aperture, and this will cause aberrations. The aperture should be in the objective Iif the objective is much shorter in focal length than the tube lens, as is usually the case) or between the objective and tube lens.