Bellows Length per Knob Rotation

[joshmacro]

Hi everyone. I can't seem to find a resource (here or elsewhere) that indicates the length per full rotation for focusing (aka adjustment) knobs on bellows systems. I am specifically interested to understand the length of a full turn of rotation for the focusing knobs on the Canon FD Bellows and Olympus OM Bellows. However, it would be nice to see a comparison with other systems (Pentax, Contax, etc.) to see if other bellow systems have greater precision that the Canon and Olympus bellows. Also, I assume that the length per rotation is the same for all 3 (front, rear, entire unit) focusing knobs, but I'm not even sure of that. Any guidance appreciated. Thanks.

[rjlittlefield]

For the Olympus OM Auto Bellows, HERE, I measure 25 mm per turn. All three places are the same.

--Rik

[joshmacro]

Thanks Rik. How did you extrapolate 25mm per turn from this site? I had gone to this site previously and couldn't glean that info.

[rjlittlefield]

Oh, sorry, I didn't mean that I got the measurement from the website.

I referred to the website only to identify the type of bellows.

The measurement was physical, using digital caliper and a marked knob on one of the units that I own.

--Rik

[joshmacro]

Got it. Thanks for the manual check Rik.

I assume that other bellows systems are similar. If anyone has specific info on others it would be great to know.

At 25mm per rotation it seems impossible to manually stack even at lower magnifications. A 2x mag at f5.6 is a focus step of around 155 µm. You would need to turn the focus knob on the Olympus bellows 161 times within a full rotation.

[rjlittlefield]

Stacking by hand is definitely challenging with this equipment. Above 1:1 it can be simplified by stepping with rear bellows movement. This is because front and rear step size are related by a factor of magnification squared. At 2X, the rear step size is 4 times larger = 0.62 mm. Round that down to 0.5 mm and you can eyeball it on a mm scale accurately enough.

--Rik

[joshmacro]

And at 3x the rear step is 9 times larger, and so on?

[rjlittlefield]

And at 3x the rear step is 9 times larger, and so on?

Yes, that's correct. So at 10X NA 0.25, where DOF is nominally 0.0087 mm at subject, that corresponds to 0.87 mm in rear movement.

This approach does, unfortunately, lead to some limitations. When using a 10X objective, it is quite common to step by 0.005 mm for 300 steps in front. But the corresponding 300 steps at 0.5 mm in back would be 150 mm of rear movement, which introduces a lot of variation in magnification and exposure. With special settings, stacking software can handle the variation in magnification, but some thought is required about how to handle the variation in exposure. So, as a matter of practice, stepping by rear bellows movement is limited to relatively short stacks, say 50 or fewer frames.

If you're very dedicated, it is possible to shoot a deep stack in multiple sections, by using rear movement to step focus from frame to frame, then periodically reset the rear distance and move the front standard a small amount to reach the next section. But I emphatically do not recommend that approach. It requires a heroic effort, to do a job that would be simple with a screw rail, or better, a repurposed focus block from a microscope.

By the way, if the rear step sizes that we've discussed here seem remarkably close to each other (0.5, 0.62, 0.87), despite the wide range of magnifications, that's because the nominal rear step depends only on the effective f-number. The relationship is that rear DOF = 4*0.00055*Feff*Feff . We like to keep the effective f-number in a fairly narrow range to manage diffraction, so the rear step ends up being kept in a narrow range also.

--Rik

Edit: to correct the error identified below.

[joshmacro]

By the way, if the rear step sizes that we've discussed here seem remarkably close to each other (0.5, 0.62, 0.87), despite the wide range of magnifications, that's because the nominal rear step depends only on the effective f-number. The relationship is that rear DOF = 0.00055*Feff*Feff . We like to keep the effective f-number in a fairly narrow range to manage diffraction, so the rear step ends up being kept in a narrow range also.

Sorry, this last bit threw me. If we have 5x mag at F/2.8 then the step size is 0.025 mm. The effective f is 16.8. So is the rear step size: (A) 0.62 mm = 0.025 x 5^2 or is it (B) 0.16 mm = 0.00055 x 16.8 x 16.8.

[rjlittlefield]

By the way, if the rear step sizes that we've discussed here seem remarkably close to each other (0.5, 0.62, 0.87), despite the wide range of magnifications, that's because the nominal rear step depends only on the effective f-number. The relationship is that rear DOF = 0.00055*Feff*Feff . We like to keep the effective f-number in a fairly narrow range to manage diffraction, so the rear step ends up being kept in a narrow range also.

Sorry, this last bit threw me. If we have 5x mag at F/2.8 then the step size is 0.025 mm. The effective f is 16.8. So is the rear step size: (A) 0.62 mm = 0.025 x 5^2 or is it (B) 0.16 mm = 0.00055 x 16.8 x 16.8.

My apologies. I botched a bit of algebra and the corresponding numeric check. The 0.00055 should be 4*0.00055 = 0.0022. So then the rear step should be (B) 0.0022 x 16.8 x 16.8 = 0.62.

Thanks for the cross-check.

--Rik

[joshmacro]

Below is a table related to the limitations of performing rear bellows focusing stacks: (1) The length of the rear step and (2) The number of images in the stack. I think I did the rear step size correctly, but let me know if it looks off.

Assumptions:

• Assumed 20mm image depth
  Assumed a rear step below 0.5mm becomes difficult to turn the knob with accuracy on a bellows. So 0.5mm is the minimum rear step.
  Assumed that the image stack needs to be 50 images or less.
  Assumed that keeping the effective Fstop less than 20 would be optimal.
  Assumed a limit of F2.8 as the widest aperture.

So for each magnification an Fstop was backed into to make sure the above limitations were not met. The result was this table which is intended to show the Fstop limits for rear bellows stacking. The red numbers indicate when limits have been reached.

[rjlittlefield]

All those numbers look right.

The red region indicates an area where it becomes worthwhile to consider switching over to using a microscope objective as the main lens. A typical 10X NA 0.25 finite objective is equivalent to an f/1.818 ordinary lens, which gets you back to effective f/20 at 10X. But even so, the number of frames that you can reasonably shoot using rear bellows movement is much smaller than you'd typically choose to shoot at 10X if you were stepping the whole system in units of 0.0088 mm.

--Rik