2

I'm working on a project to create, essentially, a servo-powered zoom lens, using a (D)SLR zoom, a zoom ring gear, and a stepper motor. After mounting the zoom ring gear, I'll attach a pinion gear to the stepper motor shaft to control the motion of the zoom ring.

ENG/cine-style servo lenses that do this by design usually feature:

  • Parfocal optics. This means that the lens holds focus through the entire zoom range; a videographer can zoom into a subject, focus, and then zoom back out without needing to adjust focus. Most modern (D)SLR lenses aren't parfocal (they're varifocal) because they don't need to be (autofocus and still photography obviate the need to hold focus), and that makes them simpler, cheaper, and lighter.
  • Linear zoom throw. As you move the zoom ring on many (D)SLR lenses, the field of view doesn't change linearly. Through some focal length ranges, the field of view changes at a different rate than it does through others. A linear zoom throw allows for smooth, consistent zooming that doesn't draw attention to itself.
  • Constant aperture. The lens doesn't necessarily need to have a fixed aperture, but it needs to hold a constant aperture through the zoom range so that the image doesn't darken while zooming in.
  • High price tag. This is mainly why I'm working on an alternative solution, though I also prefer the additional control that I'll have over a stepper motor in experimental shoots.

With that in mind, I want to find a (D)SLR lens to act as a budget alternative to a servo lens. It would need to have:

  1. parfocal optics
  2. linear zoom throw
  3. constant aperture
  4. full frame 35mm coverage
  5. a mount that can be adapted to Sony E-Mount (on an a7S) - fortunately, most mounts can
  6. the longer the zoom range, the better (ideally, at least a 2×)
  7. affordable price

I define "affordable" as somewhere in the ballpark of $300-1800 for a new or good used copy. Below that, it's probably not of very good quality; much above that, and I might as well just buy this and call it a day:

http://www.sony.com/electronics/camera-lenses/selp28135g

Weight, brand and speed aren't as important to me, as long as it's not of poor quality. I also don't need auto aperture or focus.

Does such a lens exist, or is this a pipe dream?

I've found various lists of (D)SLR lenses that are supposedly parfocal, but I don't know if they meet or approach these other criteria; I can (and will) research everything but the linearity of the zoom throw, since that's not at all a commonly-listed specification. I guess I'm mainly interested to know if anyone has experience with any lens(es) that they know to be parfocal, and can confirm that they have a linear zoom throw.

  • 2
    If you're using a stepper motor to control the zoom, there's probably a micro controller somewhere in the mix. If that's the case, then you should be able to easily accommodate non-linear zooms simply by programming the controller to vary the zoom rate. – Caleb Feb 2 '16 at 22:51
  • @Caleb: Thanks - I've considered that, and it's certainly possible, but I don't think "easily" is quite accurate. ;-) It adds another layer of complexity. I would need to discover - either through research or (more likely) experimentation - the actual behavior of the zoom throw, which can be complex. I'd need to accommodate that in my code to linearize the throw while simultaneously applying easing if/when needed. It's an option to keep on the table if I find a lens that satisfies every requirement but that one, but it would save a lot of time to be able to avoid that approach altogether. – Bungle Feb 2 '16 at 23:09
  • 1
    Parfocal, linear zoom throw, constant aperture, etc. OR affordable. You have to choose one or the other, but you can't have both. – Michael C Feb 3 '16 at 5:45
  • Michael Clark, I don't believe that yet. Sounds nice, but I need evidence. For example, the Canon 17-40mm f/4 L (<$500 used) has most of what I'm looking for, but I haven't verified the linear zoom throw (just one anecdotal report), the focal range isn't very large and it has major distortion issues esp. wide, so I haven't sprung for a copy yet. There is a ton of affordable old Canon FD, Minolta MD, Pentax K, etc. glass out there from back in the MF days when "parfocality" was more of a design concern. Can you honestly dismiss every single one of those lenses? Keep in mind $1800 isn't peanuts. – Bungle Feb 3 '16 at 7:46
  • 1
    I'm not sure if this helps, but I've not yet seen a single modern SLR zoom lens that isn't parfocal through the zoom range (someone please correct me if I'm wrong here). One other thing of note (especially if AF isn't required) is that a few of the one-touch zooms from the 1980s/90s might meet your criteria. I'm thinking of lenses like the 75-150mm f/3.5 Nikon E-series zoom. Although it is only a medium tele 2x zoom, it is constant aperture, cheap and has excellent optical characteristics. – HamishKL Feb 3 '16 at 10:09
5

This reply to @Caleb's comment kept growing and growing into an off-topic answer. Maybe you still find it useful.


After mounting the zoom ring gear, I'll attach a pinion gear to the stepper motor shaft to control the motion of the zoom ring.

A linear zoom throw allows for smooth, consistent zooming that doesn't draw attention to itself.

"stepper motor" and "smooth" don't go together well.

Stepper motors make steps (hence the name ;)) and never a continuous motion. There's no way around the fact that the torque (which is the thing that causes the motion) is applied in discontinuous steps. You can smooth the motion with various efforts to some degree, but such efforts take away the inherent advantage of the stepper motor: its simplicity. And even if you manage to smooth out the motion to a desirable degree, you basically created a regular motor, so why not use that in the first place?

Take a look at the various stabilising gimbals. They often use brushless motors. What would the footage look like if they used a ste-ep-ep-ep-epper motor instead?

And should you ever want to turn the motor driver off to allow manual rotation of the zoom ring, you'd always have to disengage the gear that connects the stepper motor mechanically to the lens, because the steps of a stepper motor are noticeable when rotating it by hand.

I don't like to be that guy, but to some degree your question reads like this:

There are these expensive lenses that have the properties A, B and C which all make them expensive to manufacture. Now I'd like to duct tape a stepper motor to a lens with the same properties and have it for cheap. Any ideas how?

I'm not seeing how merely requiring the existence of the same properties of an expensive lens enables you to build a cheap one.

It adds another layer of complexity. I would need to discover - either through research or (more likely) experimentation - the actual behavior of the zoom throw, which can be complex.

Chances are that lens companies do actually do some research and development. I'm afraid now it's your turn. You have some engineering to do that goes beyond "I just buy components that happen to full fill all my requirements".

I also don't need auto [...] focus

I think you do want auto focus, though, but not for auto focus in the classical sense. That's because automation of both zoom and focus enables you to control either one, but also both at the same time, depending on each other.

When you calibrate a lens, you obtain knowledge about how

  • focus changes, depending on focal length (varifocus)
  • zoom changes, depending on zoom ring rotation (nonlinearity of the ring)

Knowing these static errors, you can compensate for them during operation. As suggested by @Caleb in a comment, you can compensate the nonlinear zoom ring pretty much directly.

In order to turn the varifocal lens into a parfocal one, you'd have to automatically adjust the focus depending on the current focal length (and the relationship between them obtained during calibration).

If you did the inverse (adjusting focal length according to focus), that would prevent any focus breathing the lens might have.

In conclusion:

  1. With the ability to rotate and measure the position of both zoom and focus ring on a lens automatically, the hardware becomes general purpose and applicable to (pretty much) any lens.
  2. Both requirements (parfocal lens, linear zoom throw) are abstracted away from the lens itself and are now a job of your software that controls the two rings, which allows you to use lenses that are neither parfocal nor have a linear zoom throw and thus cheap.
  3. Given that the lenses deviate from the original requirements in a static manner, they can be calibrated. That means that the software that controls the automated rings of the lens is also universal and all that needs to change in order to use a different lens is to use a different calibration.

By making parfocality1 of the lens and linearity of its zoom a property of the system surrounding the lens, the lens does not have to have these properties any more, which enables you to use plenty of other (cheaper) lenses.


1 that's totally a word

  • "parfocality" – it's a perfectly cromulent word. – scottbb Feb 3 '16 at 2:08
  • Sorry, multi-comment response inbound here, but I don't know of a better format. – Bungle Feb 3 '16 at 2:13
  • Thanks for your response - I sincerely appreciate you taking the time to think it through. It is useful, but also more than a little patronizing. I do understand that stepper motors have limitations inherent to their design, that I have some engineering ahead of me (that's kind of the point), that lens companies do R&D, and that the cost of said R&D is amortized in the price of the resulting lenses they release. – Bungle Feb 3 '16 at 2:14
  • Unfortunately, I think the cynicism that's evident in your answer is all your own. If I wanted to join in, I could just as easily characterize your answer as reading like "So you don't want to pay big money for a lens that already does all of this. You want something for nothing. Easy, just start with a lens that does none of this, spend a ton of time engineering those features back into the lens, and voila." – Bungle Feb 3 '16 at 2:14
  • I'm looking for a lens that's somewhere in the middle of two extremes we've now characterized: one that has most of the features I want, none of the features I don't need, and thus a price that's more friendly to my budget than $2500. I'll then spend a limited amount of time building in the feature that I want to control even more than the $2500 solution allows (i.e., motion control of the zoom ring) while spending as little time as possible reinventing the wheel. – Bungle Feb 3 '16 at 2:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.