[regaljohnston]

hi,

i bought a lens thinking it would be compatible with the zoom button on my camcorder, unfortunately it isn't sad smiley...

i've been looking at building an servo motor to control the zoom function, this will be better for me because i'll be able to place the control stick/trigger exactly where i need it on my camera rig.

i have a few of specifications however...

-it will need to be as quiet as possible (for obvious recording reasons).
-it needs to be variable speeds.
-able to ease in and out of zoom smoothly/seamlessly.
-the control needs to be operatable by my thumb alone, a left/right thumb-stick or seasaw peddle would be ideal.

i've drawn up a very basic schematic -
http://img96.imageshack.us/img96/248...troldesign.jpg
(the control stick will go on the handle of the poleroid chest stabilizer)

that's how far i have gotten, however-

i don't know how to source a compatible joystick/tumbstick/seasaw peddle which will allow me to operate with smoothness and precision.

i don't know if i've picked the right servo motor (really don't know about these things).

when i buy the motor and control stick will it just plug and play???

please help me, let me know what your approach would be. any help would be massively appreciated.

[doorknob]

It looks like an interesting project.

The servo that you have spec'ed looks like a typical RC servo. I'm not really an experienced RC user, but I know enough to be dangerous.

Normally typical RC servos are controlled by a radio control receiver, which in turn is controlled by a radio control transmitter. The joysticks are connected to the transmitter unit.

I doubt that you will want to use radio control for this application, which means that you will have to find some other way to control the servo via the joystick. You may be able to find a purpose-built servo control board that interfaces directly to a joystick, although I have no direct experience with such a thing. Otherwise you may be looking at a DIY project (for example, by using an Arduino microcontroller or similar as your servo controller). I would not count on it being plug-and-play, but then maybe someone else will chime in and recommend a kit that will do exactly what you want.

The other concern(s) that I would have with your plan have to do with the belt-drive setup. How many degrees of rotation will be needed to go from minimum to maximum zoom of your lens? Typical RC servos have a limited angular response, with hard stops at either end of travel. You can, however, find servos that are designed for continuous rotation, but that brings its own set of challenges.

The servos have a variety of linkages that you can use with them. If you can get a pulley-style linkage for the servo end of the belt drive, then you will either need a large diameter pulley on the servo (possibly comparable to the diameter of the zoom ring of the lens) or else if you use a smaller diameter pulley on the servo, you will end up with what is effectively a gear ratio that requires potentially multiple revolutions of the servo pulley for each revolution of the lens ring. You will also probably need to work out some kind of encoder to keep track of the servo position over the multiple turn range that will be required. Also you will need to figure out how much torque will be required from the servo to do the job (it looks like the one that you have chosen is normally paired with a fairly small RC airplane, and so perhaps it may not have sufficient torque for your intended application).

So, I think that you may have some more planning to do. But I'm hardly an expert in this area, so maybe I have misunderstood your plans, and therefore YMMV.

[kevin swan]

look for a servo tester there are many different ones to pick from

[regaljohnston]

thanks guys, i had a feeling i'd picked the wrong motor.

i don't know where to begin choosing so a few suggestions would be greatly appreciated.

and a servo tester seems like a good route but i don't have a clue which to choose, i've prepared another drawing to expand on the controller's requirements-

http://img59.imageshack.us/img59/184...roldesign2.jpg

the control in the picture is a libec zoom controller, the zoom control here is exactly what i am after.

i believe you're suggesting a wired connection between the motor and servo tester, this seems logical.

so i need suggestions on ideal motors and servo testers.

[regaljohnston]

something like this seems ideal...

SERVO TESTER Circuit

but i'm not sure about the torque requirements, the noise, and the smoothness.

the lens is not stiff at all, and would probably turn under the weight of about half a bag of flour...

[doorknob]

Let's see if I can suggest how to do this without screwing up the physics (Note to physics majors: let me know if I have actually screwed up). To come up with a rough estimate of the torque required to turn the zoom ring, you could try something like the following:

- mount the camera/lens combination on a sturdy tripod (with the lens positioned in the horizontal plane parallel to the floor)

- cut a short length of thin string or thread (less than a foot long)

- attach a bent paper clip to the end of the string as a hook that you will use to hang small weights such as metal washers (or, alternatively, tie something like a tiny paper cup to the end of the string, something that you can use to hold small weights such as coins)

- tape the string to the top of the zoom ring and drape it over the side, with the hook or cup hanging down (yeah, I know that you won't want to get your zoom ring all sticky, but you should be able to use easy-release masking tape or something similar that won't leave a residue)

- start adding small weights to the hook or cup

- when you reach the point where the zoom ring starts to turn, stop adding weights, and weigh the weights that have already been added (including the weight of the hook or cup) on a postage scale with a readout in ounces

- measure or estimate the outer diameter of the zoom ring, divide by two to get the radius of the zoom ring in inches

- multiply the weight in oz. by the radius in inches to get an estimate of the required torque in oz-in

- add in a torque fudge factor (for example, multiply that figure by 2 or 3) - you will want to have more torque than just the minimum to turn the ring

- figure out whether you will have a mechanical advantage due to using a smaller diameter pulley on the servo than the diameter of the zoom ring - if so, you can reduce the servo's torque requirement by the ratio of the diameters (there will be some efficiency loss, but you already have used a fudge factor which should handle that)

There is discussion of something similar here: High performance stepper articles. Part 2 - How to measure torque. Roman Black - Jan 2001

(Disclaimer: I've never actually done this. If you try it, take some photos and post your results. Also note - I'm not sure of the conditions under which the specified torque of a hobby servo is defined, namely at what radius from the hub - does anyone know the answer to that?)

PS: I'm not at all sure what the weight of half a bag of flour might be, but I'm intrigued by that unit of measurement.

[john-100]

the zoom control is basically a potentiometer with a spring return to the mid position

the voltage at the moving potentiometer contact is the input to the servo amplifier that drives the motor

the servo has an adjustable dead zone to ensure the motor doesn't move
with the control in the OFF position due to small errors in the zoom control detent position or the setup of the servo

the servo control circuit linked to in post 5 can be adapted to use a radio control type of servo but depending on your application
you may find the RC servos jitter makes too much noise


John

[Al_The_Man]

You really need a form of PID control if going with a servo, it sounds like a stepper will be too noisy in your application.
A DC motor is normally perfectly silent in operation, especially at low RPM.
If you know anyone that uses any kind of microchip, you may find a answer in an Aruino or PicMicro.

A Big Magnet: Using a DC motor as a servo with PID control (part 1)
http://ww1.microchip.com/downloads/e...tes/00532c.pdf
I thought at first, you may be able to make use of a Automotive 12v DC servo used for cruise control, but that relies on a minimum rpm feedback, but it would be interesting to reverse engineer, maybe and see if it could be adapted.
Al.

[regaljohnston]

ok i did what you asked, the weight at which the lens begins turning is at 700 grams.

the radius of the lens is 1.48 cm
circumference is 9.3 cm
diameter is 2.96

[doorknob]

ok i did what you asked, the weight at which the lens begins turning is at 700 grams.

the radius of the lens is 1.48 cm
circumference is 9.3 cm
diameter is 2.96

Are you sure?

A quick lookup of the SAL2470Z (at Sony 24-70mm f/2.8 Carl Zeiss T* Alpha A-Mount Standard SAL2470Z) seems to show that it takes a 77mm diameter filter, and so I'd expect that the zoom ring diameter would be greater than that.

[regaljohnston]

i measured the circumference and it was 9.3 cm, that's definite.

i put that into an online calculator to give me the other values...

[regaljohnston]

ahh, sorry, the cm/inch thing threw me off.. ok i'll just measure it manually-

obviously those values were stupid (my bad sorry).

ok- radius is 4 cm. (based on diameter being 8 cm).

[regaljohnston]

thanks john, how would i acquire a potentiometer with left/right thumb-stick control? it doesn't even have to return to central axis, i could manually centre it...

[doorknob]

If I have done my math correctly:

0.7 kg * 4 cm = 2.8 kg cm of torque

Since 1 kg cm is equivalent to 13.9 oz. in., my conversion comes out as a torque of about 39 oz. in. required to turn the zoom ring.

Comparing that with the torque rating of the small D47 servo called out in your diagram, the page at http://www.rc-dymond.com/index3.php?productID=176 claims that the D47 is rated for 20 ounces (I'm guessing that they mean "ounce inches").

If that's the case, then the raw rating of that servo would appear to have too little torque to move the ring (absent whatever mechanical advantage you might devise for it), let alone accounting for fudge factors. But again I do not really know how hobby servo torque ratings are determined or specified, and so I could be way off base here. Also, it's not clear that such a servo would have sufficient angular travel to do the job that you need.

[regaljohnston]

that's very helpful doorknob.

so if i buy a dc motor with 2.8 kg cm of torque and connect it to a potentiometer then my problem should be solved right?

[doorknob]

that's very helpful doorknob.

so if i buy a dc motor with 2.8 kg cm of torque and connect it to a potentiometer then my problem should be solved right?

I think that you would need to be a bit more precise in what you are suggesting.

First, I'd look at 2.8 kg cm as a bare minimum value of torque rather than being the desired value of torque.

Secondly, when you talk about connecting a potentiometer to a DC motor, you are a bit off course if you mean that literally.

BTW, in looking over my earlier discussion of mechanical advantage, I may have made an error. I think that the situation of using the servo to drive the zoom ring via a belt may be identical to using a capstan at the servo in direct contact with the zoom ring (because the linear travel of the belt would be identical to the travel of the zoom ring circumference), and so I don't think that the original way that I described pulley ratios would apply, rather you would use the straight torque rating of the servo (at some assumed shaft or capstan radius) in comparison with the torque required to rotate the zoom ring (but I'm willing to be corrected).

[john-100]

both the focus and zoom controls in the picture are just potentiometers in a custom case with a custom knob !

the focus control has a three spoked capstan type of knob

the zoom control having a custom shaped knob to be operated by your thumb
over a limited range
about plus or minus 40 degrees from the detent position
not the 270 degrees normally needed to turn the potentiometer from end to end
on the zoom control you will find another potentiometer with a small round knob to adjust the maximum zoom speed with the operational control at either end of travel

John

[regaljohnston]

doorknob- "First, I'd look at 2.8 kg cm as a bare minimum value of torque rather than being the desired value of torque."

ahh yes of course i see what you mean, i'd want a motor with a range which covers between around 2kg cm torque to 6kg cm torque. does that seem right?

doorknob- "Secondly, when you talk about connecting a potentiometer to a DC motor, you are a bit off course if you mean that literally"

i did mean it literally but was fishing for a correction :stickpoke
i take it i have to place some kind of chipboard in between the dc motor and the potentiometer? some more details of how i would make this work would be useful.

---

thanks for the input john, much appreciated. if you could provide some more information detailing how a layman like myself could put those ideas into practice i would be forever indebted to you my friend!

thanks guys.

[doorknob]

i take it i have to place some kind of chipboard in between the dc motor and the potentiometer? some more details of how i would make this work would be useful.

If you use a plain DC motor you will be in for some much more complicated design work. You are looking for a controller that will let you run the motor very smoothly at a low speed in one direction, potentially over a large angular range, be able to stop and hold the motor at any point along that range, and reverse the direction on command.

To do that will require quite a bit of circuitry. In other words, "this is not the solution that you are looking for at this time".

My suggestion is for you to do some experiments with a hobby servo having a suitable torque, and a servo tester such as described in post #3 above, and a suitable power supply (perhaps an RC battery and charger). Find a local hobby shop that specializes in radio controlled models and follow their advice for choosing suitable parts.

Then play with it for a while and see how what you have actually works. Use the potentiometer of the servo tester to move the servo back and forth for now. See if you are able to move the zoom ring with a jury-rigged setup, either using a belt or a direct capstan drive. Find out whether or not it will be quiet enough, whether or not it will hold position adequately without oscillating back and forth, and so on. If it looks like it will do the trick, then the next step will involve things such as looking for a joystick-like replacement for the potentiometer, constructing a proper housing for the parts, and so on.

[Al_The_Man]

You may want to look this article over.
http://www.robotshop.com/content/PDF...n-now-0804.pdf
I could be wrong as I have never got really deep into R/C controllers, but it is my conception that the really simple type as shown above, operate the motor to a desired position and stop, you may, or may not find this a satisfactory solution?
As outlined in the link, anything above this requires constant feedback in order to make sure the position is maintained (PID loop etc).
Al.