Find all your DIY electronics in the MakerShed. 3D Printing, Kits, Arduino, Raspberry Pi, Books & more!





Ask MAKE is a weekly column where we answer reader questions, like yours. Write them in to mattm@makezine.comor drop us a line on Twitter. We can’t wait to tackle your conundrums!

ask_make_potentiometer.jpg

Paul asks:

I’m a museum educator, and working on an exhibit where I want to have a knob that you can spin around to change the color of an exhibit. I thought about using a potentiometer, but I am worried that it might break if people try to turn it too far. Any idea what I could use? Do they make a potentiometer that you can keep turning?

Interesting question. I now know that museums are pretty tough environments, so I understand your concern about it breaking. I can think of three options that might work:

  • Rotary encoder
  • Continuous turn potentiometer
  • Regular potentiometer, with a slip clutch

Using a rotary encoder is the first thing that comes to mind. These devices have a disc with markings in them that spins when their shaft is turned, and have a sensor (usually optical) that detects when the marking move by. Because the sensor part is non-contact, they are generally more robust and can last longer than, say, potentiometers. The downsides to using them is that they are more complicated to use, and have a lower resolution than a potentiomer solution. The Arduino playground has a nice example of using a rotary encoder with a microcontroller.

A continuous turn potentiometer is an interesting option; I actually didn’t know they existed until now. They certainly solve the issue of breaking by turning too far, because they don’t have a hard stop. The only issue that I can think of is that their value will will probably jump suddenly from full on to full off as they make a complete revolution, which may require special handling. Actually, there is a second issue- I don’t know where to buy them! Anyone have some pointers?

If you can’t find a continuous turn potentiometer, a final idea might be to use a regular potentiometer, but connected indirectly through a slip clutch. The idea is to use something that will limit the amount of torque that the user can put on the potentiometer, which will prevent them from turning it too far and breaking it. You can purchase a special mechanical device that will do this, or perhaps even use a loose belt that will slip when the knob is turned too far.

Good luck with your project, and be sure to let us know how it turns out!

Have a different way to do this that you think might work better? Sound off in the comments!


Related

Comments

  1. Stunmonkey says:

    Rotary encoder all the way.

    Especially in that kind of environment, I would want to avoid some of the problems with a pot – primarily MTBF from physical contact wear, and the value spikes you mention. The rotary encoder avoids both of those.

    A rotary encoder can indeed be hooked up with a simple microcontroller, or better yet directly into a PC if such a system is already running the display. You can easily limit the speed of value changes as well if that is a concern.

  2. Robert Ely says:

    Rotary Encoders tend to fair better in this situation, if possible don’t mount the dial directly to the encoder, instead use a simple 2 pulley system and a rubber belt. This saves the controller if some one puts a large amount of side to side stress on the shaft, or if some one puts all their weight straight down on it.(and they will)

    Also a Rock solid back plate is always help full. Plate steal if at all possible. Avoid particle board.

    I have worked in a Children’s science Museum and have seen the kind of damage a a seven year old can cause, i assure you its not pretty.

    Resolution is not really a problem, If you want you can always adjust the gearing on that pulley system i mentioned. 2:1 gear ratio will get you double the resolution.

    As far as Brains, anything from an AVR, to a PLC can get the job done.

    1. Stunmonkey says:

      Nice simulpost. You summed it up far more succinctly than myself though.

      Glad someone is thinking along the same lines.

  3. Stunmonkey says:

    Rotary encoder all the way.

    Especially in that kind of environment, I would want to avoid some of the problems with a pot – primarily MTBF from physical contact wear, and the value spikes you mention. The rotary encoder avoids both of those.

    I would avoid the addition of a slip clutch for the same reason I would avoid a pot. You are then just accelerating failures from wear/damage issues.

    I would further mount the encoder UNDER the user accessible control panel, connected by a short flexible mount (rubber tube, Helical Beam coupler, Oldham connector, etc.) to the actual user accessible control wheel/shaft assembly, which should be set in its own steel roller bearings on the main panel (skate bearings for instance).
    Experience has shown this will remove commonly-occurring issues with any high-moment lateral and axial shocks and hits sustained by the users wheel, leaving the heavy-duty bearing to take that, and only rotational wear on the actual components own rather fragile sleeve bearing to worry about. You will be replacing them weekly otherwise, especially in a children’s museum.

    A rotary encoder can indeed be hooked up with a simple microcontroller, or better yet directly into a PC if such a system is already running the display. You can easily limit the speed of value changes as well if that is a concern.

  4. Ken says:

    The radio in my car has two knobs that can be turned continuously. Short of taking it apart I wonder which they are. The volume knob stops incrementing when it gets to the highest allowable volume. The channel changer circulates back to the beginning (or end). I would think that a discarded or salvaged radio would provide ample opportunity to inspect this type of device (and re-use).

    Plus, I think car radios are pretty durable devices.

    Good Luck.

  5. Alex says:

    We had the problem, too.
    We solved the light changing issue with a touch sensitive dimmer switch.
    It has a stainless steel surface and no parts are moving as it is touch sensitive… it is one of the best solutions for public places.

  6. Anonymous says:

    I vote for the rotary encoder too.

    However, if you want to use a potentiometer, you can usually modify the large metal can type (the ones with the 1/4 inch shaft coming out of the 1 inch diameter body) to be continuous. These are the kind you can usually find pretty easily at Radio Shack (at least I assume you can still find them).

    Part of the case is bent inward and forms a stop that the wiper (internal rotating part) hits. You can remove the stop or modify the wiper. The cases are usually held together by a few bent metal tabs. They are not hard to take apart and reassemble.

    You can either cut the stop or bend it out of the way.

    If you do this, the potentiometer will now allow continuous rotation. However, the zone between the previous max and min values may give you all sorts of strange values (including an open circuit) depending on how it has been manufactured.

  7. Anonymous says:

    You could always just install stops on your knob or on the shaft, so it can only be turned. The micro solutions are neat, but a mechanical one will probably be easier.

  8. ehud42 says:

    Connect 2 linear pots to a wheel like a reciprocating motor. One in the X and one in the Y direction. Any micro-ctrlr can read the two inputs to determine the location of the dial. The dial can be rotated continuously.

  9. Ryan N says:

    I just bought one of these:

    http://cgi.ebay.com/Oak-Grigsby-Rotary-Encoder-128-Pulse-Rev-91Q128_W0QQitemZ390104395628QQcmdZViewItemQQptZLH_DefaultDomain_0?hash=item5ad408af6c

    It hasn’t arrived yet, but it might be just the right part for a good price.

  10. Don Martin says:

    Why not a slide pot? They work fine attached to rotary potentiometers as well (think audio fader).

  11. jammit says:

    You can do a google search for iPod parts. You should be able to find a touch controller. Another would be a PS/2 touch pad hooked up to a micro. There are plans and programs in Makezine for the Arduinio for interfacing with one.

  12. Garrett says:

    This is a continuous rotation potentiometer: http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=3382G-1-103GCT-ND

    Using an encoder is good and all, but this device is rated to 1 million rotations and gives you absolute position at all times using a simple analog input.

    1. Garrett says:

      Here is a through-hole version that might be easier to use: http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=3382H-1-103-ND

      1. Garrett says:

        Another approach is to use a special sensor that uses several Hall effect device to measure the position of a magnetic field. Here is one example: http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=356-1084-1-ND

        It works kind of like a compass. You place a small magnet on the end of the rotating shaft and position it above the chip.

  13. garydion says:

    I’m surprised no one has mentioned using arcade game control panel parts. They figured out the how to “bulletproof” controls from 7-year olds 30 years ago. Well, the controls now are better of course.

    The continuous rotation potentiometer in Omega Race has two wipers, so one is always in the linear portion of the pot when the other one “jumps”. That said, it’s a pain compared to quadrature encoders.

    For example, Tempest uses a quadrature encoder, and it’s much more robust. It’s tolerant to both axial and radial deflection, at least as much as a sugar-fueled roller-skating rink visitor can subject it to!

    For a wealth of information, Google: arcade rotary controller

  14. Anne says:

    The responses so far seem to be mostly voting on the original 3 solutions offered. The copper tape used for stained glass is conductive and could be used to make a pair of wooden disks which report the current position of the top disk over the bottom.

    Similarly, a top disk with a protrusion on the bottom could slide over a pair of plastic sheets pressing together conductive traces on them, much like a modern computer keyboard. Use serial protocols to decode and respond.

  15. Chris W says:

    You can buy pots with continuous rotation and sine/cosine outputs, but they are usually expensive. See http://www.p3america.com/pp/scb50,fp50.htm .

  16. vt-pete.livejournal.com says:

    Wow, that is a lot of posts. I, too am in the museum interactives business. I would not reccomend anything lighter than an industrial motor control for public use. We have used “Spinner” type devices form suzo happ in the past: http://www.happcontrols.com/amusement/gameparts/95265700.htm

    Very Rugged.

    If you are interested in a unique product that is plug and play, check out David Durlach ( technofrolic’s SpinBrowser):

    http://www.technofrolics.com/trade-show-exhibits-museum-exhibits/video-spinbrowser/video-spin-museums.html

  17. Bob Reynolds says:

    I have been building electronics for interactive museum exhibits for about 19 years and have used all of the suggested solutions. Most of my designs use a microcontroller. Incremental rotary optical encoders have the advantage of no absolute position. The system can be reset to any value and turning the encoder increments or decrements that value. The controller can keep track of multiple turns, so you can emulate a multi-turn pot. The trade-off is that there can’t be a pointer on the knob as the power-up position of the encoder can not be known. There are absolute position encoders, good for almost 1 turn. See usdigital.com. When we must use a pot, I have happily paid $100 each for knobs made by the Exploratorium which have an internal stop and couple to the pot with a piece of plastic tube which slips if the pot hits its internal stops. Pots then last for months.

  18. GordWait says:

    I used to know a tech who worked for our local Science World, who told me their method for a simple cheap clutch – buy a couple of small cheap magnets, glue one to the end of the Pot, and the other to a panel mounted rotating shaft (another Pot might do, glue the magnet to the rotating part of the pot that sticks out the back). Mount all together so the magnets align north to south and stick together. Turn the dial too far, the magnets slip.
    Obviously the magnets should be cheap ones, not some rare earth ones that stick too strongly..

  19. Simon says:

    We need to know more about what you are controlling. Is it a small number of discrete colour changes or a continuously variable one?

    If just a small number I would use a shaft with a cam disc on it to actuate X number of good microswitches (as someone mentioned arcade game controls are good).

    If you want continuous variance and want to use a pot (although the rotary encoder/micro solution people offered is better) then maybe a linear pot driven by a crank arrangement. Then you’d get a nice (sinusoidal) fade from end to end with no sudden jumps in value as they turn the control. You do have wear issues on the pot itself though then as other have mentioned.

    Maybe you could use an old style throttle position sensor off a car? They must be rated for many, many operations before failing.

  20. Alan S. Blue says:

    The way I would implement this is very similar to a trackball.

    One type of optical trackballs and mice uses the optics to watch the rotation of the ball.

    With a ninety degree turn in the orientation of the sensor, you can have a standard USB sensor that measures rotation instead of ‘up-down, left-right’.

    There were various USB devices that were sold commercially that did (or do) exactly that.

    Griffin Powermate is a (pricey) aluminum knob with limitless rotation. There have to be others.

In the Maker Shed