Seeking electric motor for serious relationship


I need a better motor for my curtain automation. I've tried several models, and they all only last two or three months before they strip their internal gears. So my current procedure is, "buy a new motor every two months." This is not a good system.

I figured the motors would just stall out before stripping gears. They are straining, but never reach the point where they actually stop turning. But no, eventually something cracks.

I'm driving the motors with with this board, which is 12vdc and has a 1.2a max stall current, so I've tried these motors, and stripped the internal metal gears on all of them, multiple times:

I'm kind of puzzled by this, because except for the 227:1 1.1A, they are not stalling out! They strain a bit at the end, but they keep turning. I expected that the failure mode I would be repeatedly experiencing would be that the motor is fine and the driver board melts. But somehow, even without the motor ever getting stuck, it's just shattering internally. (Applying power directly results in the shaft not rotating at all, or rotating only partially then getting stuck, so it's definitely that the gearbox has committed internal suicide.)

I need to find a quiet, fast motor with like, 350+ oz-in torque. Bidirectional, constant speed. And if it has more than a 1.1A stall, or if it needs more than 12vdc, I need to find a driver board to run it with. I don't even really know where to look. You'd think lots of people would want to move heavy things with a small box.


Tags: , , ,

25 Responses:

  1. James says:

    Never used one of these before, but it seems to meet your specs.

  2. Huh, a jwz post in my field.

    Bigger motors last longer. You're already using Pololu components, so you might as well buy one of their 37D motors: and pair it with one of their bigger controllers:

    You'd guess that a higher reduction gearbox wouldn't last as long as a gearbox with lower reduction, since the gears in the higher reduction box have smaller teeth, and there's more of them. For long mechanical life, buy a bigger motor and gear it down electronically, rather than use a small motor and gear it down mechanically.

  3. Lavel Marsol says:

    Have you tried oiling up the curtain rails? Or using a lighter weight curtain? Could the curtain rope be replaced with a pulley so the effort to move the curtain is less but takes longer to move it?

    • jwz says:

      Yes, I oiled the rail. No, it can't be lightened. Building a block and tackle sounds like a giant pain in the ass.

  4. b says:

    Saw this post recently, might be of some use (not sure if 5 cm / sec is fast enough but 1 kg is probably enough load):

  5. Jay says:

    Try contacting these guys:
    they're under different ownership now so can't guarantee that they haven't changed but they used to be more than accommodating for helping with these sorts of problems / projects, they'll have access to a ton of stuff not generally available online, so might be worth a shot.

  6. Paul Wilkinson says:

    This may be a) considered cheating and b) more money than you want to spend but maybe you need to replace the track with one designed for automation. Somfy has a very good reputation in this space and is used in lots of commercial boardrooms etc.

    The Glydea system can handle up to 132 lbs of drapery and has simply dry contact control (so all you need is a relay connected to your Arduino). The downside is it looks like it will be $800 or more (depending on the weight of your drapes and therefore which motor you need.

    • jwz says:

      I looked in to Somfy once, and all in, it would have been like $2k or more. (I have a big window. It's my only window.) So that's just a stupid amount of money, even compared to burning a $17 motor every couple of months. Also their protocol is some proprietary RF bullshit that is not "load this URL to close the curtain." So I'd be trading an Arduino motor torque hassle for an Arduino RF beacon hassle or something. Yay.

      • Paul Wilkinson says:

        I thought cost might be the problem as it sounds like you have a big window, and I am guessing that your drapes are "block out", so that means heavy, which is I guess why you keep going through motors.

        According to the manual The Glydea system has basic "dry contact" switch inputs as standard, so you don't need to work with any protocols. Just 3 pins that you short to ground for open/close/stop.

  7. Helyx says:

    from my favorite source for such things:


    or this one, which I don't think will blow your board like those others might. It seems that board will power 3 phase.

  8. Ben says:

    McMaster has a variety of DC gear motors.

    Scroll down past the $500 ones, unless you really want to spend $500.

    If you want more power, switch to an H bridge. I have used the Simple H before and had good results (won $1.125M)

  9. Andris says:

    Look up brushless servo motors. They are perfect for this application. Upside is lots of torque at variable speeds with no gears, will probably last well into the next technological epoch. Downside is they are freaking expensive.

  10. William says:

    First thought: NEMA23 geared stepper at about 15:1. Might work with a cheap arduino-compatible driver like A4988 or any of those kinds of things.

    Or buy cheaper from China.

    Second thought: windscreen-wiper motor or car-seat motor.

    If you're doing basic DC motors, don't need speed control and driver current is an issue, ditch the electronic driver and use a pair of 10A relays - super-easy.

  11. Hales says:

    For higher current (a few amps to dozens of amps) brushed DC motor control: you don't need a traditional silicon H-bridge if you are OK to miss out on variable speed control. You can get away with two SPDT relays instead, it's much cheaper and simpler.

    Wire them like this:

    +V GND
    | |
    | |
    + | . . | +
    . x x .
    . \ .
    . \ . Relay one
    . \ .
    . | .
    + . . | . . +
    Brushed DC
    + . . | . . +
    . | .
    . / .
    . / . Relay two
    . / .
    . x x .
    + | . . | +
    | |
    | |
    +V GND

    Another advantage: there's no 'wrong' position for the relays (everything is forced to go through the motor, you can't short out the driver like you can a silicon H-bridge).

    Con: I have not shown the driving circuitry for each of the above relays. You need to push current through a couple of pins on a relay to turn it on. If your control circuitry can't provide full coil current on its own (most likely) then you will need a small transistor (any common NPN will do, eg P2N2222) and a resistor on the base pin. Either way it's also a good idea to put a diode across the coil for 'flyback' protection too. Look up relay driving circuits (keyword 'Arduino' helps with searches) if you are not familiar.

    Automotive relays come in all sorts of current ranges and should be easy to get. Even scrapped ones from an a car wrecker's.

    Source: I design & make my own H-bridges. Feel free to ask any questions.

  12. M.E. says:

    The reason why the gearboxes are stripping before the motors stall is that when gears are turning, that creates additional stress on the teeth beyond the force at 0 RPM at a given torque. So if you matched a gearbox to the stall torque of a given motor, you would be oversizing the gearbox.

  13. joshua says:

    try using something with a planetary gear? spur gears are much more fragile.