SparkFun Forums

[aliciaeggert]

I need 60+ motors that spin reliably at exactly 1rpm (like the second hand on a clock). I would use actual clocks, but a cheap clock's second hand is actually very unreliable. I plan to use the "second hands" (IE arms coming off the shafts of each motor) to spell a word, like this sculpture I made in 2010: http://aliciaeggert.com/works/show.php?id=103 (which uses the hour and minute hands of 30 electric clocks to spell the word "eternity" once every 12 hours).

Since I need such a large quantity or motors and "second hands", I am looking for the cheapest (but still reliable and accurate) solution. The load I would applying to each motor is very minimal (second hands are very light), so torque is not an issue.

Would continuous rotation servos be right for this application? Is it possible to set a servo's rpm to exactly 1? If the motors are turned on/off regularly, would their stopping point need to be adjusted all the time?

If servos are not the right solution, can small/cheap gear motors be controlled to spin at exactly the speed I need?

I assume my other options are
a) stepper motors
b) a high-torque 1rpm motor and a lot of pulleys to turn the rest of the shafts...

Please let me know your thoughts. I appreciate your help!
Alicia

[viskr]

Interesting project. Why do you say clock second hands are unreliable? Cheap clocks could be a way to go though controlling the motor is a bit more complex (relays would be best). And you have the sync problem (actually knowing where the arm ends up at). Actually getting feedback on the arm position is the difficult issue here.

Stepper motors can be controlled in relative position, but most don't have a "home" state. Most likely you will need some position sensing mechanism on any motor, this is usually done with what are called quadrature encoders. Running 30 in parallel could be a difficult task, except since you are running them so slow, it could be handled just by looking at the quadrature outputs.

[Mee_n_Mac]

Would a better statement of your requirements be that once per N seconds the shafts off all 30 motors should be at the same angular position and that the motor speeds should match each other "well enough" and that this speed should approximately be 1 rev/minute ?

Here's my thinking ... a set continuous rotation servos could be fine tuned to run at almost the same speed but even if initially sync'ed up (a problem any solution must tackle) eventually the itty bitty speed differences from one to another would render the shafts out of sync. Their only advantage over a set of 30 simple motors (maybe) is that this takes long enough that you'll want to power down and reset the display before the lack of sync becomes obvious.

Steppers should all stay in sync, barring something interfering with the motion of the shafts. This might be the obvious choice but the cost of motors and drivers may be too much.

A single motor driving all 30 shafts is a good solution to keep everything in sync. Depending on the display the mechanics may be too complex or may be simple (think 1 long serpentine belt like your car has these days). Hard for "us" to tell you given we don't know the specifics of your display. My GUESS is cost is on the low side as well.

One thought that comes to mind is to cheat a little bit. 30 "second hand" simple motors might run close enough to the same speed that, if each were paused at the desired sync point for just an imperceptible fraction of a second, that they could all be kept apparently running at exactly the same speed. You would need a way to sense the rotation angle that's is the sync point but that's easy enough via switches or optical detection. This also allows an easy method to initially sync up all 30 shafts. Each motor is run until it's sync point is detected and then it's stopped. When all 30 are sync'ed, all 30 are simultaneously powered and allowed to free run. In fact this how the system normally operates, re-syncing every rotation and approximately every minute. So long as all 30 motors get to their sync points within 0.1 sec (perhaps longer ?) of each other, I doubt the pause would be noticed by a human. I think this would be mechanically simpler than the 1 motor, 30 pulleys/gears ... perhaps ... and maybe about the same cost.

[viskr]

I think cheap Chinese motor/encoder combinations from ebay are your salvation on this one.

I checked eBay looking for "motor encoder" and found a number of small ones for $4 and $5.

While there are lots of microprocessors out there with 1 or 2 quadrature encoders, 30 is something else. So you'll need to be watching those lines in parallel and the CPU will be pretty busy doing that to keep them in sync.

These cheap ones are DC motors, so you apply power and count pulses, unfortunately they probably won't all run at the same speed, so you will have to compensate for that in software. Or you might be able to compensate for that with some series resistors.

[dlotton]

My initial impulse was stepper motors, but that would likely be a fairly expensive solution.

For 'inexpensive' you might look into synchronous AC motors. At steady state, the motor RPM is synchronous to the AC waveform of the source. In theory, if you drive them all from the same AC source, they would all be synchronized.

http://en.wikipedia.org/wiki/Synchronous_motor

Here are some cheap ones for $6.95 each(looks like from China). Probably cheap enough to do some testing without denting the wallet too severely...
http://www.aliexpress.com/item/New-micr ... 86938.html

Also search ebay for 'synchronous motor'... lots of inexpensive options.

This option won't 'lock' all of the motors together, but it would be inexpensive and relatively simple compared to a stepper option. To hit your target RPM you may need to generate an AC waveform other than 60Hz wall power, depending on what options you find for your price range.

[aliciaeggert]

Thanks so much for all of your replies! It sounds like synchronous motors might indeed be the best option. I could plug all of them into a chain of power strips, the same way I powered the clocks for Eternity. I found some for less than $5 each on Ebay. I am going to purchase a few and give them a try. I will keep you all posted on my progress!

[dlotton]

Thanks so much for all of your replies! It sounds like synchronous motors might indeed be the best option. I could plug all of them into a chain of power strips, the same way I powered the clocks for Eternity. I found some for less than $5 each on Ebay. I am going to purchase a few and give them a try. I will keep you all posted on my progress!

Definitely warrents some investigation. I think many inexpensive AC powered wall clocks are essentially just synchronous motors. It could be that friction in cheap gear reduction sets in these cheap motors may cause 'slippage' with respect to the AC waveform. Worth a test at those prices.

The down-side, unlike a stepper motor setup, they can't all be easily indexed to a starting position by a microcontroller if things get wonky... but they're cheap.

Good luck.

[Ross Robotics]

First, I must say the "Eternity" clock is outstanding! Now that is what I call art. Very impressive.

Secondly, I think you shouldn't rule out steppers at the moment. You don't need large steppers like a Nema 17, You just need a stepper that will rotate slowly. You could get a small stepper like this one. Even comes with a controller. You would get them cheaper at a higher quantity..

Or you could use something like these:

[dlotton]

Secondly, I think you shouldn't rule out steppers at the moment. You don't need large steppers like a Nema 17, You just need a stepper that will rotate slowly. You could get a small stepper like this one. Even comes with a controller. You would get them cheaper at a higher quantity..
[/url]

Nice find. Didn't think you could get steppers that cheap.

Steppers require more effort than a synchronous motor(you need a brain to drive them), but they open up a lot of other possibilities also, like controlling direction and rotation speed of individual motors.

[Mee_n_Mac]

Nice find. Didn't think you could get steppers that cheap.

Nor did I.

Steppers require more effort than a synchronous motor(you need a brain to drive them), but they open up a lot of other possibilities also, like controlling direction and rotation speed of individual motors.

I was thinking the same thing. You could have an apparently chaotic mess come together once per minute to a cohesive message.

[aliciaeggert]

Thanks for the link to that stepper motor, codlink. I am going get some of those. I already ordered a few synchronous motors, too, so I will play with both and see what works better. I'll post pictures and updates soon!

[chartle]

Any thought to using one motor and a chain drive?

http://www.servocity.com/html/sprockets___chain.html

http://www.servocity.com/html/plain_bor ... 1227_.html

http://www.surpluscenter.com/powerTrans ... powerTrans

[aliciaeggert]

Hello all! So, an update:

I have been playing with a few of these 1rpm synchronous motors:
http://www.lusolarelectronics.com/small ... -p-63.html

I was hoping that when I plugged all of them into a power strip and flipped the switch, they would all start moving at once. However, that is unfortunately not the case. Each time you give one of these motors power, it will start spinning in EITHER a clockwise or counter-clockwise direction. I have plugged/unplugged each of them multiple times, and the direction they spin each time seems totally random. When I turn a bunch of them on at once (with the switch on the power strip), some will go one way and others will go the opposite way. This is obviously problematic if they are meant to look like clocks. Also, I've noticed that they don't start moving smoothly every time. They sometimes jumpstart, or sit still for a split before they begin moving. And it's important that they all start/stop in synch when I turn the sculpture on/off, so this is another big problem.

So, synchronous motors are not so synchronous! Unless, of course, I'm doing something wrong. Both wires coming out of the motor are black, and I can't identify a different behavior when I change them around, so it's hard to tell which is power or ground. This makes it impossible to wire them in series or parallel, which is the only other way I can imagine getting them all to move in synch. Is there another way to tell positive from negative that I don't know about?

At this point, I am considering using one higher torque motor with a chain drive, as @chartle suggested. That is how I made this sculpture, which spells the word NOW about once per second: http://aliciaeggert.com/works/show.php?id=114

If you have any other thoughts, please let me know. Thanks!

[aliciaeggert]

ITNEVERSTOPS_small.jpg

**Click on the images to see them larger**

ITNEVERSTOPS_I_small.jpg

[chartle]

Hello all! So, an update:

I have been playing with a few of these 1rpm synchronous motors:
http://www.lusolarelectronics.com/small ... -p-63.html

I was hoping that when I plugged all of them into a power strip and flipped the switch, they would all start moving at once. However, that is unfortunately not the case. Each time you give one of these motors power, it will start spinning in EITHER a clockwise or counter-clockwise direction. I have plugged/unplugged each of them multiple times, and the direction they spin each time seems totally random. When I turn a bunch of them on at once (with the switch on the power strip), some will go one way and others will go the opposite way. This is obviously problematic if they are meant to look like clocks. Also, I've noticed that they don't start moving smoothly every time. They sometimes jumpstart, or sit still for a split before they begin moving. And it's important that they all start/stop in synch when I turn the sculpture on/off, so this is another big problem.

So, synchronous motors are not so synchronous! Unless, of course, I'm doing something wrong. Both wires coming out of the motor are black, and I can't identify a different behavior when I change them around, so it's hard to tell which is power or ground. This makes it impossible to wire them in series or parallel, which is the only other way I can imagine getting them all to move in synch. Is there another way to tell positive from negative that I don't know about?

At this point, I am considering using one higher torque motor with a chain drive, as @chartle suggested. That is how I made this sculpture, which spells the word NOW about once per second: http://aliciaeggert.com/works/show.php?id=114

If you have any other thoughts, please let me know. Thanks!

You could also use 2 or more of the motor all chained together which shouldn't cause any problems.