a clever reduction unit!

[quadrirotor]

A clever reduction unit and very amazing!....
http://www.hdsi.net/media/principles/gearanimation.wmv
http://www.hdsi.net

[Al_Hammer]

Thanks Andre, that looks brilliant! (but the idea is almost 50 years old) :D
The theory of operation is described here http://www.hdsi.net/principles/

Harmonic drive was apparently invented in 1957 by Walt Musser
http://www.waltmusser.org/

[Jazzenjohn]

These are the cyclos we use on the robots I work on:

http://www.smcyclo.com/products/precision/fine_cyclo.shtml

[donshoebridge]

Just for reference, harmonic drives are VERY expensive (cheep models run $600+). They are used in robotic systems and machinery that requires a zero backlash speed reducer. When I worked for United Technologies Automotive (now Lear Corp), we used harmonic drives in some of our test equipment to measure torque curves on rotary switches. They are very smooth, have zero backlash, have high ratios and are expensive. I don't think this would be a good choice for a pre-rotator, because if you ever tore one up, it would be expensive to replace.

[Doug Riley]

Not knowing any better, I wonder how much power you can transmit through the small, soft teeth on the flexible element.

[Al_Hammer]

Another question- the smallest ratio I happened to see was around 30:1.
Could it be that they are not made in the the ratios that would be useful, such 3:1, or 2.58:1, etc, typically used for a prop redrive? :confused:

[donshoebridge]

I haven't had to deal with them all that much, but it was my understanding that they can only be made for the higher ratios. Since there is only a couple of teeth actually engaged, I seriously doubt that any amount of real power can be applied without ripping the ring gears apart.

The drives that our equipment had were 100:1. Because our computer/motion control system used a stepped motor, we wanted to have high resolution in our measurement data. I can't remember the number of steps per rev our motors were, but when you multiplied it by 100, 90 degrees of output rotation generated several thousand data points.

[Jazzenjohn]

Al, They aren't used for anything much smaller than 20-1 although you'd be hard pressed to find one even that small. If you're looking for something inline how about a planetary?

Don, They transmit quite a bit of torque and if you look carefully you'll see that they spread the load out over many teeth, not just a couple. On our robots the gearboxes have 3 plates 120 apart on all 5 or 6 axis and we can throw 300 kilos around anywhere in the envelope at a pretty impressive speed all day every day without problems.

As far as suitability for use in a gyro it could possibly be used for a prerotator gearbox if you could find one that was inexpensive and light enough. It would also likely need a planetary first stage. It isn't really a good candidate though.