The ideal servo for robotics

--Small is cute, but spare a little of your talent to design some robust stuff for "real world" bots please! :-)

Speed-wise I think you need to go no less than 1 rps. These motors are a bit too slow; the exporter ought to be able to supply some other gear ratios. I know there is a market for small motors for micro and nano robots, but I think it's a less significant market than desktop robots that use standard servos. If anything, people tend to ask for servos that are beefier and faster that the typical Futaba S-148 (or knockoff). That makes the target >1 rps, and >56 oz-in (at 6vdc).

If you can manage the motor and optical encoder for $20-25 that would place it at a competitive level with a $10 servo and a $10-15 (per side) encoder setup, but not a jaw-dropping price. Remember that an R/C servo doesn't need an H-bridge, so while control is not as precise as regular PWM, there's less to design, and overall less cost. RC servos are popular not so much because of their features, but because of their low cost, plus ease of use.

Linear servos: I've seen retrofits for taking a standard servo and making it linear, but unless you need the more precise movement, most people just use a bell crank, and live with the (somewhat) non-linear positioning.

-- Gordon

Thomas Arundel wrote:

A 360 degree servo sounds like something I would buy. Especially at the price point you suggest. Positional feedback is a nice feature too, but I think the resolution would need to be very high. An I2C interface is also a great feature. All in all I think you've got a good idea if you can meet that price point.

Hmm I'm starting to think that a retrofit board that converts geared motor units into precision servos might be more marketable and useful product, than a unit reliant on one particular size of motor / geartrain...

What sort of sizes and specs are you thinking of?- we can source much larger gearboxes too - high quality metal geartrains... :-)

So we're getting about 7-8 other types with different speed outputs in the next 14 days. We're also going to custom make some of these with higher speed outputs. Someone wrote to me asking for 250 RPM - if we were to do say

Interesting what you say about futaba servos. Any idea how fast they are / torque output?

What about mechanical specs like volume - is the RC servo an optimal design, or do you think it could be improved on for robot applications?

This is true. Have to be honest that most servos I've seen here in Europe are more like $15 for a standard size. I think that we might be better off developing an after-market system to turn a wide range of geared motors into servos, rather than limiting the market and tying the two together from day

Thomas Arundel wrote: : Hmm I'm starting to think that a retrofit board that converts geared motor : units into precision servos might be more marketable and useful product, : than a unit reliant on one particular size of motor / geartrain... --Neat. : What sort of sizes and specs are you thinking of?- we can source much larger : gearboxes too - high quality metal geartrains... :-) --Well I'm partial to those $20.- windshield wiper motors with the built in gearbox, just cuz they've got lotsa torque and they're easily obtainable from several surplus dealers..

I think it's do able - I think we can also get such a system to report back on output torque and motor currents for 'free' too. As for resolution, any ideas what sort of angular position accuracy / resolution one can expect from a RC servo unit?

If we were to consider a 100 rpm motor/gearbox, I don't think setting the output shaft to within 10 degrees of a set position would be overwhelmingly hard, which isn't bad given that you'd probably want to gear that down again by a factor of at least ten if you were doing something like robot arm positioning work. That means one could achieve an accuracy of 1 degree on a

Another feature would be that one could instruct the unit to exert a specified torque at its output. do you think that would be useful?

The specs I noted are those for a Futaba S-148, operated at 6vdc. Servos based on the S-148 (Parallax's pre-modified servo, for example), or modelled after it (GWS S03N, for instance) have the same specs.

The flat body is a plus for down-and-dirty mounting using glue, double-sided tape, or Velco or Dual Lock. Remember that for the US we don't have lots of metrix precision screws laying around, and I bet any motor from Asia will have metric pretapped threads. RC servos just have two or four holes for mounting with any hardware. Consider all mounting slop. With a standard motor ou must fairly accurately match up the mounting holes on the face, or else come up with some type of alternative case-mounting system, like the Solutions Cubed EZ Roller motor.

Finally, consider mounting things to the shaft of the motor. You can attach gears, hubs, bell cranks, or whatever to the servo by fashioning it against a common (and often, included) servo disc. For a motor with a shaft you need a hub of some type, typically at an additional cost. Sources here charge about $5 for a pair of aluminum (al-u-minium in the UK ) hubs for 3 or 6mm shafts.

-- Gordon

Yes I took a look at those. I'm confident that we can design some gearbox motor configs that are as strong by 1/3 or 1/2 the weight. How important is a plastic case - without it things will be cheaper and lighter. Don't suppose too many people do mud wrestelling with their robots... yet.

You guys will just have to buy one of our forthcoming metric screw sets then ;-)

Yes, we'll come up with something suitable. Someone mentioned to me that a servo with dual shafts would be very useful. I think that's doable too.

Yes - we'll start to look at puting together a mounting set for our current gearboxes straight away.

Some of us abuse the heck out of servos & other equipment: