600rpm motor, low backlash

[Hank Fletcher]

I know I've posted this question in the robotics section, but I figured I'd widen the net in case it'd be more likely to get an answer in this forum instead.

What I want is 600rpm and about 1.4Nm torque (absolute max), with a max amount of backlash of 1/2 a degree. I've been looking for this for about a year now! I want to use the motor to drive a wheel of about 60cm circumference, and be able to stop it to within a millimetre (hence the need for low backlash). Am I nuts, or is this possible? Is making my own gearbox a solution, and if so does anyone know where I might get the parts I would need? Thanks all,
Hank.

 

[Nigel Goodwin]

Assuming you could find such a thing you're going to be looking at VERY, VERY expensive, you're talking extremely high precision - have you considered you may be looking at it the wrong way?.

You're wanting an accurately positioned motor, but rather than an accurate motor how about an accurate measuring system to tell you where it is, then move the motor till it's where you want.

 

[Sceadwian]

A stepper motor with anti backlash gears and a wheel encoder should be able to do this with a well tuned PID loop. Anti backlash gears shouldn't be hard to find, it's basically two identical gears and they're spring loaded in opposite direction to give constant tension to the gear train, I see them in printers frequently.
600RPM is 10 revolutions per second, which with the 60cm circumference and 1mm desired accuracy would require an encoder or feedback method with 600 pulses per revolution (double or triple for increased accuracy) and if you triple just to get a little more accuracy that's 1800 pulses per revolution, times 10 is an 18khz max frequency putting it in the realm of possibility for a micro controller. Assuming shaft encoders are available in the 600-1800 pulse per rev range it should be practical. Regardless of weather or not it's going to use a stepper motor or standard motor you're going to have to have a very good motor controller.

 

[Hank Fletcher]

The closest thing I've been able to find so far is a servo motor from a company in the US: www.midwestmotion.com. With 1000 count per rev encoder attached they quoted me something like $300US, which is a bit out of my range while I'm still experimenting. And it needs a 24V PSU, which I don't have.

But put that aside a minute and presume that I'm using a DC motor. For position feedback I was thinking about a single turn pot on the axle at the opposite side of the wheel from the motor. I guess I need a pretty good pot with a mechanical range of about 350 degrees, and a linear accuracy of 0.14%, then using that to generate a 10-bit digital position reading... is that all unrealistic? I just don't know much about encoders and position feedback. I know roughly what they are, but I don't know how to go about finding what I need. Sorry if I haven't done a good job of explaining this. I guess I wasn't realistic in imagining that it would be fairly inexpensive to get the kind of pot I would need. Wow, that last sentence could really be taken out of context!

I think part of my challenge is that I have a very specific idea of what I want to accomplish, but I'm still learning a lot about how to get there. I'm coming from an intense music background, much more than engineering. After reading the comments about newbie posts on this forum, I guess more members like to get a lot of info out of newcomers rather than vague questions, so here it is...

I'm trying to build the right arm of a trombone robot. The hardest part of building any automated musical instrument is keeping in mind that mechanical noise must be kept to a minimum. The next hardest part is copying motion that humans have had hundreds of years to develop. There's some video on the internet out there from some engineering students at Hosei University who made a trombone robot, but their strategy for the right arm was all wrong.

To keep the weight of the non-trombone parts to a minimum, my design is for a large wheel (or spool, if that helps you visualize this better) with a slightly larger circumference than the operational length of the trombone slide. Lines attached to the handle of the slide will pull the slide in opposing directions. The motion of these lines is redirected and attached to the spool so that as one line wraps, the other unravels. This keeps enough tension in the lines to keep them in place on the mechanism, and also readily available for instantaneous motion transfer from the wheel/spool in either direction. I hope that makes sense, but if not I'll draw you a picture.

I've calculated conservatively for the weight of the slide, its friction, the weight of the wheel/lines/etc, and the force required to move it from top speed in one direction to top speed in the other. The answer is about a 1.4Nm torque requirement. This takes into account that the slide has to be able to move 60cm in a tenth of a second. And it, ideally, has to be able to stop within a millimetre anywhere along that 60cm (okay, I'd be happy with 2mm). This is what it'd take to build the equivalent of a professional trombonist's right arm - I'm well aware that my first prototype might fall short of that goal, but I'd like to get as close to it as I'm able on the first try. Hence asking as many questions as I can before I get started building the thing.

 

[Hank Fletcher]

...you're going to have to have a very good motor controller.

Yep. It wasn't until I started this project that I really began appreciating the beauty and complexity of the human mind/body. The fact that the majority of humans have the capacity for this kind of power and control, and that they can do it virtually silently is an incredible testament to refinement through years of evolution. Or intelligent design, depending on where you're coming from.

I started out looking at stepper motors, but because of their speed limitations, their torque charasterics, and their greater noise, I started leaning more towards DC motors. A DC motor seems more capable to me of copying human motion. I think a stepper motor is obviously more appropriate for a CNC machine, but then no one really cares how noisy, or how long, a CNC takes to get the job done. That's part of the uniqueness of this challenge, I think. I'm all ears to any ideas though. Thanks for your idea about the anti-backlash gears, I haven't searched the net for those yet... here I go!

 

[Sceadwian]

Why so big? Why not mount a fixture on the non-moving portion of the trombone and use much smaller rubber drag wheels mounted to that to pull the slide back and forth? You could use a gutted optical mouse as a linear position encoder for the slide itself, or something similar. The drag of the slide itself will act as your anti backlash, and a decent motor controller on a standard DC motor would give you a very smooth movement. Significantly less bulky and complicated than your current idea as well as being more accurate..

 

[Hero999]

they quoted me something like $300US

That sounds like a bargin to me!
Seriously, I'm not being sarcastic at all, thet's very good value or money considering it's probably only produced in fairly small production runs.

it needs a 24V PSU

Which is a reasonable requirement and they're quite cheap to buy.

 

[Hank Fletcher]

Hey, we're of the same mind! I'd thought about using an optical mouse in the same manner, but then I started looking at the specs for an optical mouse and wasn't convince that it'd be able to keep up. A mouse is usually good for 1m/s, but I'm looking at speeds of 6m/s. Even at a slower speed, if the mouse makes a mistake, how do you catch the mistake?

So suppose you use a multiturn spool and put an encoder on that. Unless you've controlled exactly how the line wraps on the spool, you can't determine how far the slide's been moved based on your reading of how far the spool's turned. Also, with a smaller spool you have to turn it faster than a large one, and if you calculate the torque in either situation, it's the same differece. Granted, spinning a small spool faster is more consistent with the specs you generally find on a DC motor, but like I said, there's the greater complexity of having to keep an eye on how the line wraps on the spool. A large spool with a groove and a line that wraps less than 360 degrees is a much simpler design, isn't it?

 

[Hank Fletcher]

That sounds like a bargin to me!
Seriously, I'm not being sarcastic at all, thet's very good value or money considering it's probably only produced in fairly small production runs.


Which is a reasonable requirement and they're quite cheap to buy.

The more I think about it, the more I think you're right. I e-mailed midwestmotion about the backlash on their servomotors, but I can't remember what their reply was. They have their current stock list with this motor on it if you want to take a look. If they're offering a 1000ppr encoder with it, is it reasonable to assume the backlash is better than that? If it is, then the hardest part of this project is convincing my girlfriend that I ought to spend $300+ on a motor... tax refund season is coming up...

If you're looking at their stock list, the motor I'm writing about is:
MMP-S28-400A-24V BR015 T-007
Hey, come on guys, stop drooling over their other motors and tell me what you think about this one!

In a related question, anyone out there have any thoughts on these 24V supplies?
**broken link removed**

 

[Hero999]

The power supply looks reasonable value but might be a bit expensive once you factor in shipping. I couldn't find any information regarding the motor though; how much power does it use?

 

[3v0]

Hank,
Did you look into linear actuators ?
I have no idea if they would work or not.

 

[Hank Fletcher]

Hank,
Did you look into linear actuators ?
I have no idea if they would work or not.

Yes. Unfortunately, they're nowhere near fast enough if used directly. I haven't done much research on using them in a lever, though. Is that what you're thinking? They have some at midwestmotion, but even then I don't know what you would use for feedback. Maybe a string pot on the slide? Thanks for thinking outside the box!

 

[Hank Fletcher]

The power supply looks reasonable value but might be a bit expensive once you factor in shipping. I couldn't find any information regarding the motor though; how much power does it use?

Mmm... I've got some 24V solenoids here I've been wanting to tinker with, too, so that PSU's looking sweet. They say $20US shipping, and about 30 days, but even then it still looks like a steal for 350W. I've never used ebay before (it looks like the only place sure-electronics exists), but I have a "..for Dummies" book around here someplace.

The midwestmotion site could be a lot easier to navigate. If you can find the link to their current stock it has a brief description of each item's specs. There are also links elsewhere on their site for various product information. I think I'm finding conflicting specs on the servomotor I'm interested in, although they're close so maybe I'm just reading it wrong? The stock list says 24V, 5.9A, and that's for 15 lb in and 1900rpm, which I dare say is quite beastly and arguably well over my requirements. Not that I'm complaining...

 

[3v0]

Yes. Unfortunately, they're nowhere near fast enough if used directly. I haven't done much research on using them in a lever, though. Is that what you're thinking?

I try not too. Hurts too much.

Ideas are cheap. Getting them to work is another matter entirely.
Projects are easy to start but never finish without my attention.
Even when I finish a board it never works till I write software.
Why is that?

Here is another cheap idea.

One thing that moves that fast is a disk drive positioning setup. Not the modern tiny ones but rather one from a huge drive off something like an IBM 370 vintage machine. You will still have to figure out the feedback. Do not know where to find one.

 

[Hank Fletcher]

I try not too. Hurts too much.

Ideas are cheap. Getting them to work is another matter entirely.
Projects are easy to start but never finish without my attention.
Even when I finish a board it never works till I write software.
Why is that?

Here is another cheap idea.

One thing that moves that fast is a disk drive positioning setup. Not the modern tiny ones but rather one from a huge drive off something like an IBM 370 vintage machine. You will still have to figure out the feedback. Do not know where to find one.

From what I can tell, they were about $1,000,000 back in 1970. Have they come down in price since then? And will it fit in my car?

 

[Hero999]

Mmm... I've got some 24V solenoids here I've been wanting to tinker with, too, so that PSU's looking sweet. They say $20US shipping, and about 30 days, but even then it still looks like a steal for 350W.

Sorry I wasn't really thinking about the currency and power rating, it looks like you've got yourself a bargin.

I've never used ebay before (it looks like the only place sure-electronics exists), but I have a "..for Dummies" book around here someplace.

Looks good to me too. My first purchase was a 40 pin universal ZIF socket yesterday which is still in the post.

The stock list says 24V, 5.9A, and that's for 15 lb in and 1900rpm, which I dare say is quite beastly and arguably well over my requirements. Not that I'm complaining...

If that's correct (I would email or phone them to clarify that) then at $300 it a bargin!

 

[3v0]

They get smaller as they get newer. Just a matter of picking the right year/size.

Regarding price. I do not think the drives were quite that expensive. If you look hard enough you might find one for the hauling.

Best of luck on this interesting project.