Hacking Brushless Drill Motors

[davidhere40]

I've been looking for high torque, low cost motors. And I noticed that brushless drill motors may have a lot of torque for their cost. They cost like $150. And I figure if I open it up, I could find a simple way to control the speed using a very low power, cheap servo. I haven't been able to find plain brushless motors and drivers for them. It's as if nobody uses them directly that way.

Any thoughts on this? Is this a bad idea for an industrial robot design?

Ignore the fact that it could be ugly using a drill directly I'm trying to find ways to cut the cost because cost has historically been a prohibitive factor for my application. I don't need accuracy or repeatability. All I need is torque and speed, and the ability to control speed a bit. I'll guide it with computer vision and a computer model.

[Soeren]

Hi,

I've been looking for high torque, low cost motors. And I noticed that brushless drill motors may have a lot of torque for their cost. They cost like $150.

Brushless drill motors??  Where do you find them?
All the drills I have had over the years and all I see in shops are brushed motors.

Torque is inverse proportional to speed in any motor.
Low cost = brushed motors.

And I figure if I open it up, I could find a simple way to control the speed using a very low power, cheap servo.

Then you'll have a chain with... A servo controller controlling the servo - acting on the speed controller - that in turns controls the speed of the motor - why not throw in some duct tape to further this really slick design

Or... You could make a new speed controller, in case you hope for this to be marketable

I haven't been able to find plain brushless motors and drivers for them. It's as if nobody uses them directly that way.

Only most R/C model flyers an quite a lot R/C cars and boats, all hard disks and CD/DVD drives etc. etc. and quite a number of hobbyists build them from raw stock material.

Any R/C shop, on- and offline, have a range of brushless motors and controllers for them.

You'd need to gear it down and you'd need an outrunner, as they are slower and higher torque than inrunners (they can be found at speeds ranging from around 700 kV and up to several thousand kV (kV = RPM/volt).

Check out these (and for price comparing to brushed motors, remember that you need a controller a well)...
http://www.hobbyking.com/hobbycity/store/__517__59__Brushless_Motors-Outrunners_by_size.html

Any thoughts on this? Is this a bad idea for an industrial robot design?

The "power drill and duct tape design"?... What do you think

Ignore the fact that it could be ugly using a drill directly I'm trying to find ways to cut the cost because cost has historically been a prohibitive factor for my application.

I can ignore that, but I'm not a potential customer!
Well, actually I don't think I can. If someone presented such a thing to me, i don't know if I would be embarrassed on their behalf or if I'd get stomach cramps from laughing - Selling takes creating trust in your product and I clearly wouldn't trust a gizmo straight out of a cartoon.
Not trying to get you down here, just telling you that I think it won't sell unless you put some effort into the appearance of the product.
And make sure it works 110% before you demo it. We have all experienced a salesman trying to cover up something that doesn't work with a "it worked perfectly just yesterday" or similar - they don't sell a lot


A prototype you keep in your garage is one thing. A demo model should be lean, visually appealing and work flawlessly.

I don't think you have much chance of selling anything with a power drill attached - industry won't have much faith in such and it will be blazingly clear that you don't know what you're doing, if you try to market such a franken-bot.
My best advice, grounded in your various posts, is to partner up 50-50 with a local person savvy in both mechanics and electronics (I get the impression that you have the software and vision system competence yourself) - 50% of something is better than 100% of nothing and the synergy from having a partner to bounce ideas around with is gold.

I don't need accuracy or repeatability. All I need is torque and speed, and the ability to control speed a bit. I'll guide it with computer vision and a computer model.

A geared down brushless motor is the smallest and lightest when comparing equal output power, but it needs a controller that is way more expensive than a brushed motor does.
A regular brushed DC motor is much easier/cheaper to control, but will be a little heavier at the same speed and torque.
Brushless are longer lasting, as there are no brushes that wear down and they don't spark, so they don't ignite combustibles.

[davidhere40]

haha. Thanks for the humorous reply You're right.

What do you think would give me the best bang for my buck then for torque and reliability?
-reliability/durability is top priority.
-Accuracy and repeatability is a very low priority.
-Is computer controlled and has access to AC power outlets.
-Speed is only important for the in/out motor. I'm thinking maybe it should be able to accelerate an 8 pound load to 6 inches per second down a flat slider? I realized the other two base motors can be slow because the length of the arm multiplies the motion.



--thanks for the advice on the brushless motors. I guess I just don't understand them yet. I especially don't understand how to convert kV to something useful to me like torque.
--I can cut the weight of the arm in half if I don't balance anything. But i figure higher weight is better than negative forces from gravity working against the motors.
--Cost should not exceed $1000 for 3 motors & controllers (at least 400 oz-in of torque each). In addition, I'll need 3 to 6 servos (brushless for industrial reliability? maybe expensive!) to handle the end effector for grabbing objects.
--In the end, I'm thinking $2000 per arm for industrial parts. It's very expensive. If I need 20 arms for the final product (likely), It will cost me $50k+.
--Brushed motors... I have the impression that these wear out fast when used a lot and require much more maintenance to replace the brushes. In my application they would be used nearly 24/7. 2000 hours is just 83 days of 24/7 use.
--Brushless drills seem to be new. Here is one for $150: http://www.toolbarn.com/milwaukee-2604-20.html?ref=base
--ps: You're right. I'll try to find some robotics help from the university nearby.

[Resilient]

Can you just use a sliding counterweight?

-------Weight-----Pivot-------------load

If you want the load to go up, move the weight away from the pivot and when it get to where you want, move it back to balanced.

Add a brake or something in there to keep it from swinging past the desired location.

[davidhere40]

.http://www.alibaba.com/product-gs/431442359/drill_motor.html

lol. That drill motor you linked to is gigantic! 300 amps. ahaha
I meant more like this: http://www.toolbarn.com/milwaukee-2604-20.html?ref=base
Here is the best google search (google shopping search) I found: http://goo.gl/MJaTM

Can you just use a sliding counterweight?

-------Weight-----Pivot-------------load

I do have a two counterweights, but for balancing. I don't think there would be any benefit to use them for motion. You still would need significant torque to the counterweights against gravity after the weight has fallen and you want to bring the arm back to zero degrees. But then, you'd need additional motors and mechanics for braking.




What about pneumatic linear actuators? I see the disadvantage as low precision. But, I wonder if it might be sufficient for my application.

[Resilient]

Can you just use a sliding counterweight?

-------Weight-----Pivot-------------load

I do have a two counterweights, but for balancing. I don't think there would be any benefit to use them for motion. You still would need significant torque to the counterweights against gravity after the weight has fallen and you want to bring the arm back to zero degrees. But then, you'd need additional motors and mechanics for braking.

It depends how far from level the arm gets. If it generally stayed close to level all you would have to overcome would be the force of friction and the inertia of the counterweight, which, the further it was away from the pivot, the less it would need to weigh. It is probably more complex than you need. But then again, I don't really know what you need.

Just stick one of these on there:
http://www.hobbyking.com/hobbyking/store/__5142__Turnigy_80_100_B_130Kv_Brushless_Outrunner_eq_70_55_.html

And I am sure you can turn whatever it is you need to turn.

Or what about modifying a camera pan and tilt system like this:
http://servocity.com/html/pt-2100_pan__tilt_system.html

Have you calculated how much torque you need to generate? What kind of loads are on the arm? How quick do they need to move and stop?

[davidhere40]

Wow. that looks like a really powerful motor for very little money. It says it requires esc 130A. I looked it up. Seems to be the motor's controller, but maybe meant for wirelessly-controlled RC airplanes? It seems to cost $115 on froogle.com.
If geared down properly, it might be perfect. But, I don't understand the ratings it specifies. Especially because it specifies RPMs per volt. I don't know how to translate that into torque. Especially because I think it operates at high speeds? So I'd need to gear it down a lot? Where would I get the gears? lol. lots of questions. And then I'm not sure how much power that gives me. It says 6500 watts max power! That's an insane amount of power. I have 18kg-cm servos that consume like 24 watts max power. So, I can't even imagine 6500 watts. With the same efficiency and torque as the servo, it would be 360 times more powerful. LOL.

So, yeah I have done estimates on the weight that would have to be moved. I'm estimating that the load will be about 8 pounds (balanced as if on a flat surface) driven by a either a belt drive, or a rack and pinion. It can accelerate slowly, but needs to move at about 6 inches per second at top speed. I'm guessing. I mean, it definitely can't be slow. 2 inches per second would be too slow because it would take 24 seconds to extend the arm 4 feet. If it had to do that for every pick cycle (which is sometimes possible) it would be an extremely slow robot.

In addition, when the arm pick's something up, the weight will increase 0.5 pounds. Also, the items it picks up, may resist with a force of 15 pounds. So, it will have to contract the linear slider with 15 pounds of force at almost 0 RPMs.