Electric motor power increase/cooling question

[eKretz]

OK, I have a project I'm working on, where I need to increase the available stall torque of some existing motors, it's important not to change the motor size if at all possible. The motors are already having some issues with heating up and losing power after about 30 minutes of continuous use. They are DC brushed motors, (ball-bearinged) of the 550 variety and run at 24V, about 4000 RPM no-load, and about 4 amps at stall. Most of the time they are running near stall, and work in pairs. They are already running a pretty minimal air-gap. Will switching to neo magnets increase torque without too much current increase? What about switching to a larger motor (If I can convince them) and running a limiting resistor? Would I get more torque than the smaller motors? I'm sure this would help combat the heat/power loss issue since there would be less current running through larger gauge wire, wouldn't it? I'm hoping someone has some insight so I don't have to go through a whole bunch of experimentation/prototyping.

[qno]

I think it is not completely correct.

Torque is related to the current in the motor.

When you say the motors are running near stall do you mean that the rpm of the rotor is very low?

What kind of motor is it ?
What is the application?

[Kremmen]

Torque of a DC motor is directly proportional to both armature current and field strength. A permanent magnet motor torque would be increased by changing to stronger magnets. At the same time the corresponding RPM would go down.
Changing the magnets of a finished motor is not practical however, unless one has a well equipped machine shop at one's disposal. I suggest other solutions.
Please describe the application in more detail, and also try to specify the amount of increase needed. That will make it easier to suggest improvements.

[strangelovemd12]

In practice I  think motors like this tend to get replaced or air/water cooled rather than internally upgraded.  As Kremmen said, it is a real pain to do anything to a factory wound motor without a pretty nice shop setup, and in some cases impossible even with that.  Any chance of getting clued in as to the application?

[eKretz]

Well, I have a well-equipped machine shop at my disposal, so no problem at all there. (I worked as a machinist for about 20 years until Sept. 2011, at which point my back was getting so bad it was hard for me to make it to work much at all, let alone try to bend, lift, twist, etc. as a machinist must). The thing I'm trying to do is develop some more reliable and stronger motors for a sim racing wheel with force feedback so I can help out the others with this wheel and .  The motors are only turning when the steering is being turned, so quite a huge percentage of their on time is spent at low RPM/high load or stall. The stock wheels have a current limit with their stock power supplies of about 5 amps. I've got some aftermarket motors in a wheel now, and they work great but need cooling to prevent power loss. If I could at minimum limit this power loss without elaborate cooling I would be at least happy with that. We're trying to keep the torque level equal or higher than stock, but switching to a larger motor (with its larger shaft diameter) will entail a lot of extra work and/or parts...hence cost. More power draw would be no biggie, as people could just buy a better power supply, but then we have the heat issue again, and even worse. A small RPM drop wouldn't kill the application, as long as it isn't too drastic.  My first idea was the neo magnets, then I wondered about saying damn the torpedoes and switching to the bigger motor but limiting it to produce less heat loss. 

Right now, the stock motors heat up and weaken considerably after only 20 minutes or so. The stock cooling setup is aluminum fins wrapped around the motors' flux rings and airflow through the fins using (2) 80mm fans. This is not adequate, and there seem to be a lot of motor failures. The aftermarket motors I tried are a good bit stronger, but draw more current and also have the heat problem, since pretty much no motor is made to stand up to almost continuous stall conditions. However, they aren't failing and have been in service nearly 3 times as long as the stock ones.

[qno]

Have you considered a geared motor

[eKretz]

I'm worried that a geared motor would give too much resistance when turning, which would make quick corrections difficult and also maybe add too much inertia/flywheel effect to the wheel. Don't have any to try out ATM.

[NiHaoMike]

Connect it via a spring or viscous coupling.

[eKretz]

It really needs to be a direct connection or it doesn't feel right at all. I had a slipping belt once and I felt it immediately.

[G7PSK]

You also get more torque for the same or smaller current by having a larger diameter on the armature.

[Psi]

Can you not add some vent holes and pump air through to cool them down.
Then you can push a little more current through.

The RS550 is a pretty old motor
You would be better off using a brushless RC motor with position feedback however it would require a custom controller for your application.

[eKretz]

You also get more torque for the same or smaller current by having a larger diameter on the armature.

This is what I was thinking with the larger motor but current limited idea. I thought it might keep the heat down also. I am running ducts between my motors with vented cans now, and it works, but is too complex and expensive to do en masse I think. Brushless would be great but the cost would be too much I think, with the controller and two motors per wheel.

[C]

Have you looked for a DC Torque motor
http://www.moog.com/literature/MCG/brushtorquemo.pdf] [url]http://www.moog.com/literature/MCG/brushtorquemo.pdf[/url]

Difference from normal dc motor.
Note a lot of poles are used.
A lot more commutator segments.

C

[Psi]

brushless motors are pretty cheap and already use rare-earth magnets, check out hobbyking.

But the typical brushless ESC requires fast rotation to detect the position from the back EMF and fire the right coils at the right time.
Since your application needs torque at near 0 RPM you would need either an encoder or continuous rotation pot on the motor or some HALL sensors in the armature (its static) so that the ESC could detect position.
I think it would solve your issue though, much more power at low RPM due to being 3phase and rare earth magnets.

Some washing machines use 3phase brushless outrunner motors to get super high torque at low RPM (they drive the drum rotation directly, no gearing at all)

http://www.flickr.com/photos/58137560@N08/6919756494/#

[eKretz]

@C: The torque motors are interesting. It might be possible to replace the twin motors with a single one of those. The motors I'm currently running are 7-pole helical slotted motors, so they are decent, but getting too hot. The stock motors were 4-pole straight slotted.


@Psi: I'm not real sure how I could piggyback off of the original controller to run a brushless ESC. The current setup is using a Microchip PIC33 controller but I have no way of accessing or modifying the firmware. Replacing the original controller would probably entail so much work I might as well just design a whole new FFB wheel, it would seem. Can you give me any direction or point me to a resource to help me learn more about trying this?

[C]

You may be reading "POLE" wrong from that link. A 20 pole motor in that link would have 20 magnets in the field.

A normal DC motor's commutator is normally much smaller in diameter than the armature, A DC torque motor's is the same or all most the same diameter as the armature with the smallest segment possible. This allows the electrical magnet field of the armature to be rotated a very small amount due to the very high segment count. Add to this the high pole count and you have very fine control of the magnetic fields.

This from memory, cobwebs here.
Had cause to replace the armature of a torque motor over 30 years ago that was eating brushes. This motor was the hub motor of a 14" diameter 1" wide mag tape reel, containing 9000 feet of tape. It was moving some weight. Even with this extreme motor in the hubs, the tape still had to be ramped up to speed. I think someone said once the tape deck was 200k$, so this motor may have been the class act for the time.


 The armature looked like a 4" Toroidal transformer.  The armature segments were thin, maybe 3/32", so a small number of wire wraps between each segment. I think a part of each  brush segment fit inside the Toroid. Was potted in something clear. A large number of brushes were used.   


I think a big part of the trick is to make a copy this motor is the armature brush segments. If you can make these segments in your machine shop and a normal toroid would work, I can see someone making one.   


C

[eKretz]

Yeah I definitely hadn't interpreted that as having 20 magnet poles. Very interesting. The 7-pole motor I have has 7 comm segments but only the 2 permanent magnets.  You have definitely given me something interesting to contemplate. Thanks! Got any pictures similar to the motor you are describing?

[megajocke]

What is the current belt/gear ratio? I got the impression you are using a 1:1 connection. Possible torque (limited by heating) of an electric motor is roughly proportional to its volume given a certain cooling arrangement and flux density. What's the rated torque of your motor and how much do you need?

Having lots of commutator segments might be beneficial to reduce torque ripple but doesn't make the motor stronger really.

Have you considered that if the gear ratio is increased by 2:1, the required torque (and hence current) required for a certain feedback force will halve which would decrease the power loss to 1/4 of its original value?  Overdriving the motor voltage/speed-wise if needed is probably less of a problem than current/torque wise. But I don't really see why you'd need a force feedback steering wheel to spin at 4000 rpm.

[C]

That tape drive was pushing a lot of limits compared to a high end commercial drive of the time.

Sorry no  pictures.

Think of a normal auto transformer that is rotating and acting like an armature with A continuous wire wrap instead of bringing out taps for the line connection.
With two brushes, you get a N pole on one side and a S pole on the other side. This would work with your 2 permanent magnets. Just changing the brush setup to use 20 brushes would allow 20 magnet poles in the field. I do remember that half of the brushes were mounted on a insulator ring and it used two brush rings. Your first thought would be the brushes were way to small to handle the current, but all the brushes for a ring were in parallel.

A different way to describe it is Rotary Converter that has no slip rings and a bunch of brush pairs.
http://en.wikipedia.org/wiki/Rotary_converter

A very high pole count makes the motor act more like a stepper motor while the high segment count and construction removes the detent action of a stepper

For what the motor did and was, it really looked too simple. 

If you are really interested, I think I could describe a commutator segment that a machinist could create and would work. The same commutator could also be used for a self built auto transformer.

C


 
 

[eKretz]

What is the current belt/gear ratio? I got the impression you are using a 1:1 connection. Possible torque (limited by heating) of an electric motor is roughly proportional to its volume given a certain cooling arrangement and flux density. What's the rated torque of your motor and how much do you need?

Having lots of commutator segments might be beneficial to reduce torque ripple but doesn't make the motor stronger really.

Have you considered that if the gear ratio is increased by 2:1, the required torque (and hence current) required for a certain feedback force will halve which would decrease the power loss to 1/4 of its original value?  Overdriving the motor voltage/speed-wise if needed is probably less of a problem than current/torque wise. But I don't really see why you'd need a force feedback steering wheel to spin at 4000 rpm.

The reduction ratio from the motors to the wheel is somewhere in the neighborhood of 125:1 through a double pulley reduction. So the wheel at max speed when freewheeling is only hitting maybe 250-300 RPM. The pulleys are already pretty much as far as I can go with respect to diameter, much smaller on the smallest pulleys and the belt will slip; much larger on the bigger ones and they won't fit in the housing. Aside from that, stall is stall and will draw the same amperage regardless of the ratio, won't it?

@C: I would be interested in seeing a pencil sketch of your idea  if you  could draw one and post a photo of it.

[C]

First drawing
http://en.wikipedia.org/wiki/File:Commutator_-_Sectional_View.png

Note that in this pic you have a way to clamp all the commutator segments together and a way to mount it to a shaft.

Note the black slots on the left side for the wires to be placed and soldered. This part will not work well.

second drawing

outside of commutator looking towards shaft center.

To fix this slot problem, drill two holes a little larger then the wire from the outside to the shaft center through the segment where the slot is in the first picture. Then use a saw to cut from one side of the segment removing the connecting medal between the two holes.

If my drawing uploads properly, the gray is the two drill holes and the blue is the saw cut. The red is the copper wire.

By using the saw slot, you can slip the wire in the slot and keep wrapping the windings with out cutting the wire. By pulling the wire tight you are also pulling the segment tight against the core.
Note that I expect the outside of the core to be very close spaced wire raps. This means that the inside of the core will have more layers of wire.

In the first photo to the left of the wire slot is, a fin needs to be added. The fin is less than the thickness of the wire. The fin helps the torque transfer to the communatator while holding some distance for the additional wraps not through the hole.

Third drawing

Looking at the side of a segment.

Note that the fin extends a small distance into the ID of the core to help center the core and the commutator.
note also the segment clamp area in the first drawing. The left side needs to fit through the ID of the core after the core has the wire wrapped.

One segment piece with out the saw cut serves as the start & ending point for the continuous wrapped wire.
Potting with the correct material will add strength.

I would think that the thinest width of segment with insulator would be three wire diameters.
 
I have probably missed something, but this may give you an idea.

I would guess the human labor involved is the reason these are not common and also the high cost of an auto transformer.

C