PWM for High-Torque Motors

The problems usually arise when there is too little inductance, the I2R losses in the driver get excessive.

Yes, you should be able to PWM a starter motor, but the MOSFETs had better be very hefty- hundreds of A honest rating, maybe more.

It should work. It might help to drive the field and armature with separate PWMs. I think classic starters were series motors, optimized for starting and nothing else.

Starter motors are brutes.

Reactance helps!

Yeah, they were wound such that which ever polarity they were connected they would always turn in the same direction. Some had the direction of turn stamped onto them, too.

How does it *help* exactly??

For PWM it smooths the current impulses to a more even level rather than literally on/off.

Yes, I neglected to mention about that. They are optimised for starting and are pretty much useless for anything involving more than around 30s operation at a time, given they have no cooling fins or vents to speak of.

That is also what I was wondering about. What is the motor going to be used for ? As you said unless the appiation is for running it for less than 30 seconds and then giving it a 10 or more minuite cool off time it will burn out in a short period of time.

Motor inductance lowpass filters the ripple current out of PWM, so you don't need separate inductors. At modern switching frequencies, a motor winding should have lots of inductance, but not enough to affect motion dynamics.

What do you want to do with a starter motor?

Depends on the load. It would be fine at reasonable loads.

The brushes might wear out or get hot from friction if run long-term. Starter brushes are big and crude, with strong springs.

A series motor can be weird. The theoretical unloaded RPMs is infinite, which might be a reason to drive the field and armature separately.

The bearings may not be up to spinning fast and long either.

Occasionally a starter drive will hang up, so that the spur gear doesn't get kicked out of the ring gear when the engine starts up.

Makes very entertaining howling screeching noises, for awhile.

Cheers

Phil Hobbs

Brushes are designed for a certain current density (think that is the correct term but not sure). Too little current can be as bad for the brushes as too much as far as brush wear.

I'll email you tomorrow.

For starting a car, you don't need any control except on/off. And, you don't make the motor good at keeping itself cool. So, what's the non-automotive application that benefits from PWM?

Proprietory ideas abound. Why not use your imagination? There are many possible applications for a high-torque motor operating in short bursts. I found mine, now it's over to you.

No reason why not, though they are usually series wound, whereas a separate field might be better from a control pov. Pwm for higher power dc motors has been done for years and things like exercise treadmills often use 1-3HP permanent magnet motors running at >

100 volts. The higher the voltage, the lower the ir losses, all else being equal...

Classically, for a burst one uses a regular-old-motor and a flywheel. An electric clutch doesn't need PWM, and doesn't dim the lights when engaged.

<snip>

High torque short bursts? You need a Marinov motor...

Interesting. I'll take the guy's word for it that it works. He says some things that don't make sense.

"One can see immediately that the baI1- bearing motor has no back tension because there are no magnets, and the magnetic field of the current in the "stator" cannot induce electric tension in the metal of the "rotor".

Thus the firm conclusion is to be drawn that the mechanical energy delivered by the ball-bearing motor is produced from nothing, in a drastic contradiction to the energy conservation law. "

There is "back tension" because of the resistance that creates the heating. I assume he means no inductive counter EMF, but that's not what makes a motor work so who cares? The statement that the motion is produced from "nothing" is absurd and violates no energy laws. Maybe he is speaking creatively, intending to mock what others might say?

This is a very low impedance device. I suppose the impedance can be increased by making the bearing balls smaller while keeping the radius of the race large. At some point this becomes counter productive to the lifespan of a high torque motor. The author also posts some errors. He claims the two motors have equal ohmic resistance, but says the larger motor draws a correspondingly higher current commensurate to the torque. If the current is higher, would not the resistance need to be lower? If he didn't use the same voltage in each case there is no comparison.

Two other statements offered apparently without evidence. --- The driving force is higher for bigger bores, as the curvature of the races is less. The driving force is greater for bigger balls, as their curvature is less.

I don't see this. A lesser curvature essentially lowers the leverage at a given distance from the point of contact. A greater rotation is required to accommodate the expansion. This would perhaps allow a larger speed of rotation, but either with no more force or with less force, but maybe I'm not grasping the full effect. I suppose since the expansion interaction area is greater this could result in a higher force. As the result is not so obvious at first glance, perhaps some measurement is in order. A lot would depend on the rate of cooling of the working area. I wonder if a thermal insulating layer might be of use? I'm also curious as to why box bearing don't exhibit this effect. Maybe there is too much area of contact lowering the electrical resistance and raising the thermal conductivity to a point the effect is nearly absent.

I wonder if there is any practical use for such a motor? I'm guessing it is not very efficient. It might however, be useful at the very low end of motor size where torque is very important to overcome stiction.

I'm familiar with that system but AFAIK it requires the flywheel to be running *all* the time. I know they don't consume much energy to keep 'em running once they're up there at speed, but nevertheless, they do still need to be powered constantly. Plus a flywheel by it's very nature has to be *big* and *heavy* so that's a non-starter for my purposes, I'm afraid.

April the 1st was a month ago. I can't see any basis for this thing working as described.