"Back to Motor School," Feature Article, December 2003

FEATURE FOCUS: Electric Motors

back to motor school

Build a bunch of different motor styles in identical frames for comparison's sake and get ready for hands-on learning.

by Paul Sharke,
Associate Editor

Here's one of those seemingly obvious questions that you'd think would have been settled by now. With no single vendor making all motor types, how do you go about comparing them to find the best one for your design?

If you happen to be asking this question of 42-volt machines and are in the automotive business, Motorsoft Inc. of Lebanon, Ohio, may have your answer—or, at least, the means to produce one. The company makes a kit that contains four breeds of identically sized motors.

According to Motorsoft's president, Jim Hendershot, the four three-phase motors that make up the kit—an ac induction unit, a brushless permanent magnet synchronous type, a switched reluctance design, and a synchronous reluctance model—if grouped alongside a dc brush motor and a stepping motor, would constitute most of the known motor universe. Could it really be that easy—the whole motor game in a take-along six-pack? Read on.

Long an industry staple, "The ac induction motor, until recently, was a fixed speed machine," Hendershot said. Gears stepped their speeds up or down and helped sustain a multibillion-dollar power transmission industry. With the advent of microprocessors and power electronics in the mid-1970s, flux vector drives developed for controlling the speed and torque of the ac induction motor—an advantage that, until then, only dc machines had enjoyed.

Auto suppliers are examining venerable dc motors, such as this one in a fuel pump from Visteon, to find instances of better efficiency and control.

Speed control of the ac machine was a matter of varying the frequency of the ac waveform, something that wasn't practical to do before microprocessors. There were experimenters that played around with variable speed ac motors before that invention, Hendershot said, but nothing developed commercially. Mostly, ac induction motors ran at fixed speeds near 1,800 and 3,600 rpm at the confluence of pole count and line frequency.

Direct current machines used brushes that ran on commutators or slip rings to feed electricity into their moving rotors. Varying speed was simply a matter of controlling voltage going into the machine. For almost a century, the ac machine dominated fixed speed applications and the dc machine ruled the variable speed world. If the cheap, simple, rugged, and reliable ac induction motor had any shortcomings, it was its lack of speed and torque control. Some might say that wasn't really a motor fault—only a long, lonely wait for the appropriate electronics to come along.

But the electronic revolution also profoundly influenced several other motor types that had been awaiting their chance to dance. Stepping, brushless permanent magnet, switched reluctance, and synchronous reluctance motors—previously wallflowers—suddenly looked a lot more appealing in an era of cheapening power electronics and improving permanent magnets.

DC machines using brushes to commutate the incoming direct current had a major disadvantage themselves: The carbon or metalized brushes wore out, created dust and arcs, and were troublesome from a maintenance man's perspective.

Placing permanent magnets on the rotor severed this mechanical link, which the dc machine had needed to power its rotating electromagnets. Electronic commutation dispatched the mechanical switching needed to race the current around the poles.

New Market

By Hendershot's estimate, the average car today leaves the factory with some 50 to 75 motors on board. They are almost entirely brushed dc units—the very same kinds of motors that require so much attention from the maintenance department.

"When was the last time you had to replace a motor in a car?" Hendershot asked. With the exception of a couple of starter motors years back, I couldn't think of a single instance. So much for unreliable motors with brushes.

Stepping motor (top) produces incremental motion without needing feedback, a boon to disc drive makers. AC induction motor (above) can acquire speed and torque control by operating through a flux vector drive.

Actually, the automotive sector is practically committed to switching to 42-volt bus systems, for many reasons. One is that "current costs more than voltage," according to Hendershot. High current has greater potential to damage mechanical elements such as switches. It takes much wire—85 pounds of copper in the average car—to carry high currents. Doubling the voltage halves that current.

Tiny 12-volt alternators are struggling, too, under the demand for more juice. Transistors that drop as much as 1/2 to 1 volt across their junctions will add up quickly to strain the energy available from a 12-volt system.

The move to 42-volt systems will open up cars for new developments, such as active suspension, four-wheel steering, and traction and braking control. It may end up replacing many of those reliable, if dated, dc brush motors.

Great Beginnings

The idea for the kit began with Ned Mohan, a professor of electrical engineering at the University of Minnesota, who had noticed a shrinking enrollment of EE students studying power engineering. Mohan said a workshop he began in 1994 has since led to a new course in electric machines and drives that's being taught in 25 universities. Class size has swollen, too.

The new course, developed with funding from a National Science Foundation grant and help from NASA Glenn Research Center in Cleveland, combines analysis and control of electrical machines in a laboratory setting that includes hands-on learning. Mohan collaborated with Motorsoft to develop a university-lab scaled dynamometer and dc generator kit. The kit includes a handful of motor types of identical physical sizes. Mohan picked 42 volts for them because it was a safe voltage for students to work with and because that was the voltage toward which the auto industry was steering.

Each of the motor set's four machines comes in an identical package, measuring 80 mm over its outside diameter and using a 35 mm stack length.

"The same inverter drives the ac induction motor, the brushless permanent magnet motor, and the synchronous reluctance motor," Hendershot explained, calling the drive "topology" a "standard H bridge with six transistors, each with a diode."

The fourth member of the group, the switched reluctance motor, needs a different drive topology, one that has not been available as a standard bridge. That's one of the things that has kept this motor type away from widespread use, Hendershot said, even though the motor itself ought to be the least costly of the group to manufacture.

The switched reluctance machine uses no permanent magnets or wound coils on its rotor, and relies instead on a magnetic field in the stator alone to produce motion. The brushless permanent magnet machine uses permanent magnets on its rotor, and electrical commutation. The ac induction motor uses wire windings or rotor bars in the rotor, relying on current in the stator windings to induce a magnetic field in the rotor.

A Smart Pump

Auto-industry supplier Visteon Corp. of Dearborn, Mich., introduced a brushless dc fuel pump in 2001. The pump relies on electronics for commutating the rotor, eliminating reliability issues stemming from brush wear. The new pump can double the life of traditional fuel pumps, the company says.

According to Visteon business development manager Dave Frey, the pump was designed to deliver fuel to an engine in closer accordance with demand. The pump measures fuel temperature and pressure, and relays this data to an electronic module responsible for controlling the power train. The module uses this information to optimize engine performance.
Instead of bringing current into rotor windings through brushes to produce torque, control electronics develop and rotate magnetic fields around the stator to move the motor rotor.

A fuel pump suggests a promising place for a switched reluctance motor, Hendershot said, because its major limitation is that it is too loud. Such a motor probably wouldn't work that well inside the passenger compartment operating a heater blower because of the noise. But in the fuel line it might be just fine, he said.

This is exactly the kind of information designers using the Motorsoft kit would be able to determine firsthand, Hendershot said.

By going from a brushed dc motor in a fuel pump assembly to a brushless design (upper right), Visteon engineers doubled motor life and created a pump that responds to engine demand.

Dan Jones, a motion-control consultant based in Thousand Oaks, Calif., always asks, "What's the application?" when someone questions him about picking the best motor.

Stepping motors, for instance, can be noisy and limited at high speeds. But they don't need a position feedback device such as an encoder or a resolver. They're only available in sizes up to a few hundred watts. They're mostly used in automated assembly machines and in packaging equipment, where low-cost, fairly accurate positioning control is desired. But they also move disc drives and similar devices in office automation. They probably won't be showing up in cars any time soon.

Direct current brush motors, besides having brushes, suffer from lower torque density than brushless dc. They're priced comparably with step motors. They are used extensively in sizes up to about a kilowatt, although some industrial applications still use dc brush motors in the hundreds of kilowatts.

Neodymium rare earth magnets make brushless pm motors expensive. But the ability of these machines to pack more poles into a given volume make them able to produce more torque than a 4-pole ac induction motor with the same power. For positioning, they need an encoder or resolver. They can reach very high shaft speeds.

Although simple and rugged in its design, the switched reluctance motor can be noisy. Fixing that requires more complex controls, which increases the price of the overall system.

Finally, there's the ac induction machine, which Hendershot called "the original brushless motor," and an industry darling. Flux vector drives make these machines variable speed.

For any of these machines, cost is an obvious consideration. The dc brush and stepping motors still rank among the lowest-cost machines, but brushless motors are making strides in that direction.

"Flux density keeps going up as cost keeps going down," Jones said. When the patents begin running out on neodymium-iron-boron magnets in the near future, and more offshore sources begin opening up, there should be an accompanying dip in pm motor pricing.

Once you move away from dc brush motors and ac motors running off the line frequency, controls become increasingly elaborate and make up a greater portion of a motor system's cost. For the classically trained mechanical engineer who'd prefer not to delve into electronics at this time, here's a chance to forget everything you ever learned about gears and kinematic linkages while repeating the mantra "change is good."

Thought Control

You can't get much simpler than the controls for a dc brush motor. Commutation, the control of the current that produces the torque, is built in. Speed control comes as a simple potentiometer that varies the voltage to the motor. Add an encoder and a controller, and you get positional control in the form of a dc brush servomotor.

A servomotor, according to George Ellis, a senior scientist at Kollmorgen of Radford, Va., "is any motor where the drive controls position or velocity based on a feedback sensor."

Inside the controller of our dc brush servomotor, for instance, electronics compute the position/velocity error. The drive then generates a current to produce torque that eliminates, or at least reduces, that error.

An ac induction motor will run quite happily without any form of controls whatsoever, other than what's needed to take it safely from zero speed on up to its nameplate rpm. These "controls" are more properly termed motor starters. They are designed to handle the inrush of current that a starting motor demands without tripping thermal overload protection, while still providing that protection if the motor stalls or short circuits.

Add flux vector control to the ac motor circuit, and you have a way of varying the frequency and magnitude of the current going to the motor. This is what determines the speed and torque of an ac induction motor. In the simplest terms, what the flux vector drive does is convert incoming ac line voltage to a dc form, and feed dc pulses to the ac motor in a form that approximates a sine wave and creates a frequency that delivers the desired motor speed, and a current that controls motor torque.

DC brush motor provides precise speed and torque control for machine tool makers, conveyor builders, and even designers of fixed-grip ski lifts.

According to Ellis, ac machines controlled through flux vector drives usually run open loop, with no feedback. It normally wouldn't qualify as a servomotor. The stepping motor is another non-servo variety and, in a sense, the one that started all the commotion.

Jones said that about 80 percent of the motor industry is concerned with controlling speed, while the other 20 percent worries about controlling position. But it's the positioning side that's called the "bleeding edge."

Usually, the stepping motor incorporates many permanent magnets into its rotor, although sometimes it uses no magnets there at all. It also incorporates two pairs of windings in its stator. Its controller "steps" the motor to any desired position by a square waveform that typically advances the rotor by 200 steps for every complete revolution. The wave form can be divided into smaller half-steps or microsteps to produce even finer movement of the rotor. Again, no feedback is used to tell the controller where the rotor is.

Because the other motor types lack the mechanical commutation of the dc brush machine, they have to be commutated electrically, Ellis explained. The drive figures out how to direct current into the motor windings in a pattern that will produce torque. For brushless motors, that usually takes the form of sine-wave or six-step commutation.

According to Ellis, the ac induction, the brushless permanent magnet, and the synchronous reluctance machines "all run on sine waves, more or less," and drives are manufactured that can control all three—as in the case of the Motorsoft kit.

That's not to say that any drive can control all three. Many are specific to one or two types of motors. The one machine this drive can't control is the switched reluctance motor, whose applied voltage "looks nothing like a sine wave," Ellis explained.

The switched reluctance machine is "highly nonlinear," he said. It requires many special techniques for reducing noise and torque perturbations.

ME 101 Cancelled?

It may have just dawned on you that certain portions of the grand field of mechanical engineering may be going the way of the T-square. Fear not.

That's easy for Jones to say. An EE by training, an ME by trade, he's been in the motor business for more than 40 years now. He's seen most of these developments unfurl in real time. For those of us with some catching up to do, he utters but one word: mechatronics.

"Successful mechanical engineers today simply have to be comfortable with electronics," Jones said. "It's become such an integral part of the motion-control world that it can't be ignored; it can't be left for the double E's to take care of, either," he added.

Yet, the new motors and drives practically reek with the scent of opportunity. Think about a mechanical world in which wear affects but two bearings in every motor, a world where gear backlash and shaft windup are but noteworthy relics. Isn't that the stuff of dreams?