Motor City

Martin Eberhard, Co-founder of Tesla Motors, October 18, 2006

A bit about motors, magnets, AC and DC, and weird little widgets called IGBTs, all the while trying to answer some more of your questions. (And yes, we are opening a Detroit office, but that’s another story.)

Nikola Tesla was a strong believer in AC (alternating current) as a means of distributing electricity because it was more efficient than DC (direct current, as favored by Thomas Edison), and because it was easy to step an AC voltage up or down using a transformer made of nothing more than a stack of steel sheets and some coiled-up wire. Plenty has been written on this subject. I recommend the following books.

Empires of Light: Edison, Tesla, Westinghouse, and the Race to Electrify the World by Jill Jonnes

Tesla: Man out of Time, by Margaret Cheney

Wizard: The Life and Times of Nikola Tesla : Biography of a Genius, by Marc J. Seifer

The Tesla Rotating Magnetic Field by Thomas Commerford Martin and Nikola Tesla

The Complete Patents of Nikola Tesla, by Jim Glenn

Rewind all the way to 1888: Nikola Tesla invented the polyphase AC induction motor (US Patent numbers 381968, 381969, 382279, 433700, 433701, and 555190). Tesla was interested in such a motor because it was simpler, and because it could be driven directly from AC transmission lines or from a dynamo without need for rectification. Tesla worked to perfect the AC induction motor, and most of the motors we use in plug-in appliances and equipment are directly derived from his work.

AC Induction motors have several advantages over DC motors: they have no field windings or permanent magnets of any sort; they have no brushes or commutators to wear out; they can be highly efficient. Tesla himself recognized that AC induction motors could work well in cars. However, they were never used in production cars in his time because it was nearly impossible to convert DC from batteries into AC to drive the motor. (The vacuum tube was still pretty neat stuff –silicon was most useful as a beach topping, no decent diodes, no transistors, etc.)

So AC induction motors found their way into industry and into appliances, using AC as generated by dynamos. For the last 100 years they were mostly designed and optimized to work at a fixed frequency: 60 Hertz here in the USA, the frequency of the alternating current in every electrical outlet in our homes. (This frequency -- still used today -- was chosen over a hundred years ago by Tesla.)

(For the less technical, what does AC mean? AC means that the current changes direction smoothly back and forth like waves on the beach – flowing in, then back out. If you could listen to 60 Hertz, it would sound like a low, smooth tone. Higher frequency means higher pitch; those annoying beeps from microwaves and other electronic goodies are around 2,000 Hertz. Your hearing probably tops out around 20,000 Hertz.)

Electric cars used simple-to-control DC motors all the way into the 1990’s, See a long list here.

DC motors don’t need much to make them turn – hook ‘em up to a battery, and they go. Maybe you played with a dc motor yourself in school, in a slot car, whatever. If you take one apart, you will find the same thing whether it is big or small, old or new: a rotor (the part that spins), made up of a bunch of wire wrapped around a frame on a shaft, and a casing that has a couple of magnets attached to it. The wires to the motor attach to “brushes” that ride on a dohicky on the rotor’s shaft, called a commutator. The brushes and the commutator conspire to run current through the coiled wire of the rotor (creating magnetism), and flipping the direction of the current (and therefore the magnetism) back and forth as the rotor spins. There has been plenty of optimization over the years – fancy “rare earth” magnets, better commutators, even tricky elimination of the brushes in (cleverly named) brushless DC motors.

JB warns me:

“Be a bit careful here. The brushless DC motor is actually an AC motor. You need an inverter to drive it! A better name for this type of motor would be: “synchronous permanent-magnet AC motor.” Most people ask why we don’t use a brushless DC permanent magnet motor, like all hybrids today use. We chose not to use one of these mainly because of the wide efficiency plateau that we get from an AC induction motor (which means high efficiency over a wide RPM and power range), not because it is really a DC motor.” (Some of you have asked about the use of niobium or other advanced magnets in our car. If we used a brushless DC motor, we’d have such magnets. But we don’t :-) ) Even today, practically all hobbyists who are making their own electric car use DC motors because control is so easy.

But Nikola Tesla was right: an AC induction motor is inherently more efficient, lighter, simpler, and more reliable than a DC motor – and would make a better electric car if only it were feasible to convert the battery’s DC into AC easily.

Fast forward to 1986. Transistor technology has come along and been refined enough that it is possible to create high-power AC from DC without a lot of wasted energy. Aerovironment was developing the Sunraycer, a solar-powered racer for GM. Al Cocconi (the “A.C.” of AC Propulsion), was working for Aerovironment, at the time. There, he figured out how to gang a whole lot of MOSFETs (fancy transistors) together to make an “inverter” for the Sunraycer so they could use a light weight AC induction motor.

But this was not your ordinary inverter. If you bought an inverter for your car or RV to power household appliances, that inverter would simply make good old 60-Hertz like what you get from your wall outlets. In contrast, Cocconi’s inverter created a variable-frequency AC waveform for the motor. (In the same era, several others created similar variable-frequency inverters for EVs and other applications – decent power transistors changed everything.)

Why variable frequency? A little digression into how an AC motor works:

The AC electricity driving the motor powers the stator (the stationary windings around the spinning rotor) and creates a rotating magnetic field. Tesla’s original motors used 3 “phases” of AC to drive the motor: 3 wires to the motor, each with the same frequency AC, but at a different phase. (Sorry – I can’t think of an easy way to explain this!) Lower-power motors today use a single phase, which (annoyingly) use two wires.

Motor engineers coined a concept called “slip,” which is the difference in rotational speed between this rotating field and the rotational speed of the rotor. The torque of the motor is proportional to the slip. So – if you want a certain amount of torque from an AC motor, you measure the speed of its rotor, and adjust the AC frequency to cause the magnetic field to rotate the right amount faster than the rotor (or slower for regen braking).

An AC induction motor is sometimes called a “squirrel cage motor” because the working part of the rotor looks like one of those cages that pet rodents run around in – a shaft with two metal rings connected together by a bunch of metal bars. (Note: there are generally no wire windings in the rotor of an AC induction motor.) Early on, Tesla figured out that he could fill up the squirrel cage (where the squirrels might run) with a stack of steel laminations to increase the power of the motor.

Tesla mostly used copper to make his squirrel cages, but had a difficult time fabricating them. For this reason, Tesla came to advocate aluminum for the rotor instead of copper, even though this reduced the motor’s efficiency considerably.

As noted above, AC motors designed for appliances usually run at one speed. Some of you have commented that we should use a Continuously Variable Transmission (CVT) to match our motor speed to the desired speed of the car. This would be true if we ran our motor on a fixed frequency.

But we don’t. Like the GM cars, and like other AC electric car motors, we feed the motor with a variable frequency AC waveform, using frequency to regulate torque and therefore speed.

Skip to 1988: The team at Aeroviroment got the GM contract to create the Impact (the precursor to the GM EV-1), so they designed a custom AC induction motor to go with Cocconi’s MOSFET-based variable frequency inverter.

The folks at Hughes/GM didn’t like the large number of MOSFETs that Cocconi used, and proposed instead to use the new-fangled IGBT transistors like those from International Rectifier. You can read about all this in the book The Car That Could. The Hughes/GM engineers liked these better because they were easier to control and many fewer transistors were needed. According to legend, Cocconi was at first resistant to using IGBTs rather than the MOSFETS he already understood. True or not, the EV-1 used IGBTs.

Fast forward again to 1992: Al Cocconi started AC Propulsion to make EV motors and matching inverters. AC Propulsion developed their Power Electronics Unit (PEU) using IGBTs similar to what Cocconi learned about from the guys at Hughes/GM.

Fast forward once more to 2003: Tesla Motors’s Power Electronics Module (PEM), in turn, uses a similar kind of variable frequency, IGBT inverter, based on what we learned from our friends Al Cocconi and the rest of the team at AC Propulsion, as well as from what JB and I had learned in our own careers as electrical engineers. Over the last 3 years, this PEM has been refined and improved by Tesla’s team of electrical, firmware, and manufacturing engineers.

At the same time, Tesla’s motor engineering team developed our own custom 3-phase AC induction motor – based on Tesla’s patents, based on the EV-1 motor, based on the AC Propulsion motor. Like Tesla’s motors, the EV-1 motor, and the AC Propulsion motor, ours gets its incredible efficiency largely due to its copper rotor.

We’ve studied the EV-1 motor carefully. The technique they used to construct their copper rotor was not great, resulting in suboptimal efficiency, and (I suspect) low manufacturing yield.

We have studied AC Propulsion’s rotor manufacturing technique. Their process creates a motor with much better efficiency. But there is quite a bit of hand labor and tweekmanship in the process, and it would not work for the production volumes we forecast at Tesla.

We studied other companies who cast copper rotors like Favi. But their process yielded rotors with lower efficiency than AC Propulsion’s.

The first Tesla Motors prototype motor

So we set out to create our own copper rotor fabrication process. It took us a few years, but it worked: our rotors are readily mass produced in our own factory in Taiwan, and their performance is quite nice. (How we do it is a secret. I don’t keep a lot of things secret from you, but this is some of our secret sauce! That’s why we didn’t outsource the construction of this piece.)

Here’s the cool thing: if you handed one of our motors to Nikola Tesla, he’d recognize it immediately as his own invention. Nice job of optimization, but clearly his.

That’s why we’re Tesla Motors.