I was just watching a mega factory video and wondered why they use an AC motor which requires a power inverter instead of DC which may be powered directly from their DC battery? Introducing an inverter means more cost (weight, controller, etc).

Are there any reasons for that? What are the differences between an AC and DC motor that may have lead to this decision? Also does anyone know what kind of motor is used in other electric cars?

  • 11
    \$\begingroup\$ Tesla explained their choice of AC instead of DC teslamotors.com/blog/induction-versus-dc-brushless-motors \$\endgroup\$ Commented Jan 24, 2014 at 13:26
  • 59
    \$\begingroup\$ If they used a DC motor, then they couldn't very well call the car a "Tesla," could they? \$\endgroup\$
    – user51119
    Commented Aug 11, 2014 at 12:58
  • \$\begingroup\$ @PJ Doland, I know what you did :P \$\endgroup\$
    – Kirk
    Commented Dec 7, 2014 at 6:26
  • 1
    \$\begingroup\$ The reason is cost. The cost of magnets is too high. They can now make cheap car and sell 3x price compared to competitors and maximize profits. You can see that Hyundai Ioniq is almost twice more efficient than Tesla (see EPA site). They rely on the fact that normal people will just consider max distance per charge, features and price. Higher price, the better, so others won't have what you have and you can claim that yours is better based on price and result of marketing efforts of Tesla. Everyone is happy. Apple did same and look how well they are doing... \$\endgroup\$
    – yurtesen
    Commented Jul 23, 2017 at 21:09

10 Answers 10


You're asking about the technical tradeoffs surrounding the selection of a traction motor for an electric vehicle application. Describing the full design tradespace is far beyond what can reasonably be summarized here, but I'll outline the prominent design tradeoffs for such an application.

Because the amount of energy that can be stored chemically (i.e. in a battery) is quite limited, nearly all electric vehicles are designed with efficiency in mind. Most transit application traction motors for automotive applications range between 60kW and 300kW peak power. Ohms law indicates that power losses in cabling, motor windings, and battery interconnects is P=I2R. Thus reducing current in half reduces resistive losses by 4x. As a result most automotive applications run at a nominal DC link voltage between 288 and 360Vnom (there are other reasons for this selection of voltage, too, but let's focus on losses). Supply voltage is relevant in this discussion, as certain motors, like Brush DC, have practical upper limits on supply voltage due to commutator arcing.

Ignoring more exotic motor technologies like switched/variable reluctance, there are three primary categories of electric motors used in automotive applications:

Brush DC motor: mechanically commutated, only a simple DC 'chopper' is required to control torque. While Brush DC motors can have permanent magnets, the size of the magnets for traction applications makes them cost-prohibitive. As a result, most DC traction motors are series- or shunt-wound. In such a configuration, there are windings on both stator and rotor.

Brushless DC motor (BLDC): electronically commutated by inverter, permanent magnets on rotor, windings on stator.

Induction motor: electronically commutated by inverter, induction rotor, windings on stator.

Following are some brash generalizations regarding tradeoffs between the three motor technologies. There are plenty of point examples that will defy these parameters; my goal is only to share what I would consider nominal values for this type of application.

- Efficiency:
Brush DC: Motor:~80%, DC controller: ~94% (passive flyback), NET=75%
BLDC: ~93%, inverter: ~97% (synchronous flyback or hysteretic control), NET=90%
Induction: ~91%: inverter: 97% (synchronous flyback or hysteretic control), NET=88%

- Wear/Service:
Brush DC: Brushes subject to wear; require periodic replacement. Bearings.
BLDC: Bearings (lifetime)
Induction: Bearings (lifetime)

- Specific cost (cost per kW), including inverter
Brush DC: Low - motor and controller are generally inexpensive
BLDC: High - high power permanent magnets are very expensive
Induction: Moderate - inverters add cost, but motor is cheap

- Heat rejection
Brush DC: Windings on rotor make heat removal from both rotor and commutator challenging with high power motors.
BLDC: Windings on stator make heat rejection straightforward. Magnets on rotor have low-moderate eddy current-induced heating
Induction: Windings on stator make stator heat rejection straightforward. Induced currents in rotor can require oil cooling in high power applications (in and out via shaft, not splashed).

- Torque/speed behavior
Brush DC: Theoretically infinite zero speed torque, torque drops with increasing speed. Brush DC automotive applications generally require 3-4 gear ratios to span the full automotive range of grade and top speed. I drove a 24kW DC motor-powered EV for a number of years that could light the tires up from a standstill (but struggled to get to 65 MPH).
BLDC: Constant torque up to base speed, constant power up to max speed. Automotive applications are viable with a single ratio gearbox.
Induction: Constant torque up to base speed, constant power up to max speed. Automotive applications are viable with a single ratio gearbox. Can take hundreds of ms for torque to build after application of current

- Miscellaneous:
Brush DC: At high voltages, commutator arcing can be problematic. Brush DC motors are canonically used in golf cart and forklift (24V or 48V) applications, though newer models are induction due to improved efficiency. Regnerative braking is tricky and requires a more complex speed controller.
BLDC: Magnet cost and assembly challenges (the magnets are VERY powerful) make BLDC motors viable for lower power applications (like the two Prius motor/generators). Regnerative braking comes essentially for free.
Induction: The motor is relatively cheap to make, and power electronics for automotive applications have come down in price significantly over the past 20 years. Regnerative braking comes essentially for free.

Again, this is only a very top-level summary of some of the primary design drivers for motor selection. I've intentionally omitted specific power and specific torque, as those tend to vary much more with the actual implementation.

  • \$\begingroup\$ Wear/Service (BLDC) is the lifetime for the magnets not limited (eg. due to temperature)? \$\endgroup\$
    – jippie
    Commented Feb 17, 2013 at 8:36
  • \$\begingroup\$ @jippie, I've clarified the efficiency estimates; I agree it was unclear as written. As for magnet lifetime, I've never heard of magnet life being an issue in these applications (as long as the motor is never run close to its demagnetizing current), though that doesn't mean that there isn't minor degradation with age. \$\endgroup\$
    – HikeOnPast
    Commented Feb 17, 2013 at 17:46
  • \$\begingroup\$ How hard is regenerative breaking for brush motors? If one drives a brush motor with a full-wave bridge switches between "forward" and "reverse" with a suitably-biased duty cycle, then trying to drive the motor at a speed slower than it's presently turning will regeneratively brake it. Also, I was wondering it it would be practical to make a cross between a BLDC and brush motor by a DC-powered electromagnet in the rotor rather than a permanent magnet? Delivering power through solid rings (not commutated) would seem like it should avoid arcing issues. \$\endgroup\$
    – supercat
    Commented Aug 11, 2014 at 15:06
  • \$\begingroup\$ Just to clarify, "Induction: Constant torque up to base speed, constant power up to max speed" is only with proper control - the motor itself provides a torque which is everything but constant with speed; proportional with slip from the synchronous speed in the region of interest. ece.umn.edu/users/riaz/animations/vf2.jpg \$\endgroup\$ Commented Aug 11, 2014 at 22:43
  • \$\begingroup\$ @MisterMystère: Off-topic: Great nick! \$\endgroup\$
    – sergiol
    Commented Feb 8, 2015 at 0:13

...and now why Tesla uses induction motors

The other answers are excellent and get at the technical reasons. Having followed Tesla and the EV market in general for many years, I'd like to actually answer your question as why Tesla uses induction motors.


Elon Musk (cofounder of Tesla) comes from Silicon Valley (SV) thinking, where "move fast and break things" is the mantra. When he cashed out of PayPal for several hundred million, he decided to tackle (space exploration and) electric vehicles. In SV-land, time/speed to get things done is everything, so he went looking around to find something he could use as a starting point to get a jump start.

JB Straubel was a like minded engineer (both space and EV) who reached out to Musk shortly after Musk made his interest in space and EV public.

During their first lunch meeting, Straubel mentioned a company called AC Propulsion that had developed a prototype electric sports car using a kit car frame. Already in its second-generation, it had recently switched to using lithium-ion batteries, had a range of 250 miles, offered lots of torque, could go 0-60 in under 4 seconds, but, most germane to this discussion, used -- you guessed it -- AC Propulsion (induction motor).

Musk visited AC Propulsion and came away very impressed. He tried for a few months to convince AC Propulsion to commercialize the electric vehicle, but they had no interest in doing so at that time.

Tom Gage, the president of AC Propulsion, suggested that Musk join forces with another suitor consisting of Martin Eberhard, Marc Tarpenning, and Ian Wright. They agreed to merge their efforts, with Musk becoming chairman and overall head of product design, Eberhard becoming CEO, and Straubel becoming CTO of the new company which they named "Tesla Motors."

The Answer

So there you have it, Tesla uses induction mostly because the first viable prototype that Musk saw used it. Inertia (no pun intended... ok, a little) explains the rest ("If it ain't broke...").

Now as to why AC Propulsion used it in their Tzero prototype, see the other answers... ;-)

If you want the full story go here or here.

  • 2
    \$\begingroup\$ Your link is a very Musk-centric view of the roots of Tesla. AC Propulsion was actively shopping for a partner to commercialize the T-Zero, and Tesla was the net end result: en.wikipedia.org/wiki/Tesla_Motors#History_and_financing. \$\endgroup\$
    – HikeOnPast
    Commented Feb 17, 2013 at 21:12
  • \$\begingroup\$ Yes. Musk, originally, just wanted to invest in ACP and have them build the car, but Gage had no interest in turning his company into a major Car OEM. I incorporated your reference and expanded that section to clarify. \$\endgroup\$ Commented Feb 18, 2013 at 3:52
  • 1
    \$\begingroup\$ Not to mention that the copper squirrel cage cost does not fluctuate as much as Neo magnets (source almost solely by China). Plus induction machine parameters do not tend to change over time as much as magnet properties. \$\endgroup\$ Commented Nov 5, 2014 at 20:58

It's hard to say what the engineers' exact reasons were without being on the design team, but here are a few thoughts:

  1. Both motors require similar drives. Brushed DC motors can run directly off a battery but the type of motor you are looking at in an electric vehicle is a brushless DC motor. The drives for an induction motor and a brushless DC motor are very similar. The control of an induction motor is probably more complex in general.

  2. DC brushless motors have magnets in the rotor. This is more costly than an induction rotor with copper. Additionally, the magnet market is very volatile. On the other hand, an induction motor will have a lot more heat produced in the rotor due to I²R losses and core losses.

  3. Starting torque on brushless motor is generally higher than on induction motors.

  4. Peak efficiency of brushless is generally higher than induction motors but I believe I read somewhere that Tesla gets a higher average efficiency with their induction motor than they would with a brushless. Unfortunately I can't recall where I read that, though.

  5. A lot of people are researching switched reluctance machines now. The last few motor conferences I've been to have been all about switched reluctance. They don't require magnets and the efficiency on these types of motors looks promising. Everybody wants to get away from magnets in motors.

So, as I said, I doubt anybody could answer your question except for the engineers at Tesla. But my best guess is that it probably has something to do with my point 4) but I don't know that for sure. I'm sure the volatility of magnet prices has something to do with it too.

  • 21
    \$\begingroup\$ Everyone wants to get away from magnets ... ironically. :) \$\endgroup\$
    – Kaz
    Commented Feb 17, 2013 at 6:03
  • 5
    \$\begingroup\$ I would not underestimate the effect of #2. A sudden spike in rare earth prices could play havoc with the manufacturing costs. \$\endgroup\$
    – JustJeff
    Commented Feb 17, 2013 at 19:36
  • 2
    \$\begingroup\$ fwiw, I think the control of an induction motor is probably simpler than DC brushless; with the latter you have to have some way to know the orientation of the rotor, so you can align your field accordingly, whereas with the induction motor, all that matters is the speed you are rotating the field at relative to the speed the the rotor is turning. \$\endgroup\$
    – JustJeff
    Commented Feb 17, 2013 at 19:37
  • 1
    \$\begingroup\$ Well, I was talking about vector control of induction motors rather than a simple V/Hz control. Tesla would need to use the former, rather than the latter. \$\endgroup\$
    – Eric
    Commented Feb 17, 2013 at 20:25
  • 1
    \$\begingroup\$ Considering that $10 brushless speed controls for RC aircraft manage to sense rotor position from the windings, I don't think it's an issue for a luxury car. \$\endgroup\$ Commented Feb 18, 2013 at 14:43

The answer comes from Tesla staff people themselves on the article Induction Versus DC Brushless Motors

This part is particularly notable:

In an ideal brushless drive, the strength of the magnetic field produced by the permanent magnets would be adjustable. When maximum torque is required, especially at low speeds, the magnetic field strength (B) should be maximum – so that inverter and motor currents are maintained at their lowest possible values. This minimizes the I² R (current² resistance) losses and thereby optimizes efficiency. Likewise, when torque levels are low, the B field should be reduced such that eddy and hysteresis losses due to B are also reduced. Ideally, B should be adjusted such that the sum of the eddy, hysteresis, and I² losses is minimized. Unfortunately, there is no easy way of changing B with permanent magnets.

In contrast, induction machines have no magnets and B fields are “adjustable,” since B is proportionate to V/f (voltage to frequency). This means that at light loads the inverter can reduce voltage such that magnetic losses are reduced and efficiency is maximized. Thus, the induction machine when operated with a smart inverter has an advantage over a DC brushless machine – magnetic and conduction losses can be traded such that efficiency is optimized. This advantage becomes increasingly important as performance is increased. With DC brushless, as machine size grows, the magnetic losses increase proportionately and part load efficiency drops. With induction, as machine size grows, losses do not necessarily grow. Thus, induction drives may be the favored approach where high-performance is desired; peak efficiency will be a little less than with DC brushless, but average efficiency may actually be better.

Permanent magnets are expensive – something like $50 per kilogram. Permanent magnet (PM) rotors are also difficult to handle due to very large forces that come into play when anything ferromagnetic gets close to them. This means that induction motors will likely retain a cost advantage over PM machines. Also, due to the field weakening capabilities of induction machines, inverter ratings and costs appear to be lower, especially for high performance drives. Since spinning induction machines produce little or no voltage when de-excited, they are easier to protect.


ALL rotary electric motors are AC motors. Every one of them.
Also, at heart they are essentially doing the same thing. The difference is how the DC gets turned into AC and how it gets used to then produce a standard result.

The only motor that is electronically DC is the brush motor. The DC is turned into AC by the rotating commutator and fixed brushes. Apart from that motor, all others are going to need some form of DC to AC conversion. The brush motor is generally unattractive becuase the mechanical DC to AC changer (commutator) is relatively expensive and relatively short lived.

So, for a Tesla or other electric vehicle the choice is not DC or AC, but, what form of AC motor best meets the design aims cost effectively.

The Tesla will use what it does because it met the design goals most cost effectively.

The downvotes suggest that a number of people agree with Marcus and think that the above answer is nitpicking. A little thought and a look at my answers in general may suggest a lack of understanding on the downvoters part.

All rotary electric motors are AC motors

  • If you think this point is nitpicking then you need to think harder about what an electric car does overall.

Let's see if the downvoters have the guts to read the following and then remove their downvotes. For myself it matters not. To the extent that you mislead other people it matters much.

ALL rotary electric motors require a controller to apply AC to the motor in some manner.
The distinction between AC motor and DC motor is useful in some contexts but in an automobile that is a closed system that starts with a DC energy source and ends with a rotary electric motor the distinction is false and not useful. The car is a closed system. Somewhere in the system there is a controller that converts the DC to AC in some form. It matters not whether it is mounted inside the rotor stator or rotor, inside the motor shell, attached to the shell or somewhere else in the car.

In a brushed "DC" motor the "controller" is a mechanical switch mounted on the end of the motor shaft. This controller is \named a commutator but it is functionally a controller that takes DC and creates a chase it's tail AC magnetic field as far as windings in the motor are concerned.

A permanent magnet rotor wound stator "Brushless DC motor" is very similar functionally to a brushed DC motor, with the commutator being replaced by electronic switches and sensors which take the supplied DC and apply it to various fields so that they can chase their tail as the rotor turns. Again it's an AC motor with a controller. Just ask any winding. The sensors are within the motor proper and the switches may be adjacent to the motor proper or remote.

A squirrel cage induction motor adds a degree of complexity by using the rotation of a nest of low impedance windings inside the stator field to induce voltage in the rotor bars and to make a magnetic field which rotates the rotor so that it chases the rotating AC field applied to the stator windings. Again, it has monodirectional (but sinusoidally varying) DC during any portion of the drive sequence. It is as much a mixed DC and AC system as any other.

One could reluctantly describe variable eddy current drive motors - more of the same but different. It's an AC motor with a controller producing it from DC.

The distinction being made is irrelevant and trivial. The real question is "why does Tesla use this particular form of motor rather than some other one". That this is not just semantics but a lack of understanding is shown by the wordin

  • ... which require the power inveter, instead of DC which more direct from they DC battery. Introducing inveter mean more cost (weight, controller, ect.) ...

The only "DC" motor which does not require some form of inverter or electronic switching system is the mechanical brushed motor. These are so unsuited to the task of light weight variable speed drives that there will be few if any used in modern electric car designs. ALL other styles of electric motor which have no inverter will have some eletronics in lieu of an inverter.

I said ROTARY" electric motors are AC motors because one can arguably produce a brushless DC motor linear motor with switched DC only operation, although this would make inefficient use of the copper and magnetics. You could do that with a rotary motor but no real world motor in volume production would do so.

  • 8
    \$\begingroup\$ This doesn't answer the question, it just nitpicks the term DC motor, which everyone knows is has raw DC input, mechanically commutated, then alternating current through windings. \$\endgroup\$
    – Macke
    Commented Feb 17, 2013 at 15:19
  • 3
    \$\begingroup\$ (I coulnd't downvote you, too low rep.) Yes, you're right in that "all rotaty motors require commutation" and had you explained that "connecting a non-commutated DC motor (should one exist) directly to a battery, it will move at most 90 degrees." and following up with "the commutation required by a 'DC' motor is mechanical, and inefficient" I'd be fine. However, being snarky about DC vs AC when we all know what the OP asked about doesn't earn you upvotes. (Being snarky in the follow up doesn't help either). However, seeing beyond the question and explain the greater scheme will. \$\endgroup\$
    – Macke
    Commented Feb 18, 2013 at 12:01
  • 2
    \$\begingroup\$ Actually, all electric motors are NOT DC. However, +1 anyway because your point needed to be made and people were being too loose with the terms AC and DC applied to motors. Pure DC motors are possible and have been made, including the very first one ever. Nothing in the physics requires the current to go back and forth. \$\endgroup\$ Commented Feb 18, 2013 at 12:45
  • 2
    \$\begingroup\$ I like @Russell and his points are all correct, but his application is wrong. A "DC" motor does not mean there is only DC inside the motor. It means that at the system boundary you supply DC ("the drive"). According to this faulty logic, there is no such thing as a DC-DC up-converter because there is no way to maintain pure continuous DC and produce an output greater than the input. The internals of a device are not germane to its nomenclature in many systems including electric motors. Confusing students/posters with correct, yet irrelevant, physics obfuscates understanding. \$\endgroup\$ Commented Feb 18, 2013 at 13:28
  • 2
    \$\begingroup\$ @Russell: No, I meant what I said. Motors in common use may be all AC inside, but it is possible to make a motor (something that causes rotary motion for as long as suitable voltage/current is applied) with only DC voltages and current, both inside and outside. Look up how the very first motor was made by Faraday (maybe Davies?). No AC anywhere. Quite ingenious actually. Later we found other ways that were more effective, and those use AC, but the physics does not require it. Some "disk" tachometers have been made on this priciple, and would work backwards as motors. \$\endgroup\$ Commented Feb 19, 2013 at 19:46

DC motors can't match the power density of Ac machines. The maximum field strength even the best magnets can achieve is 2.5 tesla across the air gap and in order to do this require some serious engineering, particularly if you want then to rotate fast so your power density is high. Induction machines quite comfortably produce 3+ tesla without all the grief of magnets and silly tolerances. They obviously don't do this as efficiently DC machines but who said sports cars where efficient flat out? Kg for kg the AC induction machine is the most powerful of all machine types when controlled buy a sophisticated inverter and running at high rotational speeds.

  • \$\begingroup\$ I am a newbie to this matter, but you seem you are contradicting yourself. "DC motors can't match the power density of Ac machines" and "They obviously don't do this as efficiently DC machines" seem to be contradictory statements to me. \$\endgroup\$
    – sergiol
    Commented Apr 4, 2015 at 0:50
  • \$\begingroup\$ @sergiol -- It's not... and this is a common result -- even the human body behaves this way. Most systems are not their most efficient at their maximums. For example, you can run much further at a jogging pace than at your maximum sprinting speed. Usain Bolt can sprint faster than you, but he uses more energy per kg of body mass than you do to do it. \$\endgroup\$ Commented Jul 21, 2015 at 11:18

The real reasons why they use induction motors for their cars are:

  1. induction motors are cheaper
  2. induction motors don't need a lot of maintenance (no brushes)
  3. induction motors are lighter in weight
  4. the new technology to control the speed of induction motors are now available (variable voltage, variable frequency) and easy to mass produce

IMHO, AC Propulsion (Tesla Motors) uses AC because a mechanically commutated DC motor that meets the high "turn down" ratio of a vehicle application is more complex than an electronically commutated AC motor. Without that high turndown ratio the physical size of the motor producing just raw torque would be prohibitive. The induction motor rather than the PM motor is not only more financially stable, but also more stable from a engineering viewpoint. Magnets can and do get damaged. The electromagnet field coils in the rotor, not so much and as they demonstrate, the energy density is similar.

I take great exception to the apparent consensus that "All Electric Motors are AC" and I base my argument on a single pole move, not the full revolution the motor.

Within a single pole move the only time AC is truly required is when it is necessary to induce a current flow in a parasitic winding, as in the rotor of induction motors. Otherwise, only commutation is necessary.

This argument can best be shown by observing a motor at stall. Only motors without PM or wound fields, which are induction motors, need AC to generate the field current which creates the reactive magnetic field.

All other motors only need to provide DC to the stator to generate full torque at stall. Wound field motors often use AC to generate the field but will also do just fine with DC, probably with even more torque than when on AC.

My PM "servo" motors may be chopping the DC to control power but they are only chopping the DC, not inverting it with every chop. Put a mechanical commutator on the AC PM servo motor and it will work on DC. True, not as efficient but not because of the lack of a sinusoidal waveform. It will also be limited in top speed without a mechanical brush advancer.

Spend some time considering the stall properties of a doubly wound motor, an obviously "AC only" motor, when supplied with DC and maybe you will be able to understand my argument. Only when you want to push each pole in addition to pulling it do you have to provide AC, otherwise DC is all you need and often all you are using, even if the power supply is AC.



All: Brushed machines are limited to perhaps 48V to avoid arcing. In contrast, a brushless machine can easily run from a 240V battery, with voltage boosted to 480V or higher by a DC boost converter positioned between battery and motor. With such high voltage, similar to that used in most of today's hybrid or plug-in cars, the losses of the speed control are minimized in relation to the total power transferred, thus promoting high efficiency.

  • 1
    \$\begingroup\$ I feel this is overly vague, as i have worked on quite a few 270 VDC rated brushed, mechanically commutated machines. Granted all of them were larger than a car by themselves. Be more specific, maybe? \$\endgroup\$
    – user39962
    Commented Feb 12, 2015 at 23:31

Actually, Tesla use synchronous electric motors, which uses both AC and DC. If the motor used just AC it would be a asynchronous induction motor, which is a unpredictable motor to use in vehicles due the slipping in the electromagnetic field when a voltage is induced in the rotor (The output-speed is slower than the rotation of the electromagnetic field. Formula: Revolutions per minute=Frequency * 60 / Pole-pairs per phase - Slip in speed).

In a synchronous motor has a AC magnified stator coil (like a conventional induction motor), but it also has a DC magnified rotor (unlike a induction motor). By doing this the output-speed can reach the theoretical max speed (the sunchronous speed) which makes a predictable and effective motor to use in vehicles. (Formula: Revolutions per minute=Frequency * 60 / Pole-pairs per phase).

Tesla can then explote this and use an ESC (Electronic Speed Controller). An ESC is a circuit board that inverts som of the DC power from the battery to AC power, changes the square-waves to sinus-waves, changes the frequency and amplitude in line with the signals from the gas pedal, and sends the processed power to the stator. It also changes the amplitude of the DC power to the rotor in line with the AC power to the stator.

  • \$\begingroup\$ Source of your first phrase, please? \$\endgroup\$
    – sergiol
    Commented Feb 8, 2015 at 0:17
  • 1
    \$\begingroup\$ Tesla uses a (relatively) conventional asynchronous induction motor. The inverters are pretty much identical to those required for a permanent magnet motor (BLDC). The control is a bit different but not dramatically so. They do indeed have slip. \$\endgroup\$ Commented May 8, 2015 at 19:50

Not the answer you're looking for? Browse other questions tagged or ask your own question.