Author Options:

Overpriced Stuff: Electric Motor Speed Controllers Answered

Commercially, speed controllers for electric motors (electric car size) are very expensive, like on here.

That page has a controller capable of 144v at 550a peak (about 90 horsepower peak) for $1250. I just designed a speed controller with parts off of digikey.com for about $200 that could handle 144v at 690a continuous so it could probably get over 180 horsepower peak.

All things considered, their's costs over 10 times as much per amount of power.

This seems like a BIG markup. I can understand shipping costs, labor costs, etcetera, but something seems a little messed up here.


I am not sure that the design is quite right, and haven't done it myself, so criticism is good.


I've been thinking about if a 12v to 230v (DC to AC) inverter could be modified. There should be an oscillator somewhere in there and possibly it could be replaced by a variable one for speedcontroll of a motor. Just an idea.

I really don't know much about AC...

Oki .. basically AC is current that alternates (hence Alternating Current) between positive and negative ,actually switching polarity, in a low frequency, usually about 50 hertz.
I'm not sure at all wether the idea would work. I just figured that an inverter already change 12 volt DC to 230 volts AC (or 110 volts) at up to 3 kilowatts. An ACmotor, such as a vacuumcleaner, uses those 50hz to alternate and rotate, instead of with a mechanical commutator. If I'm right this is done with an oscillator somewhere at a low voltage, exchange/modify that for a squarewave oscillator that can be controlled and you would have a PWM curcit that drives an AC-motor at a controlled RPM from 12v batteries with 90% efficiency in the transfer without making changes to the motor.
As I've said it's just an idea, I could be wrong for any number of reasons and I'm probably never going to try it out.

I know what AC is, I'm just clueless as to a lot of the formulas, etc. I think an AC motor generally would only need special circuitry to start itself, and thats the part I really don't understand.

Oh, I hardly ever bother with formulas. AC-motors work pretty much like brushless motors.

If the actual running performance range is the same as BLDC, then what I don't get is the messed up ohm's law. I know for AC, volts times amps doesn't quite equal watts, and if I don't know the wattage then I can't figure the HP.

Yeah, tell me about it. *sigh* Ohms law, though useful, has a lot of shortcomings especially with high voltage AC. But it's the best we've got atm so if your wattage is about right just leave it at that. I think it's because the effect is not linear and at higher voltage/amps/resistance while Ohm's only gives a reliable value in a linear curcit at constant and low voltage/amps/resistance.
Whenever I need to use formulas I reluctantly look them up, put'em into Calc (excel on MS) or some other mathprogram. I then take the answer I get, tweak the formulas and values until I get an answer that aprox equals my measured values.
Frankly I've been frustrated with Ohm's since childhood and it's mostly because it's only an approximation and the materials we use are only precise down to 1-5%, after that it's all quantum.

Can I get your design? I'm currently converting a 1980 Fiat Spyder Thanks

Here they are.

I think I got most of the stuff on there, and a lot of the parts aren't critical as to which one you use, but the really big MOSFET is here. There are some others on there that are almost as powerful and that you don't have to order two of, and those should work fine too. If you have any questions, just ask. I'll get back to you with programming for the microcontroller (for a picaxe 08M).


N-channel MOSFETs are usually used for low-side switching. No offense, but it looks like you have a long way to go before you'll have a working speed controller. I suggest you start with some cheaper MOSFETs to do "proof of concept."...

I don't even actually plan to do this. I just design stuff. About the most advanced thing I've actually done is to build my own computer.

. So, assuming westfw is right about the MOSFETs*, you're just spreading bad designs around? . If you haven't built it and tested it, how do you know that it "could handle 144v at 690a"? . I just hope the others haven't wasted too much time with your "design." *Not a bad assumption to make - judging by his previous posts, westfw knows what he's talking about when it comes to electronics (and many other things). I've seen no evidence of super-human mental powers, but he's pretty damn smart. You would do well to listen when he speaks.

144V: the parts that would have to handle this are: voltage converter: capacitor easy MOSFET even easier MOSFETs Step 1 easy Step 2 (main) is rated 100v drain-to-source. 144 if that was the only thing on the circuit would be bad, but which do you think has more resistance, MOSFET or motor? 144v on the circuit should be nothing to the MOSFET because of it's almost insignificantly low resistance (in comparison to the motor). The only part that would have to handle 690a would be the main MOSFET, and if it can't handle 690a, then why is it rated for it? And I admit that I am not 100% sure what I am doing, so I want to be told if I am doing something wrong, so I can correct it and not do it again.

You have to arrange for Vgs, the voltage between the gate and the source, to have reached the ON threshold. In a low-side switch, the source is tied to ground, and this is relatively easy (although many high power MOSFETs, including the one you reference, have Vgs-on of about 10V, so driving them from a 5V microcontroller can be a small challenge. In a high-side configuration (with the load in between the source and ground), the gate voltage will need to be 10V above the supply voltage to keep the MOSFET on when the load is running.

You didn't mention power dissipation. Max current and max power dissipation are two separate issues in power semiconductors. You can only sustain the max current IF you manage to dissipate enough power that the junction(s) temperature stays "reasonable." With Rds-on of 1.8 mOhm and, say, 600A flowing through the MOSFET, I calculate that you need to dissipate about 650W, which is no small feat. Does your current cost estimate include a big heatsink and fan?

The circuit you show has no provision for turning the MOSFET off; you'll charge up the gate to Von (maybe) through your secondary mosfet, but there's no discharge path to a Voff when it's time to stop. A fair amount of attention in high-power mosfet circuits goes to making sure that the turn-on and turn-off events happen FAST, because in-between the MOSFET is going to be operating in a linear region where Rds is much higher, and the power dissipation will leap-up (linearly with Rds-on) in neat little spikes, and fry your MOSFET if you aren't careful.

And of course, at 600A, your interconnections and even your wire become things that you have to pay pretty careful attention to.

In short, a speed controller at this sort of power level is one of those things that SOUNDS easy ("just use a huge MOSFET"), but quickly gets more complicated as you have to take into account all the annoying real-world behaviors that exist.

There are some "starting points" in the "DIY electronics" portions of RC Forums, though of course most of those are for far lower currents than you'd encounter in a full-sized car.

To get 10V from a 5V microcontroller is why there's the secondary MOSFET. The main MOSFET is rated for 2500W power dissipation, and the electronics part of this actually only costs $170. Good point, I need some kind of pull-down between the MOSFETs, and when I want it to, it needs to handle quite a bit of current. Thanks for telling me this stuff!

Oooooh... the indentation didn't come out right on that, I think you can see what I mean though...

There are P-channel high power MOSFETs out there, just check on digikey.

I'm working on the schematics, I'll get back to you as soon as I can.

Yeah, I could use those too. I have enough stuff to make a pretty neat EV, (except the speed controller).

Thanks! I'll have a good look tonight. I have all the stuff to make an electric motorcycle, and most of the stuff to make an electric Firebird.

The bike is likely, I don't know if I've the time to do the Firebird. My kids may inherit that project.

. Most commercials units will have current limiting, ramping, and other functions that DIY units usually omit. I don't know if that makes them worth the extra money or not, but there are good reasons for them to cost more. . . Keep us updated on your project. Sounds interesting.

Reasons, yes, good reasons, I don't think so. Current limiting is the main one that my design is missing, but the main reason to limit current is to keep stuff from overheating, and since most everything has a thermometer on it, I really don't see the point. Most of the other stuff has to do with the microcontroller, which makes it really cheap to add them if you want them.

. Unless you are aiming for all-out performance and aren't worried about reliability, I suggest you add current limiting. Ramping is a good idea, also. Heat is the #1 enemy of motors (and most other electrical components, including the controller). . For proof-of-concept, CL is probably not that important. For everyday use, it is.

I just looked up current sensors on digikey. 1000 amps? 200 DOLLARS. The cost of the rest of the project itself. :-(

. No need for a 1000A sensor. . Ever hear of a CT (current transformer)? . If you know the load resistance, computing the current from the applied voltage is trivial.

. Oops. I don't think a CT will work well with PWM.

Possibly. I think a lot of that depends on whether your motor has a thermometer on it too.

Also, remember that commercial products must comply with FCC, CPSC, and numerous other rules, regulations, and compliances. This all adds cost to both the initial engineering of the product and to the final construction of it.