Motor Controller, why not bigger? 12 Fet vs 18 Fet

So I have been talking to Paul at em3ev.com. I am on the verge of placing an order with him and leaning towards buying either an 8t or 10t Mac 55/1000 motor and wheel system. Planning on using a 50 volt 15ah battery. I want to oversize the controller in case I ever decide to get a bigger motor or run more batteries. However, Paul seems to be steering me towards the smaller 12 fet 40A Max current Infineon controller and I can't figure out why. Everything I have done with electric motors it has always been a rule of thumb to oversize the controller. I have indicated to him that I would like to go with the 18 fet 60A Infineon instead. Am I missing something here? Would there be a good reason to go with the smaller controller besides cost and a little bit lighter weight?

There may be other reasons, but the default current limit is higher on the 18 fet (50A) vs. 12 fet (40A). I think you want to keep the 8t/10t motor in the 25-30A continuous range.

Paul's official recommendation appears to be to use the 9 fet controller (30A) for both of these motors: http://em3ev.com/store/index.php?route=product/product&path=36&product_id=138

Granted these are programmable controllers, so you would think they would behave the same if you programmed the same lower current limit.

But, the 18 fet is probably going to have a lower ohm shunt, which means the instantaneous current draw will be higher, even though they may average out to the same current level. Maybe this is problematic?

Try asking him what the basis for his recommendation of the 12FET over the 18FET, and what are the disadvantages of the 18FET over the 12 FET.
To be cynical his recommendation could be because he has no 18 FET's in stock? A good reason to recommend the 12FET as a supplier. (NOT saying that is the reason, just that you can't know his reasoning unless you ask, and it could be anything- something that he may think is critical to your application that may not be important to you).
Ask.

With a MAC, you're going to be abusing the $#!% out of it with even a 12FET controller. They will rapidly turn to large boxes of melted junk at the power levels an 18FET can handle. A 9FET is all they need. Its a geared motor, not a direct drive. The won't take huge power levels.

Yep. It's not a 2000w motor.

You could take a big strong controller with many fets, and then limit it to less amps. That would put little strain on each fet. But really, blowing controllers has never been a big issue with me. Melting motors is fun, but I never blew a controller unless I hooked it up backwards or something. I do let my controllers get plenty of breeze though.

Just save the money for a big controller for later. An affordable one is all you need. Melting down a motor won't save you money.

That seems fairly clear!
Or get a programmable controller and use suitable settings, it can be reprogrammed if/when you upgrade.

So I am missing something obvious here, and not sure what it is. In the RC world you always oversize the speed controller. And I have never heard of blowing a motor because the speed controller is too big. The brushless motor on my helicopter puts out about 4 hp and spikes up to 7. I run a 12 S 5ah Lipo battery pack on it with 120 Amp speed controller. If I swap my speed controller out for a 150 Amp speed controller, like a friend just did, it will just run cooler and be a little bit heavier, no greater chance of blowing the motor. So why in the E-bike world does a bigger speed controller = shorter motor life?

Oh I asked Paul too and he said he would sell me the 18 fet controller but not warranty anything. I am not trying to start a battle or argue with anyone, just understand the reasoning. Thanks guys.

Plus the 18 fet is a lot bigger. Second motor second controller. You need a 12fet. in your draw if not. Mac. has it's limit's. Plus save the battery and you have more money for a C.A. and torque arm.
Try a D.D. to 3000 watt's.

Your RC chopper has a way to cool the motor and the blade spins fairly free until it loads up. Easy on the motor and the controller. The bike loads up at the start feeds the motor a bunch of amps just to get it moving then spins at far less load once at speed. Tough on both the motor and controller. Ebike motors have no way of getting rid of the extra amperage put to them while accelerating from a stop or on hills other than to waste it by heating things up. Toasting a motor is simple feed it to many amps, don't help it up a hill or travel at low speed for extended periods. Geared motors have less mass to dissipate the heat so one needs to pay attention to the way they ride as well. That being said I have 10K on my BMC V2S running 60-72V with an Infineon 12fet without burning it up and we have plenty of hills around here. Bigger controllers have more amperage available than the motor can waste in heat so there tends to be smoke instead of extra movement. You might also be being steered towards a softer start controller, in the 9 or 12fet, as well which would go easy on the gears as banging the motor abruptly with to many amps can also do damage there. Ebikes require matching up the components pretty evenly to get the performance and longevity one wants.

Chrome Dome said:

So I am missing something obvious here, and not sure what it is. In the RC world you always oversize the speed controller. And I have never heard of blowing a motor because the speed controller is too big. ... So why in the E-bike world does a bigger speed controller = shorter motor life?

Click to expand...

Well - the short answer is - these aren't RC motors.

There are two issues: heat and the freewheel clutch.

• Unlike DD motors, MACs and other gearmotors have an 'internal' motor and planetary gears. The extra mechanism and intervening air pockets put them at a disadvantage when shedding heat. So - you will need to limit the controller anyway to keep from melting the motor. You simply don't have the same lattitude to overdrive a gear motor as you do with DD. A 12 FET controller is perfectly adequate to melt a MAC without upgrading....
• More importantly, MACs have an internal freewheel clutch which can only withstand a certain amount of torque before seizing (becoming permanently engaged) or slipping/chattering - both failure modes deform the clutch surface. A 12 FET controller can push or exceed the limit of the clutch, which is why Paul doesn't generally recommend them. An 18 FET controller can supply simply abusive current levels.

There are means to mitigate some of these effects by re-programming the controller or by adding a Cycle Analyst V3 to ramp power application, but if you have never set this up before, there is a strong possibility that you will toast something while experimenting. In any case, an 18 FET is so over the top that there is little meaningful way to use any of the extra current capacity it affords.

IMHO dramatically oversizing the controller based on some possibility of a future change to a bigger DD motor will become a self-fulfilling prophecy.

RC controller Amp values and Ebike controller amp values aren't the same thing. And the motors are rated differenty.


A 25 amp RC controller is rated to handle 25 amps worth of heat, but it will let more through. there is no hard limit. If you have a Heli that has a 25 amp controller, you generally have a motor/gear/prop combo that will pull a maximum of 25 amps in a climb, and much less the rest of the time..Its possible to pull 30-40 amps or more out of it recovering from a fall or if you strap a camera to it, but that can burn out the motor, blow a gear, or melt the controller. So you upsize the controller to handle the load.
A Heli motor is spinning a blade, so it's running near it's most efficient RPM and has plenty of back EMF to keep it from over drawing current. if you Jammed the blade so it couldn't turn, and then gave it full throttle, you would expect either the motor to vaporize it's windings, or the controller to burn out almost immediately

With a bike, the amp limit is a hard limit. A 40 amp controller will not let any more than 40 amps go through. when the motor tries to draw 41 amps, the controller overrides the throttle input and holds it at no more than 40. A bike motor has to work from 0 RPM when just starting out, way down in it's near zero efficiency range, and with no back EMF to choke the amp draw, its coils are basically a dead short for the battery to pour hundreds or thousands of amps into, melting everything in no time. The controller stops that from happening, allowing the motor to get up to speed without catastrophic meltdown. A geared motor is especially susceptible to this. while they make starting and acceleration torque easier than a direct drive motor, they have much less mass to absorb heat, and they have no way to get airflow into the motor to aid in cooling. Generally, you want a controller to cap the max power for a geared motor at 2 to 3 times it's rated nominal power.

The 4110 FET used in many Ebike controllers can handle 120 phase amps each. So in a "little" 12 FET controller, the 4 FETs per phase could handle 480 phase amps peak. WAY overkill from the 30 amps the controller will be rated at.

Drunkskunk said:

The 4110 FET used in many Ebike controllers can handle 120 phase amps each. So in a "little" 12 FET controller, the 4 FETs per phase could handle 480 phase amps peak. WAY overkill from the 30 amps the controller will be rated at.

Click to expand...

The continuous rating of 120 amps is at 10V source to drain. http://www.irf.com/product-info/datasheets/data/irfb4110pbf.pdf

Go melt some motors, and then you will understand and agree.

I know squat about RC stuff, but I know a 1000w rated hubmotor will draw 2500-3000w easy if you offer it to them. But they will melt if you run that kind of load very long. Select a controller for the Mac, or adjust settings on it, so about 1500w is the max and it will last a lot longer.

48v 30 amps is not too out of line, but 60 is just too much for a Mac in 26" wheel. Stick to 20 amps if you really want a long lifespan.

My 2 cents:
*RC motors are cooled by the propeller and strong airflow, hub motors are not well cooled - as such max acceptable losses are far lower on a hub motor
*Hub motors are very torquy, so they can take demand from the controller a huge amount of current when needed (like starting from standstill); RC aren't torquy, they compensate thru high rotation speed
*Hub motors are bigger, heavier and produce the torque thru a high pole-pair count, not thru all-out current like RC motors; RC-type motor design are good if they can be cooled and at high rotational speed, Hubs are good for low RPM and high starting torque.

...all this to say (and to add to what was previously said), if you but a high FET-count controller you allow very high currents to be transited from the battery to the motor with potentially bad consequences for both (ebike batteries have usually lower current rating than RC Lipo). To sum up, not knowing your technical comprehension level, I would also advice you to buy a lower FET-count controller; if you feel you understand what's this all about get the 18FET but absolutely the first thing to do is to plug the programming cable and limit DC and phase currents in the controller to something reasonable for the MAC (30Amps) & your battery.

Any controller just needs as many fets as it takes to reliably supply the current it's speced for. If you use an 18 fet 50A controller, then it's possible you might burn up the motor by giving it too many amps. The 12 fet is limited to 40A. 2000W max at 50V. I'm with you though. An 18 fet will have less stress on the fets at the same amperage. You can also limit it by removing one or more of the shunts so it maxes out at ~ 40A, or 20A, or whatever you want to do to it. I'm assuming he assumes you want a plug and play controller and the 12 fet would be the one you wouldn't have to mod to keep the battery amp draw down.

Nothing wrong with using an 18-FET.

Order the optional MAC motor-temp sensor, and hook it up to a CycleAnalyst, and then limit the amps to the motor at 95C/200F, then...check the amp-log to see how many amps is causing 95C/200F. If you are getting that temp at 40A or less, going to 18-FETs was un-necessary. Its bulkier,and more expensive. A few guys around here have melted the motor using a 12-FET, so...even if nobody explains it to you with a persuasive theory, something is still causing that.

The MAC geared hub has no air circulation, so it responds well to oil cooling (filling 1/3rd of its interior with ATF after sealing leak-points)

Exactly, it's not the number of fets that is the issue. But if you give a hub motor too many amps under too much load for too long, you melt em.

Lots of arguments about what the wattage limits are for a given motor, because people ride so differently. But in general, you give a motor a big enough load, and no reasonable amp limiting, and you will make so much of 3000w into heat it melts the motor.

Turn on a 1500w space heater. See how fast the element glows when 1500w is turning into heat?

Lessen the load, such as riding 30 mph on flat ground in no wind, and offering a motor 3000w has no effect, since load at 30 mph is likely to be only 1000w, and only 200w of it making heat. Motors can radiate away 200w fine. Increase load a ton by riding 40 mph, and many hub motors will melt in 15 miles or so.

If you include a reasonable amp limiting, then no reason a big controller can't be used. It's merely big, and expensive for the job at hand. I just don't see the reason to spend the money, till you do have a big motor.

I agree with all the comments most controllers especially ones sold by sellers such as Paul, are already running current ratings sufficiently low to keep the controller nice and cool most of the time.

The difference between rc and these is that Rc is traction limited because of the huge power to weight ratio of not having a payload.

Some of the motors in RC are better for example the chinese feigoa motors would melt very easily if run too long at high amps. The better motors normally ones made in germany and usa have efficiency so much greater that they would just come off a run warm instead of burning hot.

However remember in RC you run for a short periods of time races are still only 5 minute sprints whereas a long ride on an ebike can be a couple of hours run an RC car for a couple of hours and without adding a fan even top end motors will begin to melt.

Finally Payload my bikes weighs less than i do by a big margin yet still has to carry me up hills etc if you were to take a payload proportionally similar and put it on an Rc car it would struggle to not overheat in even 5 minutes.

If you want to run a 18fet controller at its 85% capacity forget about all the em3ev options. You'll burn the motors.

If you wanna truly use a 18 fet you need a heavily modded 9c/clyte/magic pie hub or a cromotor and be done with it.

So I think I understand. It is basically a difference in methodology of rating speed controls. Does this sound correct: The current (Amp) rating on e-bike controllers is the maximum the controller will allow through it, while in the RC world the current rating is the maximum it can handle before damaging it. Also I think like someone else said e-bikes see high current at very low motor RPMs, where RC stuff it typically is at high RPMs. And high current and low RPMs are hard on motors.

I have a 18 FET controller with IRFB4110 FETs and the current limit is 42A, which is not much. I have a 3000W rated hub motor, and 108V nominal voltage. The power output is easy on the motor, and the controller has only 2 shunts but there's a place for a third, so I guess it could easily handle the plus shunt. The phase current limit is 100A. As I read in the comments the controller could handle the more amps, my BMS is rated 60A, so I only want to increase the controller's amps to that level. Will my controller handle it, or the phase amp will be increased too and will possibly crash? What could go wrong in the controller if the shunt's resistance is too low? I fried a controller once, but I checked and replaced the broken FET and it worked will, but not for so long, so I guess not only the FET was damaged.

jordanjozsef said:

I have a 18 FET controller with IRFB4110 FETs and the current limit is 42A, which is not much.
<snip>
and 108V nominal voltage.

Click to expand...

Then you have about 4300w of power out from the controller already.

and the controller has only 2 shunts but there's a place for a third, so I guess it could easily handle the plus shunt.

Click to expand...

It might. Sometimes they're fine doing this, sometimes they're not. I added a shunt to a generic 15fet and doubled it's current peak, with so far no ill effects, but it only sees taht current for a second or two during startups from a stop, most of the time. Dunno how it would fare if it had to do that continuously.

my BMS is rated 60A, so I only want to increase the controller's amps to that level.

Click to expand...

Well, in theory adding another shunt to a pair, if the shunt is identical to the other two, should cause it to be half again as much current, so 40 / 2 = 20, 20 * 3 = 60. What it actually ends up as you'll have to verify....if it's higher than the BMS limit, for long enough, it'll trigger the BMS to shutdown.

If yo\u don't know what the shunt values are, you can't know that any shunt you replace it with will be the same....but you could "guess" and use a shunt out of an identical controller, and it should be the same.

I used one out of a different controller, and ended up wiht a lot more current than expected, so the one I used must've been a lot lower than the originals.

Will my controller handle it, or the phase amp will be increased too and will possibly crash?

Click to expand...

If the controller only measures battery current via the shunts, and does not have current sensors on teh phases, then phase current increases by the same ratio as battery current becuase tha'ts all the ocntroller does (limit battery current, not phase current).

Only way to know if the controller will handle it is to try it, and see if it blows up. :/

Then you have about 4300w of power out from the controller already.

Click to expand...

You know, people always want more, and since I'm in between the safe limits, why not? I guess my 3000W hub motor could handle 6000W bursts without problem, because my 300W hub motor handled the 4300W too. Me with the bike weights around 100 kgs (220 lbs), so the weight is not a problem.

It might. Sometimes they're fine doing this, sometimes they're not. I added a shunt to a generic 15fet and doubled it's current peak, with so far no ill effects, but it only sees taht current for a second or two during startups from a stop, most of the time. Dunno how it would fare if it had to do that continuously.

Click to expand...

I don't know either, but I'm not planning to double the current, because my 18650 cells aren't designed for high current, on full throttle the voltage drops by 6 volts roughly (which could be high, but since it's high voltage I have to divide it by 28 to see the drop for each parallel cell groups, which is 0.21V, I don't think that's much). I have 10 connected in parallel, so 6A / cell could heat up the cells in minutes. I'll install a thermistor between the cells.

Well, in theory adding another shunt to a pair, if the shunt is identical to the other two, should cause it to be half again as much current, so 40 / 2 = 20, 20 * 3 = 60. What it actually ends up as you'll have to verify....if it's higher than the BMS limit, for long enough, it'll trigger the BMS to shutdown.

Click to expand...

Well, my analog ammeter only measures 50A, so I'll have to purchase another one to go above it or else I'll have no idea what I'm doing. I have some 36V 300W controllers I can take the shunt out and then calibrate them to reach 50A. Also that 42A might not be true, my first controller said 18A, but it was actually 14A when I measured, for this 42A controller, I don't know yet, but if it's below the rated current limit (let's say it has 35A limit), I'll definitely feel the difference if I mod it to 50A, since 1A give me roughly 100W more power, so it would be 1500W increase in power. But if the 42A is true, I'll still have 800W increase.

If the controller only measures battery current via the shunts, and does not have current sensors on teh phases, then phase current increases by the same ratio as battery current becuase tha'ts all the ocntroller does (limit battery current, not phase current).

Click to expand...

Aren't the FETs in parallel so it could handle more phase amps that 100A? They are rated 180A each, so that's 540A. The ratio of the phase and battery current is 2.38 (42A battery and 100A phase). So if I divide the maximum phase current it can handle with the ratio, I'll get 226A which is way above my tiny 60A current. Once I blew a controller with the shunt mod, but it was only a 6 FET controller went from 14A to above 30A it lasted for a few seconds. Then I replaced the broken FET and it worked again. So I think I'll purchase some irfb4110 FETs before going crazy with the shunt mod.

fets are in parallel, but you can't actually get the rated current out of them, for a number of reasons. There's a bunch of discussions over the years that probably explain better than me, but basically:

the wire bond in the fet is one limit--it's not actually capable of as much current as the die is spec'd for, and the spec sheet may rate based on the die and not what the package itself can actually do.

another limit is temperature. the current you can push thru a fet decreases with temperature, and you can only get so much heat out of one so fast. under perfect lab conditions a fet can take more current than it really can in the real world conditions inside a poorly designed controller like most generics.

another limit is even distribution of current. the fets aaren't exactly the same, so current doesn't flow evenly thru them, and some take more than others. if you run the system at the limit of what one fet, times however many there are in one half of a phase bridge, and one fet takes more than it's share, then that fet will overheat faster and be less capable but still draw more and heat more and runaway until it blows up.


another limit is gate drive--if it's not turned on and off correctly, a fet won't be low resistance and current heats it up faster. gate drive in these controlelrs is basic, and not ideally designed. even an ideal design is still not perfect in realtiy,



also, the current flow thru the phases is different than that in the battery, because of pwm / etc.

regarding the meter, i'd recommend a wattmeter with external shunt (like the cycle analyst uses), so you can monitor current along with power usage, voltage, wh, ah, etc. useful for troubleshooting and experimentation.

Well, I put a third shunt meanwhile and I'm using it like that for a few months with an extra heat sink and fan, because it can get hot during a full throttle acceleration. It draws 54A, and it works well, I just need to improve the cooling system. I want to remove the ribs on the controller's box, so I'll have a flat box, so I can mount bigger heat sinks and I'll add CPU fans too and will control them via Arduino, so they will only turn on when it's necessary.