Motor Current Limiting: More Power and Less Heat!!!

[eP]

Do you know what is the reason the startor is builded with many thin steel sheets ?
The same is true for 50/60Hz transformers.

The stator's Fe loss is quite different than winding (Cu) loss.

I was assuming that when you said "stator" that you really meant "rotor" (where the current flows) so I was trying to communicate with the language you seemed to have presented.

The stator is just the permanent magnets.

Sorry i'm still thinking about brushless motor.
At brushed magnets are at stator.

So my stator was BLDC stator as you can find at RC outrunners or motor hubs.

So what are "trying" to say about the "permanent magnets"?

The motors we use don't manually energize those magnets so there is no current wasted doing anything to them. They just "sit there" and behave magnetically and cause the BackEmf.

You seem to be off on what we call a "wild goose chase" focusing on something that doesn't appear to be a major factor in the formulas that motors use to figure out their powerbands.

BackEmf = VoltageConstant * Rpms / RadiansSec

Current = ( ActualVoltsUsed - BackEmf ) / Resistance

For very thin brushed rotor's sheets above could be the true. But cheap brushed rotors are not designed for very high rpms or big rpms span.
Even Randy knows brushes have strongly limited life at high rpms. Thicker sheets are cheaper (less sheets per rotor, cheaper material (more to lose) ), so if you try high overvoltage the motor the brushed rotor (Fe) loss will rising much quicker then linear with voltage (or rpms).

For a little overvoltage (maybe 33% is still a little i dont know that at this case) the above equations could be still good.

 

[fechter]

The armature (or stator in a brushless motor) is made of many thin laminations to prevent current from flowing in the iron. Anytime you have a varying magnetic field passing through a conductor, a voltage is induced. If there's a path for the voltage, you get a current.

By laminating the core, you still get the area needed for the magnetic flux, but the laminations are electrically insulated from each other, which breaks up the flow of current in the core. If you used a solid chunk of iron, the magnetic flux would be the same, but the iron would act like a shorted winding and create drag on the motor.

Even with laminated cores, within the thickness of each sheet, (and in the copper windings) a small amount of current is generated which results in loss. These are called eddy currents.

In addition to eddy currents, when you put a piece of iron up to a magnet, it stays slightly magnetized. Try this with a nail or paperclip. Every time the magnetic field is reversed, energy is lost in overcoming this residual magnetism. This is called hysterysis. Hysterysis and eddy current losses together are called core loss.

Core loss is a function of motor speed. Using special alloys and very thin sheets will minimize core losses.

In addition to core loss, you also have windage loss, which happens as a result of the rotor acting like a fan and pushing a lot of air, even if it's trapped inside a motor housing.

If you run a 'normal' motor at 8,000 rpm, the no-load current will be excessively high due to the core and windage losses. That's if the rotor doesn't fly apart. The motor may also overheat even with no load, since much more heat will be generated in the core.

If there were no core or windage losses, then you could just scream the motor at redline and maintain efficiency.

On my old brushed Vego motor, I was running over 6,000 rpm on a motor rated for 3,000. This allowed a much higher power output for a given current, but the commutators failed anyway due in part to running a higher voltage (more arcing). I almost expected it to fly apart, but it never did.

 

[eP]

Core loss is a function of motor speed.

It is not a linear function (generally - at wide rpms span ).

 

[safe]

Sorry i'm still thinking about brushless motor.

That would explain the confusion... I've been talking about a brushed motor... :wink:

 

[safe]

If you run a 'normal' motor at 8,000 rpm, the no-load current will be excessively high due to the core and windage losses. That's if the rotor doesn't fly apart. The motor may also overheat even with no load, since much more heat will be generated in the core.

If there were no core or windage losses, then you could just scream the motor at redline and maintain efficiency.

On my old brushed Vego motor, I was running over 6,000 rpm on a motor rated for 3,000. This allowed a much higher power output for a given current, but the commutators failed anyway due in part to running a higher voltage (more arcing). I almost expected it to fly apart, but it never did.

My spreadsheet says that the no load at 36 Volts is 1.5 Amps and at 96 Volts it's a little over 4 Amps. That part I can live with. The commutators seem like the biggest area of potential failure on a brushed motor. Beyond a certain speed the brushes would have difficulty actually getting current through and so the efficiency would go down in a way that the formulas don't normally content with.

So a whole new set of factors of losses definitely seem to come into play.

Overvolting from 36 Volts to 48 Volts seems "conservative" and if combined with "Motor Current Limiting" to keep the heat under control it should be okay.

Beyond 48 Volts and you have difficulty finding controllers that match, so it seems like "for now" that's about as much "tweeking" as I would want to go.

A brushless motor at really high voltage AND using "Motor Current Limiting" AND gears seems like a good choice if you can gear it down enough to make it usable...

What if you took a 600 Watt brushless motor and went about doing all this sort of "trickery" to get about 2 hp?

 

[fechter]

What if you took a 600 Watt brushless motor and went about doing all this sort of "trickery" to get about 2 hp?

Hmm... actually I get over 2.5kw out of my 600w motor. Works for me I don't think it would be happy at 2.5kw for more than a couple of minutes, but 1.5kw would be close to the continuous rating. It's running at about 5,000 rpm. It would be more efficient to slow it down to 3,000, but that would reduce the power quite a bit.

You should be fine running a 36v motor at 48v with the right current limit.

 

[safe]

Hmm... actually I get over 2.5kw out of my 600w motor. Works for me

Could you list:

Motor - Brand - Data/Link
Controller - Brand - Data/Link
Voltages
Amps

 

[fechter]

Originally a BMC 600w internally controlled scooter motor modified for external control and forced-air cooling.

Similar to ones sold here:
http://powerpackmotors.com/

Controller: modified Crystalyte (IRFB3808 FETs) 60v, 90amps max. Hopefully also available from Powerpack in the near future, only with 4310's, which can handle up to 90v.

There's a planetary gear reduction unit that can fit some versions of this motor used on Currie USPD drives.

 

[TylerDurden]

Gear that down and you get more power at low vehicle speeds.

What gear ratios?
How much torque to still accellerate?

I guess you haven't figured that out. :lol:


I will happily pay the shipping costs, if you want to send me your last burned-up Unite... I could rewind it and use it. :lol:

 

[safe]

I will happily pay the shipping costs, if you want to send me your last burned-up Unite... I could rewind it and use it. :lol:

Actually you wouldn't be able to. It was the commutators that burned out. Seems that one of the metal pieces that the brushes rub against wiggled loose from the rotor and began to give me less power. Even while cold I was running it and it sounded bad and you could see inside that there was a lot of arching going on. So after 1000 miles on that motor and after a series of rides where it got rather hot it simply started to give performance that was not so good anymore.

Then I took it apart and tried to sand the commutator and brushes (which appeared dirty) and see if I could make things better (which of course it didn't and now I realize that was a dumb idea) and it was while testing after this that it got so bad that I gave up on it.

The rotor no longer has smooth contact areas for the brushes... and I think my tinkering only made things worse.

So it wasn't in "real life" the way you are picturing it... it's not like I ran it up a hill and it burned out, it was more of a sequence of events that forced me to abandon it. The power seemed to be declining over the last hundred miles or so.

And I had a spare motor (1200 Watt) that took only 10 minutes to install and replace it.

So it wasn't a simple "burn up" like you think... and might in the end be attributed to a bad spot on the rotor that might not have been manufactured that well in the first place... combined with heat and my own tinkering... and 1000 miles of riding too...

Brushed motors have a problem with needing "brushes"... :wink:

 

[safe]

Controller: modified Crystalyte (IRFB3808 FETs) 60v, 90amps max. Hopefully also available from Powerpack in the near future, only with 4310's, which can handle up to 90v.

Any chance that such a brushless motor / controller combination would be able to do "Motor Current Limiting"?

 

[fechter]

Sure. Using the Allegro current sensor, the output can be sent to a peak detector (basically a rectifier) so it will work with AC.

 

[safe]

So the "dream machine" would be:

1. Brushless Motor (improves efficiency)

2. Overvolting (increases rpms, more power)

3. Forced Air Cooling (reduces heat)

4. Motor Current Limited Controller (reduces heat)

5. "Base" Current Limit Raised for Peak Power (increases heat and power)

6. Multiple Speed Hub (expands the useful powerband)

The result of all this would be light weight, powerful, able to handle heat and efficient. Hmmmm... except for the gears and motor side current control this is pretty much your machine already Fechter. 8)

 

[Malcolm]

Hey Safe you stole my wish list for my next machine . Just add full suspension and a custom alloy box frame to hold 5kWh of Life batteries and you'll have people queuing up at your door. Only problem is I want to use the Mars brushless, and I can't find a geared hub (or even a small 2/3-speed gearbox) that will handle that sort of power.

 

[fechter]

Gears would be nice.
The Nuvinci hub might be able to handle the power
They also make one in a scooter wheel already.

The externally controlled BMC motor is pretty good for the price. It's nearly as efficient as a volcano motor below 1kw.

 

[safe]

How is the motor current limiting implemented in that controller?

I suspect he was simply pleased with his own motor and not contributing anything more than "gee wiz guys, my motor is really cool". I'm not certain he even understands what "Motor Current Limiting" might mean in the context that I've been presenting it.

But I could be wrong...

 

[fechter]

I'd guess Randy's controller is using battery current limiting.
But since we can't actually buy one, it's sort of a moot point.

 

[baltipower2]

FWIW, here's the link to the manufacturer of Randy's controller. They claim it's max continuous output current is 20A, and peak output current is 35A.

http://www.slmti.com/drives/drives.asp?product=SSC024D16

Mick Sheehy

 

[safe]

Definitions:
• Closed Loop Control: Speed remains constant under varying load conditions.
• Open Loop Control: Speed remains constant when load is constant, and moderately varies when load varies.

Hmmm... so what exactly are they saying?

 

[Lowell]

Closed loop = cruise control. Probably works the same as the DrainBrain closed loop speed control.

Open loop = normal throttle operation.

 

[safe]

Closed loop = cruise control. Probably works the same as the DrainBrain closed loop speed control.

Open loop = normal throttle operation.

So that's really not anything related to "Motor Current Limiting". The core idea of "Motor Current Limiting" is to forcebly "ruin" the powerband so that it's only good in the peak power and efficiency areas. By reducing "torque/power" in areas that produce heat it forces the rider to keep the bike going by being in the correct gear all the time.

I was thinking of a mechanical "boot" that would hang off the back of the bike and simply kick the rider in the ass when they were in the wrong gear, but that ended up too complicated! :lol:

 

[fechter]

Closed loop means the throttle input is controlled by feedback from some measurement. Usually it's the motor speed, so it acts like cruise control.

You could also have a closed loop that uses motor current as the measurement.

Open loop means there is no feedback from any measurement, and is how most cheap controllers work. The human is the feedback loop.

 

[eP]

By reducing "torque/power" in areas that produce heat it forces the rider to keep the bike going by being in the correct gear all the time.

I was thinking of a mechanical "boot" that would hang off the back of the bike and simply kick the rider in the ass when they were in the wrong gear, but that ended up too complicated! :lol:

You should start to think in which way contoller will recognize is the room for right gear to change, i'm sure.

As an example lets look at the case when rider using only 1/4 of the max. power at flat (rpm are 50% of the limit and current is also 50% of the limit) at flat. And the light slope occured.
Lazy rider don't want increase the throttle and forget to switch te gear ( or even he dont know he shoul reduced gear at that case ).
The speed falling twice to 25% max, and current is close to the limit. If speed is still falling it is to late to force the drider to change to lower gear by limiting the motor current.
If controller will activate the limit at inapriopriate higer gear the speed and rpm will drop at fast rate - but is already low and the rider will be must reduce to the lowest gear if he want regain his low speed before the limit occur.

So the right controller should give to rider the right signal a long before the current reach the limit (if fact - when rpms started falling below 50%). This is my .

Best regards

 

[safe]

You should start to think in which way controller will recognize is the room for right gear to change.

Yes, this is really the "central idea" that "Motor Current Limiting" addresses.

There are many ways to "know" what is going on. You can have an ammeter and focus on it while you plow into a car. :lol: Or you could have some sort of light that goes off that tells you when to shift. Or you could somehow have a computer "of sorts" that matches the motors powerband to the speed of the bike and shifts automatically.

You could do a lot of things...

But the EASY way to force a rider to shift is to simply "take away" torque/power in places he should not be. A rider can "feel" when things are "right" and if they are wrong they can shift until they figure out how to ride. The bike will be LESS forgiving to a stupid rider compared to the existing motors that allow you to do anything (I can always run in any of three gear choices) but this "strict" behavior will force the stupid rider to BECOME a smart rider because he has no choice.

Discipline will be "taught" just like it is on a race motorcycle...

 

[eP]

You could do a lot of things...

But the EASY way to force a rider to shift is to simply "take away" torque/power in places he should not be. A rider can "feel" when things are "right" and if they are wrong they can shift until they figure out how to ride. The bike will be LESS forgiving to a stupid rider compared to the existing motors that allow you to do anything (I can always run in any of three gear choices) but this "strict" behavior will force the stupid rider to BECOME a smart rider because he has no choice.

Discipline will be "taught" just like it is on a race motorcycle...

This way it will be disturbing discipline instead helping.
This way the rider will be forced to deep gear reduce because if the power will be cut off the speed and rpms will drop faster.

And this way it will be stupid discipline as it will be tolerate riding at wrong gear to the moment when will be to late for the gear reduction (so deep reduction will be needed).

A smart controller should recognize cases when reduction is needed in spite of the current limit is not yet reached.
This way bad rider coud avoid deep gear reduce and deep rpms/power falling.

This way controller could help for more convenient riding.
Your simple "blind" controller will be "taught" and unpleasant as electric shocker.
Maybe this way the bad rider could learn faster but i dont think you looking for solutions for teaching so much lazy riders.

Am i wrong ?

Regards