**Important** reality check on motor, voltage, current etc.

[donob08]

Eric

I think Luke's point was that while the technique presently used limits Ibatt, it has no control over Iphase. I phase is mainly determined by Vbemf. Ibatt limiting really doesn't solve the high motor current problem. As Luke has been saying it is Iphase that matters. The only way to cut Iphase is to cut Vbatt. The buck converter doesn't do that. It cuts the voltage to the motor but allows the current to go to the level Vbemf defines. If we wanted the buck transformer to do it all, we would have to set D so that D*Vbatt/Rphase was at the level we wanted to limit current to. Since Rphase is very small 0.1 ohms we would have to make D very small and the speed represented by D*Vbatt would be near zero, nearly matching what Vbemf is being generated by.

I understand that when a person is tied to a way of thinking it is hard to break free. I've demonstrated that here myself. But the point of the thread is "Think about Iphase, don't be fooled by Ibatt limiting". I think I have done that.
We need to get Iphase down. I developed the equation for Iphase during battery current limiting to help us understand how to do that.

At low speed, with battery current limiting
Iphase = SQRT (C)* Vbatt/ Rphase alone. D doesn't even enter in to it.

Don

 

[rhitee05]

I think Luke's point was that while the technique presently used limits Ibatt, it has no control over Iphase.

This is not true. The Xie Chang controllers only measure Ibatt, but since the controller knows the duty cycle it can make a fairly accurate guess of the phase current. This is how they implement phase current limiting. It's not exactly true phase current limiting, but a very close approximation and effective as such.

I phase is mainly determined by Vbemf. Ibatt limiting really doesn't solve the high motor current problem. As Luke has been saying it is Iphase that matters. The only way to cut Iphase is to cut Vbatt. The buck converter doesn't do that. It cuts the voltage to the motor but allows the current to go to the level Vbemf defines. If we wanted the buck transformer to do it all, we would have to set D so that D*Vbatt/Rphase was at the level we wanted to limit current to. Since Rphase is very small 0.1 ohms we would have to make D very small and the speed represented by D*Vbatt would be near zero, nearly matching what Vbemf is being generated by.

This is also not true. Your argument doesn't even pass the sniff test - if BEMF is the only factor which influences the phase current, why bother with PWM at all? It should be obvious that this cannot be true. I direct you back to the expression I derived a while back: Iphase = (D*Vbatt-Vbemf)/Rphase. Phase current is a function of both duty cycle and BEMF. To look at that expression another way, phase current is simply given by Ohm's law, I=V/R, where V is the difference between the controller output voltage and BEMF.

 

[liveforphysics]

AFAIK, only the last generation of infinion chip controllers have phase current limit calculation ability.

The formula might not be perfect yet, but I think Don has a handle of why I made the thread now, and I'm happy with that.

 

[donob08]

Eric

I'm thinking this should be my last note in this thread. I don't want to bicker with you.

The reason we use PWM is to control speed. That is a well understood concept. If it also could control Iphase, that would be a nice freebe. Unfortunately it cant by itself. The value of Vbemf is of equal importance and we can't control that except perhaps by adding variable mechanical gearing so that motor speed can be made higher when vehicle speed is still low. Still at some point Vbemf would be very low. The fact that Vbemf is a function of motor speed is what makes the controller situation different from a buck converter. It is a very important difference.

I don't know if you read my writeup but it started with your expression Iphase = (D*Vbatt-Vbemf)/Rphase. That is the basis of my derivation. That relationship and D = Iphase/ Ibatt limit during battery current limiting, is all that you need to prove that Iphase = sqrt(C)* Vbatt/Rphase when Vbemf = 0 . Also we can see Iphase is close to that when ever Vbemf is small. So Iphase = sqrt(Ibatt limit?(Vbatt/Rphase) * Vbatt/Rphase and for Vbatt = 100 volts, Rphase = 0.10 ohms, Ibatt limit = 50 amps, Iphase = 223.6 amps.
I can't find anything wrong with that calculation. And I wouldn't call 223.6 amps a controlled current.

When Luke talks about "to the left of the red line" I believe he is talking about the area where Vbemf is negligible.
My words were definitely loose. D*Vbatt and Vbemf are both factors. But when Vbemf is 0 or negligible its ABSENCE is really the only factor of importance. D*Vbatt/Rphase is going to represent BIG current. I think that is what this thread is about.

if BEMF is the only factor which influences the phase current, why bother with PWM at all? It should be obvious that this cannot be true. I direct you back to the expression I derived a while back: Iphase = (D*Vbatt-Vbemf)/Rphase. Phase current is a function of both duty cycle and BEMF. To look at that expression another way, phase current is simply given by Ohm's law, I=V/R, where V is the difference between the controller output voltage and BEMF.

 

[John in CR]

Don,

Are you really still stuck on BEMF? PWM can and does limit Iphase on controllers programmed in that manner, which I would assume to be most. As stated before some even have Iphase limit as a user setting, and that limit is accomplished through the controller's only available means, PWM. Even if BEMF was zero this control will still work. While BEMF helps make things easier on the controller as rpms increase, Iphase is still limited by PWM up to the "red line", which is the first point where full duty is possible and the limits imposed by BEMF come into play.

Fight your way out of the formulae and derivations and establish a foundation regarding the concept to make the numbers more meaningful. That concept is that to the left of the red line, BEMF is part of the equation, but it's not the controlling factor. The effective voltage the motor sees is the means of control, which is accomplished through PWM.

 

[donob08]

I should say PWM can be used to limit Iphase. When it is used in that way it can no longer control speed. It can control one or the other NOT both.

If we want to make D such that DVbatt-Vbemf = const Iphase limit for Iphase greater than I phase limit.

(DVbatt –KVConst * Speed)/R = Iphase limit

I'm writing Vbemf as the speed/voltage constant of the motor times its speed.

DVbatt = Iphase limit*R + KVConst * Speed
Or
D = (Iphase limit*Rp + KVConst * Speed) / Vbatt

So you see D needs to be made a function of speed in order to control Iphase. When D is set in this way we will have Iphase limited, that means torque is limited. If the loading on the bike gets to be bigger than this torque can match, the speed will fall.

If we are riding along and the torque load changes speed will fall. If the torque increase experienced is related to speed, say like riding through mud or water the system will act like Viscous Damping. Speed will fall until the torque from damping falls to what Iphaselimit can provide.

On the other hand if the increase in torque load is constant at any speed, say as caused by going up an incline. The torque provided by Iphaselimit will be inadequate at every speed. The speed will fall until it is zero. The system will decelerate because the force caused by Iphase limit - force of weight moving up an incline will be a negative number. Since F = ma, and therefore a = F/m with F a negative value, the bike will decelerate to a stop.

 

[John in CR]

...On the other hand if the increase in torque load is constant at any speed, say as caused by going up an incline. The torque provided by Iphaselimit will be inadequate at every speed. The speed will fall until it is zero. The system will decelerate because the force caused by Iphase limit - force of weight moving up an incline will be a negative number. Since F = ma, and therefore a = F/m with F a negative value, the bike will decelerate to a stop.

FAIL. I go up hills every day. The difference is after this thread I know not to go up the big ones slowly at less than full throttle. If you insist on losing your way because of the trees, I'll just leave you with that compass I gave you before, and you can find your own way out to take a look at the forest.

Have a nice day,

John

 

[ZapPat]

I don't know if you read my writeup but it started with your expression Iphase = (D*Vbatt-Vbemf)/Rphase. That is the basis of my derivation. That relationship and D = Iphase/ Ibatt limit during battery current limiting...

Some dislexia again Don? :|
D = Ibatt / Iphase, not D = Iphase/ Ibatt

And you also made a slight error in this equation (I'm correcting it even though it is useless to us in this discussion):
Iphase = D * (Vbatt-Vbemf) / Rphase, not Iphase = (D*Vbatt-Vbemf)/Rphase

However, this second equation is not usefull for this discussion about Iphase current limiting, and I don't understand why you are making things so complicated. The first equation is all the controller needs to approximate Iphase, because it knows D and Ibatt already:

So the solution simply is:
Iphase = Ibatt / D , where both Ibatt and D are already known to the controller.

Basic phase current limiting in controllers is as simple as that, please don't look further!

Pat

 

[ZapPat]

I should say PWM can be used to limit Iphase. When it is used in that way it can no longer control speed. It can control one or the other NOT both.

It is true that a controller cannot purely control both speed and torque from the same throttle, but you are maybe forgeting that microcontroller-based motor controllers aren't necessarily limited to all one or all the other. For example, you can use both relationships at the same time in your firmware so that your throttle acts as either a speed limiter or a torque limiter depending on which variable has to be limited at any moment for a given throttle position, speed and Iphase. I believe kelly controllers can do this already, but I'm not sure since I don't have one (but I seem to remember other's posts about this feature).

 

[sabb]

This needed bumping because i had to look for it
also maybe the title needed more caps so it stands out better than the other one

 

[TylerDurden]

This needed bumping because i had to look for it... also maybe the title needed more caps so it stands out better than the other one

Try using the subscribe button at the top (or bottom) of the page, for topics you wish to follow. They will be listed in your User Control Panel> Manage Subscriptions.

 

[sabb]

So I am new but I don't think I am stupid

So I wanted to ask if some of my preconceptions are reality or not

a) With a permanent magnet motor you have a fixed magnet strength to work with
b) In the winding your magnetic field is based on amperage in the coils multiplied by number of turns
c) I thought/think that at some point raising the strength in the windings is wasted because torque doesnt rise or at least rises at a much reduced rate

if these are correct then when you reach this torque saturation point any amperage beyond this is just a good way to burn out a motor

so clearly we need to limit phase amperage but it will be different for each motor and winding

Btw I tried not to use rotor and stator because it depends on design and none of thoughts apply in a non permanent magnet motor so I use windings and magnets

Thanks in advance for any comments

Stephen

 

[sabb]

Tyler

Thanks

But I think my point was made in the other thread when the originator of this thread thought that thread was this one.

 

[rhitee05]

a) With a permanent magnet motor you have a fixed magnet strength to work with

Correct, although some advanced control techniques make use of what's called "field weakening" which, as the name implies, allows you to use some of the field current to oppose the magnets (lower magnetic field allows higher speeds for the same motor, although with lower torque). But, in most cases yes, the field strength is constant.

b) In the winding your magnetic field is based on amperage in the coils multiplied by number of turns

Also correct. The construction of the core has some effect as well.

c) I thought/think that at some point raising the strength in the windings is wasted because torque doesnt rise or at least rises at a much reduced rate

Yes, the magnetic material in the core (stator laminations, for example) will saturate at a high enough field strength. The field will continue to increase, but at a much slower rate so you're not getting very much more torque for a lot more heat. The saturation point depends on the material, design of the laminations, number of turns, etc. so will be different for different motor designs.

Btw I tried not to use rotor and stator because it depends on design and none of thoughts apply in a non permanent magnet motor so I use windings and magnets

Your second two points also apply to non-permanent magnet motors as well.

 

[sabb]

Eric

I understand field weakening from my days working on non permanent magnet motors as the field power on a shunt or compound wound dc motor could be reduced to raise speed at expense of torque.

I did not know there were ways to do this on a permanent mag motor but dont think that is going to happen on a off the shelf hub motor

When i said torque was related to phase amps time number of turns I was talking about again an off the shelf motor where we dont have control over the lamination quality or shape just the turn count we buy or wind and the amperage we control

and the saturation I was expecting was when the strength of the magnetic field from the windings starts to exceed or far exceed the strength of the permanent magnets

Which is why I did not think they applied

So I would rather keep this as practical as possible

In a permanent magnet hub motor geared or direct drive i want to get as much torque as possible but I dont want to push past the point of torque rising reasonably well with the phase amps because i think it is a waste

Then I think it would be usefull to figure out the max torque from each motor (this seems like it might have been done at ebikes.ca)
and then figure out the phase amps needed to achieve that for each known winding of that particular motor

also if I understand anything about this thread we also need to think limits for each controller

After that what phase amps to put in your controller is just a matter of look on chart for motor, look on chart for controller, find lower of two number and set phase amps (edit maybe think about driveline or plastic gears in geared hub motor too)

This should affect everyone that can set phase amps because phase current is often higher than battery current when you are pulling off the line and have battery current limiting in effect and most setups will hit battery current limits with no bemf

I need sleep now but will prob not get any thinking about this too much

Stephen

 

[John in CR]

Let me tack on another question. Is the stator's saturation limit affected by the windings or just the iron? eg Do all X5's have the same limit, or would an X5302 be different from an X5304 in terms of stator saturation? Also, a saturation at 90A was mentioned for the 9C, but aren't some people running 90A battery side limits with those motors?

 

[liveforphysics]

Let me tack on another question. Is the stator's saturation limit affected by the windings or just the iron? eg Do all X5's have the same limit, or would an X5302 be different from an X5304 in terms of stator saturation? Also, a saturation at 90A was mentioned for the 9C, but aren't some people running 90A battery side limits with those motors?

The saturation is all about flux levels in the tooth. The iron has a limited amount of magnetic field it can guide.

So, the flux in the tooth is about the current*number-of-turns. So, if the 2-turn can handle 200amp phase current before it saturates, then the 3-turn saturates at 150amps, 4-turn would saturate at 100amps, the 5-turn at 80amps, 6-turn at 67amps, etc etc.

 

[rhitee05]

The saturation is all about flux levels in the tooth. The iron has a limited amount of magnetic field it can guide.

So, the flux in the tooth is about the current*number-of-turns. So, if the 2-turn can handle 200amp phase current before it saturates, then the 3-turn saturates at 150amps, 4-turn would saturate at 100amps, the 5-turn at 80amps, 6-turn at 67amps, etc etc.

Exactly correct. I believe this means that a given motor design is capable of a certain amount of torque, although different windings will require different amounts of current to reach that level.

I did not know there were ways to do this on a permanent mag motor but dont think that is going to happen on a off the shelf hub motor

The field weakening is entirely a function of the control algorithm, so is possible on any BLDC motor (or any permanent-magnet motor). Look up field-oriented control for details. The only hardware requirements are enough processing capability to do the math, and the ability to measure phase currents directly. So, you won't be doing it with a normal controller but with a more capable controller you could do it with any motor.

and the saturation I was expecting was when the strength of the magnetic field from the windings starts to exceed or far exceed the strength of the permanent magnets

The strength of the magnets does obviously affect the amount of torque produced, but it doesn't set the saturation limit. Mathematically, it's probably easier to think of this as a function of the Lorentz force (due to current in a magnetic field, F=ILxB) rather than due to two fields attracting/repulsing each other.

After that what phase amps to put in your controller is just a matter of look on chart for motor, look on chart for controller, find lower of two number and set phase amps (edit maybe think about driveline or plastic gears in geared hub motor too)

That would be a good approach, although with the number of motor and controller variants out there it would be a pretty big chart!

 

[sabb]

So more questions

If we take an example like a 530X hub do we know at what point the lamination core saturates and is this the point where the torque cure changes drastically?

If the lamination core saturates first then magnet strength doesn't matter that much I had thought that the magnets would be the limiting factor.
If this is the case then improved core design would be only way to make motor achieve higher torque.

Did I also understand that the simulator at ebikes.ca does not simulate the saturation and that some of the torques at zero speed are too high as stated?


And Two Thoughts

The chart of motors with max phase amps is not so hard because if we know the numbers for one of the 530x line of hubs we can calculate based on turns what each of the others would be.
This means we need real test number for one member of each line of motors.

If I got the whole thread correct then when I go to buy a controller I am trying to buy a controller that will survive the PWM stage phase currents.

Thanks again
Stephen

 

[Miles]

If the lamination core saturates first then magnet strength doesn't matter that much I had thought that the magnets would be the limiting factor.
If this is the case then improved core design would be only way to make motor achieve higher torque.

Unless the core morphology is suspect (bottlenecks), you could change the ratio of iron to copper, which would lower efficiency and hence continuous torque, or use a core material with a higher saturation point. Or start with a bigger motor size....

 

[sabb]

Miles would increasing copper fill to the max possible help?

I thought if the core was saturating then nothing but changes to the core would be helpfull.

Do different Lamination materials have large differences in where they would saturate, Earlier in life I helped with building brushed rc motors and our choice of lamination materials where more driven by hysteresis losses at the higher rpms. It would take a much bigger punch press for a hub motor lamination


Stephen

 

[Miles]

Miles would increasing copper fill to the max possible help

It would help efficiency but wouldn't affect the saturation point, AFAIK.

So

I thought if the core was saturating then nothing but changes to the core would be helpfull.

I agree.

 

[Miles]

Do different Lamination materials have large differences in where they would saturate, Earlier in life I helped with building brushed rc motors and our choice of lamination materials where more driven by hysteresis losses at the higher rpms.

I think you can get much higher flux densities with Cobalt alloys but the cost rarely justifies its use. Using a larger core to keep the flux density below 1.6T usually makes more sense.

 

[donob08]

I have one last idea I keep forgetting to offer. If the Cycle Analyst had an input for throttle voltage, throttle position would be a good indicator of the value of D in the PWM control. Since the Cycle Analyst knows Ibatt it could determine a good approximation of Iphase. If the Iphase were displayed, the rider just has to unwind the throttle until Iphase gets to an area he is comfortable with.

With an ICE most drivers don’t lug down the engine by asking for more torque than the engine can supply at present speed, because they can sense the lugging down. With the electric system just provide a signal of motor problems, overheating. The “fixable” part of the high Iphase problem is when the operator is trying to get more speed out of the motor than it can accelerate to at the speed it is running. Displaying Iphase seems like a simple fix and still allows being able to make the decision to allow short term heating in order to get up a hill.

I think using Iphase = (DVbatt – Vbemf)/Rphase tells us an awful lot. If we write Vbemf as KVConstant * Speed,
Iphase = D/Rphase – (KVConstant/Rphase)* Speed then you can see how the linear negative slope (KVConstant/Rphase) predicts the change in Iphase after Vemf becomes dominant, that is the motor is out of saturation.

 

[Miles]

Do different Lamination materials have large differences in where they would saturate...

Ref: http://www.cartech.com/techarticles.aspx?id=1624