Windings and torque dilemma

[pdf]

Background: I've been trying to understand how winding count, current, and voltage affect low speed hill climbing ability in DD hub motors. I've been focusing on torque because a bike at a standstill would seem to be an extreme case. If the bike has enough torque to start up a hill, the ability to feed power and shed heat will determine the equilibrium speed up the hill. I want to dabble in hub motors, but want to do so in an educated way; I have a BB drive currently but just bought a 9x7 that was on sale at Cycle9 because... well, because I could. Seemed like a good place to start experimenting.

Dilemma: I quote the following from ebikes.ca (who I think everyone recognizes as leaders in the ebikes area):

Is there a tradeoff between speed and torque?
No, there is a frequent misconception that the hub motor windings are similar to gears in a conventional drive chain, and that as you go to a lower speed you get more torque or visa versa. This is generally untrue. If you were to increase the speed of your motor by increasing the battery voltage, then both the speed and the torque of the motor go up. If you increase the speed instead by leaving the voltage the same and going to a lower winding count motor, then you do get more torque at higher speeds, but ever so slightly less torque off the line because of secondary losses in the wiring and controller.

Contrast that with a recent posting on a thread here that compared 9c motors of different windings in a table of torque and speed vs count:
http://endless-sphere.com/forums/viewtopic.php?f=31&t=24515&start=45#p355632

I would have concluded that both speed and torque are strong functions of winding count, which seems to be contrary to the information on the ebikes.ca web site. In fact, they are acting exactly like gears; the torque goes up in direct proportion to the winding count and the speed, down. I have assumes it was max torque in the table I cited but maybe not. I suppose it could be the torque at max speed although I don't know why you would want to know, except to calculate power required.

Any insight is appreciated.

 

[Drunkskunk]

Justin's statment on the website is true. But only up to a point. In general, a higher wind count motor won't make any significant amount more torque at the same voltage as a low wind count motor, at the same voltage. As an example, and taken from the Ebike.ca site's simulator. the 2807 makes around 33 lbs torque at 36v 20A. The much faster 2805 makes 30lbs torque. Not much dfiffrence.


But what that doesn't take into account is the efficancy at which they can produce that torque, or the potential of torque the slower motor has when the voltage is increased.

 

[Toorbough ULL-Zeveigh]

u know how sometimes certain noobs will get sloppy or lazy & will talk about amps when from context it's clear they actually mean amp-hours?
it's my feeling that this confusion about torque & winding-count is the result of a similar units-dropout that's taken root over time.

this incorrect statement:

I would have concluded that both speed and torque are strong functions of winding count

can be made true with the addition of the required missing units.
both speed per volt and torque per amp are functions of winding count.

the amount of torque a motor produces (not the torque per amp) remains fixed for a given size motor.
and by 'given size', a motor's size or ultimate capability (maximum torque it can create) for an otherwise same configuration regardless of winding-count is essentially dictated by the total amount of copper it contains.

that's my only insight, hope it helps.

 

[dogman dan]

While this is pretty dang unscientific, I have ridden a few different motors using similar volts, amps, etc to tend to agree with Justin.

When riding in really extreme conditions, like dirt at 15 degrees, I don't feel a lot of difference between two windings torque. But you notice right away that one motor heats up more than the others. It's the efficiency thing. For climbing steep hills, which hub motors hate, the best bet is slower and using higher winding count motors. But you can still burn out motors abusing even the slower wind motors.

My rule of thumb, is pick a speed, then pick a motor that doesn't exceed that speed at the voltage you have chosen. Pointless to choose a 5303 that goes 30 mph when you intend to ride 18 mph.

 

[John in CR]

For a given motor, assuming an equal amount of copper in the windings, the maximum torque does not change with a change in the number of turns in the windings. What does change is the amount of torque per unit of current and as does the rpm per volt. Reducing the turn count increases the rpm/volt (Kv), but decreases the torque per amp of current. Because the copper is thicker for each turn and the wire isn't as long due to few turns, the windings can handle more current, making the motor faster for a given voltage, but still capable of the same maximum torque as long as the controller can deliver the greater current required.

That sounds great at first, but here's the problem when it comes to hills or heavy loads. You have to be able maintain a higher speed, or the motor will be operating in a much lower efficiency range and create lots of waste heat which will burn up the motor. To maintain a higher speed requires more power.

Our motors reach peak power output at 1/2 of no load speed. At that point and above, the motor is operating in it's range of good efficiency. In addition, at and above that max power point is where the controller can operate at full duty instead of current limiting, and that is easier on the controller. I've found that a good rule of thumb is that as long as you can maintain 1/2 of your bike's maximum speed on the flats while going up a hill, then you aren't over stressing your system. That's where the higher turn count motors are better going up hills, because they hit their maximum power point at a lower speed. The power required to climb hills at a give speed increases directly with the grade of the hill. The power required to climb a given hill also increases directly with speed. That means the higher turn count motor can safely climb a steeper hill, because it needs to maintain a lower speed.

If you're thinking, "no problem, I'll just use the faster low turn count motor, and just use less throttle so I can safely climb steeper hills at a lower speed.", I'm sorry but that won't work. Partial throttle up steep hills is guaranteed to either blow your controller due to current limiting or melt your motor. The lower throttle position imitates a lower voltage, which moves the max power point lower, but the shift in the performance graph isn't a direct match.

Take it from someone who lives in a mountainous area and has explored the performance limits using hub motors on hills. If your bike can do 30mph on the flats, then you better be able to maintain 15mph up a long hill or the hill is too steep or your load is too heavy and you better put some real effort into pedaling to get up to that 15mph. Also, only use 100% throttle going up significant hills. While you are exploring the limits of your bike's system, after long climbs immediately stop and feel the hub motor cover. If it's too hot to hold your hand on it for more than a second or two, then you need more speed up the hill. The same feel test goes for your controller, though it's thermal limits are lower.

 

[liveforphysics]

Controller load is a big difference.

For example, a 5302 and a 5304 both have equal copper fill, yet one has twice the number of turns.

Let's say both motors on are identical bikes both climbing a hill side-by-side at 25mph. Both motors are drawing say 50v at 50amps (2500watts). So the current load from the battery and power-in on both motors is identical.

However! There is a huge difference going on inside the controller. If the 5302 motor has half the phase resistance, so to get the same input power, it only needs to get half the voltage. This means it's PWM% from the controller is 50% lower than the 5304 bikes controller. Since both controllers have identical input power, yet one is outputting half the voltage, for the equation to balance it means the 5302 bike's phase current is 2x higher than the 5304 bike.

The controller does not feel battery current, only phase current, so in this situation, both bikes are chugging up the hill together drawing the same amount of battery current/voltage/power into the controllers, but one controller is trying to deal with outputting TWICE the current.

In the above example, this means the identical 50amps at the battery could easily be 100amps phase current for the 5304 motor, and a controller melting 200amps for the 5302 motor. Also, to avoid wire resistance losses, the 5302 would need twice as big of phase wires etc.

With a big brushless hubmotor, you're burst (NOT continuous) acceleration/torque is effectively always just limited by what the controller can supply. So, recognizing a fixed controller phase-amp limitation, using a motor with additional turns enables higher torque to be produced in real-world situations. In a perfect world situation where the controller has no phase current limitations, we could all be running 1-turn motors with 00-awg cables for phase wires, and enjoy great torque everywhere and great top speed. However, in practice, high turn count motors enable some pretty cool things for applications where you want a controller to be the size of a cell phone, and get rock solid reliable monster torque, with the only downside being sacrificing speed range.

So, ultimately it simply comes down to application. In practice, picking the motor with the most turn counts that is still able to achieve the top speed your application requires results in the most reliable and robust, wire/controller/system package, but it's not because of the motor itself, but the effects that motor has on the rest of the systems reliability.

 

[John in CR]

Liveforphysics,

While in your example both motors have the same output power to go up the hill at the same speed with the same load, and we're in 100% agreement on what goes on inside the controller, I don't believe the input power will be the same. I'm under the impression that the 5304 at 50V and 50A will not be running at the same efficiency as a 5302 at 25V and 100A, which is what the controller will be simulating.