Motor Efficiency Question (theory)

[ProEV]

I am trying to understand what affects motor efficiency. Please feel free to correct any misconceptions I might have.

I am speaking of motor efficiency as the motor’s ability to change electric energy into mechanical energy.

In other words, an 90% efficient motor would lose 10% of it’s energy converting electric energy to mechanical.

My understanding is that for any motor RPM, there is an ideal current to produce the highest efficiency at that RPM.

There is also, of course, a specific RPM/Current combination to give the highest efficiency that the motor is capable of.
So, for maximum motor efficiency, at various speeds, the current should be set to the RPM.

What would change this curve?

Does Voltage, aside from when the voltage is too low or too high to allow the motor to achieve that RPM at that current?

Does Load, aside when the load is too low or too high to allow the motor to achieve that RPM at that current?

What about motor temperature? Would that change the RPM/current curve?

 

[spinningmagnets]

The hotter the copper windings, the more resistance they have, and efficiency drops. Therefore, it is useful to spec a motor that is big enough to easily do the job, but small enough that it is affordable, fittable, and as light as possible within the performance constraints.

A large portion of the 10% IN-efficiency in your 90% efficient motor is waste-heat. There are several sources in the motor that convert battery watts into heat.

 

[liveforphysics]

Anytime the copper losses, or the resistive heating losses of the motors phase windings are experiencing more loss (heating) than the core losses of the motor, the efficiency improves by increasing RPM. Anytime the core-loss (Iron hysteresis, eddy currents, windage drag, etc) exceeds the copper losses, then the motor is more efficient to spin it more slowly with increased phase current.

Unless you're powering a dentist drill, odds are you ultimately want to turn something like a wheel with reasonable torque at a few hundred to a thousand RPMs. Picking a motor that costs more stages of gearing power conversion quickly accumulates more powertrain system losses than a slower motor that makes high specific torque. Direct drive or a single power conversion torque multiplying stage is generally the best compromise for systems chasing efficiency (hence why almost every solar race car uses direct drive hubs).

 

[ProEV]

Thank you for your replies. Let me see if I am understanding them correctly.

liveforphysics's explains that the right current for best efficiency at a specified RPM has to do with the balance between core and winding losses.

spinningmagnets points out that the hotter the copper windings get, the more resistance they have. This implies that the balance between core and winding losses will change and thus the best efficiency current at a specified RPM will change with motor temperature.

I wanted to explore this more, so I went to ebikes.ca and their great motor simulator tool (Thanks Justin and company!).

https://ebikes.ca/tools/simulator.html

I set Human Power to 0 and X axis to RPM. Otherwise, no other changes. I selected Run Simulation Set, with Throttle as the variable (since this will vary the current), between 80% to 90% and 11 simulations and checked All Data. I downloaded the results and dropped them into Excel.

Next, I clicked on Show Advanced and changed motor temperature from 60 C to 200 C. I then ran the same simulation set, downloaded the data, and put it on another page in my Excel file.

Using a fairly randomly chosen 285 RPM, I discovered that at I got best efficiency at 16.06 amps (80.958%) for the 60 C page. (Row 7975 for those of you playing at home).

Finding the best efficiency for the 200 C page, I found that the best efficiency was less, and it was at a different current: 78.465% at 14.25 amps.

So, to my question: What about motor temperature? Would that change the RPM/current curve?

The answer is yes. So, if I want to match my current to RPM for best efficiency, I need to correct for Motor Temperature.

I also asked: Does Load change this curve?

I go back to the Motor Simulator Tool. I reset Motor Temperature to 60 C. I set Grade to 30% since going uphill will increase the Load. I run the same Simulator Set, download, and put it in a new page.

The load has increased from 436 W to 3211 W but the efficiency and current for that efficiency has not changed.

Answer: Load does not change the RPM/current curve.

My other question was Does Voltage change the curve?

I reset Grade to 0. I change Battery to Custom Battery and accept the default of 38 Volt, .2 Ohm, 8 Amp hour and run the simulator set. Download, put in new page.

Then I modify only the Custom Battery Voltage to 36 volts, run the simulator set, download, and put in new page.

This time I use 249 RPM.
For 38 Volts, I get the best efficiency which is 79.749% at 18.15 amps. (Row 3071)
For 36 Volts, I get the best efficiency which is 80.254% at 14.70 amps. (Row 4273)

Does Voltage change the curve? Yes.

So my conclusion is that if I want to accelerate matching the current to the RPM for best efficiency, I must also correct for Voltage and Motor Temperature but I can ignore load.

Does this make sense? Have I missed anything else I should be correcting for?

 

[liveforphysics]

Iron saturation of the stator has a knee where torque is still produced with increased current, but on a decreasing slope with respect to the model scaling 1:1 with increasing loads.

If you want to take it a step further and think about the machine rather than the numbers alone, you're making a force on a lever arm. You can make that force at the tip of the lever a stronger pull by increasing amp-turns on the tooth, but only up to iron saturation limits, with diminishing returns approaching this threshold.
Gears can simulate a longer lever arm packaged into a smaller radius, but with losses to pay for each stage of power transfer.

Increased motor radius buys you this longer lever arm, with a longer array of force producing EM to drive that longer lever.

The losses that are mandatory in a motor are nill (>99% has been done in solar car motors). The losses we accept are due to a series of compromises in packaging size, cost to mfg, materials selection, etc. Due to these compromise choices to make a product that works with the budgets of folks making ebikes, and the efficiency plots and loss curves reflect this.

 

[ProEV]

It is interesting to get more of an idea about the physical design that controls the motor's performance.

I have never had an oppertunity to work with a 99% efficient motor. I wonder how much is due to a better build such as closer tolorances and better materials, and how much is down to having the motor operate in a very specific performance window?

I am running a Astro Flight 3220 4 turn motor (https://www.astroflight.com/3220-astro-brushless-motor-details.html) which claims up to 95% efficiency. To my eye, it is a solid well made motor. I use a single gear to the rear wheel.

I am racing in Electrathon. We have a rule limit to battery capacity/weight, so efficiency is important. The more energy that gets turned into motion, the faster you will go during the race.

The most fun tracks have a variety of corners. High speed. Low speed. Off camber. Dimishing radius. Our speeds vary corner to corner which means the operating RPM window is large. It is my experience that efficiency is far from 95% at lower RPMs.

I run a simple program that reads RPM from the controller and then limits maxium current to the ideal for efficiency at that RPM. As a driver, I only have to focus on modulating my throttle based on how much traction my tires have because full throttle will be the right current.

To improve that program, it sounds like I will need to take into account voltage and motor temperature.

 

[BalorNG]

It is interesting to get more of an idea about the physical design that controls the motor's performance.

I have never had an oppertunity to work with a 99% efficient motor. I wonder how much is due to a better build such as closer tolorances and better materials, and how much is down to having the motor operate in a very specific performance window?

Both, obviously.
And of course, without gears you'll have considerably higher losses outside of this narrow performance window.

With solar racing (which can be done entirely in simulation to be fair, there is very little direct competition between riders involved AFAIK) that is done in very controlled environments you can fine-tune your motors and transmission to operate in most efficient regine 99% of the time. Electrathon seems very different.

Using deraileur gears with motor seems like the most efficient way if you ask me. This is something that is still unrivaled in highly competitive environment where engines are also 'power limited' and operate with highest efficiency in narrow RPM gap - human power.

Since electric motors are not AS sensitive to changes in load/RPM, you don't exactly need 12 speeds, but a wide range 8 speed cassette will likely be more than enough. You can even easily set up automatic gear shifting using a servo and arduino.

 

[liveforphysics]

I wouldn't call it a narrow envelope.

This is a cheaper version of the Cisro:
https://renew.org.au/wp-content/uploads/2018/12/marand_high_efficiency_motor.pdf

They offer a 97% steel back-iron version and a 98% Halbach version.

This is a LEAF motor efficiency plot. Its pretty easy with a single speed to stay above say ~94% for your whole race once the vehicle is moving.

 

[ProEV]

Using deraileur gears with motor seems like the most efficient way if you ask me. This is something that is still unrivaled in highly competitive environment where engines are also 'power limited' and operate with highest efficiency in narrow RPM gap - human power.

Since electric motors are not AS sensitive to changes in load/RPM, you don't exactly need 12 speeds, but a wide range 8 speed cassette will likely be more than enough. You can even easily set up automatic gear shifting using a servo and arduino.

I can see why gears are important for human power. A number of teams have tried using deraileur gears in Electrathon, but I do not know of any team who has had success with them. The extra friction seems to offset any advantage in gearing. The additional complication is also a challenge in a racing environment.

This is a LEAF motor efficiency plot. Its pretty easy with a single speed to stay above say ~94% for your whole race once the vehicle is moving.

I am impressed by how efficient the LEAF motor plots out. Even at low speeds and high torque, it is over 90% efficient. I have not been able to find an efficiency to current plot for the Astro Flight 3220 4 turn motor we are running, but I am pretty sure it falls off much more, out of it's happy zone.