When does efficiency matter?

I've read a few threads, usually related to range, and people bring up efficiency of the motor(s). Some folks say a motor's peak efficiency is at full speed, and some say a percentage of that, for instance. Sometimes the discussion is related to efficiency of a particular motor through it's operating band, and other times it's discussed while comparing motors. I've always equated efficiency more like power converted to motion vs heat; but not correlating to range, which I believe is really only a function of the rider's throttle hand and/or willingness to pedal.

Playing with the Grin simulator, it does appear as though the efficiency curve is always ascending up to top speed. This can be seen in System A, which I modeled for my bike. System B is the system in each case, except adjusting the amount of throttle being applied by the rider.
Notice efficiency went up, not down, because the efficiency curve isn't static, but shifts based on the motor input voltage (speed).

Throttling down even more and efficiency peaks (at around 30.3mph) then starts decreasing.

Finally, throttling down further, the efficiency drops back down to being almost equal to the efficiency at full throttle.

The third simulation really shows the disconnect between efficiency and effect on range. The efficiency in both cases is nearly identical, yet system A travels 21 miles on a charge, while system B travels 137 miles. Basically, if you want range, it doesn't matter much how efficient the motor is, relative to the impact of your throttle hand has. This also seems to be somewhat consistent with my motor temperature testing, since I've found throttling down to a certain speed range (more efficient?) allows for longer steep climbs without overheating, and slower or faster than that speed makes the temps rise more quickly.

I'm going to play around with one of the Grin motors next, since those have the temps modeled.

I think this is also explained nicely in one of Justin's presentation videos.

I've always thought drag (wind resistance/drag coefficient/etc.) was predominate to such a degree that motor efficiency really didn't play much of a part in things.

Diggs said:

I've always thought drag (wind resistance/drag coefficient/etc.) was predominate to such a degree that motor efficiency really didn't play much of a part in things.

Click to expand...

Exactly. The impact on real world observations/performance is negligible due to all of the other factors.

Motor efficiency matters to the degree that the waste fraction is discharged as heat into the motor. That means when you're optimizing components, a more efficient motor can be pushed to higher power before overheating.

This isn't relevant to me personally, because I like to have lots of headroom in my components to handle more power than I ever ask of them. But if you are either trying to maximize the performance of a given system, or shrink the mass of a system to attain a given level of performance, small percentage gains in efficiency can yield large net benefits. At the same number of input watts, an 82% efficient motor makes twice as much heat as a 91% efficient motor. In principle that would allow twice as much maximum power, or half as much mass of motor to do the job, at the higher efficiency.

It can also be a very subjective thing, and you need to define the system as a whole. Or selectively define it. A commonly referenced value is that even modern cars have an efficiency of only 25-40% when measured from the amount of energy in gasoline, versus how much energy is used to make the car go forward. The rest goes to heat and noise. Historically that hasn't mattered, both because gasoline has been traditionally given a free pass, and also because, it's so energy dense, so what if you waste 60-80 percent of its energy?

Or solar panels. The new ones you can get commercially are about 22% efficient. A couple decades ago, they were 10% efficient (when measuring the total possible energy available in a given area of sunlight being absorbed). So, more than double increase. That's pretty cool, but if you just focus on how there's another 75% out there that you're not getting... Anyway, just a matter of perspective.

Electric motors seem pretty efficient (~80%) compared to solar panels (~22%) and LEDs (~15% at 100 Lumen/Watt).

Gruesome said:

I think this is also explained nicely in one of Justin's presentation videos.

Click to expand...

The video is worth a watch, but at the same time the tldr is "stop obsessing over efficiency."

My "rule of pedaling" to you know you are "throttled efficiently" if pedaling a tiny bit harder lowers your amps. If adding pedal power doesn't drop amps, I know the motor is still in a "generating excess heat" condition.

Gruesome said:

Electric motors seem pretty efficient (~80%) compared to solar panels (~22%) and LEDs (~15% at 100 Lumen/Watt).

Click to expand...

There are theoretical maximums that don't compare to each other, just like for Carnot cycle engines, wind turbines, etc. They should all be compared against their respective theoretical maximums rather than against each other. After all LEDs can't eat gasoline, solar cells can't eat wind and so forth.

Reasons you should care about efficiency:
Inefficiency eats into power delivered, range, and continuous motor power.
It also possibly wastes a small amount of fossil fuel resources.

By optimizing for efficiency you also optimize for power, and therefore enjoyment, which is what we actually want when we're done calculating things.

Chalo said:

There are theoretical maximums that don't compare to each other, just like for Carnot cycle engines, wind turbines, etc. They should all be compared against their respective theoretical maximums rather than against each other. After all LEDs can't eat gasoline, solar cells can't eat wind and so forth.

Click to expand...

No, they absolutely should be compared. Comparing them is a good way to show that there's quite a bit of room for improvement for photovoltaics and for light emitting devices. Electric motors on the other hand seem a relatively mature technology. Doesn't mean they can't be improved, but not by a factor two.
Wind turbines couple very weakly to the wind, and they have to, otherwise people would really complain....

What really matters in all this is real world examples....and I will share a few of mine:

I used to own/ride a 2011 Stealth Fighter. It went through various iterations over the years from it's stock battery (52V, 20AH) through to a LiPo setup (66V, 25AH), but I was always having to monitor/manage range due to it's relatively low efficiency (for and ebike) of 20-30Wh/km. Swapping in a more efficient Leaf motor and keeping it actively cooled with internal fans helped, but I was still limited in where I could get to, having to load it up in the car to do more adventurous rides 20-50km from home.
I got sick of that so I built a 52V 49AH battery. That reduced my speed a bit (which further helped with efficiency) and with 49AH on tap, I was never able to get near it's capacity, even riding 20-30km to get to the start of trails I wanted to explore. In theory, it had over 200km range. I never got near that though....and that was the point....I didn't need to worry about it, or it's efficiency as I 'solved' that problem with a massive capacity battery.

Another example is my BikeE CT. I built a 52V 17.5AH battery for it back around 2016. Now 8 years later it has over 600 cycles and barely manages 12AH in the cold. Rather than replace it years ago like I should/could have, I put a large wind deflector (solar panel) on the front, changed controller to a slightly more efficient Grinfineon (from an old Infineon), and changed the motor from a 27mm Golden motor with 0.35mm laminations to a more efficient RH 212. The solar panel adds more in aero efficiency than it does in charge. On a typical ride I see about 4-15Wh added from the solar panel, but the cruising speed (40kph) efficiency gain is about 50W continuous lower consumption on flat ground from the aero benefits.
The combination of all that means that I can still manage about 50-60km range even with the degraded battery. This has effectively extended the life of the battery by at least 50%. I will still need to replace it at some point, but for now this works better.

Cheers

neptronix said:

Reasons you should care about efficiency:
Inefficiency eats into power delivered, range, and continuous motor power.
It also possibly wastes a small amount of fossil fuel resources.

By optimizing for efficiency you also optimize for power, and therefore enjoyment, which is what we actually want when we're done calculating things.

Click to expand...

I agree, but to have those results you need system efficiency, not motor efficiency. When it comes to range, motor efficiency doesn’t get you there, since where the motor is most efficient isn’t where the bike is most efficient (if defined by most distance from a battery charge, rather than the coolest running motor).

E-HP said:

I agree, but to have those results you need system efficiency, not motor efficiency. When it comes to range, motor efficiency doesn’t get you there, since where the motor is most efficient isn’t where the bike is most efficient (if defined by most distance from a battery charge, rather than the coolest running motor).

Click to expand...

Motor choice is the largest variable in system efficiency; it's bigger than the effect of battery, controller, and wiring!
Most ideal is if you pick a motor where your cruising speed lines up with the sweet spot in the motor efficiency curve. This of course requires some futzing with the ebikes.ca sim or implying things from a dyno graph, but the right choice for your scenario can vary the efficiency by 3-10%.

Maybe the goal isn't maximizing range in all cases, what we optimize for depends on the rider and their preferences ^_^

If fun goes up with speed, and range goes down with speed squared, then there should be an optimum for any given rider's brain chemistry, frontal cross section and preferred posture

neptronix said:

Motor choice is the largest variable in system efficiency; it's bigger than the effect of battery, controller, and wiring!
Most ideal is if you pick a motor where your cruising speed lines up with the sweet spot in the motor efficiency curve. This of course requires some futzing with the ebikes.ca sim or implying things from a dyno graph, but the right choice for your scenario can vary the efficiency by 3-10%.

Maybe the goal isn't maximizing range in all cases, what we optimize for depends on the rider and their preferences ^_^

Click to expand...

Absolutely correct. Based on how the efficiency curve is defined by Grin, it includes the motor and controller losses (but no battery). So the motor and controller should be two areas susceptible to failure due to heat, if the battery doesn't die first. Sadly, while we'd all like more efficient motors, the choices are narrow.

Preface for the simulations below: with respect to changes various parameters, which to change and how much to change them is based on looking at the definitions and making some guesses. The changes were large enough that they should have some visible impact on the results (e.g. 2X, 3X, 1/2, 1/3 changes). Not saying that the changes are realistic.

Controller and battery efficiency are represented in their respective internal resistance parameters.


Conclusion - There are a lot of significant benefits to better mosfets and cells, but assuming they can perform per the simulator parameters, those benefits don't really impact the simulation.

I don't know where to start with custom motor parameters, and it's hard to replicate the Leaf using the custom motor parameters, but I can come close.

Adjusting the winding resistance and inductance make a noticeable difference in performance.

At equivalent speeds, the improved motor results in a range increase as well.

Interim conclusion: Improving motor efficiency can potentially have a large impact on torque and speed. However, I have to read more on how to make reasonable adjustments to the parameters since my arbitrary changes are likely costly, difficult to achieve and generally not feasible, but used to see the effect of the changes.

The biggest free performance gain, both for speed and range, is to lean forward.
Automatic 24% increase in range, 5% increase in top speed.

Or normalized to the same speed for an apples to apples comparison, a 38% increase in range.

Conclusion - the aerodynamic benefit at higher speeds, provides a greater performance gain than many other factors combined. (Too bad I don't have the guts to ride a recumbent trike in traffic).

Gruesome said:

If fun goes up with speed, and range goes down with speed squared, then there should be an optimum for any given rider's brain chemistry, frontal cross section and preferred posture

Click to expand...

Does fun go up in a linear fashion with speed or is it a more complicated function?

biphaeboh said:

Does fun go up in a linear fashion with speed or is it a more complicated function?

Click to expand...

I think that depends on the rider. But if I'm having the basic math right, then as long as one (fun) rises and the other (range) falls, there should be an optimum for any weighting function of the two. It could be going very slow or very fast, but there should be one.
You need to ride and find out!

When I'm pedaling I want a good bicycle experience. I've pedaled tens of thousands of miles on pedal bikes. With an ebike I want the power of the motor underneath and invisible. On a motorcycle I want all of the power under my fingertips.
There is this elegant beauty of a bicycle where all the power is invisible to the user.

Gruesome said:

I think that depends on the rider. But if I'm having the basic math right, then as long as one (fun) rises and the other (range) falls, there should be an optimum for any weighting function of the two. It could be going very slow or very fast, but there should be one.
You need to ride and find out!

Click to expand...

Need to map out the shape of the curve. Obviously it will differ between riders.

gromike said:

When I'm pedaling I want a good bicycle experience. I've pedaled tens of thousands of miles on pedal bikes. With an ebike I want the power of the motor underneath and invisible. On a motorcycle I want all of the power under my fingertips.
There is this elegant beauty of a bicycle where all the power is invisible to the user.

Click to expand...

Yeah, I don't want my ebike to feel like a motorcycle. Bicycling and motor cycling are different aesthetic experiences.

E-HP said:

I've read a few threads, usually related to range, and people bring up efficiency of the motor(s). Some folks say a motor's peak efficiency is at full speed, and some say a percentage of that, for instance. Sometimes the discussion is related to efficiency of a particular motor through it's operating band, and other times it's discussed while comparing motors. I've always equated efficiency more like power converted to motion vs heat; but not correlating to range, which I believe is really only a function of the rider's throttle hand and/or willingness to pedal.

Playing with the Grin simulator, it does appear as though the efficiency curve is always ascending up to top speed. This can be seen in System A, which I modeled for my bike. System B is the system in each case, except adjusting the amount of throttle being applied by the rider.

Motor Simulator - Tools

Our ebike motor simulator allows you to easily simulate the different performance characteristics of different ebike setups - with a wide selection of hub motors modeled, and the ability to add custom batteries and controllers and set a wide variety of vehicle parameters you'll be able to see...

ebikes.ca

Notice efficiency went up, not down, because the efficiency curve isn't static, but shifts based on the motor input voltage (speed).

Throttling down even more and efficiency peaks (at around 30.3mph) then starts decreasing.

Motor Simulator - Tools

Our ebike motor simulator allows you to easily simulate the different performance characteristics of different ebike setups - with a wide selection of hub motors modeled, and the ability to add custom batteries and controllers and set a wide variety of vehicle parameters you'll be able to see...

ebikes.ca

Finally, throttling down further, the efficiency drops back down to being almost equal to the efficiency at full throttle.

Motor Simulator - Tools

Our ebike motor simulator allows you to easily simulate the different performance characteristics of different ebike setups - with a wide selection of hub motors modeled, and the ability to add custom batteries and controllers and set a wide variety of vehicle parameters you'll be able to see...

ebikes.ca

The third simulation really shows the disconnect between efficiency and effect on range. The efficiency in both cases is nearly identical, yet system A travels 21 miles on a charge, while system B travels 137 miles. Basically, if you want range, it doesn't matter much how efficient the motor is, relative to the impact of your throttle hand has. This also seems to be somewhat consistent with my motor temperature testing, since I've found throttling down to a certain speed range (more efficient?) allows for longer steep climbs without overheating, and slower or faster than that speed makes the temps rise more quickly.

I'm going to play around with one of the Grin motors next, since those have the temps modeled.

Click to expand...

The whole motor efficiency and range thing can be a trip. It's wild how efficiency curves climb with speed but then plateau or dip. You're spot on about the rider’s throttle hand being a huge factor for range. That Grin simulator is pretty dope, though—it really helps visualize how tweaking the throttle can affect efficiency.

Another face of (in)efficiency is heat generation. When riding long steep slopes, a less efficient system will overheat or shut down before reaching the top. This cannot be resolved with a bigger battery like the range issue.

I was wondering if the way to avoid the overheating in the above situation is to run full throttle? Running at highest possible rpm for high efficiency. It seems counter intuitive to me.

Avner.