Hub Motor Design Simulations

@Hummie, you would need two boards, one using the same motor as a hub drive and the other (wound to a proportional higher KV) within a geared system. You could simply do a comparison, riding parallel @ same speed. You don’t want to compare apples to oranges…So you use the same motor. A thermistor inside the motors would tell you a lot about losses. A range test will also tell a lot. This would be the ultimate comparison in a real world environment.

Frank

Yeah, i agree on that, its a compromise, the only solution on that case is to go to a geared hub like Stary, or implement a geared transmission like @Nowind did on their builds

I think the main goal we have to achieve is reliability, without that our boards are just toys, mine right now is pretty efficient running on a 230kV 10S with 3:1 gear ratio, i get 8Wh/Km on my commute, full throttle for 90% of the time and with two big hills, i’m weight something like 80Kg + backpack, bit i literally can’t afford($$$) the amount of belts i’ve been using

Yeah, i have access to something of that type, will look into it, thanks, for now i will focus my effort on using VESC, my codding abilities is only on basic to intermediate level, maybe in the future or with one of yours ESC’s

It’s a lot of work but it’ll be worth it. In the end, stary has the best in wheel motor. Even if it’s not a real hub and is super loud

I hope so. The noise in the Stary is the major drawback, but with a few tricks such changing the gear profile and making it work partially submerged in oil would do the trick to quiet it down

I am starting to take a look on geared setup that i said, it will be necessary a dumb gear in the middle so the motor does not colide with the truck and i have to test if it can fit a dual motor

Also, i’m diggin a little deeper in Emetor but getting really weird results and its a lot of work to test a new design, you have to run each speed separated

Edit: just a test to see what a geared setup would look like, this is with a 190Kv 50mm motor and geared to 35Km/h

1 Like

Is there any real data on their motor? Anyone dyno it? Maybe someone can send me one to dyno

I wish. But inrunner are more efficient than inrunners and gears are also a win.

If you do, i would be impressed. They have a one wheel drive.

Inrunners being more efficient I guess you mean iron less. Core less. I think most efficient would be coreless. Id like to see how much torque u can get out of a coreless motor. Or skip torque and use its speed and a gear to get the power could you? A high speed inrunner coreless (or not coreless) geared hanger. 180mm or so of motor space

Coreless is a pain in the ass cause it’s totally submersed or immersed in epoxy. A lot of heat buildup. I toyed with the idea.

I would say, gears are badly needed and a good planetary gear box is well expensive. These gears will always be noisy unless they run in oil and have curved teeth. It gets a bit complicated and expensive in my eyes. Belts are simple, efficient, quiet and for the $$$ of a good planetary gearbox you can buy a lot of replacement belts and pulleys. A good belt cover would sort out most problems.

But its also really fun to think of new solutions. Otherwise we would still be in stone age. My feeling is: The Outrunner motor is pretty much at its peak now and and still we see to much inefficiency and problems. A major design tweak would need to be implemented. Most customers don’t understand the Hub motor issues and will buy 1:1 hubs anyway, especially if they are cheap. It’s o.K. when they are taught that hubs go hand in hand with compromises. Basically a matrix would help to make the decision “geared or hub”. This way advantages and disadvantages of both systems would get known and you can make your decision on facts rather than hype. The simulations @anon94428844 does are really helpful. @anon94428844: Could you simulate the same motor (with proportional different KV) for 1:1 Hub and a 1:2.5 geared setups? Maybe that would shine some light on the issue and will help to develop better hubs.

Frank

2 Likes

Frank, of your motors, they are made to custom parameters? Like stator format, etc?

One think that I imagine happens is most of the motors uses readily made stators and geometry to lower costs, maybe the outrunners that use are no that efficient, but the gating mask the drawbacks of that

One step would be teardown a bunch of motors and simulate them, or assembly a dynamometer and test, them we can compare, I think this quest to a usable hub can lead to even beter geared system, optimizing the motors to our specific use to have peak efficiency where we need and a custom torque curve, something on the shape that of the Mellow hub

Hi Pedrodemio, our motors use an existing Stator, but that’s about it. Everything else has been reworked and optimized. We tested the motor on dynamos at different voltages and loads. 12S setups are at peak performance close to 90% efficient, so you won’t be able to optimize that. The efficiency curve is very flat already. Torque is linear to input power. These motors are already very very efficient.

Anyway, design a hub, rewind it to higher KVs and use it within a geared system. You will see that there is a square relation between copper losses and torque@RPM As Benjamin said: double the torque, decrease the RPM by half and your losses will get 4xhigher. There is nothing you an do about hat except building a bigger motor. You can optimize things up to a certain point, sure. Magic won’t happen unless you come up with a radical new motor design that will spin the Magnets at 14x lower ERPM rate. That’s gearing again.

Frank

2 Likes

Good work, do you have a picture or something about the dynamometer used? If it’s not something confiditial could you share the results for one of your motor?

Thanks

When you say the motor is 90% efficient on 12s how would the voltage effect the motor efficiency? Frank I guess you meant a design that will spin magnets at a faster rate at the end of what you wrote When you say you can make torque more linear…how is this possible other than through the esc? The motor as I understood will have a torque to speed and efficiency graph that’s pretty set in stone with these motors and it’s more so what speed you’re at that decides where you on on the graph, as your vedder quote/law states.

Tesla car motors are potted. It need not add heat.

http://www.evtechexpo.com/media/news/2015/02/26/how-proper-applications-of-thermally-conductive-materials-will-improve-motor-power-density/ A coreless motor need not be potted anyway

Cost per 100mm wheel?

You can make the torque linear in both ESC and Motor. You just need to know how the ESC behaves and tailor the motors performance to that.

I don’t know what most of this means… But I hope to at some point in the future, so, from my future self, thank you!

300$ for two pints of the best stuff I could find. Buy larger go cheaper

http://www.lord.com/products-and-solutions/electronic-materials/thermoset-sc-320-THK-thermally-conductive-silicone

I have never see a motor that inherently had a more linear torque curve. What would make a motor design have such a thing?

Huh, not bad price wise. Could try doing it at home. Might be great as an alternative to Starys inrunner as a mod.

Well consider it from a calculus perspective. If the motors peak torque is something like 50Nm at 0.000001RPM, then the curve drops steeply and plateaus before approaching zero then that segment of the curve is spot on.

The window/frame of reference were concerned with is 0-4000 RPM with hub motors.

Plotting a DTC curve is a lot easier than an SVM.