Hub motor RPM limits? ( bearings? )

[neptronix]

Folks, i got an interesting idea and i am curious..
It seems that the faster you spin an electric motor, the more power it makes per a given amount of magnet and copper.

I'd love to take a fairly bone stock, super low turn count hub motor, dump tons of volts into it, and run it through something like a 2:1 gear reduction via chain to the rear wheel.

Seems like you could theoretically run 2x the constant rated wattage this way. It would be like lacing a motor up to a 13" wheel, but running it in a 26" wheel effectively.

I'm wondering if there would be any limitations in terms of bearings being shredded by this? or anything else?

[Farfle]

The eddy losses, and hysterisis losses start stacking up quick at higher rpm. If you want to do it, just look for a motor that has a good design, a low pole count, and thin lams, JohninCRs mini motor is a good candidate.

[neptronix]

The eddy losses, and hysterisis losses start stacking up quick at higher rpm. If you want to do it, just look for a motor that has a good design, a low pole count, and thin lams, JohninCRs mini motor is a good candidate.

Well dang it, now you opened up more questions

What's a good motor design then? low pole count and thin lams?
Just me, or does that not really explain most hub motors? .. they seem chock full of poles, as many as possible. And laminations... i don't know..

The ebikes.ca simulator shows tremendous amounts of torque and stupidly high continuous power handling running a 16"-10" wheel with a high turn count hub motor. Like a 404 doing 3700w constant, a 5302 doing >3500w constant, a HS3548 doing ~3000w constant, and a 9C 5 turn doing ~2600w constant. all motors producing within 150-250ft-lb of torque. Holy crap.

[Lebowski]

It's kind of what a Mac is doing, running a small internal motor at high rpm combined with a 5:1 ratio...

[John in CR]

Yep, Farfle nailed it. The practical rpm limit of our hubbies is determined by the iron and the electrical rpm. I've found the common designs having 51 slots, 46 magnets, and .5mm laminations are limited from a practical standpoint to somewhere around 900-1000rpm. Sure you can go higher with better cooling, but efficiency falls off fast enough that it's not worth the effort. At the high end it's all about how well the stator iron handles the moving magnets with alternating poles passing by, which causes the eddy currents and hysteresis losses in the iron.

John

[John in CR]

Check with Justin, but it's pretty obvious to me by simulating high voltage and tiny tires that the iron losses aren't adequately considered in the simulator.

John

[spinningmagnets]

Stock bearings are pretty cheap, and upgraded bearings are fairly affordable (if you want to experiment).

The customers for the popular hub-motor kits are too picky about the lowest possible purchase price. One example is how a temp probe would be easy to install into the stator during assembly, and the probe itself is only a couple dollars, and any customer could choose to add the temp read-out when and if they wanted. Why don't they all come with a factory installed temp-probe...simply because it saves a couple of dollars on the purchase price.

What I'm getting at is that Johns fat 6-inch DD scooter hub has laminations made from a higher quality of metal, and the lams are thinner so there are more of them in the same space. Most hub-kits don't bother because it would cost more, and customers would buy the cheaper hub from a competitor. Johns hub was designed for a 48V scooter (like a 50cc Vespa), but it would work fantastic as a non-hub using anything from 48V up to 100V.

If you up the volts to get higher RPMs, and then gear it down to drive the rear wheel (like the Stokemonkey), you shouldn't need lots of amps. I imagine you'd get lots of torque from doing that. I think the biggest issue for any practical system is the cost of a reasonably-sized battery at the higher volts. If you price out several 72V options, I think you'll agree. Get a 100V Lyen 12-FET, one of Johns hubs ($500-ish with shipping from CR), and two 48V 15-Ah Pings (96V)...and you will have plenty of issues to iron out long before you have any trouble with the 45-MPH bearings that the scooter hub comes with.

[amberwolf]

Assuming you have a controller that can do it, then the easy way to find out if a typical hub motor can do what you want--run it at that higher voltage and RPM, under a load (dyno? brake?) high enough to cause it to draw the current you expect it to at the actual riding conditions. See if it starts heating unduly, and you'll have an idea if eddy currents are a real problem yet or not.


Catch might be finding a controller to do the voltage you want, and the commutation rate you want. Liveforphysics would be one of the best people to do this testing, as he already has access to a dyno and likely controllers that can handle whatever is needed, as well as power supplies and/or batteries to run them. Just need to get him the motor you want to test.

[John in CR]

SM,

I don't know about Pings with these. The factory controllers do 50A peak, and the matched 18fet controllers I sell with the motors are 70-80A peak. Running lower current serves no benefit. It reduces acceleration performance, resulting in a bit more heat, ie lower overall efficiency because a longer time is spent accelerating. For steep hills or big loads, simply put it in low to keep current under control and efficiency high.


AW,

The testing is already done. Everyone who tries for high speed with a hubbie runs into heat problems and it's not all load related. It took a while for it to click for me, but since they're designed for 48V or less operation it makes perfect sense that lesser iron is used. It's not needed. Hell, most of the manufacturers (the 2 most popular on ES being the best examples) don't even bother to dynamically balance their motors, since they aren't designed for sufficient rpms to make a difference. My 9C was pathetically out of balance.

John

[Miles]

It took a while for it to click for me, but since they're designed for 48V or less operation it makes perfect sense that lesser iron is used. It's not needed.

The limiting factor is mostly the lamination thickness at 0.5mm. For your example, at 1000rpm, the fundamental frequency would be 383Hz.

More iron usually means more iron losses...