kubark42 said:
My application is to swing a propeller, and I'm asking in this subforum because the torque and RPM requirements are similar to motorcycles.
Ah. What size propeller, at what RPM? (it wont' mean anythign to me, but it will to others here familiar with aircraft of various types, and they may be able to help you more directly knowing the specifics, even potentially giving you a "better way" of doing something, if you're open to that). And what drive mechanism from the motor to the prop?
What voltage system at what expected current?
At the time I wrote the original post, I hadn't yet learned that hubmotors come in two variants, one
direct-drive and
the other geared. I am specifically asking about the geared variant, as it is what's required to get the high efficiencies at high torque/low output RPM. Apologies for the confusion.
Makes sense. But the best of the geared hubs is meant for about 1kW, maybe 2kW, and some have been pushed a little harder than that for short bursts. Part of that is because of heat dissipation problems (see paragraph after next), and part of that is mechanical limitations.
The best geared hubs I know of are the GMAC, the MAC, and the BMC. The Ezee probably follows after that. Then it's a bit of a jump down to the next level, like the BPM and such. There aren't any I've ever seen made for motorcycle power--the ones I've seen have all been DD hubs, and the best of those I've seen in person are QSMotor, or customized versions of their motors by other companies or individuals.
There aren't any geared hubs I've ever heard of that are likely to do anything near 10KW...unless you perhaps completely replace the planetary assembly in there, including their roller/ramp clutches, with something that's capable of dealing with the massive heat and torque you're going to have to deal with.
A geared hub has three layers, with air gaps between each, and each of those air gaps is an *EXCELLENT* insulator for heat, with the part that really needs to be cooled at the very center of the tootsie pop...er, motor.
So it is very easy to cook a geared hub with overheating. You could greatly open up the outer shell (which may mechanically weaken it) to allow airflow onto the middle layer, which is the planetary and clutch and rotor bell with magnets. So those would be a little cooler, depending on the airflow you generate thru it. But the stator is inside that rotor bell, and getting airflow thru that is harder, especially since AFAICR it is counter-rotating to the outer shell, so any holes you have in the bell will be chopping up the airflow thru it and disrupting it, so no nice easy way to help force flow thru there.
It's not impossible...just not easy.
Some people have used various liquids to fill geared hubs to help them cool faster, but this does add drag (whcih increases the power required to spin at any particular RPM, and thus increases the waste heat), and requires better seals on the bearings and motor wiring to prevent leakage than what these come with. I dont' know how mcuh good it would do at the power levels you're after. There is a thread by Justin_LE about Definitive Testing of HEating and Cooling of Hub Motors that has actual testing of various things, not just speculation, so you can check that out, and the http://ebikes.ca/tools/simulator.html can also help you see some limitations of various geared hubs that have been characterized (it's not perfect, but it's very useful as a baseline).
I'll assume that you're going to be using the spoke flange of the hub, or the disc brake mounting holes, to bolt a sprocket to to use a chain or belt to then drive your prop? (because you won't be able to mount a prop directly to the housing of hte motor--I doubt it has the strength for it, and it probably has far too much runout in various places and would shake the prop to pieces).
Theoretically you can spin a small motor really fast to get higher power out of it, then reduce the RPM to torque via an external reduction (maybe your application wouldn't want the reduction,b ut rather keep the speed?)....but motors will have increasing losses (internal heating, wasted power) the faster you spin them beyond their design. I am not sure, but I expect you would have to spin one of these hubs something like 10x or more faster than it's designed for to get the kind of power you want out of it, and even if it managed to mechanically hold together (no promises
) then the heat from the losses could be high enough to damage it (winding insulation, hall sensors if you're using them, magnets, plastic gears if you don't replace them, etc).
I had originally discarded using planetary gears because custom gearing is prohibitively expensive. (Most OTS planetary gearsets are seen in car transmissions, where they wind up being too heavy and impractical for my application.) A planetary gearset is very capable of carrying high torques, though, thanks to its balanced construction. This a balanced load is also a lot easier on the motor bearings. However, 100Nm is an awful lot more than most applications deal with, and the simple plastic gears in a 300W bike motor won't survive.
That's the problem--there aren't any all-metal planetaries in any geared hubmotor I have ever seen, or any of them with a clutch that's all that great. (I think the clutch problem is one reason the GMAC by Grin Tech has no clutch, and thus can't freewheel--the motor is always engaged with the shell, just like a DD hubmotor).
So you'd probably have to have the gears made for you, or choose a motor that has metal gears (of the quality you need) already available (if there is one; some motors have offered the ability to change to one steel gear but leave the others plastic (because they're incredibly loud, and wear the sun gear faster, probably due to poor tooth/etc design, using all three steel gears has only been done by a few DIYers over the years that I can recall, and I think that might actually be in the Cyclone middrive rather than a geared hub).
You might also have to have the sun gear (often, maybe always, part of the axle itself) remade to match the new gears (if you ahve to change their profiles, etc), and the ring gear too (sometimes part of the shell, but usually a separate "track" that slips into the shell).
That ring gear may be another issue if you need a lot of torque. Some geared hubs have had failures where the ring spins inside the shell, having broken loose from whichever method was used to pin it to the shell. I dont' recall which ones.
Without being a mechanical engineer, and without knowing the details of your setup and intent, I would guess that it is going to be "easier" to use an external planetary reduction (that *is* designed for the task you have for it) on a motor that is designed for the power level (and necessary cooling) that you're after. I could be wrong...and to the hacker in me, it would be very interesting to see an ebike geared hubmotor driving a prop on an aircraft with really high power (relative to the design intent of the motor)...but I just think heat buildup and mechanical issues in a geared hubmotor are going to be problems that will take more re-engineering than it will be worth.
P.S. I'm drawing a blank with what previous query you're referring to. Perhaps you've confused me with someone else? https://endless-sphere.com/forums/viewtopic.php?f=30&t=109899 is the only other topic I've started here ever. I didn't sense much interest in having the technical discussion I hoped for, so there wasn't a need to beat a dead horse. I got better traction with the same question at https://www.rcgroups.com/forums/showthread.php?3800343-Extracting-maximum-power-from-a-BLDC.
Ah. That is the thread.
Many people at least come back to comment on the answers they get, if they have any interest in the help they got; and typically when they dont' come back it means they abandoned the idea and/or the forum, or they just don't care. That's not universal...but it's common.
FWIW, there are a lot of times someone creates a thread asking a fairly technical question that is going to need a lot of work to answer, and I (or others) will start with a basic reply (often asking or implying questions for the OP to answer) to see if the OP is going to participate in a discussion, before spending all that time and work answering completely, or researching things for them, etc. When the OP doesn't come back, at least that time wasnt' all wasted, taht way.
(most of the time, they don't...they may start a new thread about a different idea, or never come back, or evne start a new thread about the same idea because they didn't like the answers they got the first time, etc)).
Most of that has nothing to do with you...sorry.
Just thought you should know why I had said what I did. (I guess I'm too old to "get" the typical internet behaviors and expect more sociability than the average person has to give.
)
so while I can help "keep it real" 
thanks a bunch....
