Combining 2 hub motors into one?

DanGT86

100 kW
Joined
Sep 6, 2012
Messages
1,179
Location
Saint Louis MO
I ended up with 2 identical 8" scooter motors. I know they arent the best motors in the world but they were free so I'd like to experiment with them. So my plan is to machine a new shaft and combine them into one motor. I'm figuring 2 controllers is best.

So here are my questions about doing so:
  • How close together can the magnet rings and stators be before they interfere with each other's field in a substantial way? Assuming I run too controllers here.
  • Is there anything to be gained by timing the windings and the magnets' poles to each other since the motors are identical
  • Is it possible to parallel each winding to its partner on the other stator and run them with one controller as if the windings were one

In the pic below the magnets are shown. Looks like I can get them a little over an inch from each other without too much trouble. I believe things get a little weird at the ends of the turns so I'm curious how bad of an idea it might be to try and sych everything as if it was one motor. This seems like a good excuse for me to play with a dual ESC but I like that 2 super cheap 6fet controllers would probably be fun at the practical limit of these motors anyway.
motor section.jpg
 
Mechanically, you might look at Farfle's magicpie motor thread, where he combines two of them to make one motor (though he rewound it as a single double-width stator).


For your setup, electrically, the two motors will be in sync, because they are mechanically attached and magnet rotors rotating identically. I would recommend for simplicity's sake to align the rotors such that the magnets are lined up with each other. I don't know whether aligning them with fields opposing or matching will work best; that's only one magnet position different so you can experiment with that if your rotor-to-rotor attachment method allows that.

Also for simplicity's sake, setup the stators so they are lined up the same with each other.

I don't know how the end-turn fields will affect each other, whether they are run by just one controller or two separate ones. That's something you could experiment with, using different distances from each other (but this would complicate your mechanical build, requiring different spacers between the rotors, and different spacing on the axle for the stators, for each test).

For operational power/etc consideration having the two motors mechanically connected like this will be effectively the same as having them separately on the same vehicle with separate controllers.


You can parallel the windings and run from one controller...as long as everything in both stators is line up the same. If they're not, then what they try to do may not line up with the rotor magnets at the same time and it may not operate correctly or at all. For this case you would also want to ensure the stator magnets are lined up as the same polarity, too.

But you'll have half the phase resistance and twice the current needed to drive each phase; not sure what it will do to the inductance / how the controller will react to that (so it will need to be a controller designed to at least drive a single motor that big ).

There have been some dual-stator motors actualy built before, I think even commercially, so it's been done successfully.






DanGT86 said:
I ended up with 2 identical 8" scooter motors. I know they arent the best motors in the world but they were free so I'd like to experiment with them. So my plan is to machine a new shaft and combine them into one motor. I'm figuring 2 controllers is best.

So here are my questions about doing so:
  • How close together can the magnet rings and stators be before they interfere with each other's field in a substantial way? Assuming I run too controllers here.
  • Is there anything to be gained by timing the windings and the magnets' poles to each other since the motors are identical
  • Is it possible to parallel each winding to its partner on the other stator and run them with one controller as if the windings were one

In the pic below the magnets are shown. Looks like I can get them a little over an inch from each other without too much trouble. I believe things get a little weird at the ends of the turns so I'm curious how bad of an idea it might be to try and sych everything as if it was one motor. This seems like a good excuse for me to play with a dual ESC but I like that 2 super cheap 6fet controllers would probably be fun at the practical limit of these motors anyway.
motor section.jpg
 
I was thinking so much about the stators earlier I may have been overcomplicating it. When you mentioned magnets it clicked. All I really need to manipulate for my dual controller experiment is the rotors' position. So if I build in enough adjustment to clock them by one whole magnet width I can try any position from poles aligned to poles opposite. I suspect the difference would be negligible.

As for unifying the windings for a single controller that seems like a way more complex issue. If I had to guess I would suspect that the stator teeth should be positioned identically such that the fields are aligned as if they were one giant winding. I guess its just the magic and the end turns that is a total mystery to me.

Consider this hypothetical scenario for using one controller.
  • Stators and magnet rings are precisely mechanically aligned and magnet polarity is matched.
  • I test one stator to find the hall and phase combo that works best. I mark the phase wires A,B, and C.
  • I test the 2nd stator to find the phase combo but I don't move the hall wires from the first stator test. I mark the phase wires A,B, and C
  • I parallel both A wires both B wires, and both C wires.
Will this be effectively paralleled for use with one controller?

Or do I need to actually align the stators such that the winding patterns match each other perfectly? I'm assuming these are wound identically but I don't know for sure until I open the 2nd motor. Do the employees/machines at the factory start winding on the same tooth or at random?

It's hard for me to visualize this. At first I thought it would require matching the exact winding pattern. But now the more I think about it the field repeats every 3 teeth right? So the physical path the windings follow might be irrelevant as long as the energized stator teeth are the same across the 2 stators?
 
It should be that simple, yes. :)

You *should* just be able to set it up with say, the "halls on top" lined up with each other, both facing the same way (not facing each other, or facing away from each other), if they wound the motor starting from the same point, in the same way.

Will require a test to find out, though.

If you have magnets aligned same on both rotors, you can test without a controller.

Once it's all bolted up, but separated phase wires by stator, use a small current (amp or two) on the "first" phase pair of one stator. The rotor should "jump" to that alignment.

Disconnect that and now put the current on the same pair of phases of the other stator. The rotor shouldnt' move from the previous alignment (if it stayed in place during the wire move) if the two stators have the same winding pattern for that phase pair.

Do this with a different phase pair, one stator at a time, and again it should move to alignment on the first one, and not move on other stator.

You don't need to do the third one as it's eliminated by the first two, but you can if you want just for experimentation's sake. :)
 
Cool.

It would be nice if one controller ran the whole thing because I have several controllers laying around here. Sadly they are all different sizes. Buying a dual motor VESC is a $200-300 solution for my "free" motors.
 
If you only have two motors, well... you're going to need at least two wheels. I don't see the need to conjoin these twins.
 
Chalo said:
If you only have two motors, well... you're going to need at least two wheels. I don't see the need to conjoin these twins.

Sorry, I forgot to mention its going to be used as a mid-drive. My kids have one of those mongoose dirt-bike "bicycles" with the brushed currie/razor looking motor. It's obnoxiously loud and surges at low speed because we are torturing it by running it at 40v when its supposed to be 24v. Id much rather run it in the realm of brushless with regen like I am used to.

I will be machining the tire mounting sections off these motors to make them smaller and lighter.
cut motor 2.jpg
motor cut.jpg

calab said:
Why not have one bigger motor that has two sets of windings?
;)

That is my logic as well. I just don't want to go through the trouble of combining them if there is some electrical reason they will be way less efficient.

These motors were designed for 350watts at 36v and about 750rpm. I can run my 10s packs in series and they go about 1500rpm. I have some ferrofluid here already from another hub project. I also have a controller that will be fine at 30-40 batt Amps and 20s. 1500RPM allows me to gear them for a top speed around 25-30mph using off the shelf bike parts. 22t chainring and 52tooth rear cog.

So my thinking is that if I combine them for one controller I can have a pretty reliable 1500-2000amp peak minibike without spending any real money. If combining them for one controller doesn't work then I am buying 2 smaller controllers or a VESC and it becomes harder to justify putting labor into these motors. For example, I could spend that controller money on a $250 Lightning Rods small block motor and have an absolute monster with the controllers I already own.

The combined phase winding parallel project is pretty intriguing to me though.

I have a freewheeling bottom bracket setup in my parts bin as well. I think one of these motors in a mid drive setup would be a nice little crank drive system. I'm sure I can find a use for them sometime. Solid little motor that is reasonably waterproof and should be pretty reliable I think.
 
As long as the rotors and stators are aligned properly, you can parallel the windings and just use one of the hall sets. You can put the rotors as close as you want. Just make sure the rotor magnets are aligned the same before joining. One way to do this is to apply a steady DC (like 12v) to a pair of phase wires and let the rotor align. Do the same to the other one using the corresponding phase wires. This should align the magnets the same on both.
 
Cool. So assuming I grab 2 phases at random to do this test. The positive and neg from my DC source makes those 2 phases unique and the 3rd phase is not used in the test. So does this mean all the info I need to parallel them is all gained from just one test?

So just to check my understanding here. I first align the magnets so the polarity matches between the 2
pos of 12v source I call phase A
Neg of 12v source I call phase B
Leftover unused phase wire I call Phase C

do the test on both stators and when the combo is is found that produces the least amount of movement between the 2 rotors I have a definitive answer on which wires will be A B and C?
 
DanGT86 said:
Buying a dual motor VESC is a $200-300 solution for my "free" motors.

Free-parts projects are often enough the most expensive ones. ;)

(cost of parts (and labor / time!) for adapting suboptimal parts one "already has" to a job instead of just getting the right ones...)
 
DanGT86 said:
Cool. So assuming I grab 2 phases at random to do this test. The positive and neg from my DC source makes those 2 phases unique and the 3rd phase is not used in the test. So does this mean all the info I need to parallel them is all gained from just one test?

So just to check my understanding here. I first align the magnets so the polarity matches between the 2
pos of 12v source I call phase A
Neg of 12v source I call phase B
Leftover unused phase wire I call Phase C

do the test on both stators and when the combo is is found that produces the least amount of movement between the 2 rotors I have a definitive answer on which wires will be A B and C?

The phase wires will have colors, so just use the same colors/polarity on both of them to align the magnets.
Second test is to power up the hall sensors and write down the signal states. 3 hall signals, so one or two will be high. Just make sure the hall signals look the same on both (you could do this to align the rotors without powering the phases too)
 
Back
Top