Definitive Tests on the Heating and Cooling of Hub Motors

[spinningmagnets]

Every engineering choice is a compromise. It's possible to make a leaf-blower very quiet while being electric, you just have to design it while using quiet air-flow principles, and compromise on other features.

Second-hand performance evaluations are used as currency here on ES, and I am guilty of that too, but...there is no single "silver bullet" E-bike system that is perfect for all applications. I have grown to be fond of non-hub builds recently. And yet, I recognize that a simple hub-kit that is well-matched to its application is often the less-expensive option, along with being fairly trouble-free and pretty quiet (if quietness is important to any given E-biker, some hubs are quieter than others, YMMV).

That being said, it's a given that for some E-bikers, adding some type of cooling is more trouble than its worth. However, I am just as certain that adding some type of cooling feature to some E-bike motors is very valuable.

Justin is performing an incredibly useful service by performing empirical data testing, and then publishing it (rather than keeping it "in-house"). If anyone fears that all future E-bike motors will have cooling that adds to their price, and your particular application does not require the additional feature...rest assured than non-cooled hub and non-hub motors will remain available at the slightly lower price.

 

[migueralliart]

I just thought about this idea and it might be a stupid one but nevertheless here I go;

How about instead we drill out a small passage on the axle that is solid. It could be a small one or perhaps a new design axle but hear me out.

You could drill holes in both sides of the covers and pump air thru that axle hole which will be released inside the hub and then out of the hub thru the covers.

Alternatively you could coat the innards of the motor with a heat emissivity coating so that it helps out

The problem is trying to find a small enough source of compressed air. Perhaps it can be activated via a temp switch inside the hub so that it only works when it gets past a certain temp.

Alternatively this could be done in future motors with water. Say for example you have an axle that's designed with two passages and a hollow stator center part that acts as a radiator. You could pump water thru one side it will cool the stator center by moving along the legs of the stator and then exit on the other side. You could theoretically use a small water pump similar to what they use on windshield washer pumps and route the water to a somewhat small radiator outside. Now this might sound a little extreme BUT it might be an appropriate way to cool hubs for motorcycles and not really bicycles.


Just my 2 cents.

 

[madmaxNZ]

I just thought about this idea and it might be a stupid one but nevertheless here I go;

How about instead we drill out a small passage on the axle that is solid. It could be a small one or perhaps a new design axle but hear me out.

You could drill holes in both sides of the covers and pump air thru that axle hole which will be released inside the hub and then out of the hub thru the covers.

Alternatively you could coat the innards of the motor with a heat emissivity coating so that it helps out

The problem is trying to find a small enough source of compressed air. Perhaps it can be activated via a temp switch inside the hub so that it only works when it gets past a certain temp.

Alternatively this could be done in future motors with water. Say for example you have an axle that's designed with two passages and a hollow stator center part that acts as a radiator. You could pump water thru one side it will cool the stator center by moving along the legs of the stator and then exit on the other side. You could theoretically use a small water pump similar to what they use on windshield washer pumps and route the water to a somewhat small radiator outside. Now this might sound a little extreme BUT it might be an appropriate way to cool hubs for motorcycles and not really bicycles.


Just my 2 cents.

I'm not convinced on the air system (may sound a bit like an air hose with a hissing leak), but I like your thinking about the coolant system with a radiator. If you haven't seen this water setup...check it out:
http://www.endless-sphere.com/forums/viewtopic.php?f=6&t=26029&start=175

 

[migueralliart]

That's exactly what I was envisioning except with a couple of changes;

The stator center part has to be made to for this purpose not adapted.
Lesser diameter tubes.
Smaller radiator.
Temp switch for the pump.
Radiator placed on downtube and a long narrow profile.

Going with a setup like this will be almost IMPOSSIBLE to smoke a motor due to temperatures with a system that's designed properly.


The trick would be to make the stator center part hollow so that the cooling cavities are integrated into it.



EDIT;

Now that I thought about this idea a bit more it can also be accomplished by machining a disc shaped to fit the stator center or the majority of it and then use thermal paste and bolts to adhere it to the stator. Then you'd have the line coming out of the disc and the axle shoulder. This way it can just be an addon.

 

[2handy]

There is quite a variety of oil coolers made for motorcycles that are long & thin.

 

[John in CR]

That's exactly what I was envisioning except with a couple of changes;

The stator center part has to be made to for this purpose not adapted.
Lesser diameter tubes.
Smaller radiator.
Temp switch for the pump.
Radiator placed on downtube and a long narrow profile.

Going with a setup like this will be almost IMPOSSIBLE to smoke a motor due to temperatures with a system that's designed properly.


The trick would be to make the stator center part hollow so that the cooling cavities are integrated into it.



EDIT;

Now that I thought about this idea a bit more it can also be accomplished by machining a disc shaped to fit the stator center or the majority of it and then use thermal paste and bolts to adhere it to the stator. Then you'd have the line coming out of the disc and the axle shoulder. This way it can just be an addon.

That kind of stuff has been tried even with pretty good efforts, but with relatively poor results. If you want to do liquid cooling then that liquid needs to flow over what's getting hot, the copper and the stator steel at the copper. Heat transfer is a relatively slow process and there's too much resistance if you have to wait for it to transfer through the through the metal of the stator to get to your reservoir of liquid to go to the the radiator is a fail because the delta T's we're working with are too low.

For a liquid cooling system to work really well, then you need to combine the oil fill method of the past year with pumping the hot liquid out to a radiator of some type. It shouldn't be too difficult to make changes to the motor guts to make it safe for something like ATF. Ideally we'd need a long overdue sensorless system that really works, so we can get rid of the halls and their wiring, leaving just magnet wire, stator lams, and magnets inside the motor.

John

 

[Sancho's Horse]

There needs to be some modifications to the wires themselves as fechter and johnrobholmes and probably others have alluded to, but you can get a good seal if...you can get pressure equalization. That is where I am really excited about this thread.

If Justin et. company have researched an adequate valve solution from Gore (which they are the company likely to have the valve know how), then this is a big deal for oil fill strategies.

Justin mentioned some bearing leakage in his other maker faire thread, but I really think even this can be effectively reduced...with some interior mods, and the largest share of the oil actually residing outside of the hub. I believe this enough to be spending my time slushing aluminum lately. Aaaargh matey! Available tech we be! Until proven. Aaargh!

 

[migueralliart]

I think filling the entire hub motor is overkill. Didn't you just say that all you need to cool is what is hot (near the copper)?

Perhaps is better to update the first or second post of this whole thread with the temporary findings.

 

[bigmoose]

Two more thoughts on oil cooling. It is pulling the heat out of the end turns of the windings. Then for a radiator, think of the already available frame tubes.

 

[madmaxNZ]

I think filling the entire hub motor is overkill. Didn't you just say that all you need to cool is what is hot (near the copper)?

Perhaps is better to update the first or second post of this whole thread with the temporary findings.

I can't help thinking from some of the posts that some poeple are thinking the whole hub has to be filled over the axel or that trying to seal the motor is difficult.
Mine is filled to a level below the axel (approx 2/3 the Radiuus of the hub, approx 200ml). This has dramatically improved the cooling for my purposes. I'm not doubting Air holes, but I can only comment on what I have experienced since this is all I have tried.
I get a very small amout of seepage from my breather bolt hole, that forms a "centrifical film", (bolt/breather pictured here in earlier post PAGE 7). This is nothing to wipe off the hub after a ride; my bike is usually covered in mud or dust, so this little bit of oil film is a non issue for me. I guess for a commuter/town bike that is kept otherwise clean, it may seem like an "embugerance". My H4080 bearings and lip seal do not leak, as they only have oil splashing over them in operation. I store and transport the bike in the upright position, but if left laying on its side, I would imagine oil would come out around the axel or breather hole.
I've got a design in mind regarding the breather to eliminate this seepage: I would drill a small hole along the centre of the axel, on the freewheel side, to near the centre of hub, then tap another hole down to meet it; these form the vent to the outside. Thread in a "walking Stick" shaped pipe with a flare on the hook end (long enough to reach up as far as possible inside the hub). Oil dripping down would run over the hook and not be inclined to travel back up the flared opening. Any oil that did manage to get into the pipe would drip out the end of the axel on the freewheel side.

My build so far: http://endless-sphere.com/forums/viewtopic.php?f=3&t=50133

 

[itchynackers]

Max, that is a good idea. I've been running 600 ml of oil for almost a year now (almost half full). My perception is that as the motor heats up, the air inside expands. This will cause the oil or air (or both) to find the weakest spot to leak from. On my bike, this is the axle (unless I press my relief valve button). It is minor leakage like yours. Not enough to worry about, although it would be nicer if it didn't leak. Its such a small amount, that I don't even relieve the pressure too much anymore. But the few times I do, I can hear the hot air rushing out for a second. Ultimately, it would be nice to have a gore-tex type snap in automatic valve. If I knew how to get a few, I'd probably try them.

 

[arkmundi]

My proposed solution to over-heated hub motor:

Really, after considering various options, I believe the best solution is let the thermistor shut the controller down (Heinzmann setup), get off my ebike and use a spray-bottle to cool the motor more quickly.

 

[robbie]

There needs to be some modifications to the wires themselves as fechter and johnrobholmes and probably others have alluded to, but you can get a good seal if...you can get pressure equalization. That is where I am really excited about this thread.

If Justin et. company have researched an adequate valve solution from Gore (which they are the company likely to have the valve know how), then this is a big deal for oil fill strategies.

Justin mentioned some bearing leakage in his other maker faire thread, but I really think even this can be effectively reduced...with some interior mods, and the largest share of the oil actually residing outside of the hub. I believe this enough to be spending my time slushing aluminum lately. Aaaargh matey! Available tech we be! Until proven. Aaargh!

Unfortunately, most of the ePTFE ("goretex") vents that we've seen are not suitable for being continuously coated in oil. Sourcing a product with a hydrophobic exterior coating and an oleophobic interior coating would be ideal. The hub motor on the bicycle that I toured down from Vancouver with oil filling had an average temperature of 30-40 Celcius lower than Justin's sealed motor on hill climbs. However it lost oil the entire time as the vent I installed has failed to work.

How are you envisioning the oil being placed outside the hub? I'd prefer to have the mass of oil in as much contact with the copper core as possible! On a side note we found that filling the hub with oil (150mL) had a minimal effect on the no-load current.

 

[zappy]

Hi Robbie, do you remember any of the numbers for loss? At what sort of speed or revs?
On a 9C that has a 27mm wide stator plus approximately 10mm of extra width either side in the side plates and has a internal diameter of 200mm, that would theoretically have a 5mm thick layer of oil on the perimeter if it actually all flung to the out side of the hub's internals. Being a machinist by trade and spending many hours watching coolant being squirted down spinning around things i don't know if this is really the case, particularly at hub motors fairly low revs? Has any one looked inside a oil filled rotating hub? I think at a lot of the hubs operational speed there will tend to be a big puddle at the bottom of the hub allowing the uppermost windings to miss out, this will vary a lot based on RPM, fluid viscosity and internal surface finishes(internal paddles?) etc.
If motor manufactures can keep oil inside a internal combustion engine that has multiple outlet shafts, dozen of different mounting flanges made of different materials, vibrating at wild frequencies then getting hot, cold then hot again you would think ev people have it easy.
It seems most old cars often used to leak a bit of oil, it was only after the 70's tougher emission standards came in manufactures started pulling fairly high vacuums in the bottom end and started running what they call "wet" inlets did the leaks dry up(code for all oil leaks are now constantly burnt and are harder to see). Anyone who has done a bit of race engine building can probably attest to after ditching the factory emission crap and using forged pistons(more piston slop), heaps of boost, more revs etc your new expensive engine leaks like a sieve and tends to fill its oil catch can. Now days most race engines run a large dry sump pump or dedicated vane type vacuum pump. As a vacuum is hard to arrange and the air volumes miniscule compared to an IC engine a small downward pointing internal labyrinth type baffle pickup with either a maze of downward pointing baffles or stainless steel wool pads vented via a very small tube that exits with the other wire bundle.
I agree with John, if we are to get real serious about improving our energy density taking the heat out side of the motor via small pumps to a small radiator seems the way forward, although WAY more complicated for most who insist simplicity is key. We would need a 2 stage pump, something like this http://www.ebay.com.au/itm/Honda-Oi...-Trail-50-Minitrail-Monkey-Bike-/310426179159 electrically driven, 1st stage for oil scavenge/pick up to a tiny oil tank/ oil air separator(maybe part of the frame) into a radiator and into the 2nd stage lower volume higher pressure for the internal oil spray bars. fairly low flows can remove a lot of heat. Just like a Toyota prius, GM volt engineers or IC race engine builders have already worked out. To most this will all sound too complicated and they just want to go for a nice ride in the country with out overheating. That's OK, silicon chips are complicated too and if we had to all understand and then construct them our selves it would be hopeless, and we now all need these to make our bikes work. It would of been so easy to sit on your couch and say brushed motors are good enough but i think because of people with more insight we are now in a better place, going further, faster more reliably and with less maintenance. So just give good engineers time to sort it out, miniaturize it and make it seamless for the masses. Sorry if this is out side the scope of this thread.
Zappy

 

[flathill]

Unfortunately, most of the ePTFE ("goretex") vents that we've seen are not suitable for being continuously coated in oil. Sourcing a product with a hydrophobic exterior coating and an oleophobic interior coating would be ideal. The hub motor on the bicycle that I toured down from Vancouver with oil filling had an average temperature of 30-40 Celcius lower than Justin's sealed motor on hill climbs. However it lost oil the entire time as the vent I installed has failed to work.

How did the vent you installed fail? Clogged? Leaked? What type was it?

Good to hear to no load current is unchanged with ATF and it results in a 30-40C advantage.

 

[itchynackers]

No load currents won't be unchanged. I documented this a while back in this thread about halfway down the page...
http://endless-sphere.com/forums/viewtopic.php?f=30&t=38039&hilit=650ml+1300ml&start=25
No-load current will increase up to about 150ml, then it will remain unchanged with more oil. Presumably, this is because the effective friction surface (between magnets and stator) is saturated at 150ml. This seems to mean more is better, up to at least the axle level (where I saw another small increase in no-load current). I guess we'll have to wait for Justin in order for this data to be believable.

 

[John in CR]

No load currents won't be unchanged. I documented this a while back in this thread about halfway down the page...
http://endless-sphere.com/forums/viewtopic.php?f=30&t=38039&hilit=650ml+1300ml&start=25
No-load current will increase up to about 150ml, then it will remain unchanged with more oil. Presumably, this is because the effective friction surface (between magnets and stator) is saturated at 150ml. This seems to mean more is better, up to at least the axle level (where I saw another small increase in no-load current). I guess we'll have to wait for Justin in order for this data to be believable.

I believed you and was about to make a post referencing it. I hated to see you stop at 50% when you planned to go to 100% fill though, because it would have answered Zappy's question better about what the oil actually does.

 

[robbie]

No load currents won't be unchanged. I documented this a while back in this thread about halfway down the page...
http://endless-sphere.com/forums/viewtopic.php?f=30&t=38039&hilit=650ml+1300ml&start=25
No-load current will increase up to about 150ml, then it will remain unchanged with more oil. Presumably, this is because the effective friction surface (between magnets and stator) is saturated at 150ml. This seems to mean more is better, up to at least the axle level (where I saw another small increase in no-load current). I guess we'll have to wait for Justin in order for this data to be believable.

I completely agree that no load current won't be unchanged. I don't have numbers in front of me, but I seem to recall the no-load current increasing about 200 mA at 48V on our Grin Motors after approximately 150mL of ATF. The relationship between oil fill level and the no load current will certainly be dependent on the internal geometry of the hubs. I'd go spin it up right now, but I have no idea how much oil is left in there with the amounts that have leaked out!

Edit: I went and spun it up because it was bothering me. Unloaded draw was 1.45 A, 47.6V, 490 RPM. I seem to recall this being ~1.6A in the shop, 3 weeks ago at 5am after being up all night. We'll get concrete numbers once the dyno testing is resumed.

It will be quite interesting to continue the hub motor cooling experiments once we're back in the shop. The oil cooling was a total last minute thought, and is certainly not optimized. However the data we've collected so far suggests that the cooling performance is not negligible.

How did the vent you installed fail? Clogged? Leaked? What type was it?

It was a sample vent from another project. MicroVent was the manufacturer. I've not had a chance to disassemble it, so I'm only speculating that the ATF is blocking. Maybe pooled in the internal vent recess. It's effectively this product http://www.microvent.com.cn/eng/cpzs/showcenter.asp?id=494.

I know that during our testing if liquid water pooled on the vents, the vent was blocked.

Good to hear to no load current is unchanged with ATF and it results in a 30-40C advantage.

The effect was minimal at 48V, not unchanged

 

[flathill]

You got me. I meant to write basically unchanged


You might try a more rugged all metal breather vent used on gearboxes or air compressors
An oil separator would probably be overkill
A little weepage out the vent is better than out the bearings
With a piece of foam no drips but ugly so hopefully you wont need it

I dont think you have to worry about water getting in if you angle it down or put a rain hood on or get a capped version
Corrosion shouldnt be as much or a problem with atf coating the innards anyways

 

[izeman]

i filled my hs3540 with 120ml atf. sealed the axle, didn't swap the (no liquid sealed) bearings. made a little breather hole with a 6mm allen bolt screw with a 2mm hole. NO leakage at all. not even an oily mist at the breather hole.
cooling is better now, even though it never can be enough. it's still easy to overheat the motor going uphill and pusing it hard for a long time.
i will do some testing these days, and then add some more oil to see if it changes somehow (to the better?).

 

[hjns]

What about a 12V airpump pumping air into the motor near the axle and have it flow out at the perimeter after exchanging heat with the windings? Needs a large diameter bearing (asHoV did with his Cromotor) and a tube between the hole in the bearing and the pump.

Otherwise cheap, less holes, just need a signal from the CAv3. The higher the motor temp, the lower the motor amps, and the faster the airpump blows.

 

[izeman]

KISS! i like these brainstorming ideas, but as a technical engineer i must say: KISS. keep it simple. the more parts you add, the more parts can break. the bike's getting heavier and heavier, and you need more power to ride it (not THAT much, but still ...).
i think passive aircooling done right (and with a lot of testing - as justin began doing some time ago) will be the best choice for most of us, or simple oil cooling like i did it (because if it didn't work i can still drill the covers, but not vice versa).
the parts that get hottest is the rotor, just between the spoke holes opposite side of the magnets. we should attach cooling fins (maybe with a kind of propeller effect) and see if this helps. maybe justin will do something like that.

 

[Sancho's Horse]

Hey not really meaning to trod so heavy on this thread...I was sooo hoping for a valve.

Anyways, simple must be effective, and effective must be explored thoroughly. Nothing wrong with that.

I agree...keep oil on the windings to be an effective strategy.
So how much oil do we need? 1/3 full? 1/2 full would be cool if it didn't put so much oil over the axle.

But why not get the half full volume and keep it well under the axle level by allowing the oil to spin out of the hub using the centrifugal effect?
It spins out very well, even at surprisingly low rpms. I checked. I mocked some pie plates, used clear tubing and lexan. It does pool...at very, very low rpms. However, my mock-up lacks a rotor, so the splashing dynamics are missing. I suspect that rotor splashing will lead to pooling at a bit higher rpms.

Anyways...I want the space between the spokes to live up to its radiator potential. I am cutting designs into thin sheets of aluminum and sandwiching them together. I started working on an ES design, it looks really cool...I think so at least. Would love to show it off ...if you guys are cool with that. If you are interested in it...its yours.

I also built a model to circulate oil. I will show this off later...I need to get some sleep, and some things I prefer to show finished.

 

[pendragon8000]

I've gone and drilled 27x12mm holes around perimter of one side and opened op the holes i had in the disc mount to 4x19mm and 2x16mm

thermistor temp increased to 110ºC and place on the back edge of stator instead of the 100ºC i had at the top.
so now i think for sustaned stop start trafic in cool weather 6kw is fine on the H3525 100v 60a 26"
if i thrash it for 10 mins it will click off, where as before in warm weather it would trun of in a couple mins at 6kw.
thanks for the help guys.

from my post:
http://endless-sphere.com/forums/viewtopic.php?f=6&t=47452&start=200#p763023

 

[Merlin]

...i wish some updates from justin

iam not sure to drill or oil my hub ;(