Definitive Tests on the Heating and Cooling of Hub Motors

[hillzofvalp]

I'm kind of surprised that over 15 years of ebike hub manufacturing have passed and yet this simple heatsink part hasn't been made until now.... It's kind of sad.

 

[spinningmagnets]

Its only been this past month that someone has produced a 35mm stator direct drive hubmotor with an aluminum stator support as a heat sponge. The major producers have been asked over this past year, and they have been willing, but some wholesaler would have to pay up front for 50 or 80 units before they would produce a batch.

So it comes down to the customers, what would they be willing to pay for? Sometimes there's a "group buy" to prime the pump, and hopefully show wholesalers that there genuinely is enough interest that...if a US based retailer stocks it, customers will buy it. And then...sometimes the "group buy" cannot get enough investors to make the first batch.

I have to hand it to sketchism, he is putting his money where his mouth is...

 

[hillzofvalp]

Well he has some business opportunity for that risk... As the die for that tool will cost <$3000 and the per unit cost likely <$2. $50 a set of 4-6 would be quite a nice little low-risk business. Happy to see it.

I'd like to share my findings on ferrofluid and skateboard hub motors soon.

 

[sketchism]

I'm kind of surprised that over 15 years of ebike hub manufacturing have passed and yet this simple heatsink part hasn't been made until now.... It's kind of sad.

i think the fact that i come from a computer science and automotive background helps a bit, heatsinks and radiators lol, the idea isn't new but when i researched and found most of the hubs we use share a common outer diameter i saw an opening to help out.

So it comes down to the customers, what would they be willing to pay for? Sometimes there's a "group buy" to prime the pump, and hopefully show wholesalers that there genuinely is enough interest that...if a US based retailer stocks it, customers will buy it. And then...sometimes the "group buy" cannot get enough investors to make the first batch.

I have to hand it to sketchism, he is putting his money where his mouth is...

That was one of the biggest calls to make to decide whether to invest in these, i only want to eventually make back the outlay and have them as cheap as possible while doing so, i could have done pre orders to get the capital to go into production but it would have taken forever and i really wanted to begin proper testing ASAP. once myself and the beta testers have conclusive results hopefully some of the e-bike shop guys and dealers will grab some batches at cost and i wont have to be sending them out myself lol.

Well he has some business opportunity for that risk... As the die for that tool will cost <$3000 and the per unit cost likely <$2. $50 a set of 4-6 would be quite a nice little low-risk business. Happy to see it.

A little more because i opted for a more thermally conductive Aluminium (AL6063-T5 - 209 W/m-K), extra polishing on the mating surfaces, CNC machined mounting holes, M5 Nylock nut/bolt fixings and and black anodisation to make them as efficient as possible, but it was the damn MOQ's measured by the linear extruded meter or tonne that really jacked up the initial outlay haha

I'm more than happy to give the .dwg and .Stp files out if anyone wants to have their own CNC machined (or maybe 3d printed from alloy) but from memory the price was about $250USD for a single CNC unit at a machinist, the only condition would be i'd prefer no one take the designs and go into commercial production lol but i also want these to be cheap enough that it wouldn't be viable to do so instead of just grabbing a bulk batch from me direct, i have no interest in being a retailer for them once i've proven the concept and product lol once the development is done i'm bored.

 

[parabellum]

I'm kind of surprised that over 15 years of ebike hub manufacturing have passed and yet this simple heatsink part hasn't been made until now.... It's kind of sad.

Also remember, there is no sense sinking rotor without thermal connection to stator. This improvement is next logical step after FF implementation, which was made less then 1 year ago on this kind of hubs.

 

[hillzofvalp]

deleted

 

[liveforphysics]

CNC machined mounting holes, M5 Nylock nut/bolt fixings

I like that your objective is a premium product. Head clearance to fit with Nord-loc washers? Nylocks are generally ineffectual in my experience.

Thank you for making these. I would love to help you beta, and have access to a motorcycle chassis dynamometer and Flir equipment.

 

[hillzofvalp]

deleted

 

[sketchism]

CNC machined mounting holes, M5 Nylock nut/bolt fixings

I like that your objective is a premium product. Head clearance to fit with Nord-loc washers? Nylocks are generally ineffectual in my experience.

Thank you for making these. I would love to help you beta, and have access to a motorcycle chassis dynamometer and Flir equipment.

Thanks Luke, I'd be honoured to send you some of the Beta versions to test to destruction haha, your videos were some of the first to bring me across from ICE to EV's

I just spoke to Nord-Lock and their M5 pairs have an øo of 9,0mm

Which gives us about .5mm clearance to fit, they should work well so i've ordered a batch to test too.

 

[Cowardlyduck]

Also remember, there is no sense sinking rotor without thermal connection to stator. This improvement is next logical step after FF implementation, which was made less then 1 year ago on this kind of hubs.

Actually I would argue that isn't necessarily the case.
If you measure the temps on the magnet ring after hard riding it easily gets as hot if not hotter than the side covers. Sure there's not direct thermal connection to the heat generating windings, however if the magnet ring has become heat-soaked (which doesn't take much) it sure isn't going to help the stator windings to cool. Having a cool magnet ring is a helpful contribution to keeping the stator cooler even without FF IMO.

Also, I don't totally understand it, but someone else mentioned heat generation from the magnets themselves from the magnetic field changes, so it would help with that also if that's a consideration.

Something else I was thinking about which I'm sure would have already shown up if it was an issue, but does the thermal expansion of the hub have any potential to cause any problems with that heat-sink ring? I imagine it will need to be fitted pretty tight to make good thermal contact, but if the magnet ring expands could that make the heat-sink explode off under pressure?

Cheers

 

[hillzofvalp]

Would love to hear what you guys think about my above posts I know they are long? Do none of you know what exact fluids (specifically statorade A) have been tested so I have a reference? I'm talking to ferrotec today to try to get samples of better stuff.

Re: thermal expansion of ring causing explosion. I would venture to guess this would only happen if you overtigthened it and then thermal cycled it 1000 times it might start to crack. Personally I would use a gap pad like parker chomerics 579, forget the screws, and use the heatsink and the pad as spring and just have latching/hooking features on the heatsink so all you have to do is click it together... no screws required.

 

[parabellum]

Would love to hear what you guys think about my above posts I know they are long? Do none of you know what exact fluids (specifically statorade A) have been tested so I have a reference? I'm talking to ferrotec today to try to get samples of better stuff.

Maybe it is because the ferro particles gradually loose their magnetic properties with raising temperature? Then less magnetic fluid is displaced by cooler fluid to the outside and can escape the magnetic field, if it does not cool quick enough. Passive ferro fluid cooling principle works that way, but fluid is contained by gravity or sealed. There are few papers and demonstration on this mater online.

 

[hillzofvalp]

deleted

 

[parabellum]

Actually I would argue that isn't necessarily the case.

You are right, every bit helps, it just gets to much of investment with to low return. About 3 years ago I run GM 1000 hub windings black, running in a river, thankfully insulation on phase wires melted and shorted before the magnet wire insulation gave up.

 

[hillzofvalp]

deleted

 

[fechter]

If you just use a little extra ferrofluid, any that splashes out will eventually run back down into the gap as long as the motor is semi-sealed.

 

[hillzofvalp]

I'm not sure that's 100% accurate. You would need a pretty heavy drip and large pool to get it down the cover, and act of God to get it back over the magnet spacer. Problem with my motor is that even if that were true, in order for it to "run down" by gravity, it has to run down a spinning rubber seal on the bearing. I think that in practice, the fluid congeals very quickly if just a small droplet. I think there might be an issue with the materials of the motor causing FF to congeal, which is obvious that it is hitting covers and congealing from the pictures I've seen in this thread.

I think that the more and more the ferrofluid touches a porous surface such as enamel, aluminum, steel.... laminations.. etc.. .the colloidal structure breaks down as constituents get ripped out of solution and stick to things.

edit: I think what might be exacerbating the issue is that the phase current goes down at higher RPMs, so most of the fluid at this point is more attracted to the magnets than to the stator.. this may be wrong though.

 

[ecycler]

SO I ran another test at 10000erpm but this time I added 1mL additional fluid and, more importantly, used some of that wall putty to block the ends of the stator. This stopped 99%-100% of the leaking. However, as I raise the erpm to 17-18k, then it starts to leak again, but it's pretty obvious it is coming from the magnet steel (OD) of the motor instead of like before when it was shooting axially off the end of the stator. This means that at this rpm, the centrifugal force is stronger than the magnetic force holding the fluid and/or the fluid that is normally inside the stator teeth crack moves to the air gap and in the cracks between the magnet spacing.

Possible solutions could be combination of:

1) increase height of the magnet spacer or add shoulder so that it contains the overflow.

2) Increase magnet strength

3) Increase magnetization of fluid, and better colloidal structure such as aforementioned APG L23

4) pot ends of stator with very low durometer, very high tear resistance (>300ppi) silicone, oversized slightly so it makes a seal against the magnet ring.

5) Increase the gap between magnets so there is more of a reservoir to deal with overflow of the fluid, Or make reservoir contained inside the spacer ring (think I might try this first)

...as I was catching up on the thread my thoughts were as such:

1. I agree with you and believe what you are seeing can be explained by the inertia of the ferrofluid and centripetal force overpowering the magnetic attraction between the fe2o3 particles and the magnets at the boundary of the ferrofluid blobs.

2. Few of us have played with ferrofluid in anything but low rpm hub motors on ebikes. When I get a little further I will be testing this on a larger scale at about 3,000rpm on an 18kw motor for an ePPG I am starting to build - using some of your batteries you sold me before you moved, actually!

3. I believe at these higher temperatures and rpms that you are playing with you, you could actually be separating some of the fe2o3 particles from the carrier fluid as well. Does the 'splatter' that is coming off the magnets appear to be less concentrated with ferric particles?

4. I think you are on the right track to some N42SH or N45SH magnets and sourcing some different types of ferrofluid.

 

[hillzofvalp]

http://liquiglide.com/

 

[macribs]

Hm. I can't see any easy fix for the FF at the moment.

If swapping FF for ATF or motor oil, how can one assure no leakage from axle? The inpro seal have been mentioned before.
Could that maybe be an easier path to cooling? Seal the axle and side covers and fill up with regular oil? Should be possible to get a good seal.

So if FF has problems, so have motor oil. I recall some saying about filling the hub with ATF causes drag. Has that drag ever been quantified?
How much drag are we talking? And could it be that the drag caused by the oil, could be off set by cooler motor and less resistance? As temps goes down/up so goes resistance.

I wonder if the heat sink and oil filled hub might be an easier route for hard core cooling then struggling with hard to get FerroFluid that is kind of a "dark force" as we don't have enough knowledge about the various qualities of FF, how FF will work over time.

With regular oil we have 2 problems. Leak and drag. Leak can possible be plugged. Drag - what are we talking about? How much peak power do we possible loose due to drag? Efficiency might go down - but a workaround would be to add more batteries. Will the drag caused by the motor/ATF oil cause so much drag that we will notice it in real world performance? Will we be "slower off the line" with less acceleration and top speed?

At least heat sink and oil filled hub would make for best cooling, as there would be more thermal masses inside the hub to transfer heat to the heat sink.

 

[hillzofvalp]

I think the drag is offset by reduced losses elsewhere..

You could do it but you'd need to do a custom ebike motor or heavily modified. I think partial fill sealed is better bang for buck.. And refill every so often and deal with the mess every so often.

But I also think there is an ideal ferrofluid and manufacturing process.. which we will hopefully find soon.

edit: Here's what I really like about skipping ferrofluid in favor of the oil. If you do a partial fill, you're prettty much golden as long as you don't tip the motor over for a long time. High Rpms will do wonders for dispersion on the magnets, especially lots of starting and stopping. If you have <.01mm fit on the races and the OD of the bearing is pressed in, it is extremely unlikely for oil to leak.

 

[macribs]

I hope you are right. Cooling without leakage and messy oil stains is by far the best choice if we can actually achieve it.
So far I guess we are not there yet for FF.

If I could park the bike in a garage with plenty of cardboard under the rear wheel I would just do partial fill the hub with oil and refill when temp rises.
But as I take the bike inside the apartment I really can do without the oil stains. Another benefit from using oil is there will be no corrosion inside the motor.

I have drilled the side cover to equalize the pressure to avoid moisture to enter, so maybe next motor will be oil filled.
At least then I can avoid any epoxy, paint and vanish. Just fill her up and seal her tight.

 

[Ohbse]

I just remembered something and then did another test with similar result.

...

What are the dimensions on your motor that you're testing with? RPM?

Something small diameter and running at high RPM is going to be subject to some wildly different forces than on a relatively large diameter hub motor with a 25 inch tall rolling diameter.

I would be honored to test on my QS in 17" rim. I have a batch of both A and B formulations of statorade and plenty of battery to get the motor up to temp

 

[hillzofvalp]

Small 40mm ID... 3000 rpm. 3000 rpm with a larger hub is going to be way worse. You'll have to do the math but I'm guessing something closer to 1000rpm in a large 200-300mm hub are the same forces that I'm seeing

 

[amberwolf]

Tape might be bad, once it's adhesive has vbeggun to deteriorate, especially depending on it's winding/layering direction. It could begin to fray or peel at an edge or end, then get rolled up between stator/rotor, causing at the least extra friction, and at worst jamming the motor or cuasingn enough uneven pressure on a magnet edge to cause one with weakened glue to pop loose (unlikely, but playing devil's advocate).

If it's always wound so that it is in the same direction the mtoor will be used, and the motor does not need to ever reverse rotate, it poses a lot less risk of unwinding.