Good! Someone finally oil cooled a hubbie!! HS3540

[deVries]

Prolonged exposure of the epoxy to the oil could be an issue though?

No, not with the right epoxy. I helped spec the epoxy for a gent here who repaired a motorcycle alternator where the magnets were epoxied to a steel ring, and bathed in hot engine oil...

Can you tell us what epoxy that is or the type to get

Thanks!

[bigmoose]

Here is my bonding epoxy thread:
http://endless-sphere.com/forums/viewtopic.php?f=16&t=27270&p=394183

[Teh Stork]

These results are really nice. I plan on running a maximum of 6,5kW through my HS3540. I had planned on drilling holes and possibly having a pressurized water mist sprayer, but this seems like a better solution. Then again, looking at his video of him riding - accelleration and deceleration in city landscape over and over, dumping 3kW into the motor at every intersection. It's not a surprise he toasted it. I'll definitely run it under conditions far more forgiving than he did. I plan on running fast tho, 70 kmh.

Isn't it possible to program the controller, or bootstrap a uC, to plot current regulation as a function of speed? As long as you don't exceed max power out (witch depends of speed*torque), you should have great efficiency and minimal heating. Everything above this will be heat? So feeding the motor 2kW should only generate a measly 240 watts of heat (using 88% motor efficiency). Am I missing something?

[parabellum]

Have filled Mxus DD sensor hub with Motul SAE 0W40 oil. 150ml
Did not got anything thinner in stores around, so I was choosing between 3in1 and 0W40. For next one will try to find thin spindle oil.

Normally whenever I get home hub covers stars heating first, going up 10- 20C then cooling down.
As original at 60V 24C and up
freewheeling 1.2A

I sealed sensors and covers with black silicone.
Edit: I also covered the windings with varnish for windings 2 times afterwards. That stuff is thinner based, so silicone does not like it.

After oil fill 60V
24C- 2.7A
50C- 1.7A
After oil filling, the temperature was 50C on arrival and started drooping from start on.

Other improvements: pulled 3x 1.2mm magnet wire trough axle for every phase, for~AWG 12, they was AWG16 before.
Fail: Forgot to pull temp sensor lines

Some porn.

[Burtie]

Have filled Mxus DD sensor hub with Motul SAE 0W40 oil. 150ml

Hi parabellum, any updates on how this is working out with use ?


This oil cooling sounds like a good idea to me, so I had a crack at it too.
I also modified an MXUS DD hub as follows:

1) Added a temperature sensor in the windings near the top of the motor ( away from the oil pool )

2) used silicone gasket sealent to fix the covers and seal the axle slot.

3) Added a threaded filler hole in the side cover for the oil.

4) Filled the hub to various levels with mineral ATF (Dexron II) and measured the no-load current @ 36v

Temperature sensor.jpg (232.27 KiB) Viewed 918 times

Sealant.jpg (146.21 KiB) Viewed 918 times

Fill level at 400ml.jpg (194.06 KiB) Viewed 918 times

At 400ml fill, the oil level is just below the axle, where the filler hole is in the above photo, (you can just see the red coloured oil inside the hub).
With this much oil, the motor makes a satifying, just audible ,sort of sloshing noise when it is spun up.


No oil, no load current = 1.05 A
150 ml oil @ 18 degrees C = 2.0A
200 ml oil = 2.0A
250 ml oil = 2.1A
300 ml oil = 2.15A
400 ml oil = 2.15A

I will leave the oil level @ 400ml and install it on my bike to test the thermal effects.
The bike currently has an identical motor fitted, including a temp sensor (just no oil), so the comparison will be interesting.

Burtie
edit: corrected reference to Dexron II

[Alan B]

Nice experiment. Would be great to record a time temperature trace of the two motors on the same course.

[parabellum]

Sorry did something stupid here.

[parabellum]

Hi parabellum, any updates on how this is working out with use ?

You know, I forgot to install temp sensor wires, having 5 temp sensors on the shelf from HK. So temperature measurement is relative.
My system:
BMX 20" bike
15s 9A Lipo pack
50A limited 15 fet infineon controller.
95kg rider.

I discharged the pack going up and down same stall hill repeatedly, WOT up and regen down, in some points i need to pedal uphill to avoid stall the motor and kill controller.
Motor outside gets to hot ho touch, over 70C, thermal shearing works nice since it is not getting hotter on stops.
After short stops I could fry sausages on my HS R5303 under similar conditions on the burned paint stinking motor covers.
I believe this motor is capable of 2kw continuous in low speed condition and 3kw or more at higher speed.

[parabellum]

At 400ml fill, the oil level is just below the axle, where the filler hole is in the above photo,

How do you plan to make breathing, if your motor is half full of oil?

[Burtie]

I dont really see a problem with the motor not being vented.

I guess there may be some small migration of oil out through the axle seals or through the cable cores (see Justins latest video) due to heating.
Or perhaps some small water ingress via the same routes due to cooling.

The oil level is normaly below the axle seals, so this should prevent any major leakage.

At least the metal parts will be protected from corrosion by a coating of oil

[parabellum]

If it is tight and oil can not travel fast enough, there is risk of pushed out seals. Anyway, if you are right I would prefer to close my vent to, now it loses little oil and I need to wipe it out from time to time.

[electr0n]

So what is the verdict on oil cooling a hub motor? This is much easier than installing fans, drilling side covers and using liquid cooling.

[parabellum]

So what is the verdict on oil cooling a hub motor? This is much easier than installing fans, drilling side covers and using liquid cooling.

It improves heat transfer to ambient significantly, extra bonuses are rust protection and physical protection of closed system (as opposite to air/hole cooling). It also makes external cover cooling reasonable. Planing to make this mode to all my hubs after I get some really thin oil, 0w40 ads to much drag when cold.

[Neovin]

I was thinking for future motor designs, why could we not use small diameter copper enameled tube, I doubt you could find anything suitable for a retrofit but if you designed from the ground up you could take the heat away from the center of the problem.

[Burtie]

I got my oil cooled hub fitted, and early tests results look great.

Setup summary:
The hub is filled to just below the axle with 400ml ATF.
The temperature sensor is fitted in the windings at the top, not in the oil pool.
The hall sensors are sitting at the bottom, in the oil pool.
The hub is a 7 turn MXUS running at 66v, 4.5kw peak power.
Ambient temperature during testing was 8 to 10 deg C

My regular 9 mile round trip takes in some sections of 20% gradient. I rode the same route at the same speeds (stopping mid point) with identical motors, exept that only one of them was oil cooled.
This route, when ridden briskly, burns about half of my 15AH battery capacity.
In short:

Max measured winding temp No oil ----> 140 C, with oil ------> 55C
Finish point winding temp .... No oil ----> 106 C, with oil ------> 38C

The small amount of drag due to the oil is not noticeable when riding, and the motor runs significantly quieter.

Seems to work for me

The only downside I can see is that the rear disk brake may be prone to some oil contamination if the LH axle seal weeps.

Burtie

Edit:
Just checked rear disk after last nights ride, and it is now covered in oil

[Spicerack]

You need some form of vent for the hub to relieve excess pressure as it heats up.

Try a little bit less oil and drill a small (1mm) hole in the cover near the axle. Angle it so that it traverses from the central axis towards the outer circumference as you go from outside to inside. That way, when the wheel spins, any oil that happens to splash into the hole will be flung back inside the hub. I've done this on my mac and had no oil leak at all from the hole or axles.

For the record, the outer cases of my mac are alot hotter than without the oil so it certainly transfers heat away from the windings.

[Burtie]

Thanks andy, sounds like a good idea.
I will try drill a small angled vent in the opposite side, behind the sprockets.

[bigmoose]

Thanks for the great data on oil cooling Burtie! I think this will become standard practice for high performance motors as we go forward. I saw how much this improved the "Power Dense" motor system developed at Bowling Green State University about 10 years ago, and strongly felt it was applicable to hub motors. Your data now proves the point.

This should now end any lamination/magnet corrosion issues, and bump power handling a bunch. I will also bet thermal soak hall failures are a thing of the past also.

[Arlo1]

Thanks for the great data on oil cooling Burtie! I think this will become standard practice for high performance motors as we go forward. I saw how much this improved the "Power Dense" motor system developed at Bowling Green State University about 10 years ago, and strongly felt it was applicable to hub motors. Your data now proves the point.

This should now end any lamination/magnet corrosion issues, and bump power handling a bunch. I will also bet thermal soak hall failures are a thing of the past also.

WHat about drag at hi rpm??

[bigmoose]

WHat about drag at hi rpm??

You do the system analysis and see if you have enough AHr to overcome it. It is my opinion that we now have the capacity to accept some losses that were not acceptable just a few years ago. Using gears in the reduction of mid drives is the next area I see coming...

One thing I harp on at the "big lab" all the time is "It's all about the system..." Maximizing individual component efficiencies and capabilities does not always increase the efficiency and capability of the system. It is ALL about the system... hence my support there for multivariate optimization routines and Monte Carlo simulations.

[itchynackers]

I still don't quite understand whether the oil is just delaying thermal soak (by the oil storing the heat), or if the oil is allowing the hub to dissipate more heat.

Edit: After re-reading the thread, it seems as though the oil transfers the heat quicker to the covers (making them hotter) so more heat is carried away with the same amount of windage. Is this correct?

[parabellum]

I still don't quite understand whether the oil is just delaying thermal soak (by the oil storing the heat), or if the oil is allowing the hub to dissipate more heat.

Edit: After re-reading the thread, it seems as though the oil transfers the heat quicker to the covers (making them hotter) so more heat is carried away with the same amount of windage. Is this correct?

The second. Oil helps transferring heat to covers and they dissipate it to ambient.

[parabellum]

Try a little bit less oil and drill a small (1mm) hole in the cover near the axle. Angle it so that it traverses from the central axis towards the outer circumference as you go from outside to inside. That way, when the wheel spins, any oil that happens to splash into the hole will be flung back inside the hub

I did exactly this on Mxus DD inside disc brake screw hole with inclination to outside D inside, then tightened a screw with sprung washer (sprung part in axle/center direction), I only have 150ml of oil but it gets out of the vent anyway(maybe few droops per ride).

[Burtie]

I cleaned the oil off the disc and pads with caustic soda, and drilled a small vent on the opposite side of the hub, close to the axle.

It all works well, the brake stays clean
Just a small drop escapes from the vent each use, so will have to wipe it over occasionaly to stop things getting too messy.

Oil Cooled Test Rig.jpg (154.03 KiB) Viewed 697 times

This seems to me like a massive step forward. I can now practicaly ride this 4kw bike as hard as I like, for the whole duration of the battery pack, with no more worries about cooking the hub!

Thanks to everyone here for their ideas and suggestions
Well pleased
Burtie

[Spicerack]

Good to see. I was going to liquid cool my 9C 8x8 and run quite a few amps through it. The initial plan was copper cooling pipes around the inside of the hub, going out through a modified/fabricated side cover to a little radiator etc etc but I'll just do the oil bath first and see how I go. It only helps a little bit on my current MAC hub but that's because the windings are damaged from overheating them so a large chunk of the amps get converted to heat!