Oil cooling your hub- NOT snake oil!

Spicerack

10 kW
Joined
Jan 27, 2011
Messages
602
Location
Perth, Western Oz
Ok, so here’s my story. I cooked my poor little MAC by putting way too many amps through it. It still works, but gets hot real quick so whilst waiting for another stator to rebuild it, I decided to try oil cooling as a band aid solution.

Without oil, the motor gets up to 120 celcius and is rising by 10kilometers (6 miles). This is running 60V, 20amps. In real terms it tends to be cruising at WOT around 1300W on the CA.

I sealed the motor up with silicone on the covers, drilled a little pressure relief vent hole hear the axle on one of the covers and put 1-200ml of transmission fluid in it. I didn’t want too much as it would leak out of the axle- which I also sealed as best as I could from the outside. It’s a very basic closed oil-splash system just like most lawnmower engines.

I rode it to work today, which is 22km. The hottest it got was around 90 celcius. It was very interesting to see the temperature on the windings go up by a degree or two a second whilst accelerating from standstill, then drop almost as fast once up to speed and on WOT. Most of the time it was running at 75 celcius.

It was also very interesting to note that cruising at say ¾ throttle resulted in the temps going up- but as soon as I went to WOT at the same speed, the temperature started coming down again and quite rapidly.

I know that people have always said that it’s better to run at WOT but have never seen the results as plain as this today.
So there you have it. A simple oil cooling setup can drastically cool your motor- in my case although my windings are knackered, at least I can still use it until the new bits turn up. I would definitely recommend this over air cooling- certainly for the geared hubs as it takes the heat to the outer covers of the hub.
 
It is good to see more people discover and document the benefit of oil cooling a hub. I predict it will become the standard in the near future.

Pioneering work was done in this area by Bowling Green State University (BGSU) Professor Anthony Polumbo in BGSU Paper "PowerDenseElectricMotor-BGSU-01566299" (0-7803-91 45-4/05/$20.00©)2005 IEEE) where they increased the power of a NEMA 215 Frame induction motor used to power a bus from approximately 10Hp to 100Hp. Not all the improvements were from oil cooling, but it sure enabled much of the increase in "Power Density."

Sadly the Electric Vehicle Institute at BGSU has been dismantled and the staff scattered... right when we on the cusp of electric propulsion. They had a great staff and great vision, just too early for stable funding...
 
....or because the oil companies had their say.....

Sorry, bit of a conspiracy crazy, me. They're all out to get us!

Rode home again tonight and didn't get over 80 celcius- down to a balmy 20 celcius here in Western Oz in the evening. 20km, top speed 54km/h and average speed 35km/h.
 
My only question is, but WHERE does the heat go? (http://www.youtube.com/watch?v=wcFZIj96LwY#t=00m47s)

Seriously.
 
dozentrio said:
My only question is, but WHERE does the heat go?

Build the heat flux model of a conventional motor from copper windings > radiate + convect + conduct through air to the covers and the rotor. Then the covers and rotor convect and conduct to the atmosphere.

Now replace air convection (because the radiative component is nil) with immersion oil conduction in the middle step, which is the rate limiting step. See where the gain is now? Why do we immerse an egg in water to hard boil cook it? Just put the egg in the shell on top of the fry pan and compare cooking times. And compare that with a fried egg. Liquid conduction of heat rocks.
 
dozentrio said:
My only question is, but WHERE does the heat go? (http://www.youtube.com/watch?v=wcFZIj96LwY#t=00m47s)

Seriously.


The outside of the motor case sheds all the heat, just as it does in the stock un-oil-filled hub.

The difference is, the thermal path between the copper and the outside case becomes extremely good with oil slopping around everywhere. This means the case will be able to be much hotter for a given motor temperature inside, and the hotter the case, the more energy it can shed into the air contacting it.

Your goal in any cooling system is to minimize the difference in temperature between the thing you're trying to keep cool, and the thing that you're ultimately trying to transfer heat into. The end of our cooling path is to heat the air rushing passed the motor's outside case as we ride. The bigger the temperature difference between the air rushing passed the case and the case, the more energy it can transfer to that air rushing passed it.
 
I mentioned to farfle that this could be taken to the next step since he is machining axles for the stator. You can build an oil circulation loop into the system by injecting at the axle and scavenging from out near the windings. The hub does not have to be filled for this to work if the scavenging port is played with for fluid dynamics, or you can use a scavenge pump and gravity feed. I hope the high power guys are embracing this development!

Also the guys that are using mid mount inrunners may want to get the BGSU paper that I referenced, as it shows how BGSU/Lincoln Electric optimized the oil cooling. For the 100 Hp units they re-purposed planetaries from an automotive automatic transmission.

I was able to talk with my gear expert friend yesterday and his off the cuff high efficiency reduction was planetaries. Now most have to be custom made which is cost prohibitive for us. So does anyone know a car that has had reasonable production volume that has very small planetaries in the transmission that could be repurposed?
 
So does anyone know a car that has had reasonable production volume that has very small planetaries in the transmission that could be repurposed?

Geo Metro?

http://parts.nalleygmc.com/showAssembly.aspx?ukey_product=3122064&ukey_assembly=389549&ukey_make=1024&ukey_model=14540&ukey_driveline=0&ukey_trimlevel=14831&modelYear=0
 
I believe you want the reverse gears in an automatic transmission. They tend to be fairly wide relative to the power transfer of a bicycle. Don't know of a manual trans that would use PG.
 
Ideally, you want the runniest oil you can get to minimize viscous drag. Maybe automotive brake fluid might be better than ATF? It dissolves paint, so may not be compatible.
 
bigmoose said:
I was able to talk with my gear expert friend yesterday and his off the cuff high efficiency reduction was planetaries. Now most have to be custom made which is cost prohibitive for us. So does anyone know a car that has had reasonable production volume that has very small planetaries in the transmission that could be repurposed?
For the gear motor folks - planetaries are already in hand. Interestingly, this strategy might extend the life of gear motor planetaries by providing lubrication and direct gear cooling.

Here's a little déjà vu from last month.... :) Keyne's DH Comp Build - now with liquid cooled HX35!.

fechter said:
Ideally, you want the runniest oil you can get to minimize viscous drag. Maybe automotive brake fluid might be better than ATF? It dissolves paint, so may not be compatible.
ATF is already very low viscosity and is designed in part to transfer heat very well - in cars it gets pumped out to the oil cooler and back to cool the tranny. See section 2.1 of Handbook of Lubrication and Tribology. On the face of it, ATF looks like it may have been a good choice for this prototype.

It's super that you tried out this oil cooling strategy and it's proving successful. It would be great if this lets gear motor power ratchet up a bit - it seems to have been stuck in a design/materials snafu for a while now.
 
Very light machine oil may be suitable. I happen to have many different types of oil in my workshop but ATF is the thinnest so that's what gets used.

I still can't believe what a difference it has made- and doesn't appear to have affected top speed or amps much with the drag on the reduction gears.

AS far a planetary gears go, has anyone looked at certain reduction drive starter motors for cars? I might pull one apart from my shed stock and see what it looks like.

Oh, and don't use brake fluid- as mentioned, it strips paint so it won't be going anywhere near my motor! It also isn't really a lubricant - an added benefit for using atf in a gear reduction hubbie.
 
This is awesome. If you can actively use the side covers for cooling, then you could even add a couple of lighweight fins to help with surface area.

However, the problem to retrofit is sealing the hubs.

How much oil are you getting leaking out of the axle?
 
Very very little leaks out of the axles. I guess you could say the wires are possibly a bit wet and that's it That may be from before though. As long as you only have enough oil in the hub to fill it say to about a third- so the oil level is well below the axle, you shouldn't have any dramas. Having a little vent hole is an absolute must. I tried it without and within a minute of testing, oil started to seep out of the axle, following the air as it expanded inside and escaped.

When I replace the stator in this motor, I'll seal the axle properly from the inside and then refill it with oil. Then we shall see how much power these little MAC motors can handle... :twisted: Or then again maybe not- don't want to kill the clutch! I also have a 9C 8x8 which is going to get oil cooling and push lots of power through it. The plan for the mac is eventually to go in a custom frame as a mid drive.

I drilled a small hole (1mm ish) near the axle. I drilled from outside to inside, with the drill angled so that the hole goes from the axis towards the circumference. My logic was that when spinning, any oil that comes near the vent hole will be flung back inside the hub...seems to be working so far.

I'll try and chuck a couple of photos up- not terribly exciting though as it's a sealed hub! I did try a filtered air cooling system but that did nothing so the holes are blanked up with plastic bungs glued in. One of the bungs has a hole and rubber stopper for filling the hub with oil.
 
It would be interesting to try running a MAC/BMC with oil and a full set of the steel replacement gears. Normally only one steel gear is used and they tend to eat the outer ring gear (steel on steel - dry). With some effective lube in there and meaningful heat removal, this upgrade might become substantially more successful. Lyen sells these for $25.
 
teklektik said:
Here's a little déjà vu from last month.... :) Keyne's DH Comp Build - now with liquid cooled HX35!.
/doi/abs/10.1201/9781420003840.ch2]Handbook of Lubrication and Tribology[/url]. On the face of it, ATF looks like it may have been a good choice for this prototype.

I did follow this with interest. I still think the flooded or splash design is better as the windings lose heat directly to the oil. People go on about trouble with sealing the motor to stop oil leaks but it ain't hard. Both covers get cleaned thoroughly then sealed with proper automotive silicone. The axle bearings should be sealed already but if they aren't you can replace the bearings with sealed ones, or even just use oil seals. You'd have to machine a recess in the covers so I guess it's a bit more complicated....unless you have a lathe in which case it's real easy....

Seal up the axle with more silicone, put a vent hole in for a splash design and you're done. If the motor is totally flooded then you need to have some form of vent still. If it's a circulating system then it gets more complicated, especially getting the pipes out of the hub- involves machining the brake side cover for a much bigger axle bearing and making up a spacer to fit on the axle that the new bearing sits on, with holes in the spacer for coolant pipes, even fat phase wire etc. With this design you could even make a solid axle but we're talking serious mods here.

My simple closed system splash design works well for now and doesn't take any special tools to do.

As for the steel gear- yes that might work but I think the new grey gears for the MAC are plenty strong. The next weak link will be the clutch.
 
andynogo said:
I drilled a small hole (1mm ish) near the axle. I drilled from outside to inside, with the drill angled so that the hole goes from the axis towards the circumference. My logic was that when spinning, any oil that comes near the vent hole will be flung back inside the hub...seems to be working so far.
If there's a bit of room on the inside of the cover, an oil slinger could be affixed to the cover surrounding the bearing with the vent hole under the slinger. This is basically a thin sheet metal disc that does what it sounds like to keep oil away from a bearing and seal.
 
Yep don't use DOT 3 or 4 brake fluid. It is hydroscopic, and will rust the snot out of everthing and likely strip paint and insulation. I have a gallon of DOT5 from the days of my Porsche.... a long, long time ago. It is silicon based and very, very thin. It has been filling the submersible pump in my wet metal bandsaw for 25 years now and working fine.

I would use ATF, Dextron III or Honda ATF first; then try the DOT5 silicon brake fluid if you had perceived problems with the first two.

I would be interested in seeing pictures of the planetaries in starter motors. Might be about the right torque rating...
 
teklektik said:
andynogo said:
I drilled a small hole (1mm ish) near the axle. I drilled from outside to inside, with the drill angled so that the hole goes from the axis towards the circumference. My logic was that when spinning, any oil that comes near the vent hole will be flung back inside the hub...seems to be working so far.
If there's a bit of room on the inside of the cover, an oil slinger could be affixed to the cover surrounding the bearing with the vent hole under the slinger. This is basically a thin sheet metal disc that does what it sounds like to keep oil away from a bearing and seal.

Could do- just like automotive engines of yesteryear- but no need to do this in my case as the vent hole isn't leaking. I think someone did have this issue- but I try to follow the KISS principle wherever possible! :D
 
andynogo said:
...but I try to follow the KISS principle wherever possible! :D
Agreed! (Just saying that if weeping became an issue, a simple fix is in hand...)

andynogo said:
The axle bearings should be sealed already but if they aren't you can replace the bearings with sealed ones, or even just use oil seals. You'd have to machine a recess in the covers so I guess it's a bit more complicated....unless you have a lathe in which case it's real easy....
I don't know how MACs are built but BMCs already have double lipped oil seals.

andynogo said:
As for the steel gear- yes that might work but I think the new grey gears for the MAC are plenty strong. The next weak link will be the clutch
The BMC blue gears are very strong as well, but then, an important facet of the success of this strategy is pouring more power through the motor so gears will become an issue - if steel would run quiet and not degrade the ring gear, I would switch. As for the clutch, I have run with a frozen BMC clutch and the bike still coasts pretty freely in DD mode - even at 40mph. Just welding/epoxying (DP420) it up would not be a bad solution. I used to drive SAABs with freewheel clutches and they frequently blew up (operator abuse) - welding them up was the universal fix :)
 
I'm pretty sure the mac has sealed bearings.

Good to hear that a lockup clutch didn't affect things too much. I must admit I don't do much freewheeling on my bike. It's wot most of the way!
 
..It was also very interesting to note that cruising at say ¾ throttle resulted in the temps going up- but as soon as I went to WOT at the same speed, the temperature started coming down again and quite rapidly.

does anyone have an explanation for this effect ??

Planetary gears:
Many Toyota starter motors have planetary gear set in them
And some Honda bike starters too !
 
I'm sure someone more intelligent than me will explain. I thought that it was to do with the motor operating at lower efficiency (more heat) at lower rpms, but in my case it seems to be the same rpm and different throttle settings. Who knows.
 
Back
Top