Oil cooling your hub- NOT snake oil!

[teklektik]

Another alternative is to contact MWKeefer who has been doing oil in geared hubbies for years.

From what I can glean from his old posts MWKeefer was using BMC motors. I can't speak with authority about his particular model but the V1-V4 BMCs all use the same case with proper double lipped oil seals riding on a normal (non-slotted) shaft so sealing is not a problem - the bearing/seal arrangement is conventional and similar to any engine, tranny, diff, etc. The only concern there should be wicking in the wire which is nowhere near the challenge faced here. In this thread GCinDC is trying to seal a motor with a slotted shaft and no seals - a problem of far greater magnitude.

I wonder if there might be a way to maintin a slight vacuum inside the hub? Something like a miniature air pump with an air/oil separator. Seems like technological overkill, but would certainly keep stuff from leaking out.

I think the draw of this oil fill approach is the simplicity of the solution - no external equipment, no side cover rework for bigger bearings, or rewiring the axle. I think at the time you get to running tubes into a hub like this with no seals, an approach like Doc's water-cooling approach is likely to give a less troublesome long term solution. The heat transfer of an oil bath should be better than the stator cooling approach, but that's just another familiar engineering tradeoff...

That could work via the axle, and there would also need to be an oil-scrubber or drip-line feedback: Several industrial systems work that way to solve similar problems. When you think about it, ICEs pull mild vacuum on the crankcase via the intake manifold.

Well, yes they do, but not to solve a similar problem - it's to address pollution from blow-by not shaft seal leaks.....

 

[Kingfish]

That could work via the axle, and there would also need to be an oil-scrubber or drip-line feedback: Several industrial systems work that way to solve similar problems. When you think about it, ICEs pull mild vacuum on the crankcase via the intake manifold.

Well, yes they do, but not to solve a similar problem - it's to address pollution from blow-by not shaft seal leaks.....

Yes – definitely true; that’s a primary goal. However the technique was employed on some motors before the EPA came into creation – and it had the small legitimate effect of eliminating an extra seal at each shaft interface.

The problem with hub motors is that they are not normally vented, so pressure and vacuum can and do accumulate. The axle journals are not well finished when compared to ICE crankshafts, and the seals employed in a hub cover are for debris and moisture.

Useful diagram: From the SKF catalog, this seal exceeds the stock unit shipped with a 9C hub.

In normal engines, there will be a specialty seal having a dust lip and a sealing lip, or two seals organized in such a way to reduce leakage/contamination into and out of the system. The journals will be well-finished and polished to avoid premature breakdown. Pulling a light vacuum can reduce cost in part fabrication; it’s a gimmick that has merit in many mechanical designs where leakage is verboten.

Returning to the conversation at hand, attesting for 9C hubs (I own several), they cannot be pressure-sealed through the axle without overt effort beyond the typical DIY tinkerer on ES. A machine shop could possibly do it, fabricate a new axle, manufacture new covers, and assemble with the correct components, etc. If the periphery of the cover is sealed (completely achievable), all that remains is a method to prevent leakage to the exterior via the axle when the motor is in a positive-pressure state.

What about just venting through the axle via a capillary tube up to an accumulator, say... the size of a CO2 cartridge which is vented to the atmosphere through a porous filter? Oil vapor travels into the accumulator, the pressure is vented, and when the reverse is true the oil would draw back into the system. No pump needed. Caveat: The vent would need protection against direct moisture.

Still, for this to work properly, it would be optimum if the journal was better polished.

My ½ W. Cup empty… considering another pot. KF

 

[GCinDC]

wasn't it agreed that a vent was necessary?

if not, another idea may be to cook the oil before injecting, and then capping the tube.

 

[John in CR]

The heat transfer of an oil bath should be better than the stator cooling approach, but that's just another familiar engineering tradeoff...

The difference is that the external radiator increases the continuous power capability of the motor, while simply putting oil inside doesn't because there's no change in the outside surface area. Stock DD hubbies do a surprisingly good job of rejecting heat.

Talk about an engineering tradeoff would mean there's been an engineering analysis, but the whole oil cooling approach is based on a false assumption, that the bottleneck for heat is the layer of air inside the motor. The real bottleneck for heat transfer is from the outer shell to the environment. That's why ventilated cooling can be so effective. The result is multi-kw of cooling vs the 1kw or less that's possible from sealed covers with or without an oil fill.

John

 

[ian.mich]

The heat transfer of an oil bath should be better than the stator cooling approach, but that's just another familiar engineering tradeoff...

The difference is that the external radiator increases the continuous power capability of the motor, while simply putting oil inside doesn't because there's no change in the outside surface area. Stock DD hubbies do a surprisingly good job of rejecting heat.

Talk about an engineering tradeoff would mean there's been an engineering analysis, but the whole oil cooling approach is based on a false assumption, that the bottleneck for heat is the layer of air inside the motor. The real bottleneck for heat transfer is from the outer shell to the environment. That's why ventilated cooling can be so effective. The result is multi-kw of cooling vs the 1kw or less that's possible from sealed covers with or without an oil fill.

John

Still, it has to have some effect, as I said with oil cooling it's possible to run 6kw through a 1kw pro 901. even without oil, i have customers running 3700w through the 901 with 10awg enameled phase wires without issue

 

[teklektik]

The heat transfer of an oil bath should be better than the stator cooling approach, but that's just another familiar engineering tradeoff...

Talk about an engineering tradeoff would mean there's been an engineering analysis, but the whole oil cooling approach is based on a false assumption, that the bottleneck for heat is the layer of air inside the motor.

Here I thought I was fairly specific in talking only about two means to pull heat from the windings - the difference between wetting the windings and cooling the stator, not the means to cool the fluid. The recent context of discussion has been about running additional piping into an oil cooled motor, possible use of a sump system, etc.

There is no question (without engineering analysis) that the passive cooling from the minimal surface area of the hub cannot compare to the heat dissipation available with active (pump) cooling and the large surface area of an external radiator - but those elements of the cooling system were not addressed in my remark.

 

[amberwolf]

if not, another idea may be to cook the oil before injecting, and then capping the tube.

Won't help, as the air (and oil) can still move back and forth within the wiring harness, even within the wires inside the insulation, as the motor heats and cools. This is why it's impossible to keep water out of a hub completely, too.

You *might* be able to seal each individual conductor strand against all others, inside the wire insulation, and then each insulated wire against all others, and against the bundle insulation/jacket, and then the bundle jacket against the wire exit hole/tube/etc., but then you still have the bearings and covers. Sealing the covers including the threads and holes for the bolts is not that hard, but sealing the bearings against changes in pressure is more difficult. Even the "sealed" bearings discussed above will probably still weep a little over time, and with pressure changes will probably make that worse. Pressure differential is pretty low even at the hottest and coldest these things get, even in some deep winters vs really hard use, but it might be enough to cause wicking thru the bearing, either thru it's seals or around it where it is mounted to the axle or the cover, if that part isn't sealed, too.

 

[Kingfish]

The heat transfer of an oil bath should be better than the stator cooling approach, but that's just another familiar engineering tradeoff...

The difference is that the external radiator increases the continuous power capability of the motor, while simply putting oil inside doesn't because there's no change in the outside surface area. Stock DD hubbies do a surprisingly good job of rejecting heat.

Talk about an engineering tradeoff would mean there's been an engineering analysis, but the whole oil cooling approach is based on a false assumption, that the bottleneck for heat is the layer of air inside the motor. The real bottleneck for heat transfer is from the outer shell to the environment. That's why ventilated cooling can be so effective. The result is multi-kw of cooling vs the 1kw or less that's possible from sealed covers with or without an oil fill.

John

+1: I would have to agree with that. Introducing a cooling medium to rotating bodies is adding resistance. Even with oil mist and oil splash there will be resistance. The copper in the winds and the electric steel are design to shed heat, as are the aluminum covers. Inside the motor we have radiation, convection, and conduction all working in proximity to various levels, with air being the worst of the transports. Back iron to flanged part is partly inefficient because the surfaces are smooth, as are the surfaces of the hub covers. Therefore the boundary layer must be perturbed to increase the transfer rate from stator/magnets to the exterior.

• Big hole venting, especially if scalloped/shaped to pull air in – will certainly increase the rate of heat transfer. However to be fair, the ventilation holes will also introduce resistance. There’s no free lunch.
• Personally, I like the idea of a vent hole in the hub cover near the axle and a few on the periphery – centrifugal pump-style flow, in conjunction with a completely refurbished and varnished inner assembly.
• I would remove the paint from surfaces treated as such and instead use hard black anodize (for aluminum); it’s a durable surface treatment that better absorbs and radiates heat over paint.
• Also look at the circular surfaces that is exterior to the back iron and figure out some way to add small vanes to increase radiation; fins could be ¼ inch/6 mm high by 1 or more rows, and can be straight of beveled. It’s a natural rotating radiator waiting to happen and no one is talking about it.

Still, I like the exercise of oil cooling; it’s good to explore every option, even if for the future.
~KF
PS - Just finished upgrading the phase wires to 12-AWG on another hub. All this talk of cooling has been directly on my mind :lol:

 

[teklektik]

Useful diagram: From the SKF catalog, this seal exceeds the stock unit shipped with a 9C hub.

This ground was covered earlier in this thread.

Returning to the conversation at hand, attesting for 9C hubs (I own several), they cannot be pressure-sealed through the axle without overt effort beyond the typical DIY tinkerer on ES. A machine shop could possibly do it, fabricate a new axle, manufacture new covers, and assemble with the correct components, etc. If the periphery of the cover is sealed (completely achievable), all that remains is a method to prevent leakage to the exterior via the axle when the motor is in a positive-pressure state.

What about just venting through the axle via a capillary tube up to an accumulator, say... the size of a CO2 cartridge which is vented to the atmosphere through a porous filter? Oil vapor travels into the accumulator, the pressure is vented, and when the reverse is true the oil would draw back into the system. No pump needed. Caveat: The vent would need protection against direct moisture.
KF

Earlier posts here discussed the positive pressure problem and suggestions to address it. The specific motor under discussion already has a vented side cover and is experiencing no positive pressure. (The vent is visible on this image from two pages back.)

 

[Kingfish]

Are we talking about the leaking of oil from the motor? And with the penetration of oil into the bearing? I’m pretty sure that’s part of the contemporary discussion for the last two pages.

I shall try to be clearer; in my observation the problems are:

1. Replacing the stock dust ring with one capable of retaining oil is fraught with issues unless the mating surfaces can accept a good seal. This typically employs good quality steel and surface treatment so that the journals can be ground to a near perfect polish. This prevents the seal from eroding or allowing passage of fluid form one side to the other during rotation. If there is pressure differential between the exterior and interior - leaking can occur, or if the journal surfaces are imperfect, there will be leaking. BTW – pressure can include force of liquid impinging upon surfaces through kinetic action.
2. The axles of these hub motors are made of relatively soft steel; I can drill the buggers out with a carbide bit! This is an invitation to failure if attempting to use it for sealing liquids because the surfaces cannot be made hard enough to hold the seal. You can try surface hardening although I don’t know the alloy of the axle or if it could take it. If money was no object, I’d replace the shaft with a known quantity; something deliberately fabricated to be a true shaft, and precision ground. That would eliminate the problem the shaft and seal.
3. Now we look at the balance of the entire wheel: Is the seal leaking due to centripetal forces? Balance is everything, and well, it’s not common to balance them to the specification of a flywheel or automobile wheel. Mine are definitely out of balance and I feel it above 35 mph. What do you think that seal is doing when there’s that much deflection? Imagine the interaction between the seal and the axle with the force of liquid in motion.

It’s a great discussion, truly. But there are problems with trying to make a silk purse out of a sow’s ear. Look at how the motor was made, where it was made, and how much did they spend making it, and how much you paid for it. I could sit back (and have for some time) and let this perk onward, or I could jump in and share what I know about mechanical engineering – coming from a robotics environment, working ICE design, and hands-on steam propulsion. I’m just being straight-up here.

Look at these pictures and visualize the problem. It’s a great idea! But I don’t think the product (motor) is created in a manner of sufficient quality to retain a liquid seal. Valiant effort though – very close to solving it, at least for a while.

Without intent to harm, KF

 

[teklektik]

Now that we've gotten done with the reasons why this is a Bad Idea... perhaps earlier contributers to the thread who have had success with this inadvisable cooling method can chime in with some additional assistance for GDinDC regarding the bearing types/manufacturers in their motors or offer any other factual insights as to how this implementation differs from their successful oil cooling mods.

 

[ian.mich]

From what i can deduce, and from looking at the innards of the pro 901, i think creating a 6000W motor with INOX MX3 oil cooling it goes a little something like this
1. open both side covers, drill 2mm hole near the axle and lead a small tube out of it, point the tube towards the axle so gravity while spinning won't let the oil go out through there.
2. leave in halls, take out the black insulation around the current wiring, and take the phase wires out, silicone the shit out of the halls.
3. cram 3x 10g solderable magnet wires down the axle and solder, silicone, and heatshrink.
4. replace the current bearings with sealed (Edit: i take that back, these stock bearings are sealed), cover the part of the axle that the bearing sits on with automotive silicone, and apply as much as possible around the bearings as well.
5. cover the seating for the motor covers with silicone and put the covers back on.
6. leave it for 24-48hrs to let everything dry properly, and then empty a can of MX3 into the little tube.
7. 18 fet 4110 greentime controller and 24s lipo, and life's good.

Only way to make this work perfectly would be some truly sealed bearings.

 

[Sancho's Horse]

IP 67 connectors for the wiring harness issues, but the wheel bearing issue is pretty significant

 

[FastDemise]

I have been running mineral oil for a couple months now. My setup is in my signature and recently went to 20s at 60amps battery and 150phase. That's roughly 4.2kW peak and 2-2.2kW continuous. Running errands ~10miles with mid 80 degree weather both the controller are just warm to touch. I prefer mineral oil due to my understanding that it's perfect for electronics and it has no detergents or cleaners added that will do any long term harm to anything in my motor. I just filled my motor with a modest 50-60mL and been happy with the results ever since. I strongly strongly recommend mineral oil (preferably the laxitive kind) cuz it's easy to get a hold of and cheap.

But how did you seal it?

I just simply applied a light coat of Red RTV to the cover lips before putting them back on to keep the covers from leaking. I didn't seal up the phase/hall wires going into the center of the axle so that's my breather tube.

This RTV:
http://www.permatex.com/products/automotive/automotive_gasketing/gasket_makers/auto_Permatex_High-Temp_Red_RTV_Silicone_Gasket.htm

 

[ian.mich]

I have been running mineral oil for a couple months now. My setup is in my signature and recently went to 20s at 60amps battery and 150phase. That's roughly 4.2kW peak and 2-2.2kW continuous. Running errands ~10miles with mid 80 degree weather both the controller are just warm to touch. I prefer mineral oil due to my understanding that it's perfect for electronics and it has no detergents or cleaners added that will do any long term harm to anything in my motor. I just filled my motor with a modest 50-60mL and been happy with the results ever since. I strongly strongly recommend mineral oil (preferably the laxitive kind) cuz it's easy to get a hold of and cheap.

But how did you seal it?

I just simply applied a light coat of Red RTV to the cover lips before putting them back on to keep the covers from leaking. I didn't seal up the phase/hall wires going into the center of the axle so that's my breather tube.

This RTV:
http://www.permatex.com/products/automotive/automotive_gasketing/gasket_makers/auto_Permatex_High-Temp_Red_RTV_Silicone_Gasket.htm

but the no. 1 problem with oil cooling is the bearing seal, how did you deal with that?

 

[fechter]

but the no. 1 problem with oil cooling is the bearing seal, how did you deal with tha?

If the bearings are replaced with ones that have adequate seals, then the OD and ID of the bearing would need to be sealed with RTV when installed. It might be possible to use an O ring on the ID.

To seal the end of a stranded wire, I've used solder to fill in between all the strands (tin the end of the wire), making it one big solid strand at the end, then heat shrink over it to make a seal.

Adding cooling fins to the OD of a hub motor could be done by slicing off a bunch of short sections of aluminum "U" channel and using RTV to glue them to the outside of the hub (between the spoke flanges). RTV is a reasonably good heat conductor and withstands very high temperatures.
I did something similar to my old Zappy brake drum, but in an axial configuration:

 

[GCinDC]

seems like an important point to note/restate: the bearing was rated to 100C. i was running the motor for a while at 110-120C, so surely the bearing was over 100C. the leaking occurred thereafter... perhaps if it had been kept below 100C all the time it wouldn't have leaked? another experiment to try...

just for the record, this 19" motorcycle rim with a heavy dual sport tire is an impossible combo for this motor. i first had it on a vented 9C and it got up to 100C going downhill. :lol: so i knew it was a hog, and would need a lot of cooling to run even for short periods. just so we're clear that this is not some ideal, efficient sustainable platform. it's an overheater. :lol:

before this heat wave, and going over 110C, the oil cooled motor worked for several weeks.

i should also note that it took some serious hits too. those jumps, and certainly some poor landings, god knows what those did to the bearings. :lol: all those cross section pictures made me wonder..

oh, and the wheel is horribly unbalanced.

in any case, i'm riding the same wheel vented for now, w/o any fancy scoops, scallops, fins or fans, and the motor only got to 82C on my mostly downhill inbound commute. it was raining however and ambient has dropped 20 degrees to only 80F. :lol:

 

[FastDemise]

Like fechter mentioned just don't overheat it. I ride a reasonably low power setting and the motor gets just to 120f in my 15mile trip to college averaging 20mph with bursts to 35mph. I use it more to keep my little 6fet cool at 2kW constant than the motor.

 

[GCinDC]

oh yeah, one more thing: that brake cleaner is absolutely phenomenal!

working w/ oil is totally nasty, and the few times i've touched stuff w/o gloves i've regretted it. the oil seems like it will never come off!

i yanked the oiled pads out of my 4 piston calipers, then with a couple paper towels as a backer, i sprayed the open calipers with the brake cleaner and i swear i didn't even need to wipe them. the stuff dries up and there's no oil. it's nuts.

i cleaned up a sidecover the same way the same way and simply could not believe it.

as i mentioned, i had figured that residual oil had caused the sealant to separate from the stator but nope. the brake cleaner got it all so the sealant could fully adhere.

i've been carefull to keep it away from the windings tho. is there a danger it will dissolve the insulation?

 

[teklektik]

To seal the end of a stranded wire, I've used solder to fill in between all the strands (tin the end of the wire), making it one big solid strand at the end, then heat shrink over it to make a seal.

Clever - eliminate the need for a seal. I like it.

i should also note that it took some serious hits too. those jumps, and certainly some poor landings, god knows what those did to the bearings. :lol: all those cross section pictures made me wonder..

oh, and the wheel is horribly unbalanced.

The out-of-balance wheel and impacts are unlikely to have had much effect on the bearing seal proper since it's located between the hardened races and the bearings are new with relatively close tolerances - no substantial dimensional changes/movement that could not easily be addressed with the (new) compliant rubber seal. These forces might have had an effect on the sealants, etc between the bearings and the shaft or the machined recess, but your post-mortem seems to show that you maintained good adhesion/sealing in those areas.

seems like an important point to note/restate: the bearing was rated to 100C. i was running the motor for a while at 110-120C, so surely the bearing was over 100C. the leaking occurred thereafter... perhaps if it had been kept below 100C all the time it wouldn't have leaked? another experiment to try...

Very interesting. I initially dismissed this thinking that the temperature rating was more likely associated with breaking down or liquefying the grease but a bit of Googling shows that the temp limit is specific to the RS style bearing and tied directly to the seal material. I found sites with temperature specs similar to yours and unique to seals of this type.

So, considering the oil in the bearing, the absence of leaks elsewhere, and the compromising temperatures with leaks developing coincidentally, it does seem that heat deformation of the rubber sealing surfaces is the likely culprit - which is good news.

 

[FastDemise]

I also believe mineral oil is thicker than the ATF so what leaks I have are just small oil marks on the wires but dirt collects on it and it never spills anywhere. Just to thick once it cools. Hell, the stuff is sold as a laxative and is used in many cool computer aquarium builds.

Just seems weird that when looking for good oil to use in a motor how ATF is being used and not the cheaper non-toxic mineral oil.

It just hit me I don't get much if any seepage from my wires is my rear motor has the wires face downward so the axle hole inside the motor is also facing downward so the only oil I get that leaks out is just after weeks of use it slowly traveling down the length of wire. So I don't think that counts as a leak of any sort. Just like the oil residue on a vehicle after a oil change. Just a oily shine sort of thing. Perhaps I'm just speaking too simple with my weak power compared to the many that run 5kW cont. but it works amazing for my needs and is trouble free for 2months now and 500miles.

 

[teklektik]

I also believe mineral oil is thicker than the ATF so what leaks I have are just small oil marks on the wires but dirt collects on it and it never spills anywhere. Just too thick once it cools. Hell, the stuff is sold as a laxative and is used in many cool computer aquarium builds.

Just seems weird that when looking for good oil to use in a motor how ATF is being used and not the cheaper non-toxic mineral oil.

There was a lot of discussion about this earlier in the thread, and I think ATF was chosen because it's readily available, is safe for metal/plastic/seals, has de-foaming agents, is an engineered liquid with stable viscosity over a wide temp range, retains its operating characteristics at least to 150C, is specifically intended to transfer heat, and has a very low viscosity to minimize losses due to churning.

I confess that at the time I was looking at the issue from the perspective of motors with real oil seals, not simple sealed bearings and was thinking 'thinner is better'. However, after your reports and GDinDC's experiments, perhaps an alternative with higher viscosity would offer more advantages than disadvantages - at least for certain motor constructions.

It just hit me I don't get much if any seepage from my wires is my rear motor has the wires face downward so the axle hole inside the motor is also facing downward so the only oil I get that leaks out is just after weeks of use it slowly traveling down the length of wire. So I don't think that counts as a leak of any sort. Just like the oil residue on a vehicle after an oil change. Just an oily shine sort of thing. Perhaps I'm just speaking too simple with my weak power compared to the many that run 5kW cont. but it works amazing for my needs and is trouble free for 2months now and 500miles.

There is no question that this technique is sub-optimal from a cooling perspective compared to other more complicated strategies, but as you note and GDinDC reported, it seems useful for moderate powered motors with what many would consider an acceptable level of difficulty to implement and maintain. It seems particularly attractive for folks in the geared hubbie club where air cooling is not an option because of abrasive contaminants. In those cases the clutch issue limits pouring massive power into the motor anyway, so a more effective but complex solution may not offer a worthwhile advantage over this 'pretty good' approach.

Many thanks to you and others (and especially GDinDC) for sharing information and experiences -

 

[Sancho's Horse]

I may have missed this, but is it chlorinated or non-chlorinated brake cleaner?

The reason I ask, is because it was my understanding that one of them, chloronated I suspect, when combined with oil, creates corrosive oil products. Not that this was necessarily the reason for seal failure, (type of seal trumps this) but certainly something to consider. Especially with all the wires providing nooks and crannies.

Fechter's method for wire harness protection sounds like it would do a lot to delay ingress, and the longer the solder plug the better.

 

[ian.mich]

What type of mineral oil was use, ebay is coming up with REPSOL 10W40, 15W50, and 20W50

 

[FastDemise]

This was just mineral oil from the local Walgreens. sold as laxitive at ur local pharmacy. Pretty straightforward. Nothing fancy


Edit:
http://i540.photobucket.com/albums/gg332/Fastdemise/IMAG0440.jpg