Oil cooling your hub- NOT snake oil!

[auraslip]

John, I have to disagree with liquid cooling being a band-aid. IMHO it's the future! Why have a massively overbuilt 40lb motor when a properly cooled 20lb motor will do the job?
The fact that someone here can run a standard 9c at 4.5kw and it only gets up to 60c with cooling is all the proof I need. Heck, my ventilated 9c gets that warm at 1.5kw.

The difference between running at 60c and 120c is pretty large from an efficiency stand point.

 

[Lemlux]

Brake fluid not only strips paint but it leaches copper from the linings of the brake plumbing. One of the tests used to see if it's time to replace the fluid is to measure the copper ppm in the fluid. Doesn't sound very friendly for the windings.

 

[bigmoose]

IMHO it's the future!

You know I am biased and feel this way also. The motors that were referenced in the paper I cited, I have run my hands over, and not in virtual reality. This "technology" of oil cooling works, and works well. There is a bit of windage loss, (and it was measured on the motors in the paper) but whether it is relevant or not only matters if you have or don't have the electrical power to overcome it.

You see me harping on "It's the system, boss" that's what matters. Consider what the weight would be of a 100 HP motor that did not have the "power dense" modifications. You think that would enable flight applications? The "power dense" technology ENABLED all the rest of the system. Yes, it carried a windage loss for the system; but there was no system otherwise. (If you want real power dense, we (a contractor) designed a 300HP LH2 cooled motor the size of a 3# coffee can and a 50,000 HP motor the size of a trash can. The 300HP version was verified in test... now that is "power dense.")

One of the big things is that the oil keeps the Halls from cooking off and keeps the moisture out. IMHO the Halls are dying from heat soak. When shut down, I believe there is a significant temperature spike as the copper in the middle of the laminations now sheds it's heat into the areas that were receiving airflow out at the turn ends. This heat soak now over temps the location of the Halls. The wire can take more temp than the Halls.

Some folks have worried about electronics in oil... I don't. Why? Because in grad school I worked with a 8 MeV Accelerator, that's 8,000,000 electron volts. Do ya' think any of the electronics were NOT put in a big vat of oil?

Look at the results of the guys that have engineered their installation well with oil cooled hubbies. Not one operational complaint. All raved at the new power band. There was oil leakage, but you guys figured out a great vent configuration.

Partial oil filling is the wave of the future. It will be followed with limited circulation after batteries/controllers take another step. ... at least that is my prediction!

BTW: Stick with ATF Dextron III, or Honda ATF. It is the right oil to use, it's cheap and it works.

 

[auraslip]

Good write up big moose.

Another thing that needs to be harped is overall efficiency. People running around with their windings at 120c are taking a HUGE hit to efficiency. Like a 150% increase in I^2R losses.

 

[sn0wchyld]

IMHO it's the future!

You know I am biased and feel this way also. The motors that were referenced in the paper I cited, I have run my hands over, and not in virtual reality. This "technology" of oil cooling works, and works well. There is a bit of windage loss, (and it was measured on the motors in the paper) but whether it is relevant or not only matters if you have or don't have the electrical power to overcome it.
......
Partial oil filling is the wave of the future. It will be followed with limited circulation after batteries/controllers take another step. ... at least that is my prediction!

BTW: Stick with ATF Dextron III, or Honda ATF. It is the right oil to use, it's cheap and it works.

I'm with you, but im watching this closely before doing it to any motor I own. the two issues I see are long term problems with the oil interacting with motor parts, and one big issue, the fact that you're now limited to motor temps of no more than 80 deg, since your magnets are going to be at almost the same temp as your windings. on a hot day, that could mean you have less than 40 deg delta V to cool your motor without damaging your magnets (allowing for a small safety margin).

If I ever go this route Ill be either modding my current ones, or fabing new side-covers with a much larger surface area in order to increase heat dissipation. I dont know if theres a linear relationship between surface area and heat transfer but even if its logarithmic doubling or tripling your surface area (which should be a walk in the park) should have a huge effect. adding a band of heatsinks between the spoke flanges would likely help too.

edit
just as another thought, it also makes spraying the outside of the hub with water a viable cooling option... it'd be pretty easy to have a temp alarm wired up to run a small pump that sprays the outside with water whenever temps get over, say, 75deg.

 

[fechter]

Thanks, Bigmoose. You make some good points. While I'd like to see what the windage loss looks like, it is not going to be an overriding consideration in the overall system.

I suppose adding cooling fins to the outside of the motor will help too when oil cooled.

 

[Spicerack]

(If you want real power dense, we (a contractor) designed a 300HP LH2 cooled motor the size of a 3# coffee can and a 50,000 HP motor the size of a trash can. The 300HP version was verified in test... now that is "power dense.")

And when exactly were you going to tell us about this motor? 300hp in a coffee can? Do you reckon it would fit in an ebike- and be able to keep the front wheel down?

There's alot of talk about oil affecting the internals, water spraying the outside, magnets demagnetizing etc.

ATF hasn't affected any of the components in my hub so far- I've had it apart for a look. The epoxy is still good, magnets in place, wires perfect etc. It's good enough for the electromechanical auto gearboxes in cars so it's good enough for our hubs. Some one (John?) was worried about the oil flow interfering with the operation of the halls. Well my hub still does up to 60km/h as it did before, with no apparent issues. I guess if the oil filled up with metal particles over time it may affect the way the halls operate... but probably not as any metal will stick to the magnets.

Magnets. Running them from the factory, they have an air gap of less than a millimeter. When the stator gets hot, the magnets will also get hot on the surface facing the stator. I'd bet it's above 80 celcius when your windings/stator are at 120+ which is what some people are running at. Mine runs at 75 pretty much constant with brief peaks to 90. I' be surprised if this was worse for the magnets- if they do start to fail I'll be the first to chime in and let people know. I'd also like to know where this magic figure of 80 celcius comes from. AFAIK the Curie Temp of Neo magnets is around 320 celcius. I believe the magnets may start to demagnetise a bit below that temperature, but surely not at 80? We are in the business of helping each other out on the forum which is a fantastic thing. Many minds to the problems and issues we face comes up with all sorts of clever solutions. Even with my oil cooling I'm just following on from others before me

As for water spraying- it might help but why would you? More complicated, you need temperature switches, pressure pump, water reservoir etc. You are also not taking the heat away from the source- the stator and windings. Not to mention getting water on your back tyre. Oh, that's another good reason for sealing the hub well if putting oil in it- oil on the rear tyre is worse!

As Bigmoose and others have said, why wouldn't you consider oil cooling. Car engines are water cooled. Commercial electric motors are liquid cooled. Heck, even Golden Motor sells liquid cooled electric motors! I think the advantages far outweigh the disadvantages.

 

[Spicerack]

.... So does anyone know a car that has had reasonable production volume that has very small planetaries in the transmission that could be repurposed?

Audi A4 starter planetary set being repurposed with a RC motor..

Maybe a new thread could be started for these- but yes, looks beefy! I'll try and take one of mine apart for a looksie.

 

[motomech]

I don't have much to add to this except to relay my experiences as a Japanese motorcycle technician for 25 years
As mentioned earlier, most motorcycle engines have the A.C. generator running in the engine/transmission oil and charging systems have always been the Achilles heel of the otherwise, generally reliable Nipponese power-plants. Charging system failures were particularly chronic in a long running model line, the Suzuki GS series of transverse mounted, in-line 4 cylinder motorcycles of the late '70's and '80', but were common in all the UJM's[Universial Japanese Motorcycle]of the period. The components that failed were quite regular and predictable and the trouble-shooting procedure simple.
Once the battery was established to be sound and having a full charge, a "drain" test was performed. With the ign. switch off, a multi-meter[set to Volts DC] was placed in series with one of the battery leads with the expectation of a reading of zero[no-drain]. But rectifier problems were common, accounting for more than 50 % of charging system failures. Readings in multiples of three, 4.5V, 9V or 13.5V represent the number of failed diodes used in 3-phase A.C. rectification. Rectifier/regulator assemblies are non-repairable[hermetically sealed, for our protection ]. I'm sure repairing these units would be "small potatoes for the Gurus here, except that the insides of the module was tightly packed with some sort of gel that had dried into concrete hardness, making it not worth the trouble.
The second phase of trouble-shooting consisted of checking D.C. output. Most service manuals and the manufacturers would have the tech. reading current, but in my four plus decades of working as a mechanic, I have never gone to the bother. I simply place a quality analog meter across the battery an look for a needle deflection of 1 to 2 volts above the battery's full SOC[13.8V, more or less, for a 12 V system]as the engine is reved up. No D.C. output-on to the next step.
At this point, the manufacturers would have the tech. looking at the A.C. output at the stator windings, but I seldom did this. I simply started checking continuity, one of the three stator output wires to ground. Inevitably, one of the windings was shunted, "cooked", we would say.
I really don't have any epiphany-type conclusions to draw from this tale, except to state the obvious. There did seem to be a correlation to the "freshness" of the oil and the stator failure rate. "Cooked" oil equates to "cooked" stator.

Another common failure on motorcycle engines is a leaking counter-shaft sprocket seal, allowing engine oil to migrate over the drive chain and sprockets. Not necessarily a bad thing for the lazy rider, it became an automatic chain oiler.
Which brings me to the point of this rather long, rambling post.
Why not place the vent hole in a strategic location behind the gear cluster and let it vent.
And I though my days of an oily right pant's leg was over

 

[John in CR]

Bigmoose,

In a direct drive hubbie, how much oil would you put? Would it be enough that at speed there would be a spinning river of oil over the windings, or a lot less? I'm trying to visualize the dynamics of that fluid once the motor is spinning fast enough for centrifugal force to push it to the perimeter. I've got few sealed DD's that aren't powerful enough, and don't mind trying to find the limits with one in particular. It's already strung on a rim, and I've got CA's to measure before and after consumption, along with programmable controllers to test different limits. Time is a premium, but this is interesting enough to check out, and then compare directly to my method of air cooling.

WRT to Halls, I wasn't worried about them working in oil, but instead the 10-15mph river of oil eventually breaking off one of the 9 tiny wires going to them that will be exposed to the flow.

Oil filled hubmotor may very well be the future, but they'll be specifically designed for it. OTOH oil filled hubbies just so they can be pushed further into stator saturation or bogged down with too high a load is a dead end unless it's just experimentation toward building a better motor design.

John

 

[auraslip]

I agree with the last part john. 9c/mxus is saturated pretty quickly.

 

[Spicerack]

WRT to Halls, I wasn't worried about them working in oil, but instead the 10-15mph river of oil eventually breaking off one of the 9 tiny wires going to them that will be exposed to the flow.

John

Sorry John, I misunderstood you! My wires are all set in epoxy so it shouldn't be a drama. Come to think of it, that's probably the way to go if oil cooling- set all wires in epoxy so they can't move.

I've no idea which the best oil fill is for it all to work- what I did seems to work! Next time I might try to be a bit more scientific....

 

[sn0wchyld]

...................................................
Magnets. Running them from the factory, they have an air gap of less than a millimeter. When the stator gets hot, the magnets will also get hot on the surface facing the stator. I'd bet it's above 80 celcius when your windings/stator are at 120+ which is what some people are running at. Mine runs at 75 pretty much constant with brief peaks to 90. I' be surprised if this was worse for the magnets- if they do start to fail I'll be the first to chime in and let people know. I'd also like to know where this magic figure of 80 celcius comes from. AFAIK the Curie Temp of Neo magnets is around 320 celcius. I believe the magnets may start to demagnetise a bit below that temperature, but surely not at 80? We are in the business of helping each other out on the forum which is a fantastic thing. Many minds to the problems and issues we face comes up with all sorts of clever solutions. Even with my oil cooling I'm just following on from others before me

As for water spraying- it might help but why would you? More complicated, you need temperature switches, pressure pump, water reservoir etc. You are also not taking the heat away from the source- the stator and windings. Not to mention getting water on your back tyre. Oh, that's another good reason for sealing the hub well if putting oil in it- oil on the rear tyre is worse!

As Bigmoose and others have said, why wouldn't you consider oil cooling. Car engines are water cooled. Commercial electric motors are liquid cooled. Heck, even Golden Motor sells liquid cooled electric motors! I think the advantages far outweigh the disadvantages.

how long have you been running with oil?

I think you mis understood what I was saying re water spray. I meant that you'd still have a oil filled motor, but that you could potentially spray the outside of the sealed motor with water, thereby cooling the outer covers more. this would result in a larger delta T between the windings and the outside covers, meaning faster transfer of energy (ie heat). the temperature switch would be easy... wire up the alarm output of a temp alarm to a relay that switches on the pump, rather than just a buzzer.

Same thing with high surface area covers... since the covers are the final point of interaction between the heat source and the outside world a larger surface area naturally means faster energy dissipation, much the same as above. So with a cooler outer cover, the heat will be pulled away from the windings at a faster rate.

as to the magnets, I was under the impression that they start to fade at around 80deg, just going on what I've read here on ES. I'm more than happy if im wrong, it could be that they only suffer minor temporary demagnetisation above 80 or something. And I can say for myself atleast that while my windings may at times reach 80+ deg, my covers/magnets dont go above 40. I wouldn't think the inside of the magnets would be significantly higher than the outside.

edit
heres a site that looks like it has the details...
http://www.kjmagnetics.com/blog.asp?p=temperature-and-neodymium-magnets
the magnets in my ht are 35x14x3 mm, giving a pc of just .29. thats slightly higher than these magnets
http://www.kjmagnetics.com/proddetail.asp?prod=BX8X02
which have a lower pc due to their shape and an thereby an max operating temp of 80deg before damage. so the motor magnets we have might be good for slightly higher temps than 80. It at least explains where that # comes from. the good news to take from it is that while you might suffer some permanent loss at above 80 deg, you wont loose any extra unless you exceed your previous max temp. Ie if you lost 10% by going over 80 up to 90deg, you wont loose any more in the future until you exceed 90 deg.



sorry if i sounded like I was attacking your ideas, far from it. I was trying to build on them.

 

[Hillhater]

...As Bigmoose and others have said, why wouldn't you consider oil cooling. Car engines are water cooled. Commercial electric motors are liquid cooled. Heck, even Golden Motor sells liquid cooled electric motors! I think the advantages far outweigh the disadvantages.

The oil in car/truck/bus IC engines plays a major part in transfering the heat from the internal components to the surrounding walls of the water jacket. Many vehicles now have oil coolers also to further help disperse the heat.

 

[flathill]

...As Bigmoose and others have said, why wouldn't you consider oil cooling. Car engines are water cooled. Commercial electric motors are liquid cooled. Heck, even Golden Motor sells liquid cooled electric motors! I think the advantages far outweigh the disadvantages.

The oil in car/truck/bus IC engines plays a major part in transfering the heat from the internal components to the surrounding walls of the water jacket. Many vehicles now have oil coolers also to further help disperse the heat.

Yup and even "air cooled" engines like the porsche flat 6 are really oil cooled...I mean does your car take 12 13 quarts of oil!

The magnets weaken with heat which provides unintended flux weakening which actually will increase your top speed (less back emf)

As long as they cool down when you decelerate they will recover and you'll have full torque off the line

They are spmore sensitive to being demag'd when hot with a reverse pulse (bad timing) but it is not usually a problem with radial flux outrunners

 

[Spicerack]

I think it's a case of- we're all right! I didn't take any of it as an attack- just thought some of it was complicating a simple solution. I considered the full circulating system with radiator but thought I'd give the simple way a go first- must admit I'm more than happy with it.

Interesting reading on those website, thanks. I didn't think the demag T would be that low when the curie T is 300+ but there you go. You know what, if you ain't living life on the edge, you're taking up too much room!

I've been running oil for a month now. Not long enough for long term testing obviously but enough to see the massive difference for me on the first ride, and it's repeatable. Just got back from Perth CBD. 20 minutes each way for 12k's. Quicker than by car. The motor again hovered in the 70's and only goes up when first accelerating from a stop. Comes down very quickly when up to speed.

And yes, I forgot to mention that car engines are oil cooled too. All my mazda rotaries and my Suzuki RE5 rotary have coolant rads and oil coolers on them...Sheesh, you can't get away with anything on this forum

 

[bigmoose]

John,

I would fill the unit about 1/4 to 1/3 full. It's going to splash around and such, but mostly be slung to the circumference where it will do the most good. As for Hall wiring, if it fails, stake it down with more tywraps and/or increase the wire diameter to counteract the loads of the circulating oil. We did not have Halls in the power dense motors that I referenced, but saw no erosion problems with the windings.

Here is a good summary paper on the Liquid Hydrogen cooled "power dense" motor/generators that I referenced above. http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA537468 and a picture of a 600 Kw unit from that paper. Yes that ruler is in centimeters!

 

[fechter]

Here is a good summary paper on the Liquid Hydrogen cooled "power dense" motor/generators that I referenced above. http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA537468 and a picture of a 600 Kw unit from that paper. Yes that ruler is in centimeters!

That looks like it will fit on Luke's bike.

I was thinking it would work well to drill and tap a hole in the side cover near the edge to use as a drain/fill hole. With the hole at the bottom, it could drain and with the hole about 1/3 up (just rotate the wheel), it would only take the right amount of fluid, then start coming out like a car differential. Something like a 1/8" NPT tapered pipe plug (which is really more like 10mm) would be a good size.

I was also thinking it might be good to use a piece of flexible tubing coming out of the axle along side the wires as a vent. The open end of the tube could be located higher than the axle to prevent oil from leaking out and it would be sealed at the axle. It could be bad if oil comes out of the motor and gets onto the tire.

 

[auraslip]

What about a gore-tex patch over the drain/fill/ vent hole?

 

[Grinhill]

Another idea for ventilation - the rubber plug could just have a small slit cut in the centre of it (e.g. with an exacto knife). Should only open slightly under pressure.

 

[Spicerack]

Another idea for ventilation - the rubber plug could just have a small slit cut in the centre of it (e.g. with an exacto knife). Should only open slightly under pressure.

Tried using the rubber bung as vent but it leaked. I think the angled small vent near the axle is good. I think bigmoose suggests a thin plastic tube inside the vent hole to further prevent oil seepag, or even fetcher's idea above.

 

[teklektik]

I was thinking it would work well to drill and tap a hole in the side cover near the edge to use as a drain/fill hole. With the hole at the bottom, it could drain and with the hole about 1/3 up (just rotate the wheel), it would only take the right amount of fluid, then start coming out like a car differential. Something like a 1/8" NPT tapered pipe plug (which is really more like 10mm) would be a good size.

I was also thinking it might be good to use a piece of flexible tubing coming out of the axle along side the wires as a vent. The open end of the tube could be located higher than the axle to prevent oil from leaking out and it would be sealed at the axle.

It seems this project is slowly re-inventing the solutions in place in cars for almost 100 years

Since we're using the primary lubricant/cooling fluid in cars (ATF) and the same fill technique used in sealed-oil components (differentials and transmissions), let's use the same technique used in pre-PCV engines to relieve crankcase pressure without venting oil - the labyrinth and breather tube.

The tube out through the axle that fechter mentions can be rigid thin-walled metal or high-temp plastic to minimize diameter. It doesn't really need to be routed anywhere on the outside. (Ideally, it could just be a small diameter hole through the opposite axle stub). On the inside it can go to a simple baffle attached to the stator above axle that might allow for 90-100 degrees of axle tilt to accommodate horizontal to vertical dropouts. Oil spray, dripping oil, etc on the inside of the motor will be shed by the labyrinth and will not penetrate to the inside top of the baffle where the breather tube opening is positioned. For a dedicated DIY solution the labyrinth might be simpler since the motor/dropout orientation is known.

We can also borrow the same seal solution used in cars if deeper oil fills are desired or if seal weeping becomes a problem; the current simple lipped seals can be replaced with the type used for auto main seals that have inner angled directional ridges to forcibly wipe seeping oil back into the motor instead of allowing it to slowly migrate under the seal to the outside.

 

[aroundqube]

I am interested in another possible affect from adding a small amount of ATF to a gear hub motor, other than cooling. When posting future results would members please also describe if there are audible changes. I am hoping for a reduction in noise.Thanks

 

[John in CR]

As far as testing goes, how much is a comparison of before and after no-load speed and current going to tell me? That's enough to quantify windage (oilage in this case), isn't it? Then I could just put a hole near the axle and add oil in small increments, and record no-load speed and current at each step...maybe for good measure also record current at a fixed speed (typical cruise).

My issue is that I'm not worried about cooling that much, because my ventilation approach works so well, especially for high speed riding. I believe that motors only get burned up from abuse, bogging down on hills or the power chasers pushing current levels into stator saturation. Geared hubbies are different in that they run hotter in normal operation, because they are more thermally restricted. Ventilation however comes with a big compromise,the motor being open to the environment. It hasn't proven to be an issue in normal riding, so I won't take a ventilated motor riding on the beach or trail riding across streams or through heavy mud. Oil in the motor would actually be plus for these types of riding, and if it helps dissipate heat from short load peaks without too big a performance hit, the it might be worthwhile on a couple of my bikes.

John

 

[bigmoose]

The magnets weaken with heat which provides unintended flux weakening which actually will increase your top speed (less back emf)

As long as they cool down when you decelerate they will recover and you'll have full torque off the line

Sorry, but I need to correct this. When you overtemperature the magnets you "knock down" their flux. It does not recover when the magnets are cooled. Their flux is permanately "knocked down".

http://www.ndfeb-info.com/temperature_ratings.aspx

An irreversible but recoverable loss occurs when the output falls but does not return when the magnet cools down (e.g. the high temperature takes the Intrinsic working point beyond the knee of the Intrinsic curve, causing demagnetisation) but this would be recovered if the Neodymium magnet is remagnetised. For all extents and purposes, this output is lost because the magnet will not be remagnetised during practical application... When cooled, such a demagnetised magnet will have the original Hci but a lower Br (the Br will have increased by the reversible temperature coefficient amount applied to the reduced high temperature Br).