Oil cooling your hub- NOT snake oil!

[amberwolf]

I've heard 80c is when they start to break down. But is this permanent? Can any someone quantify this? Is it immediate? Does the speed the damage occurs speed up lineally? This seems like VERY important information.

Take a look around for "Curie temperature"; that is when magnets permanently lose their magnetism. Differnet materials ahve different behaviors at different temperatures, and I dont' know what any of them actually are, but AFAIK it's a permanent loss above the Curie temperature for that particular material.

[motomech]

Yes, it exists, so perhaps "theoretical" i not the best choice of word.
I'm just saying the "Currie Temp., at 1300 deg.s F for Ferris Iron, is way above the point where things like wires and insulation and the oil itself would be ash.

[motomech]

As a motorcycle mechanic, I have replaced dozens of oil-in-bath, A.C. generator, 3-phase stators over the years, it was always the same thing. Shunts in the winding[s] due to insulation failure.[no Hall wires to melt]
When trouble-shooting the charging system, a zero reading of D.C. Volts output would take me to the stator phase wires. In the interest of expediency, instead of reading A.C. output, I would check continuity across combos of phase wires. First zero Ohm reading and the stator got tossed.
One would think, if the magnets were losing strenght. there would still be output, abeit, at a lower output.
Never saw that the few times I checked A.C. output.
For me, stators were essentially go/no go devices.

[Spicerack]

Neo magnets are a different beast to the older style magnets used in bikes (have they changed to neo yet?).

Neo Curie Temp is around 320 celcius but they begin to permanently demag below that. 320 celcius and they are totally toast. Discussed from the bottom of page 4 onwards of this thread.

[gensem]

gensem - the oil conducts the heat from the windings much, much better than air. The oil then conducts the heat to the covers. Since the covers are much larger in surface area, they radiate heat much better than ventilated windings would. Also the covers get much better airflow then the windings.

Auraslip I understand that oil will conduct alot more heat to the covers, but I wasnt comparing a dd hub motor with oil against a ventilated one.
I was just saying that the covers ll probably not be able to dissipate all that heat with you are using the motor for sometime.

[auraslip]

kingfish - that was a lovely post!

Anyways, I created an"askengineers" post on the issue oftemperature on reddit and got this reply.

Not really (this is besides the point, but if you're curious...) The magnets actually increase in temperature due to something called Eddy currents, it has nothing to do with contact or an air gap, and they heat up from the inside, outward. As long as your motor is running and producing the alternating fields which ultimately drive your bike, your magnets are going to heat up internally.

The more your magnets heat up, the less magnetic field they produce, and the less power your motor can produce. The whole idea is to keep your magnets either as cold as possible. Now your permanent magnet most definitely has extra additives that make it perform better at higher temperatures, but there's no way of me knowing exactly what that is. Needless to say, a pure Nd2Fe14B magnet has a magnetization curve that looks like this. I took that measurement myself right now with a pure magnet, and it's very accurate. As you increase temperature, the magnetization decreases (quickly!) as you approach the Curie temperature in the neighborhood of 600 K / 300o C. Now, 120o C is about 395 K, so take a look at where that is on the graph. It's at a magnetization of 110 emu/g. If you increase that to 400 K, you'll have a magnetization of 109 emu/g. This is less than a 1% drop in magnetic field. Now, this doesn't directly translate into a 1% drop in performance (your performance drop will be a little more than 1%), but it's close enough to get an idea of how the performance of your motor is effected by a 5o C increase in temperature.

Now I don't understand motors that well. Automatic Transmission Fluid gets in direct contact with the magnets in a hub motor? I wonder how that works, I'm not getting much through Google. Anyway, general intuition of mine says that even in electric cars you want adequate ventilation to your magnet so it can act as a heat sink and cool down your magnet. If sealing the motor and filling it with ATF brings the liquid into direct contact with the magnet, there will be much better heat transfer, so your magnet will cool down much better as long as the ATF is cold, and your magnet would heat up much faster if the ATF were hot. Based on what you say, it sounds like this method of introducing ATF to the magnet is suppose to work by cooling the magnets down more efficiently. If that's the case, then your bike will see slightly better performance.

As for the 80o C, that sounds like a relatively arbitrary temperature. It won't be a permanent degradation of performance at that temperature. Once your bike cools down, your magnet's performance will go back up nearly all the way. I can't quantify it without some technical data on your magnet.

With this picture attached.


He explicitly says that the magnets will return to normal after they cool down?

[Kingfish]

That sounds correct as long as the rated-temp limit of the magnet is not broached. That’s one-steep and unfriendly curve.

I liked his eddy comment and frank explaination. Good stuff; thanks for sharing
~KF

[fechter]

There are definite differences in magnet grades that affect the temperature where demagnetization occurs. It would be hard to know what grade is installed in a motor unless you bought the magnets yourself. Worst case is around 80C. Pretty much any NeFeB magnets can handle that much. They make "H" and "VH" grades which can handle higher temperatures without permanent loss of energy.
I had some magnets that I ran through a steam sterilizer (273F, 134C) and they were nearly completely demagnetized. Others going through the same cycle came out with no apparent degradation.
80C is a safe limit for most any magnets. Many will take 100C. We don't want the motor to get that hot anyway.
Loss of power due to magnet heating may be less significant than loss of power due to copper heating.

[bigmoose]

Take a look at this page of this thread, it may help with the info you are looking for.
http://endless-sphere.com/forums/viewtopic.php?f=30&t=32049&p=463988&

I don't believe magnet "knock down" will be a problem with the oil cooled hubbies.

[John in CR]

Bigmoose and others, I was crunching comparative numbers for oil cooling vs air cooling of DD hubbies using the convective heat transfer calculator here about halfway down the page http://www.engineeringtoolbox.com/convective-heat-transfer-d_430.html. I used temp limits of 100°C at the covers for oil, and 200°C for the stator with air ventilation. Do you think those are realistic, reasonably conservative, and fair for comparison? My biggest issue has always been the magnet limits for the oil cooling case.

FWIW, the result really wasn't even close for continuous operation, but I am still interested the oil approach to boost motor limits for a beach cruising and mud riding bike where oil cooling has obvious benefits, and oil finding it's way to a tire is no big deal.

For a geared hubbie oil appears to win out, because of the drastically greater surface air of the outer shell compared to the stator, along with the trouble getting air flow through the stator and air gap, so the heat transfer coefficient would lower than is possible with a DD. Some relatively small changes to the rotor bell could drastically change air cooling a geared hub, especially since the area between the spoke flanges can easily be turned into an optimized centrifugal fan.

John

[teklektik]

Some relatively small changes to the rotor bell could drastically change air cooling a geared hub, especially since the area between the spoke flanges can easily be turned into an optimized centrifugal fan.

[bigmoose]

John, frankly I do not think we can answer this with the typical analysis tools that we have available. Oil is going into mine, for three reasons. I'm a big guy with too much mass to haul around, we have now documented evidence the oil has stopped hall failures due to shutdown or otherwise heat soak. Third, it will stop the corrosion issue.

Now think about your magnet problem, compare it with the hall issue. The magnets won't be hotter than the halls.

The answer is to put a thermistor inside the hub, in a place where the oil will be when running and read out the data. That is IMHO what is going to be required to put it to bed.

Frankly, I am overjoyed with the evidence that is being reported as people try it. I believe it will be "the standard" in about 3 or 5 years due to the above reasons.

[auraslip]

So the guy responded again. Quite interesting stuff.

Okay, so we're talking about something called irreversible losses. This is when you heat up the magnet enough and cool it back down, but it won't retain its peak performance.

The magnet I work with is very similar to what they run in a Toyota Prius. But more specifically, this is the magnet used for torque in the motor. These are specialized magnets with extra dysprosium, a special element that is used specifically to increase the operating temperature, because these magnets are subjected to temperatures above 150o C. It also makes the magnet much more expensive, since dysprosium is rare. The neodymium magnet used to power the windows in the Toyota, however, don't get subjected to high temperatures so they leave out dysprosium. These magnets can't get above 80o C, or else irreversible damage occurs. To give you an idea of numbers for these window-motor-magnets, by the time you get to 100o C, you'll have a %2 permanent decrease in performance of your magnet. At 120o C, you'll have a ~5% decrease in performance. At 140o C, it's about 12% permanent decrease in performance. These magnets that have poor high-temp performance are the same magnets found in CD and DVD players, laptops, HDDs, etc.

That being said, I'm not sure what alloy your neodymium magnet is made of. As you can see, whether or not it has dysprosium plays a huge role in the irreversible loss of performance. In one magnet, it can be used safely at 150o C, in another magnet it can't get above 80o C while keeping its full potential. There are a bunch of different magnetic alloys that lie in between as well. If I got more information on your specific magnet that is used with your electric motor, I could give you more information.

But earlier you stated that you have a temperature gauge on your windings, and it's reading about 120o C. The really important quantity is the temperature of the magnet itself, and that won't be the same temperature as the windings. However, this is where my knowledge stops--- I'm not sure if the magnets typically run at a temperature above the windings, or below it.

What I take from this is that we need to find out what type of neo magnets they use to make these hubs! Until then, I think I can hang out in the 80c zone. With my air ventilated 9c I get it up to 120c pushing 4kw. With my new H series clyte in an oil bath I think I can do 4kw and stay at or under 80c.

edit: I just had one more thought: does the irreversible damage stack? As in at 100c you have 2% damage. If you take it to 100c again, do you get 2% MORE damage?

[johnrobholmes]

More than likely the cheapest grade N magnets are being used. Otherwise there would be some boasting of the grades used. Probably N35s.

[gensem]

So if you have magnets that are good for 80ºc and winding that are good for alot more temperature, adding oil to a DD ll have even less headroom.
We seriously need some tests.

[auraslip]

There has been some test. Quite inspiring results too.

IMHO 80c is a good practical cut off anyways. Efficiency is already pretty poor that high anyways.

[Spicerack]

Bigmoose hinted on something that we havent really covered. I have my temp sender on the windings- so the temp is whatever the windings are at. Perhaps a better idea is to have the sender on the windings but low in the motor so that it's immersed in the oil. This way we will know for sure what temp the oil is and therefore get a better idea of what temp the magnets are running at. At the moment my figure may be a bit higher than magnet temp as the sender is up high- not immersed in oil.

[auraslip]

That's a good point. I wonder what the difference will be. On my next bike I'm going to have a temp sensor on the motor and the controller. Maybe I'll have two in the motor; top and bottom?

BTW, does anyone know an easy/pre-made circuit for a popular themoresitor that'll use a simple lcd and run off 12v? The hobby kink sensor runs off a popular $1 lil part, but the unit itself costs $10.

[Spicerack]

Nah. Just buy a bunch of $5 bbq temp probes, dissect them and join the boards together to make a big display unit. Have four temps- oil, windings, controller and rectal.

Well, maybe not the last....but you get the idea... or one of those inside/outside units. If you're really speccy you could install one with a high temp alarm on it.

[John in CR]

Bigmoose hinted on something that we havent really covered. I have my temp sender on the windings- so the temp is whatever the windings are at. Perhaps a better idea is to have the sender on the windings but low in the motor so that it's immersed in the oil. This way we will know for sure what temp the oil is and therefore get a better idea of what temp the magnets are running at. At the moment my figure may be a bit higher than magnet temp as the sender is up high- not immersed in oil.

Once it's spinning I'm sure the oil is pretty much the same depth all around the perimeter of the shell and reaches the same point on the windings. It would be nice to see exactly what happens inside though, especially on a geared hubbie with the shell spinning one direction and the motor bell spinning 4-5 times as fast the other direction.

[d8veh]

Too much intellectual masturbation. People are trying to solve a problem that doesn't yet exist. More people should try oil cooling and if it gives problems, its those that should be solved. My understanding is that so far it's been successful. I'd try it myself, but the way I ride my BPM, it doesn't even get warm. I'll probably try it in my Cute next time I remove the back wheel if I can find a way to seal the side-plate thread that will allow me to get it apart again. Any suggestions will be welcome.

[motomech]

As an Ex Japanese motorcycle mechanic, my favorite sealant by far is the gray, semi-drying, non-silicone based stuff sold by the manufacturers.
Various names, Hondabond, Yamahabond, etc.
All the same stuff and it's perfect for that application.

[gogo]

As an Ex Japanese motorcycle mechanic, my favorite sealant by far is the gray, semi-drying, non-silicone based stuff sold by the manufacturers.
Various names, Hondabond, Yamahabond, etc.
All the same stuff and it's perfect for that application.

You'd think there would be "improved" stuff by now, but you're right, Yamabond worked great on motor side covers and gearcases.

[zombiess]

I like rightstuff

[gogo]

I think its this Yamabond 5 stuff that's suitable. It would peel off nicely for the next application.
http://www.yamaha-motor.com/yamalube/categories/adhesives/27/yamabond/359/details#productscroll

This stuff might be similar and easier to obtain:
http://www.permatex.com/products/automotive/automotive_gasketing/specialty_gasket_maker/MotoSeal_1_Ultimate_Gasket_Maker_Grey.htm