Definitive Tests on the Heating and Cooling of Hub Motors

John in CR said:
Of course the easiest route to cooler temps (other than proper controller settings) remains the least used. That is directing more air flow toward the hubmotor shell.

or a bigger motor which will have a lower electrical resistance
 
Hummina Shadeeba said:
John in CR said:
Of course the easiest route to cooler temps (other than proper controller settings) remains the least used. That is directing more air flow toward the hubmotor shell.

or a bigger motor which will have a lower electrical resistance

Another easy way, though not as easy or cheap as air deflectors near the rear wheel. It doesn't necessarily have to be bigger though, just more efficient in the operating range required. My 14kg ventilated HubMonster slaughters the competition in the temperature department for moving 200kg or less at street and highway speeds. Of course I also use a non-typical controllers settings to run 28kw peak power input and stay cool even on mountain roads, ie 1.5:1 phase to battery current limit settings.

My much heavier 273 motor gets hotter just riding at 60mph or so even on flat roads, so bigger isn't always better in terms of heat.
 
The motor vents need to be designed to pump air in on one side and pull air out on the other. Nothing a engineer couldn't figure out quickly. It would likely take me a day or two or more to figure so i'll leave it to those with the Computer programs and related skills :)
 
Hummina Shadeeba said:
Why would running such a motor amp to battery amp ratio be beneficial? Just reduced?

The typical high 2.5:1 Phase to Battery current limit ratio so commonly used is a main source of heat problems. The reason is because it too often creates high phase current when you don't want it, eg at lower speeds during ascents.
 
eCue said:
The motor vents need to be designed to pump air in on one side and pull air out on the other. Nothing a engineer couldn't figure out quickly. It would likely take me a day or two or more to figure so i'll leave it to those with the Computer programs and related skills :)

Even optimized intake and exhaust hole shapes create very little air flow through a motor. The rpms are too low for such small effective blades to move much air. If you combine exhaust holes at the most extreme radius practical with exterior blades, then you can move alot of air except at very low speeds. I've used that approach of turning my motor sheel into a centrifugal fan for years with great results. I'd argue that it's the most effective quiet form of hubbie cooling, since I can make a 10 mile run generally uphill running at highway speeds 65-75mph pushing peak power input well above 20kw and arrive with a stator temp below 70°C. With 28kw peak power controller settings pushing a 400lb all up load I've only pushed the motor to my alarm temp setting twice in the past five years. Once was blasting up a continuous grade of 20% where I accelerated hard coming out of every switchback, and the other was running a 12 mile loop with 2 significant climbs riding as hard as I dared.
 
John in CR said:
eCue said:
The motor vents need to be designed to pump air in on one side and pull air out on the other. Nothing a engineer couldn't figure out quickly. It would likely take me a day or two or more to figure so i'll leave it to those with the Computer programs and related skills :)

Even optimized intake and exhaust hole shapes create very little air flow through a motor. The rpms are too low for such small effective blades to move much air. If you combine exhaust holes at the most extreme radius practical with exterior blades, then you can move alot of air except at very low speeds. I've used that approach of turning my motor sheel into a centrifugal fan for years with great results. I'd argue that it's the most effective quiet form of hubbie cooling, since I can make a 10 mile run generally uphill running at highway speeds 65-75mph pushing peak power input well above 20kw and arrive with a stator temp below 70°C. With 28kw peak power controller settings pushing a 400lb all up load I've only pushed the motor to my alarm temp setting twice in the past five years. Once was blasting up a continuous grade of 20% where I accelerated hard coming out of every switchback, and the other was running a 12 mile loop with 2 significant climbs riding as hard as I dared.


Good job you should team up with grin tech for a collaboration build / design :)

I say that as they could help you bring it into production and you could help them design a better motor case. Do you have a performance link on it ?

It likely can be tweaked further using a engineers computer program and 3d modelling.


Im liking the sounds of it
 
John in CR said:
Hummina Shadeeba said:
Why would running such a motor amp to battery amp ratio be beneficial? Just reduced?

The typical high 2.5:1 Phase to Battery current limit ratio so commonly used is a main source of heat problems. The reason is because it too often creates high phase current when you don't want it, eg at lower speeds during ascents.

with your reduced ratio you reduce your low speed power potential with the rationale being its the most inefficient speed to be feeding big amps?
 
Cowardlyduck said:
justin_le said:
Oh man, I have SOOO many interesting new tests on Statorade from the last 3-4 months of lab work that I need to share here...
Thanks for the reply Justin!
And so awesome that this is still being tested by you!

Yeah, doing the experiments is all the excitement but formally writing and documenting the results is a little more of a chore. In case people haven't noticed though we now have way more motors listed on our simulator with and without Statorade option, and the 2017 simulator updates allow you to see the expected steady state temperature with the basic motor sim instead of having to use the Trip Simulator application.
Simulator Motors with Statorade.jpg


One of the questions that I wanted to understand better is the relationship between the quantity of Statorade added and the maximum RPM at which it has a pronounced effect. We would expect that as the motor spins faster, there is more centripetal force pushing the fluid out and that at some speed the ferrofluid would be thrown from the gap completely and would cease having much of an effect at all.
I showed data to this effect with the 9C+ motor tests last summer, please refer to this post here
https://endless-sphere.com/forums/viewtopic.php?p=1309536#p1309536

Sterling from Bend Electric Vehicles sent us a large Cromotor for simulator testing and so we opened that up to properly seal it for Statorade and so got to run a set of tests on this wider and more powerful motor.
Cromotor Open.jpg

I went in 1mL increments up to 12mL addition to see the effect at 100, 200, and 300 rpm motor speeds, and the resulting core to shell conductivity graphs are quite nice and revealing. The large rise in conductivity happens between 2-4 mL at 100rpm, 3-5 mL at 200 rpm, and 5-7 mL at 300 rpm.
Conductivity vs RPM, Cromotor.jpg


Similarly and not unexpectedly, when I looked at the no-load motor current (converted to a drag term), the first pronounced step increase in motor drag happened at 7mL with the 300 rpm test, but 5mL at 200 rpm. The 100 rpm data is somewhat suspect as the no load power measurement (on the order of 10 watts) isn't far from the +- ~2W measurement error
Cromotor Drag Torques.jpg


This data results in two important ramifications for end users of Statoarde which hadn't been fully articulated before:

  • 1 The optimal fill level for Statorade Ferrofluid depends on the application RPM. If you want to cool a motor that is mostly overheating as a result of being loaded on slow hill climbs, then even a large hub like the Cromotor only requires like 4-5mL of fluid. On the other hand, if you are cooling a motor because you are doing high speed racing, then 4-5mL won't have much effect and a larger fill quantity is required. This higher fill will conduct great at low speeds too, but it will result in a larger increase in rolling drag than the optimal fill for low speeds.
  • 2 If you don't know in advance how much Statorade to add, the technique of looking at the no-load current on the motor while adding it in small quantities until you see a detectable increase in current remains totally valid. However, you should do this at the target RPM at which you are running the motor to get the minimum (hence optimal) fill level at the speed you are running. If you are running at 60 kph on a 24" wheel, then the motor is spinning a little over 500 rpm. You would want to be looking at the no-load current from the controller at this particular speed and stop adding fluid once you see a small increase.
 
I also had a chance to do these tests with Ferrofluid quantity fill vs. RPM in small diameter skateboard hub motor wheels. We get asked all the time how Statorade will work in RC style motors spinning at high RPM's, and so I've wanted to actually characterize rather than just extrapolate. Direct drive skateboard motors in particular are prone to overheating issues and have a lot of merit for testing.

We had someone drop off an Inboard electric skateboard with damaged controller that provided a good candidate motor for these experiments. I put tape over the 3 vent holes that are normally on the outside of the hub.

Inboard Motor in Tunnel.jpg

The motors were run from up to 2500 rpm with wind speed going from 0-40 kph.
Inboard Motor Results.jpg

With 1mL, the effect of Statorade peaks clearly at around 500 rpm. With 2mL, it peaks at more like 800 rpm. And with 3mL it seems to plateau a little around 1250 rpm.

What interested me most was to see how these results compare with larger diameter motors. In theory, the centripetal accelleration on the fluid is
acc = (2 * Pi * RPM / 60)^2 * r

So doubling the RPM would result in four times the force away from the stator core, while doubling the radius of the motor at the same RPM should double the force. If we have a motor with twice the radius, we should expect the max RPM we can spin it at for the same force on the Statoarde to decrease by 1/Sqrt(2).

If we look at my previous post on the Cromotor, that has an air gap diameter of about 200 mm. The acceleration on the FF is

10.9 m/s^2 at 100 rpm (just over 1g)
43.8 m/s^2 at 200 rpm (almost 4.5g) and
98.7 m/s^2 at 300 rpm (exactly 10g)

At 300 rpm in that motor we're dealing with 10g forces on the fluid, no surprise that small volumes would flatten out and a larger fill was needed. When you spin the motor at 400 rpm the force is ~18g, and at 500 rpm the G forces at the magnets reach 28g. I've done lots of tests of standard DD hub motors at 400 rpm in the wind tunnel and with sufficient QTY Statorade definitely still works well, so forces up to 20g appear fine.

On this skateboard wheel, the air gap diameter is more like 45mm. For this we have

500 rpm = 6.3g
800 rpm = 16g
1250 rpm = 39g

On a skateboard wheel, 1250 rpm is about 20 kph / 12 mph. Not super fast, but not a useless speed either and well in the ballpark of what these hubs would be doing up a hill climb.

Going in the other sense to bigger hubs, I also tested the large diameter BionX D motor, see
https://endless-sphere.com/forums/viewtopic.php?p=1361795#p1361795

On that one, the 300 rpm test corresponds to a force of 16g on the motor.

Anyways I'm hoping that by collecting more info like this over a range of wheel diameters and rpm's, we can also come up with some good rule of thumb guidelines for Statorade fill levels based on some simple equations given

a) diameter of the motor
b) width of the stator
c) the motor RPM.

And also have a reasonable max RPM number where modest amounts of FF cease to have much of an effect as the acceleration is just too high. For instance, an 80mm outrunner motor spinning 4000 rpm? G force at the magnets is over 700g, and for sure that's way too much for FF to bridge the gap.
 
Bullfrog said:
Conclusions
A. 2.5 to 3.5 ounces of low viscosity ATF can aid in cooling a MAC IGH motor.
B. Minimal increases in the viscous drag were observed as expected but considered acceptable based on the decrease in temperatures.
C. 2.0 ounces or less of ATF does not appear to add any cooling benefit.
D. More than 3.5 ounces of ATF could lead to leakage and the associated safety concerns regarding braking.

Thank you so much Bullfrog for contributing meaningful test data like this, and establishing the benchmark fill level range where ATF has no effect to when it leaks out. I'm really curious to hear now several months later where things are at with ongoing ATF leakage out of the motor. Do you notice it slowly seeping out of the cable harness and bearings or has it largely stayed contained in the MAC motor with their stock sealing?

Recommendations
A. Conduct similar experiments in a laboratory setting where all of the variables could be controlled, monitored and recorded. This experiment was conducted as carefully as possible utilizing the available resources.
B. Utilize additional fluids to determine if low viscosity ATF provides the best benefit based on viscous drag and temperature reduction.

I'm going to be setting up the wind tunnel for fluid tests on geared motors next up once the current statorade evaporation test is complete. The Bafang G310 motors seem particularly well suited to ATF since they are an inrunner rather than an outrunner, and so the fluid just needs to make the bridge from the stationary motor housing to the spinning motor shell which is quite a small gap. I tried using a ferrofluid actually for this, with magnets glued outside of the motor to keep the fluid located around the motor shell and away from the gears/bearings, but at least the first order tests weren't as successful as hoped.

G310 Shell with Statorade.jpg
G310 Shell Gap Picture.jpg

The sealing on these motors is quite good (full O ring on the side cover plate) and I think that a regular ATF style fluid will work well, and should have a much lower relative drag increase than on say a MAC or eZee motor as there is not a high RPM outrunner motor spinning in the fluid. Instead it's just the very low RPM of the motor shell and a stationary motor.
 
Thanks for all this great info Justin!

On the ATF in geared hub front, I sort of did this for a while about a year back when I put some chainsaw oil into a '200W' hub motor and successfully ran it at 750W for a long time.
https://endless-sphere.com/forums/viewtopic.php?f=3&t=53322&start=25#p1218669
DSC_3320.jpg


Ultimately I killed it from overheating, but not from the over-power, rather a badly rubbing hub brake on the side of the hub contributed so much heat that combined with the over-powered windings this was the end result:
DSC_3393.jpg


But yeah, while it was working I was able to push 4 times the stock power through it with the oil very effectively transferring the heat to the side covers and heat-sinks.

Cheers
 
So everyone, here it is, some definitive answers to the question of "How long will it last?!?" There are people who have had Statorade in motors for a couple years and reported no decrease in thermal performance, while others have opened up hubs after a few months when they found the core was getting hot quickly again and found little Statorade left. Does it evaporate after extensive use or is this just a case of it leaking from motors that are not well sealed? And are there long term consequences to the stability and performance of the fluid when it is run regularly at elevated temperatures?

To answer this I finally committed to some long term testing over the winter when there weren't other experiments going on. For the test platform I used a MXUS 30mm rear motor. MXUS has done an excellent job in the side cover sealing on our Statorade-Ready hubs, so this way I could count on no micro leakage.

For the first setup, I put 6mL of Statorade into the motor and ran it at 100oC with the motor spinning at 400 rpm. The reason for 6mL is that at this higher RPM that is the point where slight changes to the quantity of Statorade have the most effect on the thermal conductivity. The data here only goes to 300 rpm rather than 400 rpm, but you can see how at this fill level just the loss of 0.5 mL would be quite detectable in the net conductivity, with a slope around 0.8 (W/K) /mL

MXUS Conductivity vs Statorade Fill.jpg

The test setup had the motor in the wind tunnel test station but without the actual tunnel fan running (which is a little noisy to run 24-7), so it was free spinning in ambient air. I chose 400 rpm just because it seemed that the faster the motor was spinning the more likely to agitate and stir the fluid into mist/droplet that would encourage evaporation. That's equivalent to a 26" wheel travelling 50 kph.

For the first test, I had the test computer configured to drive the motor until the core reached 100oC and then adjust the phase current to maintain excatly 100 degrees, while recording the power going into the motor. Every couple hours the data was recorded for ambient temperature, shell temperature, input power to the hub, etc. If the Statorade was evaporating or loosing effectiveness, we would see less and less heat going into the motor in order to maintain 100oC.

Here's the graph of this first test, and it's a good lesson in why you should do the full data analysis before moving on.
MXUS Evap Test at 100oC, 400rpm, 6mL.jpg

What I was looking for was a drop in the power going to the motor to keep it at 100oC, but even by the 7th day the power was totally steady at about 145-150 watts. I thought perfect, no signs of any statorade losses at all. Let's crank it to 120°C to see if it evaporates then!

However, what I didn't notice at the time was that the average ambient temperature was dropping over the course of the experiment as we got a cold spell in Vancouver at this time. Initially we were 23-24 degrees, and by the end it was more like 20 degrees. So when I later calculated and plotted the actual thermal conductivity at each data point, it turned out that it was sure enough slowly declining over the course of the test as you can see in the blue plot here. Starting off just over 4 W/K, and trending down to 3.75 W/K after 7 days.

A loss of 0.25 W/K would correspond to loosing about 0.3 mL of fluid based on the first graph. Note also that 7 days of spinning at 400 rpm is equivalent to 8400 km of travel on a bike, and that's running with the motor core at 100 c. So even though there is a measurable loss, from this data it's looking to be to the tune of 1mL / 20,000 km.
 
You pay more for an oil change every 20,000 km :)

Great to see the data. Science for the win :)
 
Justin and those interested in my MAC/ATF experimental cooling results....
The original experimental set up and results are posted here: https://endless-sphere.com/forums/viewtopic.php?f=2&t=48753&p=1337584&hilit=mac+atf+terrex#p1337584

I couldn't have accomplished what I did without the help and suggestions from Grin Tech. So please provide financial support with your purchases...here is their store: http://www.ebikes.ca/shop.html

It has been about three and a half months and I have not detected any leakage to date. When I say no detectable leakage, I mean when I visually inspect both sides of the motor before and after riding, I have not seen any droplets form on the bike or on the ground below the bike. I have been fairly surprised because capillary action is pretty amazing. When I am not riding the bike it is stored in a vertical orientation and my rough calculations indicate the level of ATF does not reach the bottom side of the axle when the bike is vertical.

I just took a white paper towel and wiped down everything for the first time and on the side where the wires exit the axle, I did see a slight discoloration on the towel that covered an area about the size of a quarter...actually it was about the size of my fingertip that was pressing on the paper towel...duh :). Based on the smell, I would believe the fluid to be ATF. There was no evidence of any leakage on the opposite side of the motor i.e. no discoloration of any kind on the white paper towel.

The disheartening part of all of this is that virtually ANY ATF on your brake rotor and you lose your brakes completely. I know because I accidentally spilled some ATF on my rotor during one of the experimental sequences where I was adding fluid and I had to clean my rotor and change brake pads to regain braking capability.

Soooo I guess my NEW recommendation is to seal the wiring/axle interface inside the motor where the wires enter the axle but it is going to be very challenging and capillary action WILL suck fluid into any opening that is not sealed...no matter how small. Sealing the area where the wires exit the axle is going to be like sticking your finger in a dike...it will help but I am afraid it won't work for long.

One of my original conclusions was "2.5 to 3.5 ounces....". I'd add to that "Do not go over 3.5 ounces because the chances of leakage will dramatically increase".

Next time...I plan to attempt to seal the wiring/axle interface inside the motor and add 3.0 ounces of ATF. A smarter person would just utilize a motor that had a Kv that was appropriate for the task. My problem is that I use my bike for riding off road with a maximum speed of approximately 17-18 mph. Even with a 12T MAC in a 20" wheel (4" Fat Tire), my top speed is around 22 mph so I am operating well below the most efficient motor rpm most of the time which means more HEAT.
 
jonescg said:
You pay more for an oil change every 20,000 km :)
That is true, although the experiment is not over!

As mentioned, I thought that from the initial cursory look at the data that there was no loss at all since I was looking just at the input power being steady. So after a week of supposedly no change I decided to crank up the target core temperature to 120 °C, which required about 30 phase amps through the motor.

I let it run this way for nearly a month. The data as I would see it on the test station looked like this here, with the stator core staying fixed at 120 degrees while the input power was fluctuating up and down to maintain that target as both the ambient temperature fluctuated and the heat conductivity was reduced. This is analogous to an ebike with a temperature sensor in the motor and a CA3 doing thermal rollback to keep the motor core from overheating.
MXUS Statorade Evap Test Data Example.jpg

Another means of doing the test would be injecting a fixed amount of power into the motor windings and then watching the temperature of the motor core vary over time. That would be more intuitive to look at, since any decrease in the effectiveness of Statorade would be seen as a trending increase in the motor temperature. However, it would not be a fixed temperature experiment so I could as easily quantify the effect of temperature on the evaporation rate.

Anyways, after an accumulation of 27 days and then analyzing the data with a ~3hr sampling point interval to derive the thermal conductivity values, this is what I got. I actually only just did the complete data analysis now so it's the first time I've seen this result too and it's pretty cool. I love how from the noisy heat and ambient temp data comes a totally clean curve for the W/K conductivity (blue line):

MXUS Evap Test at 120oC, 6mL.jpg

After 27 days (32,000 km), there was a total decrease in conductivity from just over 3.5 to 2.75 W/K. And if you extrapolate that from the slope of the conductivity vs mL fill curve in my last post, you can get that it's equivalent to the loss of about 1mL, but that's over 3 times the time span. Cranking the core temperature up from 100 to 120 degrees didn't seem to accelerate the rate of statorade loss.

I have plotted the results of a short run with no Statorade in the motor as a reference point at the end of the test data.

At this stage I wanted to open up the motor to have a look. The amount of Statorade remaining on the magnets in the motor core was definitely less than what a fresh 6mL injection looks like:

MXUS Remaining Statorade after 6mL Test.jpg

However, I'm not 100% sure that evaporation is the actual or primary mechanism of fluid loss. It's also possible that over time more and more of the fluid works its way in between gaps in the stator laminations, the copper wires, soaking into fiberglass sleeveing or any other fabric or twine holding the windings in place. Here is a picture of the actual stator core and you can see even the white cable ties have that film coating.

MXUS Remaining Statorade on Core.jpg
 
So after wiping off the remaining Statorade inside the motor, I decided to reseal the motor and resume this experiment but using a 9mL quantity of Statorade that would have full conductivity at 400 rpm, rather than being right on the slope of maximum change. As well, I figured that if the stator itself was going to absorb fluid in any cracks and fibers that this particular core would be well saturated by now, so if we did see a decreases in effectiveness with this 9mL fill then it was definitely from some form of evaporation loss.

Anyways, I'm now over 3 weeks into this latest experiment. The raw test data continues to have this form, what you see are daily fluctuations in the temperature of our building
MXUS 9mL, 120oC, Example Data.jpg



But when I do the data analysis for conductivity, we're not seeing ANY change at all so far.
MXUS Evap Test at 120oC, 9mL.jpg

It's been hovering right around 4.75 W/K for the first 24 days. That's 29,000 km. Assuming that it is being lost at the same estimated rate of ~0.3 mL/week, then we probably won't notice any decrease in conductivity until it's down to 7mL or so. That would be at least another month down the road, so we'll keep 'er running and see. If the initial loss in fluid was more the result of it vanishing/wicking into the core, then we might not see any decrease at all.

jonescg said:
Great to see the data. Science for the win

Exactly! It's the only way to get meaningful answers and improve things in the industry, but as you can see it's also quite a bit of work, so I appreciate the appreciation. ;)

What we can conclude from this particular data so far? It's that if you put just a few mL more Statorade in the motor initially than is strictly required for your target RPM, it looks like you'll be able to ride for many 10's of thousands of miles before top-up is required. That is pretty good news I'd say.
 
What I find interesting is that the motor "soaked up" the Statorade gradually over time. I'd think that in real service with varying RPMs, road shocks, and starts and stops, that the Statorade might fill all the nooks an crannies in the motor faster and that the "losses" would plateau sooner.
 
Thanks for this awesome analysis Justin.
This makes me not so bothered that I accidentally put the full 10ml in my MXUS I built recently.

I would love to see some destructive testing. What would happen to the FF if you pushed a motor beyond it's limits to the point of failure?

I've also been wondering what the effect of the 140-160C temps I'm hitting in my MXUS are having on the magnets. Any chance you can test if the magnets are any more susceptible to demagnitization when FF is being used at high temps vs no FF?

Cheers
 
Justin...Thank You, Thank You, Thank You.

I know it is a lot of work and very boring to publish all of your results (been there done that a bunch) and it isn't nearly as much fun as the experimental work.

A suggestion...take a known amount of Statorade and put it in a glass beaker in an oven at the temperature of interest. You could use a large amount compared to what really goes in a motor just to increase the total change in volume. Any loss could still be related back to a motor by calculating the change in volume per unit volume. This would define what is lost purely due to evaporation.

As far as the soaking of the laminations and other parts of the motor...I would consider that a good thing since it is only going to increase the thermal conductivity and help you get the excess energy to the atmosphere.

If you really wanted to get into the details, you could simultaneously perform the oven experiment with the surfactant, the carrier, and Statorade in seperate containers. It would also be interesting to know if the sum of the surfactant and the carrier equal the loss of the actual Statorade. I don't believe the solids are a factor but I am just guessing. All three containers will experience the exact same conditions if you put them in the same oven. Of course a lit search may turn up the answers more easily if you have access to a quality library (I don't anymore). Remember not to reinvent the wheel...just improve it. I apologize if I am preaching to the choir :).

Dang you are going to get me all interested and I am going to have to build a bike with a hub motor...been considering it for quite a while but my riding style (100% off road at relatively slow speeds) just doesn't fit a hub motor very well, unless somebody can find me a motor with about a 3KW or more output and a Kv of about 4.
 
justin_le said:
We had someone drop off an Inboard electric skateboard with damaged controller that provided a good candidate motor for these experiments. I put tape over the 3 vent holes that are normally on the outside of the hub.
are the measurements w. 0ml statorade w. or w.o the vent holes taped ? (those kind of hubs exist with " venting - forced aerothermal heat transfer" - and w.o ...) thxs
 
I have been away from my ebike for quite a while. Last summer I was in the process of lacing a cromotor for FF to test against my custom air cooled cromotor. I never finished through with the project. The wheel just needs to be trued and FF + hubsinks added.

I still believe my air cooled motor will win against a statorade + hubsink motor.

Has anything changed in the last year with statoraide or hubsinks that I should consider changing? I bought the statoraide and hubsinks last summer.
 
Warning Warning Warning...if you follow my recommendations for adding ATF to a MAC to keep it cool, eventually it will leak on the brake rotor and you will lose your back brakes.

As usual the wonderful people at Grin Tech are correct....they tried to warn me that the leakage would eventually become an issue but just like listening to my parents, professors, and friends...I had to prove it via experience. The good news is that adding 2.5-3.5 ounces of low viscosity ATF to your MAC will keep it cool with the max temps never going over approximately 85C.

The leakage didn't occur while riding on relatively smooth paved surfaces but over the last couple days I have ridden off road on some pretty rough ground with exposed roots and rocks...my guess is the rough surfaces caused the ATF to splash around inside the motor more, therefore it came in contact with the opening in the axle where the wiring exits...and eventually started to seep out of the left side of the axle and onto the brake disc.

A lot of good information in this thread...and I can tell you how NOT to cool an Internally Geared Hub (IGH) motor. For anyone riding off road at relatively slow speeds and needing high torque...my recommendation is to go with a Mid-Drive motor like the Bafang BBSHD because you can utilize the bikes gearing as well as the motors gearing to keep the motor spinning at an efficient rpm and produce less waste heat.

The next best option for slow speed/off road/high torque is a Hub motor with auxiliary cooling like Statorade, Forced Air, etc.

The last option I would choose is a IGH motor just because it is sooo difficult to cool. Mine works great when riding on the pavement at speeds where the motor is in an efficient rpm range but you have to choose the right tool for he job and I didn't when I chose an IGH motor for riding off road at relatively slow speeds with high torque requirements.
 
Warning Warning Warning...if you follow my recommendations for adding ATF to a MAC to keep it cool, eventually it will leak on the brake rotor and you will lose your back brakes.

This is why a 4mm vent hole as close to the axle as possible on the freewheel/cluster side is so important. Also, it doesn't hurt to squirt some silicon sealant around the wire housing where it exits through the center of the axle. When there is a vent issue, it blows out against the freewheel on the opposite side. Also, dont run over 60cc of ATF.
 
60cc isn't enough to bridge the gap between the motor and the housing....with the MAC and most Internally Geared Hub motors, the gap between the stator and the housing is huge compared to a direct drive hub motor. Approximately 3-3.5 ounces is the minimum required to enhance cooling on a MAC.

The vent and sealing the wiring/axle interface may be the answer. I didn't install a vent because I was concerned about contamination entering the motor thru the vent since it will breathe in and out depending on the motor temperature and I only ride off road where it is dirty and dusty.

IMO the answer is to utilize the correct tool for the job and for me that means a Mid Drive where I can spin the motor at an efficient rpm.

When you have to add additional cooling to a motor it is only a band aid that is required because you don't utilize all of the power being sent to the motor and you have to reject the excess (heat) to the atmosphere to keep from damaging your motor. Don't get me wrong...I absolutely love the simplicity of a hub motor (one moving part) and an Internally Geared Hub motor isn't much more complex but neither is the best tool for riding off road at relatively slow speeds with high torque demands.
 
Back
Top