Cooling fans inside Hub motors

[Offroader]

Interesting placement, much better than on a stock cromotor.
What kind of temp probe is it? If it is not a KTY81, like what came with your max-e, you might as well replace it to get proper temp readings with your Max-e. The Max-E should have come with a motor temp probe.

You may have a NTC 10K. If so swap it out while you can.

I'd look at samples of how people drilled out the side covers. Most people drilled close to the windings. I wouldn't go to crazy at first, you could always add more holes later. Not worth weakening the side cover like in the link below. Don't drill so many close to the axle as that will weaken it quicker.

Remember, you can always get new side covers so don't worry too much.

https://endless-sphere.com/forums/viewtopic.php?f=2&t=62803

 

[BCTECH]

all that true!
does it matter which side of the cover to drill holes ? both sides or doesn't really matter

the temp probe on the motor is KTY83/122, i am not sure if better or worse than the KTY81. I read that the KTY81 proofed to be accurate on max e

Thanks for all the tips !

 

[Cowardlyduck]

all that true!
does it matter which side of the cover to drill holes ? both sides or doesn't really matter

the temp probe on the motor is KTY83/122, i am not sure if better or worse than the KTY81. I read that the KTY81 proofed to be accurate on max e

Thanks for all the tips !

I would start with holes just at the perimeter on both sides. Size partially depends on the height of your windings. I would try and match the diameter of the hole to the height of the windings if possible.
On my HS4080 I used 20mm holes which were slightly larger than the windings size.
I've found a good way to space the holes is to put 2 evenly spaced between each side cover bolt.

Cheers

 

[Cowardlyduck]

It begins again:

In case you couldn't already tell...I'm a fan of this approach. :lol:
This time it will be 12 fans in the motor!

The HS40 modified previously still works, however I've had ongoing issues with minor axle play, so I'm reluctant to use it and looking to pass it on to someone who can actually put it to use.
My 35mm wide magnet Leaf motor is both lighter and more efficient than the HS40, so make's for a better fit for my Fighter provided I can cool it sufficiently when I ask it to perform.
So far running the Leaf motor sealed at 50 amps it has been great. Sure it's not as torquey as the HS40, but it's very close, and the weight saving is noticeable to me. It's also got a 14mm axle made from semi-decent steel, vs the 1/2" (12.7mm) cheese axles used in the Crystalyte motors.

Lot to do still though, and a lot of parts still on there way. I've got a replacement side cover coming after I discovered the freewheel side-cover looked like this:

Details in the Leaf motor thread:
https://endless-sphere.com/forums/viewtopic.php?f=30&t=66489&start=650#p1110335

I've also got bunch of wire, heat-shrink, bolts and other bits and pieces still on there way, so it will be a while before I get this up and running.

Cheers

 

[Cowardlyduck]

Made some more progress on the Leaf motor mods.
I'm still not quite finished, but getting close now:

I haven't drilled these side covers as much as my previous motor's for two reasons. One, the freewheel side cover as posted above has a pretty poor casting finish that has me a little concerned about structural integrity, and two; with the 12 fans I will be running, I should not need as big openings for air to move through at a decent flow rate.

The phase connections were a little tricky due to the positioning of my wires through the axle. I did not have enough room to slide heat-shrink over the wire, then slide and shrink over the join, so instead I cut open the heat-shrink, layered it over the joins, secured tightly with zip ties and heated until the separate layers of heat-shrink melted together. I've used this method before when I've forgotten to slide it on first and it does work, although not as neat as otherwise.

The fans are secured with long bolts this time instead of epoxy. I drilled small 3mm holes next to the 25mm holes to position the bolts, and the fans are held to each other through the stator support.

I managed to squeeze a decent amount of wiring through the axle without really enlarging it apart from some light smoothing around the sharp edges.
The wires going through the axle consist of:
3X 12awg silicone wire with the silicone removed and replaced with heat-shrink for the phases.
2X 18awg silicone wire with the silicone removed and replaced with heat-shrink for the fans.
7X 36awg teflon wire for the halls, KTY83 temp sensor (for the Adaptto), and another temp sensor for a dash display.
All the wires are inside a very thick glue coated heat-shrink layer going through the axle, so I'm not at all worried about chafing or cutting near or in the axle.

I have a spare set of halls wired up, however the positive isn't connected through the axle. I originally had 8X 36awg wires through the axle, but one accidentally pulled back through, and there was no way I was re-doing the whole thing just for that. If I need to use them at least it will be quick/easy to open the motor and solder the positive in-place.

Next I just need to wire up the rest of the connection to my controller. From just outside the axle, I will use 8awg for the phases, 14awg for the fans, and normal wires (24awg?) for the halls and temp sensors. I also need to plug all the remaining small holes with some silicone.

Oh and the push/pull config of the fans feels like it makes a difference. I've only run them on ~20V so far and I can definitely feel a lot more air being pulled through than I did with just 6 fans on my HS4080 motor.

Cheers

 

[Offroader]

I wouldn't worry about the 2nd phase wires, I made so much effort to keep the 2nd set in my cromotor and it caused me lots of work to keep them. Now after popping open my cromotor almost 6 times in a week off my bike I can have it open now in no time and realized that if I needed them I can just open the motor back up. Most likely you will never have issues anyway and the max-e can run the motor sensorless until you get home to fix them.

I may cut my wires next time I go into my motor.

cowardly duck what did you wrap all those teflon wires with? How are you soldering those teflon wires?

Does using the thicker gauge phase wire make any noticeable differences?

 

[Cowardlyduck]

I wouldn't worry about the 2nd phase wires, I made so much effort to keep the 2nd set in my cromotor and it caused me lots of work to keep them. Now after popping open my cromotor almost 6 times in a week off my bike I can have it open now in no time and realized that if I needed them I can just open the motor back up. Most likely you will never have issues anyway and the max-e can run the motor sensorless until you get home to fix them.

I may cut my wires next time I go into my motor.

cowardly duck what did you wrap all those teflon wires with? How are you soldering those teflon wires?

Does using the thicker gauge phase wire make any noticeable differences?

Very good point on the Adaptto being able to run sensorless. I had forgot about that ability.
The teflon wires are just in there own very small diameter clear heat-shrink through the axle for a bit of protection although it's possibly not needed.
Soldering the teflon wires is not too difficult, I just use some flux and it wicks up the solder nicely.

Using thicker gauge phase wires definitely makes a difference. Probably the biggest difference after cooling. It achieves two main things;
1 - Allows more amps without it being turned to heat.
2 - Acts as a heat-sink when things do get hot.

The second point is the main reason why I'm going to 8awg outside the axle as I will not normally push so much current as to need 8awg, however when things inevitability get hot the large mass of copper will suck some of the heat away from the stator.

Cheers

 

[Samd]

Great work CD
It really seams to works great I only wish the manufacturers would take note

Cheers Kiwi

Well the stator can only convert so much electricity to kinetic energy before saturation - the manufacturers know this.
All the fans so is pump the extra heat you are making out to the atmosphere faster. You don't really get any extra action out the back of the bike even if you nitrogen cool it.

Hence the efficiency curves dropping off fast on the ebikes.ca simulator.

 

[Cowardlyduck]

Well the stator can only convert so much electricity to kinetic energy before saturation - the manufacturers know this.
All the fans so is pump the extra heat you are making out to the atmosphere faster. You don't really get any extra action out the back of the bike even if you nitrogen cool it.

Hence the efficiency curves dropping off fast on the ebikes.ca simulator.

Appreciate your input Sam, but not sure what point your trying to make.
I never claimed the ability to pump increased power beyond the saturation point of any of the motor's I've added fan cooling to.

It's simply about removing the heat generated during spirited or prolonged up hill riding that would normally result in a melt down of a non actively cooled motor.

Adding additional cooling to E-bike hub motor's doesn't give us any more peak power from the motor, but does allow us to run them at higher continuous power levels for longer.
I know you already get that, so not sure why you feel the need to play down these attempts.

Cheers

 

[Samd]

Not playing down at all - it was a caution to Kiwi - That's why I quoted him.
I know you get it now!

The more I think about it - the more I see the stator width being good for about 100 watts per millimeter of width for lengthy and frequent peak bursts.

 

[Cowardlyduck]

Not playing down at all - it was a caution to Kiwi - That's why I quoted him.
I know you get it now!

Cool. And I'm pretty sure Kiwi get's it too.

The more I think about it - the more I see the stator width being good for about 100 watts per millimeter of width for lengthy and frequent peak bursts.

That's a really good way of looking at it actually, and right around the max power I plan to use with this hub, so pretty spot on.
Of course, because I can, I will see how 4-5KW feels also, but doubt it will make much difference on this hub.

On a related/side note, I was playing around tuning my Adaptto Mini-E for my HS4080, and have so far managed a max of 4.3KW from a dead stop which feels nice. It's just at the limit of lifting the front wheel which is just about perfect for regular non-suicidal use.

Cheers

 

[Samd]

Well Kiwi's not silly, he's one of my best customers.

But the point is the motor manufacturers really do get it. AC motor theory is usually covered by second year of a four year engineering degree. All the equations are known. Motors get designed to have their natural peak in the curve at the design speed. Motor designers then usually apply the heirarchy of design to flatten and lift the curve as much as the price point of the motor will allow. The heirarchy generally starts with eliminating the heat rather than moving it off the motor if a user wants to run the motor outside the design rpm / efficiency band.

It's ok to run the motor at 60% efficiency for bursts, it's just not on the designer's radar.

It would be wrong not to point out that motor designers really do know their stuff and the parameters for AC motor design are well know for nearly ninety years.

 

[Routybouty]

I keep wanting to find an ...not sure what to call it... "internal circumference" heat sink for the stator. I feel like a bunch of surface area added to the stator "rim", that could extend into the fan air-path, would increase the heat transfer significantly. I've been trying to google image something similar and haven't found anything. I'm not sure how hard it would be to create one from a standard flat heat-sink.

 

[Cowardlyduck]

I keep wanting to find an ...not sure what to call it... "internal circumference" heat sink for the stator. I feel like a bunch of surface area added to the stator "rim", that could extend into the fan air-path, would increase the heat transfer significantly. I've been trying to google image something similar and haven't found anything. I'm not sure how hard it would be to create one from a standard flat heat-sink.

I don't think such a heat-sink exists.
The best option I think would be to buy a bunch of smaller (maybe 20x20mm) heat-sinks that could be attached to the inside rim of the stator ring using thermally conductive epoxy.
However, I doubt it would be worth all the effort TBH, and I suspect the risk of one of them coming loose would be high if just epoxied in place.

As myself and others have found here, adding fans is enough...you don't need heat-sinks. The fans are continuously pulling in cool air directly over the windings, something no heat-sink will change, or benefit.

I took my fan cooled Leaf motor for a spin on the weekend. I did some pretty slow/steep trail climbing with the fans running and couldn't get the motor over 80C.
I also rode through a bunch of creaks/streams and large/deep puddles with the fans running without any issues at all.

Cheers

 

[Samd]

Heatsinks do two things - store heat, and if the shape is right they shed it faster (as air is a bad conductor of heat).
Motors that have aluminium stators store the heat better than a pressed steel stator - like a shock absorber. There's a bit more surface area there too (design dependent) to shed the heat fast. You do need some convection going on if you want to get it out of the stator at some stage.

If you are riding along the flat and you do a couple of 10kw bursts this gives you somewhere to put the heat until the motor can give it up to the surrounding atmosphere.

CD's method has a clear advantage if you want to ride up a hill and stop for the view. Many of us have to keep riding around a bit to keep air moving over the motor to get the heat off. Although, if you are making hundreds or even thousands of watts of heat, you are by definition no longer in the 85% to 90% efficiency range for a 3kw to 10kw system. Possibly as low as 40% for a smaller DD hub with a 8 kilowatt infeed.

There are radial heatsinks (search Radial Heatsink on google) with long fingers for this purpose in general motor design. Over the years on ES I've seen people make their own - even by using thermal epoxy to glue various shapes with a high surface area to steel stators.

The other interesting thread is the one run by Justin where he is using a small (about five ml) shot of ferrofluid to connect the stator to the rotor so the heat can travel efficiently to the outside case of the motor. Classic thermodynamics gives us a good feel for what to do here, Justin is dialing in the predictions with some fun tests that seem to be heading in the direction expected for ferrofluid bridging.

 

[Offroader]

Now that my EDF cooling has run flawlessly since I had it installed it is time to make it better now that the winter is here and riding is getting too cold.

1) I'm going to add some heatsinks to the center of stator to get heat quicker out of the stator. I plan on using some of the EDF air to flow through the heat sinks.

2) I'm going to add an air duct inside the motor side covers so that the flow of air more passes through the copper windings as there is a big space between the copper windings and the side cover. So lots of air is just passing right above the copper coils.

3) I'm going to upgrade the size of my EDF phase wires as 24 gauge is way too thin causing the EDF to run at a much lower speed than it could.

4) Upgrade my phase wires on my motor to 10 gauge.

5) Possible try to put an external air duct so that the exhaust hot air doesn't get pulled back into the motor as easily.

With these modifications, and especially the heat sinks to the inside of the stator I think that cooling could be like over 2x maybe 3x better. Tests are showing the stator takes in lots of the heat quickly from the windings, so adding heatsinks with forced air through them should greatly push the heat out of the stator and out of the motor.

I figure the heatsinks, which I bought already, could make the stator have like 5x-10x more surface area. This will greatly help shed the heat into the air and push it out of the motor.

I also believe that with this setup with heat sinks it could work better than water cooling if I can get the EDF to properly circulate the air through the heatsinks.

 

[Cowardlyduck]

That all sounds awesome Offroader.
Looking forward to hearing/seeing the results.

I managed to squeeze 12awg phases and 18awg for my fans for my Leaf Motor. On my HS4080 I managed to fit 14awg for the fans as it has a bigger slot.
The key is to remove the silicone outer layer and replace it with heat-shrink.

Cheers

 

[Cowardlyduck]

So I finally got around to doing some testing on the weekend.

This hill is steeper than it looks. I had to pedal in order to prevent the motor from stalling out, but kept my effort and speed consistent on both runs. This is evidenced by the very similar rise in temps.
I did also record the Regen temps, but didn't bother including or plotting them as the footage was too rough, and it took long enough for me to just do the climbing data.

I tried to keep the video concise, but for legitimacy I included all 20+min of test footage. Feel free to skip through.
[youtube]1S0WSt_dZWk[/youtube]

The chart kinda say's it all!

A peak difference of 5C isn't huge, but also not insignificant. The obvious difference is cooling time which comes into play massively during normal riding conditions that includes periods of coasting or lower power usage.

Overall, I have not seen an improvement so far in going to 12 fans over the 6 I have in my other motor. I think this is most likely due to the smaller holes I drilled in the side cover of this motor compared to my other motor. I may add more holes/make them bigger, but for now I think the cooling performance is good enough for my usage, so will leave it alone.

I would still like to do another set of tests similar to this, but under normal off-road riding conditions including many ups/downs, etc. to better gauge the normal usage performance difference. I can arbitrarily see a difference, but a proper test would still be nice.

Cheers

 

[sendler2112]

Re real world peak temps. Maybe you could find an even on road hill that is long enough to heat the motor. Maintain a power that just keeps the non fan run to less than the 100C level where the contoller starts limiting. Repeat at the same power with the fans on to see how much cooler it is. Or up/ down, up/ down a shorter hill at the same power to see what the average temps run.
.
Are those fans IP65 so you can still ride through water?

 

[Cowardlyduck]

Re real world peak temps. Maybe you could find an even on road hill that is long enough to heat the motor. Maintain a power that just keeps the non fan run to less than the 100C level where the contoller starts limiting. Repeat at the same power with the fans on to see how much cooler it is. Or up/ down, up/ down a shorter hill at the same power to see what the average temps run.
.
Are those fans IP65 so you can still ride through water?

You should have a read back through this thread...both points have been partially addressed already.
The tests I performed on the first motor I modifiedwere on a shorter hill, with no pedalling, and less than 100C.

It was with this same motor/fans running that I did this:
[youtube]iMIEGfayg30[/youtube]

With the newer (far better) setups, I've also gone through large water bodies without any issue.
I think because any water that gets in either sinks to the bottom, sticks to the outer magnet ring from centrifugal force, or get's evaporated from heat, it is not an issue for the fans.

Cheers

 

[kiwiev]

Good work Blake

You need to go commercial with this

Cheers Kiwi

 

[Cowardlyduck]

Good work Blake

You need to go commercial with this

Cheers Kiwi

Thanks!
I think your right...there may be a market...and who knows how many people are reading this thread.

Given the time it takes, and the cost of parts, I reckon $450 AUD + shipping would be a fair cost. What do you think?
That would include:
-Working with the person to figure out there needs
-Drilling the side covers with appropriately sized holes
-Painting the stator and magnet ring with electrically insulating varnish
-Upgrading the phase wires, running new wires for the fans, and replacing other non-power wires with thinner wires to make more space
-Installing temp probes/sensors
-Mounting and wiring all the fans
-Replacing the bearings with more efficient units
-Increasing the axle slot size and smoothing out sharp edges if needed

Not sure if I missed anything, or if I'm under/over selling the work. What does everyone think?

Anyone want this done to there hub motor to start things off?

Cheers

 

[Offroader]

I really love having a fan cooled motor as it makes you feel safe when pushing the high temps.

I pushed my temp in my motor to 120-130C peak temp climbing a steep hill in the woods. Normally that would worry me but knowing the airflow is circling through the magnet gap makes me know the magnets won't heat up to those temps.

I assume the stator and windings should easily take at least 150C peak temps without any issue. Seems like the halls also don't have any issues as I already overherated this motor enough to melt the epoxy holding the magnets.

 

[Cowardlyduck]

That's pretty hot Offroader!

Ironically I've never pushed any of my fan cooled motor's that hot, but I've previously pushed non cooled motors to melting point without even trying.

I wonder at what temp the copper starts to degrade permanently? I know Doc Bass has had one of his X5's up around the 260C mark once.

Cheers

 

[Artur]

It's not about melting copper, but about magnets.. You dont want to overheat them.