It's been a while since I posted here, but I've been working on solving this annoying motor overheating problem.
I fried my air cooled 9C (babied it for a bit then got impatient), constantly overheat my X5305 (which I only don't fry because it was my only remaining motor) I thought about going to a bigger motor with more thermal mass, like the X54XX, but we're trying to reduce unsprung mass aren't we?
Now I think I've pretty much solved the problem - motor overheating is now a thing of the past. Liquid cooling is the way to go. Hal did a liquid cooled hub on his albino build but never told anyone how it went, so I thought I'd give it a crack.
In some ways this is pretty low tech and very much a prototype, but it shows it is possible and definitely does the business.
I don't have a lot of time for this post, so I'll just go straight to the pictures:
The brief was, cooling tubes into motor without using a custom disc rotor. So I got a custom disc side panel made with an oversize bearing.
- panel1.jpg (100.29 KiB) Viewed 1002 times
- panel2.jpg (103.73 KiB) Viewed 1002 times
Room for two 5/8" copper pipes and 3x 8awg phase wires and 1x 9 stranded data cable
Probably the most important photo, but somehow I lost most of the ones I took of this process. This is the stator partially finished.
Note the carbon fiber inserts which when epoxied in to the stamped stator form a cooling channel. The rest is just pipes and wires and springs to stop the pipes kinking etc. Also a thermistor under the windings. Not to mention the hall sensors I added (this was one of the original sensorless versions)
Needless to say, this was a high temp curing epoxy.
Now while I was at it I did some magnetics modeling of the motor in FEMM and came to the conclusion that 3mm of iron for the flux ring would be enough and that the 7.5mm that is standard was really not required - all sounded good, but I didin't have the guts to go any further until I watched justin's hub motor presentation video and saw that he basically came to a similar conclusion. So I figured while I was at it I would start to reduce unsprung weight. I got it taken down to 3.5mm in the middle, then spent a couple of hours milling bits out of the hub flanges
- rotor1.jpg (33.52 KiB) Viewed 1002 times
- rotor2.jpg (73.72 KiB) Viewed 1002 times
In all that got it down 900g. If I did it again I'd take out a few more bits, (larger cutouts between spokes and the 1.5mm or so on the outside of the disc side flange) I'm pretty sure you could get it down to at least 1kg lighter.
Now jumping to it installed on the bike:
- radiator2.jpg (132.54 KiB) Viewed 1002 times
120mm pc cooling radiator + 120mm Fan (reasonably grunty delta fan)
- pump.jpg (89.99 KiB) Viewed 1002 times
1x 12v solar water pump - pc cooling pumps have great specs but are usually only rated to 65degC, this one is rated to 110degC. With the whole system hooked up I get a flow rate of about 1.7L per min with water, a bit less with the 33% antifreeze I use.
For those who don't know, antifreeze is also antiboil, so the 33% coolent I'm using is good from -15 to 105 degC. It also stops corrosion which is pretty important given how the cooling tubes work in the motor.
- radiator3.jpg (72.6 KiB) Viewed 1002 times
Radiator viewed from behind
- radiator.jpg (112.66 KiB) Viewed 1002 times
Mounts almost perfectly in the fork. Held on by cable ties of course
- fillport.jpg (87.17 KiB) Viewed 1002 times
Fill port (also happens to be a pressure release valve) installed in raidator
- fan.jpg (70.83 KiB) Viewed 1002 times
Pretty grunty fan.
- bike.jpg (104.19 KiB) Viewed 1002 times
Pipes and wires all over the place at the moment, different sizes in different places because the pump had different sized inlet and outlet fittings.
- motor2.jpg (144.19 KiB) Viewed 1002 times
Motor in place, I only painted the rotor as I plan to replace both side panels with carbon fiber versions - I think there is another 400g to be saved there.
- motor.jpg (81.69 KiB) Viewed 1002 times
Other side.
I'm running out of time, so I'll be brief:
I've only had it working for a while, in terms of unsprung weight, the HT3525 is pretty good. After adding stuff inside and taking stuff off the outside I came out with a nett loss of about 700g, so it's about a 6.8kg hub motor. I think I can get another 400g off with carbon side panels, and if I did it agian I could reduce the rotor a bit so a 6.3kg liquid cooled carbon fiber panelled HX35 should be possible
It certainly Handles significantly better than my X5405. It feels pretty similar to the 9C, but more torque. I have changed the shock since then so not a fair comparison.
Now what you've been waiting for - how it works in terms of cooling:
Ok, so first up I've got a 24s lipo battery pack and controller set for 65A - so about 6kW.
I've only got qualitative data at this point: There's a hill I usually include on my rides that gets the windings of a cold motor up to about 90deg C by the top, if I then ride down and do it again it gets to about 135. With my new system I rode this hill repeatedly without the fan on and after a few runs I would get the windings to about 80 at the top of the hill, but by the time I got down again it would be back down to 60. If I ran the fan it was even better, I couldn't get the motor windings over about 75 and when I stopped at the top of the hill I could just watch the temp drop at about half a degree a second!
The radiator wasn't getting that warm, so I think there is quite a large delta T between the windings and the cooling channel.
Next time I'll install a flow meter and temp sensors at the inlet and outlet of the motor to calculate actual cooling rates, but for now you'll just have to believe that it works really well!
I haven't had a chance to try higher currents, next time I work on the bike I'll try 80A (7kW) and see how it behaves, I think that shouldn't phase it too much. I think I could probably get away with a smaller radiator and fan.
gotta go...
I just don't think air cooling is going to cut it...
)
