Definitive Tests on the Heating and Cooling of Hub Motors
- Thread starter justin_le
- Start date
sketchism
100 W
el Raptor said:and it would not be good tilt a bit all fins to fan mode ?? (It's just an idea).
I did look at having angled fins as well as Pin Fins and axial, notched and solid fins, but the manufacturing method of extrusion makes this radial design less expensive to produce by a factor of two over the next nearest, with around 95% of the same efficiency in testing, In fluid and thermodynamics, turbulent flow shows a significantly increased mixing of the boundary layer and the bulk fluid making it more effective at transferring heat than a laminar flow which mounting the fins axially would achieve, and i do think the centrifugal action of the spinning fins draws air in from the sides, between the fins and out the top giving slightly more time for thermal transfer to occur.
Although i haven't tested it practically with axial vs radial fins yet so its just a theory currently, twisting or angling the fins 45 degrees to the axis of rotation has been suggested before as well which i'm not averse to apart from the extra manufacturing complexity.
These should be under the $30 mark at current calculations which is less expensive than it cost me to make prototypes from repurposed computer sinks.
here is something from the other day, "i just went for a test ride, its 21degreesC outside and i did the same hill run that will usually see me hit temp limiting at about 120C on the QS and the MXUS with FF after 5 runs from ambient, then i turn around and cruise home and its back to 90DegC when i get into the garage. (not very technical i know, but its just for my own reference usually)
todays testing saw me do the same 5 runs, and i thought something must be wrong with my thermistor, i was sitting at between 60 and 65 degrees C after the 5 runs, so i did another 15 runs up this hill, every time i got to the top it hit about 80, but was down to 70 again by the bottom, the only way i could get it above 80 was to regen all the way down again then back up at full throttle. the hottest i could get a reading was 89 degrees but as soon as i let off the throttle it headed back down.
the hotter the motor is, the faster it cools down, i know this is because of the difference in Delta T between the source and the air, but in this case its genuinely noticeable, and stops me getting above 90C regardless of how hard i tried just then
when i got home from the hottest run, 90C i took these readings to see if my heatsink had a decent thermal transfer from the stator compared to the existing side covers.
i am usually 90C when i get home from this run, today after 4X the ride, it was 60C at the stator and dropping so fast it was a bit lower by the time i took the pic.
We know heatsinks work, they're on all electronics, and we know radiators work, we're basically re-creating a VW bike motor lol, but in this case it feels like i used to be driving a car without a radiator and finally installed one, its like night and day, and i think it would allow me to run a much lighter hub at higher power levels, and only 300G extra weight from the heatsink, cost to result ratio is probably one of the most effective mods i've done, although i do think it must be done with FF as well.
I am happy to do more testing if anyone can suggest anything specific or has an idea for a test i can do to get a more accurate reading.
sketchism
100 W
Interestingly during modelling, the thinner version (15mm) is still 72% as effective as the 30mm with 50% of the mass, because it's smaller and stays hotter while under the same thermal load it has a higher Delta T between its fins and the air at all times and can dump its thermal load faster for a given load.
Once they are ready i will be more than happy to send a box to Justin for testing under the same parameters as everything else we've grown to love.
Once they are ready i will be more than happy to send a box to Justin for testing under the same parameters as everything else we've grown to love.
Hyena
10 GW
Quality work sketch! Have you tried playing with the height of the fins ? Giving the motor a big mohawk is cool but if you can get decent results from something lower profile that may appeal to some. I wonder if you'd get the same ~70% efficiency with a 30mm fin that's around half the height? Possibly it'd be even better than the taller thinner one given your thermal modelling shows the ends of the wider ones to very rapily cool.
Where are you getting them made ? I'd be keen to try a set too when they're ready.
Where are you getting them made ? I'd be keen to try a set too when they're ready.
sketchism
100 W
Hyena said:
i've been running them thru fluid dynamics software and found that this layout the fins work amazingly while riding, even cruising, but on standing the fins radiate heat into each other so spacing them 2 degrees further apart really helps there too.
ill play around on the new version with a shorter fin height and run the tests again as well, i guess porcupine hubs aren't to everyones taste haha
sketchism
100 W
emmgee said:Nice word combination: Porcupine on Statorade
Interesting work sketchism, if Justin would jump in and test the combo on the stand....
Question: Can the fins be mounted through the spokes?
More than happy to send as many to Justin as he wants if he finds them interesting enough to do the testing on, i would personally love to see him run them through their paces.
I mounted the prototypes through the spokes but had a bit of trouble, i think simply 2 spokes and letting them sit loose will give you enough clearance to slide them in and position them
version 5 below, i added a little more space between the fins for more efficient stationary radiation
litespeed
100 kW
I'll definitely be in for a set for my QS 205. So I wonder if using two 15mm spread apart would be more efficient as one single 30mm in the middle?
Is this something your thinking of doing soon? I'd rather go this route than epoxy a bunch of smaller non removable ones around the hub.
Super cool.
Tom
Is this something your thinking of doing soon? I'd rather go this route than epoxy a bunch of smaller non removable ones around the hub.
Super cool.
Tom
sketchism
100 W
litespeed said:
The preliminary orders are in with the manufacturer after long nights of 3 am discussions with about 6 factories that i narrowed down from 30 i approached, i think i've found one that will do me a quantity and tool up without leaving me with 10,000 heatsinks on the first run haha.
they are working with me to fine tune the final design now so hopefully it should only be a couple of weeks.
You have about 35mm to play with on a QS, 40 from inner flange to inner flange, then there is a small radius you can't go over of 2.5mm each side so you could probably fit 2 15mm ones with a gap between them which would actually get you an improvement over one solid one i think.
sketchism
100 W
bigbore said:
I did some testing and 3 doesn't work as well as 4 you'd still have to remove two spokes.
With 4 like this you could install and remove from most hubs while still on the bike.
does anyone else have a preference either way, thermal performance and weight is almost identical
I think most peoples preference would always be not to have to remove spokes...I know that's my preference.
Cheers
Cheers
litespeed
100 kW
I prefer the not remove spoke method myself....I'm pretty sure that is going to be 99.9% of your buyers.
How has the manufacturing bidding gone? Find a vender yet?
Tom
How has the manufacturing bidding gone? Find a vender yet?
Tom
spinningmagnets
100 TW
Having four pieces will be easier to ship. Since it only has to be installed once, I wouldn't be bothered if there were eight or twelve pieces, especially if it made it easier to mount without removing spokes. I have a 205/H35 hub with one-cross 12-ga spokes in a 19-inch moped rim.
The QS 205, Cromotor, MXUS 3000, Leafbike all seem to have the same diameter, and they are the most popular...
The QS 205, Cromotor, MXUS 3000, Leafbike all seem to have the same diameter, and they are the most popular...
parabellum
1 MW
I paused 1 project in the spoke lacing step because of this sinks, but if I can install them laced, that's awesome! I could think of retrofitting them on 3 already running bikes as well. 
sketchism
100 W
spinningmagnets said:
I've done a bit of asking around and it seems like everyone retrofitting would definitely find it more appealing to just fit them straight up on the bike, it might be a little bit more fiddly with a couple more nuts but not as much of a pain as removing spokes lol
Manufacturing wise its been interesting, some factories came back with tooling and min quantities by the tonne that needed a $20K+ outlay haha
but i think i've found a place that looks great, I have some samples coming of 4 pc and 6pc ones being made up now to find which ones fit the best, when they arrive i should have enough to send out some Beta versions for testing/feedback.
most of the full size hubs i've got the tech drawings for have that same OD around the magnet ring, i'm sure they use the same backing iron (geared and 273's aside)
Some of the tests i'm most eager to do is on the lighter more efficient hubs like the leafs where i think they'll greatly improve the power/weight ratio.
Ohbse
10 kW
Awesome work sketch, looking forward to seeing the samples. Definitely the right call on segmenting to make installation possible without lacing changes, if you can snake a three piece through the spokes I think that would be optimal. Worth testing on some smaller diameter rims with large spokes too, I'm running the QS in 17" rim as are quite a few others now and would be keen to assist with testing.
One idea that might not be feasible in reality would be to adjust the extrusion so that the multiple pieces hook into each other rather than requiring mechanical fixings at each join, you would still require a single bolt to close and clamp the final connection. Not sure if you could achieve that behaviour with only one profile of sink.
One idea that might not be feasible in reality would be to adjust the extrusion so that the multiple pieces hook into each other rather than requiring mechanical fixings at each join, you would still require a single bolt to close and clamp the final connection. Not sure if you could achieve that behaviour with only one profile of sink.
sketchism
100 W
Ohbse said:
I looked at a few ways to have them hook together and they all required at least two moulds and dies to be cast which is a huge percentage of the initial setup cost so for this Beta run it looks like identical segments are most cost efficient,
I did some testing on rims here with a 17 and 19" moto as well as standard 26" and found that even 4 segments would require the tyre and at least one maybe two spokes to be removed.
The optimum length for convenient retrofitting is actually 6 60 degree segments which slide between even the thickest spokes on a 17" rim, it works out more expensive to produce per complete sink but cheaper for the initial tooling setup as they can use a smaller extruder to produce them so the final cost will be the same.
i've also decided on a single width of 15MM which will allow one sink to be produced which can be fitted to the thinnest hubs as a single, or doubled up on the 35/40/50mm hubs (giving better performance than a single 30mm sink as well) macribs has suggested some great ideas as well which im exploring for the next version but i would really like to get theses ones ASAP.
This is so awesome Sketch. Love that your following through on getting them made. I'm sure they will be popular.
You should start a new thread for heat-sink cooling on hubs so this discussion can really take off. If you don't I will.
Don't have any pics, but I also epoxied 4 heat-sinks to a geared '250W' hub that I run at 750W with oil cooling. So far the heat sinks seem to be shaving a good 15-25C off the surface temp.
Cheers
You should start a new thread for heat-sink cooling on hubs so this discussion can really take off. If you don't I will.
Don't have any pics, but I also epoxied 4 heat-sinks to a geared '250W' hub that I run at 750W with oil cooling. So far the heat sinks seem to be shaving a good 15-25C off the surface temp.
Cheers
Marc S.
100 W
Cowardlyduck said:This is so awesome Sketch. Love that your following through on getting them made. I'm sure they will be popular.
You should start a new thread for heat-sink cooling on hubs so this discussion can really take off. If you don't I will.
Don't have any pics, but I also epoxied 4 heat-sinks to a geared '250W' hub that I run at 750W with oil cooling. So far the heat sinks seem to be shaving a good 15-25C off the surface temp.
Cheers
Since I burned-up an ATF cooled Bafang BPM with 20s on Sunday, I might have to look into heat sinking the new Bafang CST as well...
Yeah, for DD hubs Ferro Fluid + heat-sinks is the ticket, but for geared hubbies you need oil + heat-sinks. Doesn't have to be ATF either. I didn't have any ATF for mine and did not want to spend any money getting some, so just used what I had on hand which happened to be super sticky chainsaw oil. :lol: It still worked just as well AFAIK and no issues yet.Marc S. said:
Cheers
sketchism
100 W
Well, its official, the factory is paid and the first run is on its way, expected to arrive within 18 days
They worked out more expensive than i had hoped but they will still be cheaper than buying computer heatsinks and mucking around getting them to fit which was my goal for the project, for the power handling increase i don't think there could be a better option and i want them on as many bikes as possible so they'll be as cost effective as i can possibly do them for, i'll be happy to send larger quantities to shops and dealers if they're interested as well so i'm not sitting around boxing up heatsinks all day haha.
As soon as the sample run arrives i will be running some bench wind tunnel tests as well,(unless maybe Justin would like to if i send him a box of sinks, they are extremely effective with FF so it would be a great combo) i'm working on a test stand as we speak that will load the motor up so we can watch it on the FLIR, ill run three distinct tests, Stock, Heatsink and FF+ Heatsink
if anyone has any tests they be interested in seeing i'm all ears!
I'll make a thread shortly just on the heatsinks as Cowardlyduck suggested so i don't keep clogging this one up too, but it's not quite ready for a 'sale' thread as i would like to still complete more testing once the first units arrive, i will however have some extras of the first run if anyone would like the beta versions?
thanks to everyone for their input on the initial design, hopefully we make a big difference here
They worked out more expensive than i had hoped but they will still be cheaper than buying computer heatsinks and mucking around getting them to fit which was my goal for the project, for the power handling increase i don't think there could be a better option and i want them on as many bikes as possible so they'll be as cost effective as i can possibly do them for, i'll be happy to send larger quantities to shops and dealers if they're interested as well so i'm not sitting around boxing up heatsinks all day haha.
As soon as the sample run arrives i will be running some bench wind tunnel tests as well,(unless maybe Justin would like to if i send him a box of sinks, they are extremely effective with FF so it would be a great combo) i'm working on a test stand as we speak that will load the motor up so we can watch it on the FLIR, ill run three distinct tests, Stock, Heatsink and FF+ Heatsink
if anyone has any tests they be interested in seeing i'm all ears!
I'll make a thread shortly just on the heatsinks as Cowardlyduck suggested so i don't keep clogging this one up too, but it's not quite ready for a 'sale' thread as i would like to still complete more testing once the first units arrive, i will however have some extras of the first run if anyone would like the beta versions?
thanks to everyone for their input on the initial design, hopefully we make a big difference here
litespeed
100 kW
Assuming they fit I'd be interested in being a beta tester.
Let me know where this goes.
Tom
Let me know where this goes.
Tom
sketchism
100 W
litespeed said:
They will fit all hubs with the 'standard' 222mm OD (MXUS, Leaf, QS, Crystalyte etc) can be run single or in pairs
I've been doing further computer modelling and in (very rough) testing we are shedding almost an extra 1000W of heat energy from the magnet ring while underway, slightly more than my initial calculations so it will be very interesting to get these on the bikes and on the bench and compare directly. I'm definitely interesting in sending the first batch out to people willing to beta test and real world results, if you're interested definitely shoot me a PM.
Similar threads
