What can be done about motor overheating?

Get all your technical information about electric bikes here.
User avatar
CGameProgrammer   10 kW

10 kW
Posts: 645
Joined: Mar 30 2007 7:26pm
Location: San Diego, CA

Post by CGameProgrammer » Jul 06 2007 8:27pm

All I know is a poor connection has high resistance so that can create heat on it. How this would affect the motor's temperature, I do not know.

Nimbuzz   1 kW

1 kW
Posts: 300
Joined: Jan 05 2007 10:57pm
Location: Stinson Beach, CA

Post by Nimbuzz » Jul 06 2007 8:54pm

I don't know much other than this -- I would not be the tiniest bit surprised if a motor heated up because there was something amis with its phase timing.
Al
Tidalforce M750X
C-Lyte 5304, 36V, Montague
Optibike 800Li
Optibike 1100R

AlexK   1 µW

1 µW
Posts: 4
Joined: Jul 06 2007 11:41pm
Location: Sparks, NV, USA

Post by AlexK » Jul 10 2007 12:24am

Hi guys,
I have a few comments.

I don't think a poor connection outside the motor would heat up a brushless DC permanent magnet motor. An AC induction motor maybe but that's not what we have here.

A black surface will tend to both absorb more and emit more radiation than a silvery surface. That means it may give a false reading to a radiation-sensing thermometer gun. A contact temperature measurement would be more accurate. I've seen radiation based temp guns that claim to be super accurate but it is doubtful to me unless they area measuring the radiation spectrum and not just the intensity (doubtful). That aside, cool experiment nimbuzz!

At these low temperatures the dominant mode of heat transfer out of the motor is convection, not radiation so painting the motor white seems like a good idea to reduce sun heating. I've seen motorcycle engines painted black but they run hotter than our electric motors and I truly think it's for looks anyway.

A way to transfer more heat would be to increase the surface area like attaching fins to the outside. Not easily done though.

Does anyone know what type of magnets these motors use? I suspect neodymium because they are super strong and fairly cheap. They start to permanently lose magnetism at 80C. the good news is the heat is created in the windings while the magnets are affixed to the outer rotating ring which is well cooled.

Not really any solutions, just comments. BTW I'm a thermal engineer.

Nimbuzz   1 kW

1 kW
Posts: 300
Joined: Jan 05 2007 10:57pm
Location: Stinson Beach, CA

Post by Nimbuzz » Jul 10 2007 12:42am

When I was checking temp I was constantly feeling each surface with my fingers. The temp readings seemed to make perfect sense as the 150 degree car roof wouldn't let me keep my hand on it for over two seconds and the 120 degree black wheel was very warm but not hot and the silver wheel was warm at 98 degrees. It was also very logical that over time the temp crept up and went up more on the black surface. I would take multiple readings -- like 5 or 10 each time at different parts of the surface -- for what it's worth. The thermometer's manual warned about shiny surfaces like chrome and it definitely couldn't read those -- it didn't mention any other problematic colors and no others seemed to be weird like chrome and bare shiny aluminum were impossible for it to read. I've been playing around with the thermometer in cooking and stuff and it always seems to make sense. I wanted to test it a lot to see if it was reliable. It came from a reliable old motorcycle touring company with quality products, Aerostitch.
Al
Tidalforce M750X
C-Lyte 5304, 36V, Montague
Optibike 800Li
Optibike 1100R

User avatar
xyster   1 GW

1 GW
Posts: 3089
Joined: Jan 02 2007 8:39pm
Location: Visualize Rural Sheep

Post by xyster » Jul 10 2007 7:24am

AlexK wrote:At these low temperatures the dominant mode of heat transfer out of the motor is convection,
Thanks for adding your expertise, AlexK. Got a question about this: Can convection transfer heat out of an enclosed, sealed space like a hubmotor? I thought convective heat transfer required airflow out, else the convective currents will simply circulate within the enclosure.
Ebike: 5304/20", 72V 35A controller, 33AH 80V 20s15p (18650 sized cells) DIY lithium-ion pack
http://endless-sphere.com/forums/viewto ... 47&start=0
Scooter: '06 Stealth s1000, 48V 30A, 4x10ah SLA
http://endless-sphere.com/forums/viewtopic.php?t=148
Ebike: '06 Currie Mongoose, 32V 35A, 32V 22AH hybrid SLA/Li-ion pack
http://endless-sphere.com/forums/viewtopic.php?t=1010

User avatar
fechter   100 GW

100 GW
Posts: 14717
Joined: Dec 31 2006 3:23pm
Location: California Bay Area, USA

Post by fechter » Jul 10 2007 8:54am

AlexK wrote:
A way to transfer more heat would be to increase the surface area like attaching fins to the outside. Not easily done though.

Does anyone know what type of magnets these motors use? I suspect neodymium because they are super strong and fairly cheap. They start to permanently lose magnetism at 80C. the good news is the heat is created in the windings while the magnets are affixed to the outer rotating ring which is well cooled.

Not really any solutions, just comments. BTW I'm a thermal engineer.
They are NdFeB magnets, but I don't know what grade. Some are good to over 100C. Like you point out, they are well cooled under normal conditions. One way to add "fins" would be to saw up a section of aluminum U channel into small pieces and glue them to the drum with silicone or aluminum-filled epoxy. It might look cool too.

Heat from the windings will transfer to the air inside the motor, which in turn will transfer to the housing. This is a high resistance path for the heat, but you have lots of surface area.

Alex, any idea how much improvement you would get if the air inside the motor was circulated (like with a small fan), but not vented to the outside?
There is a fair amount of space between the spokes inside the motor and power could be tapped from the motor windings or from the hall sensors.
"One test is worth a thousand opinions"

AlexK   1 µW

1 µW
Posts: 4
Joined: Jul 06 2007 11:41pm
Location: Sparks, NV, USA

Post by AlexK » Jul 10 2007 10:45am

It is a high resistance path from the windings to the outside air. The hub motor relies on convection and conduction to get the heat from the windings to the outside surface. Since the only conduction path is through the bearings and axle, most if it probably happens through convection of the air inside the motor, conduction through the housing, and convection to the outside air. Not ideal.

It seems to me that it would help quite a bit to have vents in the motor housing but that would let dirt and water in so probably not practical.

The fins fechter suggests would cool the magnets most which I suspect are the most likely thing to be damaged by high heat. I wonder what the winding wire insulation is rated for because that would be the next thing to go, and would destroy the motor (and maybe a controller too).

A quick calculation shows a 50C temperature rise above ambient is reasonable for your hub motor (2500W, 85% efficient, 25 W/(m*K) heat transfer in light wind). But that would be steady state with no resting.

Nimbuzz, looks like you have done testing on your thermometer. I am curious about how good those really are.

User avatar
CGameProgrammer   10 kW

10 kW
Posts: 645
Joined: Mar 30 2007 7:26pm
Location: San Diego, CA

Post by CGameProgrammer » Jul 10 2007 6:35pm

Thanks for your posts, AlexK; they're very helpful. I spent all day Saturday carefully redoing my connections as they were very poorly done before. Indeed, it didn't appear to affect motor temperature, but it was work that needed to get done anyway.

I should note that while the side discs of the motor get hot, I haven't noticed any excess heat around the cylinder part of the motor. So I guess that's good news for the magnets? I ordered that contactless thermometer Nimbuzz recommended and it'll get here Friday.

User avatar
xyster   1 GW

1 GW
Posts: 3089
Joined: Jan 02 2007 8:39pm
Location: Visualize Rural Sheep

Post by xyster » Jul 10 2007 6:58pm

It's 102F and sunny where I am today. I took my X5 out earlier when it was about 90F, and because of this thread decided to check its temp by hand after running it up a few 2500 watt hills. It wasn't too hot to hold my hand on, but definitely much warmer than normal. Mine is silver.
Ebike: 5304/20", 72V 35A controller, 33AH 80V 20s15p (18650 sized cells) DIY lithium-ion pack
http://endless-sphere.com/forums/viewto ... 47&start=0
Scooter: '06 Stealth s1000, 48V 30A, 4x10ah SLA
http://endless-sphere.com/forums/viewtopic.php?t=148
Ebike: '06 Currie Mongoose, 32V 35A, 32V 22AH hybrid SLA/Li-ion pack
http://endless-sphere.com/forums/viewtopic.php?t=1010

sketchism   100 W

100 W
Posts: 171
Joined: Feb 17 2016 8:35am
Location: GC, Australia

Re: What can be done about motor overheating?

Post by sketchism » Feb 03 2018 12:23am

someone just linked me to this thread and it made me rememeber how far we've come in such a short time with HubSinks and Statorade!
Don't overheat your ebike hub motor - www.hubsink.com (beta testing still open)
https://hubsink.com HubSink - Hub Motor Heatsinks

https://australiandroneco.com Australian Drone Co - Aerial Videography

Nothing is fool proof to a sufficiently talented fool, trust me we'll find a way :idea:

Post Reply