Liquid cooling hub motor maybe not so difficult!!!

[Doctorbass]

Guys.. I had a really good idea last night

We always had problem with the the room on the axel and bearing for adding liquid cooling tubes on our hub motor.

Now i realized that the gauge 10 silicone wire from Hobbykin that we can already succeed to put on the axle have exactly the same diameter as the 3/16 O.D. diameter copper tube!!!

So if someone would try to watercool it while the same occasion of a rewind job on the motor that would be nice!


Think about that...:

2 copper tubes with kapton tape skin around that goes inside the hub as for two of the 3 phase wires... and one gauge 10 silicone wire as last phase wire... that make 3 electrical conductor that also are used as liquid circulation!

The 3/16" O.D. copper tube have usually 1/8" I.D. and i found that the waterflow we can achieve with that diameter of tube is still really interesting.

This 3/16" tube could be just used to cary the liquid in and out of the motor.. and inside the motor, we could use larger tube like 1/4" to maximize the overall flow rate...

 

[Hillhater]

This 3/16" tube could be just used to cary the liquid in and out of the motor.. and inside the motor, we could use larger tube like 1/4" to maximize the overall flow rate...

The MAXIMUM flow rate will be determined by the MINIMUM tube diameter. :wink:

 

[Doctorbass]

This 3/16" tube could be just used to cary the liquid in and out of the motor.. and inside the motor, we could use larger tube like 1/4" to maximize the overall flow rate...

The MAXIMUM flow rate will be determined by the MINIMUM tube diameter. :wink:

Not exactly... Yo talk about the weakest part of a chain.. I know.. But it can not really apply if you take account of the overall tube lengh

The capillarity of the liquid in the tube is offering different restriction with larger diameter and smaller diameter so having a tube all with 1/8" I.D. all over the lengh is not the same as having 10% of the tube with 1/8" I.D. and 90% of the lengh with 3/16" I.D.

That's what i meant! :wink:

Doc

 

[karma]

Guys.. I had a really good idea last night

We always had problem with the the room on the axel and bearing for adding liquid cooling tubes on our hub motor.

Now i realized that the gauge 10 silicone wire from Hobbykin that we can already succeed to put on the axle have exactly the same diameter as the 3/16 O.D. diameter copper tube!!!

So if someone would try to watercool it while the same occasion of a rewind job on the motor that would be nice!


Think about that...:

2 copper tubes with kapton tape skin around that goes inside the hub as for two of the 3 phase wires... and one gauge 10 silicone wire as last phase wire... that make 3 electrical conductor that also are used as liquid circulation!

The 3/16" O.D. copper tube have usually 1/8" I.D. and i found that the waterflow we can achieve with that diameter of tube is still really interesting.

This 3/16" tube could be just used to cary the liquid in and out of the motor.. and inside the motor, we could use larger tube like 1/4" to maximize the overall flow rate...

if you wish i have just the hubmotor for test :wink: can you make a pic of what it would look like inside?

 

[sackspeed]

My idea is to take thin ring bearings with 30 mm bore and use a axle made of 25CrMo4 tubing.
This would give a hole of 26 mm dia. normal bearings are 42mm * 20mm . Thin ring is 61806 (type)
42mm * 30mm.

 

[John in CR]

For the effort needed to implement liquid cooling, the cover modification to run a much larger ID bearing is a small effort that you shouldn't try to avoid.

The heat transfer from the stator steel and windings to a place on the stator where you can place coolant flow tubes is such a slow process that I can't imagine any possibility that it would be better than a forced air cooling rig. Air cooling can directly reach the hot parts, which have significant surface area, and the motor operation provides a natural means of directing the the air flow to the parts that need it most. Air cooling also requires less energy and space, plus implementation is far easier. Air cooling is easier, smaller, lighter, and more effective, make it the obvious choice for our hub motors. Liquid cooling is fun to think about as a mental exercise, but unless you have some operating environment where a filtered air input isn't acceptable, then I believe liquid cooling would be a poor choice.

 

[karma]

ya sure i just ordered the role of copper :wink: your probably right. i think doc's idea is worth trying the water cooled controller works

 

[John in CR]

ya sure i just ordered the role of copper :wink: your probably right. i think doc's idea is worth trying the water cooled controller works

For someone willing to invest the effort, doing it in a controller makes good sense, because you're putting that coolant flow very close to the heat source with only short distance of high thermal conductivity metal between the coolant and the FETs.

In a controller instantaneous and continuous operation are quite similar, but not in a motor, and the extreme power guys are really looking to maximize their short duration power. Liquid cooling would help primarily with the continuous limits, but the steel acts almost like an insulator for heat from a short duration burst. I can easily easy a winding burned up before the temperature at the surface of the cooling tube has risen even one degree.

As far a hub cooling goes, I love what Luke is doing with his rewind, because it leaves good air flow space in the windings. I'd completely block the holes in the stator spokes, and take air in one side of the motor and exit at the perimeter on the other. Then the forced air would flow only through the windings and air gap for an idea air cooling situation. The big stationary covers on the drum brakes on my motors would make them ideal for that alignment, because I have an easy route to force air in, and for 50W or so I could blow hundreds of cfm of fresh air directly through the windings. It would require coolant flow through hollow copper windings to surpass what is possible with air space in the windings like Luke is doing.

Hey Doc, if you want to go faster then what you need is a different motor.

 

[Hillhater]

I think Hal has the best solution to cooling pipe access and heavy wiring...

 

[karma]

yes very nice

 

[karma]

for the water lines and connectors what do you recommend? air lines?

need to make a list :wink:

 

[bobc]

A few good ideas there - combining liquid cooling of the motor with a liquid cooled controller makes sense.
I've made the suggestion before to bury a copper cooling pipe in amongst stator windings, the reasoning is:
copper has excellent thermal conductivity - steel has not (~ 1/7 as good)
the windings copper is the source of nearly all the heat in the motor
you can't blow air through the windings - particularly when they're impregnated with varnish
A very small water flow can cart away a LOT of heat: SHC of water is 4.2kJ/kgK so 100W of heat into a flow of just 1 litre/minute would only raise its temperature 1.43 degrees
(actually that's an incredibly small temperature rise - but I think my sums are right).
Using copper pipes as conductors (OP suggestion) might solve some thorny packaging problems!

 

[rhitee05]

I'm inclined to agree with John, that liquid cooling via pipes attached to the outside of the stator isn't likely to be very effective. Re-winding the motor with copper tubing for the conductors is probably the best approach, but would be quite difficult. Winding the motor without kinking off the tubing would be difficult, although not impossible. Another option would be to flood either the just windings (with some sort of secondary enclosure) or the entire inside of the hub with cooling fluid.

 

[Gordo]

Guys.. I had a really good idea last night 2 copper tubes with kapton tape skin around that goes inside the hub as for two of the 3 phase wires... and one gauge 10 silicone wire as last phase wire... that make 3 electrical conductor that also are used as liquid circulation! [/size]

The 3/16" O.D. copper tube have usually 1/8" I.D. and i found that the waterflow we can achieve with that diameter of tube is still really interesting.

This 3/16" tube could be just used to cary the liquid in and out of the motor.. and inside the motor, we could use larger tube like 1/4" to maximize the overall flow rate...

Doc;
I am sure you know there are different grades of copper tubing. What I am wondering is what the resistance is of the copper tubing vs copper wire? Did you check it?

 

[AussieJester]

for the water lines and connectors what do you recommend? air lines?

need to make a list :wink:

Again, pc watercooling compression fittings are the best shot, looks to be exactly what Hal has on his motor.

KiM

 

[Doctorbass]

Guys.. I had a really good idea last night 2 copper tubes with kapton tape skin around that goes inside the hub as for two of the 3 phase wires... and one gauge 10 silicone wire as last phase wire... that make 3 electrical conductor that also are used as liquid circulation! [/size]

The 3/16" O.D. copper tube have usually 1/8" I.D. and i found that the waterflow we can achieve with that diameter of tube is still really interesting.

This 3/16" tube could be just used to cary the liquid in and out of the motor.. and inside the motor, we could use larger tube like 1/4" to maximize the overall flow rate...

Doc;
I am sure you know there are different grades of copper tubing. What I am wondering is what the resistance is of the copper tubing vs copper wire? Did you check it?

Ies i tested it :wink:

the resistance of a 3/16" "plumbing"copper tube is nearly the same as a 10 AWG gauge wire!

I tested it using kelvin resistance measurement with mu 0-100amp 0-10V lab psu on a 1 foot lengh

Doc

 

[karma]

for the water lines and connectors what do you recommend? air lines?

need to make a list :wink:

Again, pc watercooling compression fittings are the best shot, looks to be exactly what Hal has on his motor.

KiM

i had no idea. look to be the same :wink:

http://www.koolance.com/default.php

thanks

 

[Elektrofreak]

Just a little thing...

You know that if you have a voltagedifferece and water you are doing elektrolyse and your copper will corrode... i hope it will be not that much because you are on AC but it could become a problem :?
Anyone who has an idea? Are there special water additions to prevent it?

Edit:
Thesis:
We could see that you don't need that much waterflow to cool 50 Watts. If the motortemp can rise up to let's say 50°C at 20°C ambient temperature under full load, you don't need that much water either (yes, i know that doctorbass has ver HUGE power dissipation :wink: ). Perhaps it could be enouth to drill as an example a 4mm whole into the axis of the motor, drill a 5 or 6mm drill only 1cm or so at the outer and inner surface if the axis to put a copperpipe in and solder it. I don't know whether the axis also corrodes but you also could push a small pipe all throu the axis into the motor. few (cold) water is better than no water )

[Dis]Prove
when the waterflow would be the same as in your video (250ml/min) than 5kW of motor power would result in a water temperature rise of (let's calculate)

let's say 70% efficiency => (5.000/0,7)-5000 = 2150Watts Power dissipation
Cwater = 4,19 J/K*g
250ml/min => 4,1ml/s => 4,1g/s
P=2.150W=2.150J/s

=>1K temperature rise =>4,19J/K*g * 4,1g = 17,2J/K
=>125K temperature rise for 5.000W motor power
Conclusion
a bit more than i expected ... :?
please someone validate this calculation :lol:

I hope someone understood what i tried to say above my thesis

 

[keyne]

This is probably a really stupid question as I have little knowledge of cooling...

I thought the limiting factor would be the radiator rather than the flow rate, as long as the coolant is moving through at a rate that is great enough to get it all close enough to the motor temperature? Could someone quickly explain why flow rate is so important? Thanks.

 

[auraslip]

If your using the input and out put pipe as a phase wire, wouldn't that mean they would be touching like two phase wires together? You'd have to have non-conductive break somewhere in it.

 

[bobc]

keyne - you answered your own question "as long as the water is moving quickly enough" ;^)
the slower the flow, the hotterthe water gets - hot water can't really cool things!
Pump slow enough and the water will boil, giving a hard limit on the cooling capability.
Elektrofreak - our calculations seem to agree - 2kW of losses is a helluva lot though!!!!

 

[keyne]

But if the radiator can't do it's job well enough surely it doesn't matter how fast you are pumping the coolant through the motor? If you're heating your motor at say 2kW and you can only cool at 300W I wouldn't have thought the flow rate would really make a difference? Please school me!

 

[bobc]

Whatever the radiator is & its performance was not being considered. The OP's proposal was to use thin pipe so my immediate concern was whether sufficient flow was available.
& it probably is ;^)
obviously you need sufficient flow AND a suitable radiator

 

[darkshirikens]

You know a cheaper way you can probably cool the motor is if you rig something like a fish tank filter or the things that make the air bubbles could possibly work if you can rig the power supply to power,and then tubing to get into the motor i have a more than a few ideas lol. just make sure its a motor for like a 40- whatever size gallon tank what ever you choose make sure it moves allot of water but you are going to use it just for the air

 

[AussieJester]

Whatever the radiator is & its performance was not being considered. The OP's proposal was to use thin pipe so my immediate concern was whether sufficient flow was available.
& it probably is ;^)
obviously you need sufficient flow AND a suitable radiator

Exactly, even on pc watercooling 7/16ID pipe is the smallest I have ever gone 7/16 and 1/2ID are the
standard, 100-120watts of heat is about what a modern CPU puts out IIRC been awhile haha... So yes
IMHO if you will need a high flow pump like the Laing DC3.25 7/16 hose 1/2 inch fittings (then you get no leaks
and need no hose clamps) A 360 size double pass Thermochill radiator...at least. The copper pipes in the
controller Doc's used are too small to be effective they will constrict the flow rate faaar too much IMO
But...we all learn by our mistakes haha


KiM