• Howdy! we're looking for donations to finish custom knowledgebase software for this forum. Please see our Funding drive thread

Motor designed ground-up for water/liquid cooling?


10 mW
Aug 22, 2009
Does anyone know if there are any (BLDC preferably) motors designed ground-up for being run with watercooling? Or can any of you motordesigning people give input on pros and cons?

My initial thoughts would be, that it would probably weigh somewhat less. While some weight will be structural integrity, I'd think one would save some metal weight with a liquid cooled motor, even with the extra liquid weight? More importantly, one could run it in higher power due to active liquid cooling presumably? I mean most motors usually have cont rating, 5min 2min 1min etc ratings and peaks. Could a liquid cooled run nearer to max continuously? This would result in a seriously improved power/weight rating.

Or are there parts within motors that are just not possible to get within "close enough" distance of the watercooling (close as in short enough thermal route) that will still prevent running at high powers for too long? Even if the design is made groundup for being watercooled?

Currently it seems to me that thermal limits and mechanical limits of motors are not very close to each other, considering prolonged high power use such as cruising at high speeds compared to max motor power available (for example my ICE motorcycle has topend around 160 something km/h and I often cruise at 130s, and on faster roads even at 140-150s, it's running quite near max power at those times for quite a while, even up to an hour ... and it's funny that the fuel use doesn't change that much doing that because it runs in more efficient powerband and seems like the bike has pretty low drag =P).
ewert said:
This would result in a seriously improved power/weight rating.
For the motor.... but you'd be running at less efficient torque levels and so you might need a bigger battery than you would if you had a larger motor..... So, whether it makes sense, or not, depends on the overall design objective. ......

Remember that, at max. power out, efficiency is close to 50%....

You could look at the motors made for RC boats, perhaps?

See also:
Miles said:
Remember that, at max. power out, efficiency is close to 50%....
Ahh good point that didn't remember when making the initial post. So very very true and makes the issue pretty clearcut in itself: going to thermal limits results from lowered efficiency more than higher total power output, hence it is not suitable for continuous operation anyways. Excellent. :) As in, "idea squashed" type of excellent. ;)
As an aside, yeah I had read the 14kW and the 12kW motor threads. Am very interested about the is it Emrax or Emrad or whichever 30kW (60kW max) aeronautically used motor too. I had forgotten about the Csyz (or whatever, man talk about hard name) D1g1t4l whatever (can't those guys think of any easy names? =P) though, which would point to the direction that aiming straight to liquid cooling could still be a way to get higher power/weight just as long as you already aim for silly high power/size (designing all the tech stuff that I don't know enough about that matters about total power and efficiency to high enough levels).

FWIW, my dream project so far is tilting motorcycle-trike (Tilting Motorworks, youtube it) type of thing with aerodynamic fairings, motor efficiency high till 140km/h speed and enough power to get 0-100km/h in around 5 secs. With range 120km or so with reasonably travel speeds (preferably atleast 100km/h). Need the next gen battery tech really for it though. Money not that much of an issue, chassis controller motors etc. well within budget tolerances already, but batteries weight/size still bit of an issue considering the price of anything that could do what I need...
Another Water Cooled BLDC is at http://www.goldenmotor.com


See AEVA Post
The price of 10KW BLDC motor is USD395
The price of 500A controller is USD513

Don't know what the quality or claimed performance is like in reality.
The basic max speed torque for 60V was posted (I inverted the image to see volatage plot)

Need to setup a table comparing it to the beta Colousus Out Runner.
I always wondered how Reid Welch could run his motor with oil in it. He did it for lubrication more than cooling and I think he had a little oil drip at the axle but it worked ??
Miles said:
Remember that, at max. power out, efficiency is close to 50%....

This doesn't sound right, and doesn't agree with any of the factory test reports I've gotten for any motors. Think about it, getting the maximum power out of a motor in the range where it converts half to work and half to heat, just doesn't make sense. If this is correct it must be the theoretical power limit, not real world subject to limitations of the controller and battery.

You're talking about continuous rating specs for conventional passive cooling.

I was addressing this:
ewert said:
More importantly, one could run it in higher power due to active liquid cooling presumably? I mean most motors usually have cont rating, 5min 2min 1min etc ratings and peaks. Could a liquid cooled run nearer to max continuously? This would result in a seriously improved power/weight rating.

Max. power out is at 50% of no load speed, with efficiency at close to 50%. Of course one can't approach this torque level continuously without liquid cooling, so efficiencies for normal cont. ratings would be much higher.

I don't see any need to bring battery or controller limitations into this.
I guess the question remaining would be about the hightech parts then, which I don't really have any clue about:

How much high efficiency power can be gotten out of a motor of size x so that it remains within high efficiency area (70-90% of noload rpm with proper loads from drag/rr), and would the required structural components be enough to dissipate the waste heat or would there be significant possibilities with going liquid cooling in even those situations of high-eff power?

The Motocsyz (scyz? man...) new motor would maybe indicate that yes, there would be possibilities in going with liquid cooling straight up. A single 100hp max motor would still probably have pretty damn high-eff max power too (50hp-ish?), and based on the small info on it in the 14kW thread it is wickedly small too? Gotta find some info on that motor...
From their press release:

The MotoCzysz LQ (liquid cooled) IPM (internal permanent magnet) 75kW/100hp electric motor is one of the highest torque/density motors in the industry (250+ft/lbs/340+Nm continuous 77lbs/35kg). The IPM solution generates more torque while being more efficient than any comparably sized AC induction motor. The MotoCzysz patent pending cooling system elevates the motor’s nominal power output closer to the motor’s peak power output. The MotoCzysz motor can deliver a continuous 75kW (100hp) and is 93% efficient under continuous heavy load (140C). The motor is available to accommodate either right or left hand chain drive.

I think it can be a bit confusing to think in terms of a "higher efficiency area". It is only that because the current draw/torque is limited by the bEMF to a commensurate level. You can't change the relationship between torque and efficiency......(with PMDC)

Max. efficiency is where copper losses and parasitic losses are equal.

There's a dyno curve for an electric engine. The relationships of the different curves remain somewhat similar in electric motors, even though of course the exact numbers are different? I think comparing the RPM "x-axis" graph is the easiest regarding real life mileage, since speed is governed by rpm (and gearing of course).

What I mean by high-eff "max power" is the power around the high efficiency rpm. Without a gearbox, one would probably want to have a reduction that results in the distance delivering cruising speeds being in the high-eff part of the curve. With curves such as those in the link, one could surmise that a properly optimized cruising machine probably will not gain much if anything from liquid cooling, since drag+rr (as in required power) would only mean 100% duty cycle in the max cruise speed which would definitely be in the falling max power part of the RPM area with high efficiency. At the lower rpm's, and thus speeds, the required power would be less and thus less controller duty cycle, less heat dissipated?

Am I totally off? As an aside, man electric engines are like exactly what the doctor ordered for intelligently designed traveling machines (high max power for acceleration is almost a side-effect of aiming for high efficiency cruising speeds). =)

If feeding same input power but at lower voltage / higher amps, it won't change much the curves shown like these with RPM baseline? The motor efficiency curve mostly is governed by RPM? "Can't change relationship between torq and eff" hmm... P=torque*RPM, so for different U*I but same P at same RPM must mean almost same efficiency if torq ~ eff relationship stays constant (but dutycycle and res losses change the eff peak just moves a bit to some direction)?

Mostly incoherent ramblings aloud, don't worry no need to answer. ;)

I think am finally getting some proper insight into the soul of the electric motor ... Seems like liquid cooling is not required for sensible driving, racing uses aside of course.
The heat in PM Brushless motors is mainly being generated in the stator slot copper windings.
After seeing how mass produced flouro ballasts are dipped in special varnish under vacuum to remove air with the windings, I tend to think water cooling a motor would have much greater benifit if the same was done to the motor windings.

After the varnished motor stator is dried the excess varninsh would need to be taken off to keep the tight clearence to the Rotor magnets.
One prolem is you can't easily remove the windings if you want to rebuild the stator for some reason.
Usually thats because its overheated, which chances of would be much reduced by this extra process

outrunner stator example.

As the wire has a thin insulating enamelled layer the heat transfer is very limited to blocks of copper.

The rapid step changes in current with brushless controllers produces more magnetising eddy current losses in the laminations of the stator.
I tend to think if the lamination material is of hi grade this heat contribution is small compared to the windings.
So to keep the wire insulation from melting and then causing shorted turns, which then create more heat, the varnish would work like heatsink paste improving heat transfer into the laminations.
So if the main focus is a big jump in continuous power water cooling alone has limited benifits if there is a slow release of heat energy in the copper windings.
If the motor location has limited air flow then a water cooling jacket (or core for outrunners) like most water cooled motors have is only going to give a small benifit.

It would be great coolant pipes could sit in between the windings or pole ends. Basically anywhere the is no copper winding or lamination.
I'd like to try 2 copper capilipary tube "s"ing their way between windings with the varnish that is full of powdered aluminium.

The other idea is whether ferrofluid (like in tweeter speakers) could be used in the magnetic paths inside the rotor. Something to look into.