Oil cooling a motor with transformer oil/dielectric fluid - the next frontier?

neptronix

Administrator
Staff member
Joined
Jun 15, 2010
Messages
17,407
Location
Utah, USA
User 'TBD' brought up an interesting idea of using transformer oil in a geared hub motor as a way to cool it in the bafang G310 thread.

In this thread we have someone quoting a German forum and claims that ATF eats wiring, then fork oil eats wiring and in the next sentence correlates fork oil with mineral oil. Seems to me fork oil has a range of formulations.

I had planned to use mineral oil, aka paraffin oil, also used as transformer oil for dielectric and non reactive properties around electrics with a few percent PTFE (and maybe <1% MoS2) in the little G311.

Thoughts?

From what i've read, transformers can contain wood and various plastics. Transformer oil should also be compatible with motor enamel, copper, and various metals. The conductivity of this oil would also be as low as it could possibly get by it's nature, so we could figure that there's the lowest amount of risk of shorting phase or hall wires.

I'm curious if anyone here would know more about the feasibility of this idea. If there seem to be no problems, i propose this:

1) I collect a variety of different broken geared motors from forum members for testing. Ideally, i would have a pair of each type of geared motor.
2) I take extensive high resolution pictures of each motor 'before'.
3) I buy two different types of transformer oil and fill one half of the motors with oil #1 and the other half with #2.
4) I let the motors sit for 6 months, and swish the oil around once per month.
5) After the 6 month period, the oil is drained, cleaned off, and i take another series of high resolution pictures to assess whether any materials in the motor have been damaged.

We can then have a better idea as to whether transformer oil is an ideal fluid for cooling geared hub motors.

But first.. does the idea sound feasible? :)
 
Strip motors and drop the electrical parts into an insulated container and cover with oil. Immersion heater set at say 90C and (fish tank) air pumps running continuously.

I don't think you'd see anything nasty with specific transformer oil. Hardware store mineral/paraffin oil would be interesting.
 
That would be a pretty proper test, but i do not have the equipment or space for that... maybe someone else could take this test on.

I did read one thread where someone complained about transformer oil eating plastic.

This document i found would suggest that Dielectric Fluids ( DF ) may work well:
https://dsiventures.com/wp-content/...l-in-Motor-Cooling-Part-1-Heat-Transfer-1.pdf

Base oils used in blending DFs were created by a chemical reaction, and as a result,
they contain very few impurities, compared with white oils or mineral oils. This helps
with issues such as oxidation resistance and material compatibility. Synthetic oils also
have very low pour points because they don't contain any waxes. This makes them
ideal in applications at very low temperatures.

"White Oil" is a special type of mineral oil that's undergone a refining process called "hydrotreating". Hydrotreatment removes many of the impurities found in refined products. White oils are often used in cosmetics and as medicine,
but because of their clarity, wide availability and lower cost, have been tried as heat transfer fluids, also.
Like mineral oil, white oils do not contain any additives, and therefore have a lower resistance to oxidation. White oils generally have physical characteristics similar to those of mineral oils.

Here is an interesting test that dsiventures did:

https://dsiventures.com/wp-content/...Reliability-in-Liquid-Cooled-Applications.pdf

2020-12-27 10_06_05-Window.jpg

A paper about chosing which dielectric fluid to use:

https://dsiventures.com/wp-content/...g-the-Right-Fluid-for-Dielectric-Cooling1.pdf

2020-12-27 10_07_20-Window.png

More interesting stuff found in various papers:

2020-12-27 10_13_35-Window.jpg

They appear to sell a fluid that SHOULD be compatible with every material inside our motors.

2020-12-27 10_11_30-Window.png


..this is all very interesting stuff.. and could translate to a cooling medium for controllers as well.
I have a thermal camera on order and have sent DSI ventures an email asking what would be best for our motors.
If i do not get an email back, i am calling them next week.
 
Here's Gigabyte in 2019 demonstrating a liquid cooled server with multiple graphics cards and minimal heatsinks:

https://www.youtube.com/watch?v=YyKIZPuepl8
[youtube]YyKIZPuepl8[/youtube]

As you can see, there are all types of plastics and metals in this enclosure.

This is a new dielectric cooling fluid made by 3M.
Looks like a good liquid for our motors... :)
 
White oil, mineral oil, paraffin oil, transformer oil

none of these terms have a hard & fast objective definition.

However **food grade** mineral oil as used for keeping butcher blocks lubed,

I would think be pretty well standardized as to composition and

pretty well refined, relatively pure compared to industrial versions

and not expensive.
 
I dont know a lot of these things, but it seems unlikely that atf would damage the windings in a motor.
There are a lot of solenoids and wiring in an automatic transmission. I am not sure if the solenoids are sealed so the fluid cant reach the winding, but I dont think that it is always the case. In the cases where the controllers are built in the transmission those seem to be sealed as far as I know.

There are sometimes electric hydraulic pumps too, I know of one case where that is defiantly not sealed. But that is in a controller for a dual clutch transmission, and it uses its own fluid. That is the same fluid that vw/audi uses for power steering.
 
Food grade mineral oil may be a good lead due to the low amount of impurities in it.

What i like about DSI ventures' products is that they have done the science and posted the results already. So we can skip the exploration phase of this and go straight to verifying their results.

At ~$100 per gallon, their dielectric fluids are expensive, however.. consider that a gallon of this fluid would fill dozens of motors and provides a huge benefit.. this may be the ferrofluid equivalent that geared hub motors have always needed.

I would certainly rather experiment with this first, since so much more is known about how it performs in this application.
 
I received an email back from dsi ventures this morning:

Hi David. Thank you for your interest in DSI Ventures’ OptiCool Fluids.

I believe we have the exact fluid you need. Our OptiCool Fluids have been successfully utilized for EV motor cooling for over 10 years and is their specified fluid of choice.

In regards to your query, any of our OptiCool Fluids would be fine for your application. They all have similar base oils and therefore have equivalent chemical resistance and similar thermal properties. They should all be compatible with the materials you have listed below. For chemical compatibility with general materials of construction, OptiCool Fluids are compatible with the following materials: All Metals; Plastics- Tygon-HC, Acrylates and methacrylates, Epoxy resins, HDPE, PEEK, Nylon, Polypropylene, Polyacetal; Elastomers-Viton (all grades), Nitriles, Buna-N, Fluorosilicone, Fluoroelastomers, PTFE. OptiCool Fluids are least compatible with: Urethane/Polyurethane at elevated temps, Neoprene, standard and RTV silicone, and EPDM and is not recommended for extended exposure to those materials. For seals and o-rings, viton has demonstrated the most resiliency in OptiCool Fluids and is the recommended elastomer of choice. This list is not all-inclusive nor absolute, but represents materials that have proven to be compatible with OptiCool Fluids in immersion, electric motor, and battery cooling applications over time. Be cautious of seals that were specified for glycol/water service since a majority of these are made of EPDM and will not hold up in OptiCool Fluids service.

In regards to fluid properties, I have attached a product data sheet for our OptiCool Fluids line with various physical and thermodynamic properties so that you can see how these properties vary over our fluid offerings. For your application, OptiCool-H would probably be the best fit since it is the least viscous fluid and the most utilized for EV motor cooling. Since your max operating temp is specified at 100 °C (and I assume your system will be closed), the flash point of 135°C of the OptiCool-H should not be an issue. If you need additional thermodynamic and physical property data over your fluid temp operating range, we can provide that as well. These data are suitable for input into CFD thermal models and other engineering calculations.

I hope this helps. If we can provide any additional information to help with your EV motor cooling fluid application, just let us know.

Sincerely,
David Childs
VP Technology and Technical Services
 
My analysis:

Opticool-H seems ideal for our motors because it's flash point is 135C. 100C is where neodymium magnets tend to start degrading... so we may be cutting it close to the ideal amount of thermal headroom but with a temp sensor on the immediately hottest part of a motor ( windings ), I think it is safe to use this. Without a sensor, Opticool PH5 would be ideal... accepting higher viscosity and thus motor drag as a downside.

The viscosity of Opticool-H is half that of 0W engine oil @ 100C.
The viscosity of Opticool-H is a fourth of DEX-III ATF @ 100C.
I cannot find viscosity data on ferrofluid but i imagine it is hard to calculate because the magnetic properties skew the measurement. My guess is that ferrofluid has higher overall resistance inside the motor.

In terms of thermal transfer, the thermal conductivity of all opticool fluids is right between diesel oil and gas engine oil.
Ferrofluid has approximately 25% better thermal conductivity.

The good thing about ferrofluid is that it naturally hangs out in an area where it can do useful work - between the magnets and stator. Instead of cavitating ( splashing about, creating a kind of resistance in the motor ), ferrofluid stays where it is useful. But ferrofluid should also create a magnetic resistance.. so.. one has to wonder if the overall added resistance picture is better or worse.

The opticool may cavitate a lot but it's low viscosity negates this. You could use a much greater volume than you would with ferrofluid, and end up with much better and even heat dispersion across the case.

This may actually be the best fluid for any motor.
 
j bjork said:
I dont know a lot of these things, but it seems unlikely that atf would damage the windings in a motor.
There are a lot of solenoids and wiring in an automatic transmission. I am not sure if the solenoids are sealed so the fluid cant reach the winding, but I dont think that it is always the case. In the cases where the controllers are built in the transmission those seem to be sealed as far as I know.

True, and that is a high material compatibility.. but it's been noted that ATF often eats PVC wiring insulation. So this is a big problem.
In order to use ATF, we could swap wiring insulation.. a pain in the ass job.. but i think it is worth trying in a geared motor, etc.

As you see above, i think opticool-h or some other brand of hydroisomerized paraffin would be better.
 
I put about 200mil of this stuff in my QS3k mid drive like 2-3 years ago without thinking much about it, maybe I'll open the motor this week and check if any damage occurred :!:
https://www.duralag.ch/uploads/qRexMlJd/Olykon-Techn.-Datenblatt-Oel-Additiv-D_ELow.pdf
 
Yes, lubricity is a big concern.

Since the stuff is compatible with many different fluids, perhaps adding a touch of liquid PTFE or some other high end lubricant may give the fluid adequate or perhaps even better lubrication.

PTFE ( teflon! ) is said to provide such a massive reduction in gear friction that it can cut measurable heat from a gearbox 45%.
However, using such an environmentally nasty chemical in an electric vehicle would be a bit ironic!

I am left wondering what other kinds of lubricants could be added to opticool to increase the lubricity. Any leads?
 
Chemistry is complex,

just getting one known-OK substance when so many materials in great variety of combinations is enough of a challenge

adding more factors to the mix is verging on chaotic systems,

end up relying on luck of the draw

or drastically increasing the difficulty of the required testing regime
 
The long version : https://dsiventures.com/wp-content/uploads/2019/04/Technical-Manual-OPTICOOL.pdf . Note that they claim good lubricity , and compatibility with some gasket materials, and insulation materials.
 
I am leaning towards encapsulation. I have some ER2225. I probably need to ask questions, do some research. Will probably need to replace hall and phase wires.

It would be nice if there was an easier one size turn key solution, but alas...
 
I do believe there are some threads from 2007-ish of people filling their motors with oil to help with cooling - Toyota's hybrids drip oil on the stator. I'm not certain if they are using a separate oil for that or what but perhaps it could be a starting point.
 
neptronix said:
Yes, lubricity is a big concern.

Lol Half fill it with alcohol free gasoline and suck the O2 out. Replace it with Hydrogen under pressure of about 200KPA. Mmmmm Phase change.

Lubricity. Lubricity is for valves. Grease is for bearings, and I hear bearing come with labrynth seals. Containing the grease. I wouldnt worry. Plus bearings r cheap to replace if testing.

I do like the idea, I have some experience and knowledge when it comes to heat, an asset. I considered mineral, parrafin, peanut, and looked into them all. I am certainly happy at this level (o power) with some ATF. Lasting, reliable, results easily enumerated. Cheap. Off the shelf. Reliable insulator. Petroleum based.

I more worry about the continuous current in the lead lines. Given we cannot use very high voltages. Use rubber hose over a phase line and cool that with a (conductor) medium.

I am not sure any of us are at that level with our vélo électrique though. Given available power/weight, my fellow Horatios.
 
I'm surprised that nobody has played with this stuff still. I tipped Justin off to the existrance of isoparaffin oils and as far as i know, he hasn't gotten down and dirty into it.

It may be suitable to add a lubricant additive which is compatible with the materials. A few drops of PTFE ( teflon ) comes to mind first, though there may be many other high density, high lubricity additives out there which have a better environmental safety profile.

Even if this theoretical additive may not dissolve properly into the isoparaffin, it will continuously be mixed/churned by the motor's gears and the stator rotation. It should be effective enough.

isoparaffin has better heat transfer characteristics than mineral oil and likely ferrofluid.. it should absolutely be considered for this purpose.
 
Pushing for heat shedding modifications to get more power out of a lighter smaller motor may not fit Grin's business model

Those customers concerned with utility, focus on transportation, getting a load from A to B, maybe daily commuting

are radically different than those out for fun, sporty use cases.

Just staying within a motor's design parameters, going to a heavier motor when "higher performance" is required

is going to give more dependable reliability.

 
I believe there is a solid reason Grin is not selling very high power hub motors, as well as not selling complete ebikes.
The video on the tour of the old old store/warehouse while their outside the front door, he explains why.
They do sell a lot of different lower power hub motor models to suit legislations all around the world.
I doubt they'd sell Statorade if it wasnt selling or didnt fit into their business model, diy'ers and school/univ/college ev projects.

https://ebikes.ca/statorade.html
10mL Syringe of Statorade ferrofluid for use in cooling direct drive hub motors, with convenient Luer Lock plastic nozzle for injecting into holes as small as 2mm diameter.
Approximately 5mL of this fluid injected into a hub motor will more than double the ease of heat flow from the stator of your motor to the motor shell without any noticeable increase in motor drag

https://ebikes.ca/shop/electric-bicycle-parts/motors.html
All of our direct drive hub motor offerings generally feature a built in thermistor for temperature sensing, 0.35mm laminations for low rolling drag, and are sealed with a Statorade injection port for extra cooling if required


john61ct said:
Pushing for heat shedding modifications to get more power out of a lighter smaller motor may not fit Grin's business model

Those customers concerned with utility, focus on transportation, getting a load from A to B, maybe daily commuting

are radically different than those out for fun, sporty use cases.

Just staying within a motor's design parameters, going to a heavier motor when "higher performance" is required

is going to give more dependable reliability.
 
I see a transformer winding is completely stationary and sealed convection does the fluid transfer.

Inside the motor theres no fluid transfer to a heatsink but the fluid increases the transfer of hest into the surrounding motor shell plus the oil itself will have a specific heat capacity that will increase the heat well size of the motor giving a longer burst duration possibly increasing burst if motor design is desirable and good contact is made.

Heres where i show you a trick of an ice engines coolant systems.
IMG_20210419_005300096.jpg
Isochoric specific heat (Cv) for water in a constant-volume, (= isovolumetric or isometric) closed system.
Isobaric specific heat (Cp) for water in a constant pressure (ΔP = 0) system.

Specific heat is how much energy we have to put in to a substance to get it to rise by 1c in temp high the number better.
Transformer oil has a specific heat capacity of 2.1 at room temp and pressure
100% car coolant is around 2.7 with a 50% water mix its more like 3.4
Water alone is 4.1

The table shows as we boil a liquid and the temperature increases the amount of energy we have to put in per C degree change also increases.
In a sealed system, as the pressure rises the energy the liquid then needs to rise each C increases too.

Combined together cars operate with a specific heat capacity close to 4.1 + once warmed up just as good as pure water with the bonus of winter protection and added rust inhibitors.

Its for that reason cooling loops are good for high powered equipment motors etc but even then oil still has its place in the sump where lubrication is vital.

If an existing motor could be made to work and windage not be an issue i say why not wont hurt to try buy dont go out your way making it a traditional sealed cooling loop would do the job better.
 
markz said:
I believe there is a solid reason
Yes statorade is his thing. I just meant as to his not investing time into experimenting with other options.

 
john61ct said:
markz said:
I believe there is a solid reason
Yes statorade is his thing. I just meant as to his not investing time into experimenting with other options.

Yeah who knows, high overhead means how much experimenting and to what goal when high power is not his thing. Good news is he's got vented motor specs in the motor simulator :thumb:

I'd be interested in reading more about alternative ways to cool a hub motor. The most I've done is pour some drinking water on a hub and once I dipped my hot motor into a river for a good cool down right before I was about to go up a long hill.

Also remember some mid drive motors come with liquid cooling piping ready to go! :thumb:

Oh wow, look here, water cooled piping for hub motor. Motorcycle/Moped, high power!
QSMOTOR 17X4.5inch 12000W-14kW 273 96V 120V Water Cooled DC Electric Hub Motor For motorcycle
http://cnqsmotor.com/en/article_read/QS%20Motor%2017X4.5inch%2012kW~14kW%20273%20Water%20Cooled%20Electric%20Motorcycle%20Hub%20Motor/538.html

I found this by looking at picture search for water liquid cooled hub motor
https://endless-sphere.com/forums/viewtopic.php?t=65972&start=450#p1120436
 
Back
Top