Spreading unwanted heat around

[betwixt]

I post elsewhere is about needing a heat sink but not using metal. One of the ideas presented was to use oil to conduct heat away and it piqued my interest. I invite comment from others about its merits.

My thoughts:
Good points -
Liquid oil penetrates the places other methods can't reach,
Convection currents in the oil help to disperse the heat widely,
Suitable oils are a good electrical insulator,
Oil can be chemically stable and can have a long lifetime,
They are normally waterproof!

Bad points -
Potentially high dielectric constant,
Potentially high expansion rate, may require some pressure relief method in an enclosed environment,
Properties may be degraded if contaminated, leaky capacitors for example,
Possibly flammable if overheated or ignited,
May react with plastics, including the enclosure.

I appreciate oil is widely used as a coolant in power transformers so I'm thinking of small scale applications here. Has anyone tried using it as an alternative to metal heat sinks? Maybe the final heat radiating surface is the limiting factor, its difficult to fit fins to oil!

Brian.

 

[KlausST]

Hi,

I appreciate oil is widely used as a coolant in power transformers

I tend to say: it´s more in high voltage transformers.
.. mainly because of dielectric strength

***
We have power transformers where the windings are made of copper pipes .. where water flows though.
Thus it "removes" the heat exacle at the place where it is generated.

***
And even oil cooled systems only transport the heat from A to B ... and at point B there somehow needs to be a heatsink or heat exchanger.
(so the high voltage / high power transformers with oil coolant have metal cooling fins. At least the ones I know)

***

If I´m not mistaken when sparks were generated in (former) oil filled electrics .. there was dioxine generated. Highly toxic.
I assume nowadays they use a different oil that does not produce that toxic material.


Klaus

 

[dick_freebird]

Oil, because it will circulate "for free" to the
cooling radiators of high power xfmrs and
convection / forced circulation >> conduction
when any distance is involved (like from iron
core to sheet metal skin).

That oil would be loaded with PCBs, PBBs,
whatever passes present-day muster for
flame retardant.

Probably not so much directly toxic, as
carcinogenic, but you'd have to wade
through some highly politicized material
to draw a science based conclusion.

 

[cupoftea]

I am thinking of a laptop PSU and needing oil instead of alu heatsinks...it would be really awkard for the production staff to manage in that case.
Pouring the oil in the case then somehow sealing it up, and doing this cheaply.

 

[KlausST]

oil instead of alu heatsinks

Again:
Oil is not a heatsink, it is just a transport medium for heat.

And on a flat laptop you don´t have any useful oil convection, thus you need some pump and a rather difficult to design system for oil flow.
If not designed correctly you get "hot spots" with almost no oil flow and spots with high oil flow. Fluid mechanics is a rather difficult task.

Klaus

 

[wwfeldman]

transformer oil serves as an insulator and a heat transporter - as was noted above, sill need a heat sink
it needs to be clean (diatomaceous earth as filter) and the water needs to be removed (dried paper as a water absorber)
both of which need cleaning, drying and or replacement periodically.
the case will needs bellows, or some other volume providing system, as the volume changes with temperature.
need to fill the case entirely, so that changing orientation does not leave the circuits uncovered
need to fill under vacuum (so it gets filled entirely) and need someway to be sure all the air is out

all of these are non-trivial tasks

 

[Easy peasy]

This is already done in power capacitors as well - some server farms are moving to all liquid cooling - where all the psu's are submerged - as well as all the CPU blades . . .

 

[D.A.(Tony)Stewart]

Silicone oil is better than mineral oil and more than 10x thermal conductivity than air, and transport speed up to a couple m/s improves conductivity another 10x. is easier in air.

Water is at least 3x better conductivity than oil but more corrosive and 20x higher Dk. 80 vs 4.

--- Updated Feb 11, 2025 ---

 

[betwixt]

The applications mentioned are all 'large scale' in power generation and distribution systems. The original post I referred to was much smaller, I think just a single street light. Obviously the heat produced is relatively small but it mentioned an Al heat sink couldn't be used and assuming the same rule applied to other metals as well, I wondered if oil (or some other liquid) could be used to distribute heat more equally around the surface of an enclosure, using greater outside surface area to radiate heat away rather than fins or forced air. I'm thinking on the scale of small power supplies and mains adapters. I've never seen it done and I can see it causing production headaches and increased costs but maybe it could be useful in some specific applications. I'm ignoring the end-of-life disposal/recycling problems for now.

Brian.

 

[KlausST]

Hi,

I think we are running in circles.

could be used to distribute heat more equally around the surface of an enclosure

Enclosure:
What material is made of?
It needs to be sealed, every IO needs to be sealed. Still it needs to care for thermal expanding. how much effort needs to be done there? And even if you do high quality construction (pricy) .. there will be a failure rate. What happen on failure? Let´s say it leaks from a baby monitor in an unattendent time and room.
For sure it needs to be non toxic. But now the oil is missing inside the device .. creating hot spots. Dangerous hot spots - otherwise the oil was not needed in first place.

Heat distribution:
How?
Convection? Only works good with bigger devices with fixed mounting orientation. For a flat horizontal laying PCB the convection is minimal.
Convection needs to have hot spots and cold spots to interact with each other.
Pump? a lot of effort for a small device...
Relying on the thermal conduction of the oil (oil not necessarily moving)? Then better mold the electronics. No leakage problems .. done million times.

using greater outside surface area to radiate heat away rather than fins

How?
Just a bigger enclosure? This is counter productive for metal enclusures, since fins need less materiel for better heat spreading.
More (bigger) plastics environmentally does not make much sense.

***
I miss some details. Use a fictive device and describe how you expect it to be designed and work.

Klaus

 

[betwixt]

It isn't a design idea Klaus and I agree with your comments. I was just curious if anyone had attempted and could give feedback if they had tried it. Clearly it isn't a method of increasing overall heat dissipation but in an environment where metal heat sinks can't be used, would it help to distribute heat over a larger enclosure space.

Brian.

 

[dick_freebird]

At small scale, oil fill is used in submersible pumps' motors to get heat out to the case.
With a rotating seal this is a life limiter.

 

[Vbase]

It isn't a design idea Klaus and I agree with your comments. I was just curious if anyone had attempted and could give feedback if they had tried it. Clearly it isn't a method of increasing overall heat dissipation but in an environment where metal heat sinks can't be used, would it help to distribute heat over a larger enclosure space.

Brian.

15 years ago I cooled the CPU of my desktop with water. Attached to the CPU a copper plat soldered to short copper pipe, from there 2 plastic pipes to 5 litres plastic container hanging above the computer. No pump was used, it was sufficient to dissipate 60W with temperature rise of 20 centigrade in the CPU.
It was done to silent my computer.

 

[Calm down]

Flowable petrochemical oil is not suitable for low-power or small volume power circuit heat dissipation, and container, sealing, and lifespan are all issues (petrochemical product performance may deteriorate). Low power products use more curable silicone for heat dissipation, with a higher dielectric constant, no need to worry about sealing issues after curing, and better chemical stability of silicone. Because it is also a fluid before solidification, there is no need to worry about contact area issues (vibration is required to eliminate bubbles after gluing), and the thermal conductivity after solidification is higher than that of liquid oil. We usually call him sealant.