[SOLVED] Trying to plan an oil cooler system

[stonedarts]

Been seeing a lot online about mineral oil having issues over water cooling. One being less heat transfer etc. I've seen many threads on here about similar topics but haven't seen one that is an actual match to the idea I'm working on.

So this is what I'm thinking, system will be pretty high end, haven't been able to decide on actual specs. Debating between i9 and ryzen etc. But I'm thinking about building a 15-20 gallon hex tank. Running something along the lines of a 600 liter per hour pump into a MO-RA 420 PRO.

Haven't build a rig that I can actually mess with overclocking yet so I do want this one to be capable.
So to further upgrade the radiator I'm thinking of buying like a 7 cubic foot chest freezer. Punching a small hole in the left and right sides, run power and radiator intake hoses through left side into the radiator inside the freezer and the exit hose from the other side. All space not used for the radiator and it's parts would be filled with something along the line of blue ice camping freezer blocks. That way there is nearly no actual air space and over night while it is cooling it has a very large thermal mass it can drop down to, possibly like 30 degrees so it's atleast a couple above freezing temps for the oil but has a very large stored cooling potential and the fans the. Blow air from all through the frozen pc.

Is this viable? Has anyone tried it?
Obviously I would have to reseal the holes with insulated material etc and run a thermometer to monitor that it hasn't overran the cooling in the chest.

 

[USAFRet]

Is this viable? Has anyone tried it?

Of course people have tried it.

Resulting in one of two outcomes:

1. Great for marketing, to show off our brand
2. What a PITA, I will never do this again.

 

[USAFRet]

To continue...

What do you hope to get from this mineral oil thing?

 

[USAFRet]

And if I'm reading that correctly...a combination of mineral oil AND actual fridge cooling?
This is when the GoodIdea Fairy gets smacked down by not just one, but TWO BadIdea Fairies.

 

[stonedarts]

Of course people have tried it.

Resulting in one of two outcomes:

1. Great for marketing, to show off our brand
2. What a PITA, I will never do this again.

What is the issue though? I've seen mostly the fridge problem is that it creates condensation so the actual computer can't safely go in a fridge.
And the oil always seems to be a lack of cooling. Oil gets hot needs to cool faster somehow.
What would the issue be in combination?

 

[ElectrO_90]

Thermal Dynamics - some basics
http://physicsexperiments.eu/1770/comparing-specific-heat-of-water-and-vegetable-oil

 

[Deleted member 14196]

Thermal Dynamics - some basics
http://physicsexperiments.eu/1770/comparing-specific-heat-of-water-and-vegetable-oil

nice... lol

 

[USAFRet]

What is the issue though? I've seen mostly the fridge problem is that it creates condensation so the actual computer can't safely go in a fridge.
And the oil always seems to be a lack of cooling. Oil gets hot needs to cool faster somehow.
What would the issue be in combination?

The fridge part of the issue is that the fridge is not meant for a continual heat source inside the box.
You'll burn out the compressor in short order.

 

[rubix_1011]

So, the apocalypse is surely upon us:

A single thread of mineral oil submersion and 'computer in a fridge'.

My multiple nightmares have become one.

The refrigerator idea has nothing to do with condensation - it is about the compressor not being able to handle a continuous stream of heat being produced. Fridges and freezers are only designed to remove heat (once)....and maintain. This is the reason your dad got mad when you stood with the refrigerator door open.

Oil submersion is less effective than watercooling due to the large difference in specific heat over that of oil: compare water with just about every 'oil' on this list: https://www.engineeringtoolbox.com/specific-heat-fluids-d_151.html

It also permeates PCB and is impossible to remove once it is no longer submerged in oil. Also, some oils can cause degradation of thermal compounds and thermal pads, meaning those interfaces might lose the ability to effectively transmit thermal load from say a CPU's cores to the heatsink that is dissipating that load into the surrounding oil.

 

[USAFRet]

So, the apocalypse is surely upon us:

A single thread of mineral oil submersion and 'computer in a fridge'.

My multiple nightmares have become one.

 

[Lutfij]

I think we should let this thread evolve and bear a fruit for the real world.

 

[stonedarts]

So, the apocalypse is surely upon us:

A single thread of mineral oil submersion and 'computer in a fridge'.

My multiple nightmares have become one.

The refrigerator idea has nothing to do with condensation - it is about the compressor not being able to handle a continuous stream of heat being produced. Fridges and freezers are only designed to remove heat (once)....and maintain. This is the reason your dad got mad when you stood with the refrigerator door open.

Oil submersion is less effective than watercooling due to the large difference in specific heat over that of oil: compare water with just about every 'oil' on this list: https://www.engineeringtoolbox.com/specific-heat-fluids-d_151.html

It also permeates PCB and is impossible to remove once it is no longer submerged in oil. Also, some oils can cause degradation of thermal compounds and thermal pads, meaning those interfaces might lose the ability to effectively transmit thermal load from say a CPU's cores to the heatsink that is dissipating that load into the surrounding oil.

I had planned to make it so it wouldn't really have to work as hard. By buying a freezer/fridge that is much bigger than needed. I would then fill it with fake ice type things like blue ice for camping. So there would be a very large amount stored cooling at like 30 degrees, it should take even an overclocked computer a bit of time to catch up to that even if I turned the actual cooling off. Example, say the actual radiator and all parts inside freezer take up one cubic foot, instead of using a two cubic foot freezer I use a 7-10 and all extra space is used to have frozen things to store cooling while I'm asleep and it isn't being used.

 

[USAFRet]

So...
Refilling/replacing the BlueIce things inside, every day.

To achieve what, exactly?
What specific temperature numbers are you looking at?
What does it do now, and what do you hope to get it to in actual every day use?

 

[aldan]

my question would be,what kind of super mutant computer do you have that cant be cooled by conventional means?

 

[stonedarts]

It's being built for high end gaming, web design, video editing and perhaps designing a game or two.
Being built massive overkill because why not, let's see what happens and if everyone has done the same thing already, where is the fun.

The idea would be to create enough stored cooling that it doesn't have to be replaced or refilled.
I would probably build it and run it hard for 8-14 hours and see how temps went, see how thawed the freezer is. And how fast things went up. The hope would be that the large oil reservoir and stores cooling would last 10-14 hours, but that is essentially the experiment.
I would be overclocking, but how much would depend on how well this system works. I would dial it in until the cooling supported it well enough.
And from the response here, probably then see how long that freezer lasts before it dies.

 

[USAFRet]

We eagerly await the build log and testing results.

And photos are a requirement...

 

[Azzyasi]

It's being built for high end gaming, web design, video editing and perhaps designing a game or two.
Being built massive overkill because why not, let's see what happens and if everyone has done the same thing already, where is the fun.

The idea would be to create enough stored cooling that it doesn't have to be replaced or refilled.
I would probably build it and run it hard for 8-14 hours and see how temps went, see how thawed the freezer is. And how fast things went up. The hope would be that the large oil reservoir and stores cooling would last 10-14 hours, but that is essentially the experiment.
I would be overclocking, but how much would depend on how well this system works. I would dial it in until the cooling supported it well enough.
And from the response here, probably then see how long that freezer lasts before it dies.

Are you happy with the room air temperature + 10-20-30 degrees?
If yes, then go with a conventional watercooling solution. Waterblocks can suck up lots and lots of heat.. the only thing to look for is the water temperature you get after the radiator. If the water is not cool enough, then it will stabilize at a high temperature.. for that add another rad in series, or just add a bigger rad (all copper car radiator is not that big, looks nice, a beast in cooling comparing to PC type), thicker pipes, a pump with as much pressure and flow the waterblock will safely accept.
Watercooling is as best you can get.. water carries lots of heat with big specific heat, and decent convection surface-liquid flow. So if you increase the flow of water you will get higher potential wattage of carry to the rad. (otherwise the watercooling is limited not only on the rad side but on the waterblock side as well if wattage is too high, and the water moves slow and in thin tubes and the water will rapidly heat to unwanted temperatures regardless of water input temperature.. so the block is under-dimensioned for the load, or the pump is not up to task to provide more water quatity over time over the block)

 

[rubix_1011]

It's being built for high end gaming, web design, video editing and perhaps designing a game or two.
Being built massive overkill because why not, let's see what happens and if everyone has done the same thing already, where is the fun.

The idea would be to create enough stored cooling that it doesn't have to be replaced or refilled.
I would probably build it and run it hard for 8-14 hours and see how temps went, see how thawed the freezer is. And how fast things went up. The hope would be that the large oil reservoir and stores cooling would last 10-14 hours, but that is essentially the experiment.
I would be overclocking, but how much would depend on how well this system works. I would dial it in until the cooling supported it well enough.
And from the response here, probably then see how long that freezer lasts before it dies.

You can build a very capable computer these days without all the fuss.

If this is a project more than an actual build, then that's understandable - have at it. I certainly wouldn't touch oil immersion with a 100 foot pole, but glycol chilled coolant, I'm game.