Phase Change cooling by oil

[t800]

Hello,

As some may have heard/read about cooling of whole computer by vegetable oil, which in practice works but in case comes need to change parts, things may get messy. Anyway from this idea, i'm thinking has someone tried to use same method for phase change cooling systems? Since they produce lot of heat and are noisy and won't require maintenance in years, so why not sink whole system in vegetable oil and so both cooling & noise problem get resolved in same?

Of course phase-change system procedures lot more heat than basic computer, so this means oil needs to be cooled also, what can done simple pump + radiator, so whole setup i guess would look something like this:



So point is has anybody tried this or any comments why this would not work?..or not good idea at all?


Regards
t800

[Gray Mole]

Your picture's not showing up for me, but I know what you mean.

It's somewhere between phase and chilling if you think of it that way. Closer to chilling since it's oil instead of water, but somewhat direct since it's like an extension of the evap into the oil and ONTO the cpu etc.

It can be a good system. The pump issue is awkward, making sure the pump can deal with the oil.

Water won't mix with the oil but it will condense on it like anything else, be swirled into the oil, and could be an issue over time. I would recommend an oil/water separator. Just like you'd see on a diesel engine. If you could fine a diesel filter housing and mod it into your loop, you'll find that filters for those can have water separation built in.

I'd say a light mineral oil, lighter the better/synthetic low temp viscocity but not hygroscopic would be ideal.

But it's been done successfully. No reason why anyone couldn't do it. Just keep in mind that while oil isn't a conductor and the right kind won't absorb moisture, that moisture can be an issue longterm. Oils can also become more acidic over time with oxidation and the introduction of mosture. Water separation can be pretty important in making sure that isn't a problem.

The more acidic the oil becomes, the more damage it can do to coating, the more moisure it can absorb, etc etc.

So even with water separation, it's important to replace the oil somewhat regularly. A simple PH test when you set up, then maybe monthly testing to make sure the PH doesn't start spiking.

Nothing that's difficult or expensive to deal with, just maintenance you'd need to keep in mind in that kind of system.


Gray

[t800]

Thanks for answers Gray,

Image should now work and also couple more details

but somewhat direct since it's like an extension of the evap into the oil and ONTO the cpu etc.
Basically i was just thinking to sink the phase-change part in oil, meaning evap itself would not be contact with oil.

Just keep in mind that while oil isn't a conductor and the right kind won't absorb moisture, that moisture can be an issue longterm
Ok, if this is an issue with vege oil, probably best is to make oil circulation air proof that oil is not direct contact on air expect few square inches what will be in oil box if covering that with lid.

If i find time & effort on summer, i probably will try this out and post results here.

[sdumper]

Personally I dont think it will work but you never know until you try it.
Will the pump be strong enough to keep the oil flowing? Making a leak proof enclosure might also be tricky and if you have a leak it would be a PITA.
Also wouldnt this be very heavy?

[t800]

Personally I dont think it will work but you never know until you try it.
Please i'm happy to take any comments for why it would not work?

Will the pump be strong enough to keep the oil flowing?
Vege oil viscocity is lot higher than water, but i'm planning to strong pump or real oil pump.

Making a leak proof enclosure might also be tricky Also/ wouldnt this be very heavy?

Not really, if something is water proof it is also oil proof. I'm thinking to use some ready strong(== thick walls) plastic contrainer where has lid ready. Vegeoil will surely add's 50-85% more weight, but i'm planning that contrainer would be just a bit bigger that setup itself, so that i don't need buy lot of extra oil for that.

Basically i don't think there an issue with a pump or container, but biggest concern is oil turning something else (as Gray mention) causing damage to system or turning conductive.

[biglipzit]

You have a proof of concept that will work but it needs tinkering. Like the idea of using vegetable oil, throw that out of the window. A long long time ago a group of us on the extreme overclocking forums looked into oil submerged computers back when we were unlocking Athlon xp 1800+'s with conductive pens and we cam across the problem of vegetable oil breaking down after prolonged periods of heating and cooling. It is not stable enough to do the job and through periods of heating and cooling, the fatty acids would congeal out of the glycerol basically causing half of the oil to solidify and the fat solids are very good at absorbing water.
Now that doesn't mean it cant be done. There is a commercial grade cooling oil used in transformers etc but costs an arm and a leg and that is what hardcore computers, http://www.hardcorecomputer.com/index.html , uses to cool their systems. That doesn't mean it can't be done cheap. Mineral oil would be your best bet. It is not nearly as viscous as vegetable oil and works well in a good quality watercooling pump. it is sold in small quantities at pharmacies and larger quantities at animal supply stores. http://www.amazon.com/Durvet-Mineral.../dp/B000HHLUE6 <--- There is a ink for cheap gallon bottles from amazon. This stuff is very light grade oil and very cheap and works great for what you are planning.
Good luck and keep us updated!

[biglipzit]

Okay straying off topic a very little but a really awesome vid on similar concept... http://www.youtube.com/watch?v=Z_X_hgtlJpA

[sdumper]

Personally I dont think it will work but you never know until you try it.
Please i'm happy to take any comments for why it would not work?

I dont think your going to have enough volume in a standard size case to disperse enough heat even with the external rad and fan.
Of course it depends on how big your enclosure is.

Also dont forget to monitor your high side pressure while your testing it out...

Opps I forgot to ask whats the point? I mean the compressor isnt really that loud and the external fan on the rad is going to offset the fan noise savings.

[biglipzit]

Opps I forgot to ask whats the point? I mean the compressor isnt really that loud and the external fan on the rad is going to offset the fan noise savings.

That is the real question... if you are looking for better condenser efficiency which is the only plus that I could see then why not just get a larger condenser with bigger fan or water cool the condenser...

[Gray Mole]

Ok, I had no idea what you were asking

Um, 'pointless and messy' would be the new answer to your question.

A water cooled condensor is the best choice if you want the noise to be more remote.

Ensuring a quieter compressor would be needed to meet your requirement.

The acidity issue wouldn't be a factor in that setup though, since moisture isn't a risk where it's so warm.

For this kind of need, where you want to get the noise 'elsewhere', perhaps going to a chilling system, placing the condensing unit outside could be an option?

Taking the heat out that way should work, but it's such an effort to do this, it seems like more work than the result would give you. But it would block virtually all of the compressor noise, and the external rad would get the heat out if it's large enough. Keep in mind, it will have to deal with compressor heat dump and condensor heat as well. So the system wattage plus compressor wattage would need to be kept in mind.

I'd say 200w minimum for the compressor plus the 300-400w for the system would ensure coverage, so at least 500-600w of heat removal on the radiator/oil cooler. 1kw would be a good idea really, just to make sure it's bigger than needed.

Interesting idea, seems a bit more effort than is 'worth it' but if it's worth it to you, then you can try it out.

Wouldn't want to be the one to make it though

Gray

[t800]

Opps I forgot to ask whats the point? I mean the compressor isnt really that loud and the external fan on the rad is going to offset the fan noise savings.
Yes well like i mention benefit comes from noise and cooling efficiency for condenser and compressor. What comes for normal system condenser, yes it is easy add bigger one + big fan you get it quieter, but for compressor it is different case. Yes they are noisy (specially when we talk about any big one/R404/R507/rotaries) and you cannot sound insulate them due they require good air circulation. Adding few sound dampering/insulation your phase-change cooler walls are useless when sound has direct path to out the case. So in here oil comes picture, kill all sound totally and external radiator can be made quiet.

I'm aware this is messy and big job, but pointless? Well ppl here are building massive 3-4Stage cascade systems to get few dozen degrees colder which helps your CPU to OC 100mhz more..so talk about point in there

But for me, since no one has not done it? and there is some benefit, i see this would be interesting project try.

[Gray Mole]

Cool

I mean, the heat dump of a compressor is a partial factor in evap temp outcome. So it's possible you can get slightly improved temps from it.

Your cooling unit may also smell like a 'fish and chip' shop but that's the trade off I guess

But if silencing the compressor is that much of an issue and you do want to stay 'direct' with the cooling then it's an option to try.

Sealing the case enough to have the oil in it shouldn't be a huge ordeal I suppose.

So it's mainly a pump that can move the oil, and an oil cooler that's big enough to get rid of that much heat. I don't know how much the transmission cooler kits will do but they're not too ugly and not huge.

http://www.ebay.ca/itm/10-ROW-UNIVER...item256e26c560

Small one there, I see up to 20 row larger ones, or could find really cheap ones and stack them with powerful fans.

http://www.ebay.ca/itm/High-temperat...item256b9d05bf

These guys sell a few versions of this pump. would be good to know what level of oil pumping the less expensive ones can do.

But it doesn't have to cost a fortune to set it up. Just need to keep the oil really light and make sure the pump can do the job, and the radiator can handle the heatload.


Gray

[sdumper]

I really dont think its going to be that much quieter but it would be a fun project to build.

[Gray Mole]

I think it would likely take the 'noise' away almost completely, but the vibration would still be there. Most of the noise of a recip is that 'ticky' noise and I don't think it would travel through the oil very well.

The lower omnidirectional noise still would, but it's not nearly as pervasive as the higher pitched stuff. Worth the effort? Guess it depends on how much of the noise is gone. Some, at least.

Worth it smelling funny in your room? That's a much harder question to answer

No oil is completely free of odour. Some are less disgusting. I'd go with light mineral as opposed to veggie oil. That stuff reeks when it's warmed up.

Totally possible it'll do what you want, could be a great system in fact. Just seems kinda gross.

[sdumper]

You hit the nail on the head Vibration wont go away and vibration will equal noise.

[Buckeye]

So lets try and break this down a little.

Instead of a fan blowing air through the condensor we now have oil.
Is the oil pump pointed directly into the condensor ?

Oil is now taken out and cooled with a external condensor.

We also add the heat load of the comp into the oil mix heating up the oil much more, which means the external condensor has to remove more heat load then just a condensor and a fan.

Seems to me that just using a fan blowing air into the condensor is a much better idea. Less mess to deal with and air is better at removing the heat of the condensor vs oil.

Plus you also add a very hot comp into the oil which could get very smelly if it starts to cook off the oil. I guess it depends on oil used.

[Gray Mole]

I think the target of all of this is the noise. Since I've never tried it, I don't know if it will remove the noise. It will take some, but will it take enough to make a louder/larger compressor quiet?

If it can, it's an option. Just a really odd option.

The rest could be good, could be terrible and that's all to do with getting the heat out.

If the oil reduces the noise of a larger compressor to very tolerable levels, then the rest of the noise can be completely elsewhere. If the oil cooler and pump could be placed outside, then that's the other half of the noise reduction equation.

I've never really looked at oil vs water for heat conduction. Worth asking that question though. I'm confident that enough flow will overcome any limitations, but how much flow, how much rad space to cool?

Ensuring you have a good positive flow through the tank, that forces the liquid through the condensor very well, is a big issue. That may take a more expensive/higher flowing pump.

If you keep the oil cool (and it's the same idea as any liquid to ambient cooling) it won't stink. Hot, it will stink, even mostly non-stinky oil will have a smell when hot.

But your target is the same as any system. Keep the liquid cooling the condensor as close to ambient as you can, your system will operate as well as it can.

That will be pretty heavy though Already heavy, another gallon or 3 of oil is going to make it seriously heavy.


Gray

[Stewie007]

I just stumbled on this thread. I am not trying to ruin your fun, because if you feel like this would be a fun thing to do, then you should go for it.

That said, submerged cooling with mineral oil is not going to be as successful as water. It was mentioned above that a water cooled condenser will be a superior option. I hear a lot of people here and there talking about specific heat. Thermal Conductivity is really what you want to look at here. While Specific Heat is important, thermal conductivity is what makes or breaks your coolant.

The reason water is preferred is because it has a higher thermal conductivity than oil. This means that water will pass the heat from molecule to molecule more quickly. Think about a tank of water with a heating element at a certain temperature. Then the same heating element at the same temperature in a tank filled with oil. Then lets assume you're pumping the fluid through a system to cool it. Because of water's superior thermal conductivity it will transport the heat more efficiently. Water molecules that are further away from the element will absorb heat. In the oil, this will occur more slowly. That means that at the source of heat, the temperature will be higher than the surrounding oil. In other words, you will have a sort of heat island. Now, you will of course be pumping your oil through a cooling system, and that adds an additional problem. Assuming you buy a (more expensive) pump able to handle the oil, you will need a much larger radiator. It will give up heat less readily than water (thermal conductivity!).

Specific heat, although not really related to thermal conductivity, says that less energy is needed to raise the temperature for mineral oil than water. Its near half as much for oil. So your mineral oil will absorb the heat more quickly. That is great if you can cool it effectively. For a refrigeration system, this would essentially mean that you'd have to chill it, as air would not be an effective medium. Thermal Conductivity is the most important factor here, and its lower conductivity vs water, and its lower specific heat vs water make it a poor choice for this application; problematic.

Consider steel and copper. Heating techniques for brazing of these metals is different. When brazing two copper tubes together, you heat the inner connection, and then the outer socket connection equally. This means you'll concentrate your heat on both of them with the same concentration of heat. Throw steel (stainless for example) into the equation. Think of a brazed plate heat exchanger. The steel has much lower thermal conductivity than copper. If you heat that steel connection as a copper connection it will be glowing bright orange quite quickly and easily. So we don't heat the steel directly. We allow it to heat by absorbing the heat from the copper, which will pass it on much more efficiently. Then, when you waft the flame over the steel it is glowing a deep red with just one or two passes. Give it the concentration that you would a copper joint and it will be glowing orange very quickly; in the case of brazing that is VERY BAD. However, in my example here, I am comparing for you two materials with different thermal conductivity. This is why copper is used in electrical wires and not steel (I am ignoring aluminum). The copper is your water, and the steel is your oil.

The moral of the story? Oil will retain heat compared to water much as steel does compared to copper. Unless you can cool the oil very effectively, as it will not give up heat as readily as water, you're only inviting a burnout and poor performance. This is not the same as cooling a computer chip. It might seem cool/fun, but its more trouble than its worth in my opinion.

Water cool your condenser if you want to do something fun.

[t800]

Thanks guys all the feedback,

Yes bit crunched the numbers of everything and it with oil it seems it is not good option in the end, instead it is best to replace with water.

Of course this causes first query "how about electricity", but it will be quite simple leave electrity outside of cooling box so we avoid risk getting electric shock.

You hit the nail on the head Vibration wont go away and vibration will equal noise.

Vibration will not go away, but water dampers the sound and most important you fully able to sound insulate whole system because it does not anymore need direct contact on air which equals sound.


Also i want to point out this just an project/experiment which i'm going test only in my workshop, eg controlled environment. No way i would test this in home for multiple reasons

[Stewie007]

Well, you would use a brazed plate heat exchanger, so there shouldn't be much worry about water contacting electricity.

[wdrzal]

Look @ Mytek's and my last posts in Gray Moles evap thread.

I always wondered when someone would get to vacuum insulation. Also vacuum noise attenuation, sound waves can't travel through a vacuum.and neither can heat. You would have to be sure 100% of the heat of compression is rejected to the condensor or through a auxiliary cooling loop.


http://www.youtube.com/watch?v=ce7AMJdq0Gw

With some inventive engineering one could vacuum insulate the evap,line set, heat exchangers where required. and use vacuum noise attenuation (instead of oil) on the compressor. How well it work depends on how deep the vacuum is and how well you can isolate the walls between the compressor and the outer vacuum chamber.

Walt