Will my Preliminary Design for Mineral Oil Submerged PC work?

Obviously, this has been done before.

Not sure why you'd bother to do it.

No, copyrighting is a waste of time unless you have invented something novel.

So far, you haven't.

Thanks for nothing. I started the project purely for the looks.

Mine is different in the fact that the feeds for the cooling go through the side, and the cooling is not meant to come out with the component tray.

I was wondering about the sides being able to handle the pressure, and about the acrylic expanding too much so that the oil is able to leak around the pipes or the bolts, and whether the flow of oil would be enough to keep everything cool. I already know that submerging a pc in oil will work; Im asking about my specific design.

Don't mention it.

While we're on the subject of nothing, there's nothing novel about " the fact that the feeds for the cooling go through the side".

The two questions you ask first, "In theoery will this work? Suggestions for improving the design?" hardly solicit advice on, "the sides being able to handle the pressure, and about the acrylic expanding too much so that the oil is able to leak around the pipes or the bolts, and whether the flow of oil would be enough to keep everything cool."

Good luck.

Why not just buy yourself an aquarium instead of trying to build one? I know the "looks" aren't there, but all I want is performance without the leaks. Also, rather than create potential leak sites just put your cooling fluid ports through the top with long and short extensions for outlet (hot fluid to heat exchanger) and inlet respectively, no need to cut or seal anything.

I'm always amazed that people take the hot fluid off the top and return it to the bottom, too much of a temperature gradient for me. I'd rather take the fluid off at the lower fitting and return the cooled fluid at the top, smaller temperature gradient from top to bottom with fewer hotspots. Check out the math on "counterflow heat exchangers" and LMTD (Log Mean Temperature Difference) for more details. And why submerge the pump, put it outside, much easier to get to and replace if necessary.

Just out of curiousity, how big is your pump, how many oil changes per hour do you expect, and what is the heat rejection capability of your radiator, is there a fan to cool it, or does this whole thing just run on expectations?

The You Tube video will explain a lot.

Puget chose not to use an aquarium because it couldn't handle temperature extremes very well; they needed thicker acrylic.

I choose to go through the sides with the cooling so that I can just about completely seal it except for a couple spots on the top to eliminate dust contamination and oil leakage, and also so that the top could be removed for any sort of maintenance I may need to perform without the cooling system. The idea of having the pump outside the tank is probably a good idea.

As far as having the outlet on the bottom and the intake at the top, I plan to have the circulation going that direction because computers are generally designed that way anyway. All the fans, including the gpu and psu fans which I cannot change anyway, will be exhausting upwards.

I have not picked out any of the cooling parts yet. Is there a minimum power pump you would recommend? Someone on the linustechtips forums told me a d5 pump and 2x 280mm radiators with 4x140mm fans should do the job.

Edit: Looked up the d5 pump, comes with a built in reservoir which I obviously do not need.

Ah. Maybe because it causes more turbulance which inturn causes more movement through all the nooks and crannies? So should I have the intake at the bottom and the exhaust at the top? Or should I reconsider where I put the pipes altogether?

Thinking of going with a 360mmx360mm radiator. Found this article here. Probably going with the phobya g-changer.

http://forums.bit-tech.net/showthread.php?t=258018

Also as far as having the pump outside, it would probably be much quieter and require less maintenance if located inside the tank.

fun project, enjoy!

and pls let us know how it turns out, ok?

:)

You seem to have money to spend...

Have fun reinventing the wheel anyway.

So much effort goes into additional cooling for overclocking in the gaming clique.

More air flow, bigger heatsinks and some airconditioning in summer or an open window in winter work just as well. Doesn't look "cool" but it works.

You do know that the heat will warm up the room regardless of how its removed from the PC? The oil won't get colder than the room temperature.

I realize all of that. This build is simply for the wow factor and that feeling of accomplishment when it's all done.

Enjoy.

Without bothering to see the U-tube videos, I have some questions. Do you really need to use mineral oil? It's rather messy stuff. What about water cooling? So far I'm happy with air cooling. What does the CPU manufacturer say to use to cool it? And why do you need 3 monitors - are you a 3-headed alien?

It's simply for looks. Submerging a pc in mineral oil voids all warranties. It does have some benefits, such as having no hotspots, the fans having a virtually endless life cycle, and not having to worry about dust (for the most part). The disadvantages-weight, mess, difficulty of upgrading and maintenance-far outweigh the tangible benefits not counting looks.

Well, it's going to be messy when you pull your fingers out of the oil after keying the keyboard.

lol idt ill submerge the keyboard

Gaming with multiple players is one reason. I use 2 monitors for normal PC use...bigger desktop can have 2 maximised panes open....three is a bit over the top unless the OP wants a panoramic gaming view, these have been around for a while now.

My son went through this gaming PC phase, I managed to curtail his desire to spend money, pointing out that it's getting better too fast to ever be satisfied. Girls and motorbikes were the replacement therapy....set him straight.

It's true I'll nbever be satisfied, but replacing a graphics card every other year and a system every 4-5 years isn't too much.

He is a gaming freak, and wants to have three displays for his 3-6-9 views alongside his main view.

The oil is meant for cooling off high-capacity graphics cards that usually burn up the gaming computers.

Why not make a computer that is oil-sealed so that it could withstand pressure at the bottom of the Marianas Trench?

Then make it the heart of the autonomous robotic underwater vehicle that goes surveying the ocean bottom for lost jewelry, etc. Or even jewelry that is yet to be made.

I always wanted a well-oiled Crayfish computer.

I don't get the point here: why would you put a computer in an oil filled aquarium?

You have to protect all the plastics and coatings from the oil as they are not compatible.

To achieve this you will add a layer of Silicone over everything = adding a serous additional insulating layer over everything.

The thermal system of a computer is designed to work with air as heat transfer medium. That results in small gaps between cooling faces. Oil will fill the gaps but will not circulate, unless you create a system that forces the oil to flow though these channels. Just adding a circulator will not solve the problem as the oil circulation will remain in the big voids surrounding the boards.

If you would use a light oil you risk putting fire to the system as they have low evaporation temperatures and low flash points.

You really want to build something special that improves the processor cooling?

Go for a pressurized system: put your computer in a pressure vessel: The higher density of the air will improve the heat transfer and the effect on the viscosity is lower. (one advice: fill it with Nitrogen to take away corrosion)

The only issue will be the cable transfer through the wall, but that was an as big issue with the oil filled system.

Gwen that pressurised vessel will still need to dump the heat outside.

I think a partially vented tank of liquid refrigerant with the PC boiling the contents away is sexy too but not practical.

Gwen, the point is there really is no rational technical point. It is bling applied to accessible technologies. Younger folk enjoy doing this sort of thing.....learn through their hobbies that sometimes loads of effort yields little and that wheel reinvention gets old fast.

I know but it starts whith a pimped computer and end with free energy scams.

I rather prefer to work on my hot water improvement

Like an airconditioned gaming room using the pumped heat to heat your water.

That would be practical and useful......can't have that.

I reckon a screaming vortex cooler would be a nice pimp, cool end blowing straight onto the CPU's stock dryback heatsink ( yes they have wetbacks too) Hot end directed at the window curtains of course....flaming games at lightning speed.

OP will learn alot trying.

Mineral oil is nonconductive and nonreactive with the exception of rubber, and then only to a small degree.

Read this (or simply google puget aquarium-it'll be the first link)

https://www.pugetsystems.com/nav/aquarium/mATX/customize.php

Rubber seals in my pressure fittings should be fine, although if i can replace them with silicone i will.

Rubber trim pieces on my motherboard and whatnot will need to be removed-they will dissolve over time and float around in the tank, gumming things up. As far as having hotspots, oil actually does a very good job of absorbing heat evenly-it's just not very good at dispersing it. Hence the need for an extreme radiator, such as this 360x360x60mm radiator which I will probably end up using.

http://www.aquatuning.us/water-cooling/radiators/radiators-active/15007/phobya-g-changer-xtreme-nova-1080-radiator-60mm-full-copper?c=6566

As far as the reasons-the only (or atleast the primary) reason should be asthetics. Watercooling is a much better cooling option, and as far as bang for your buck, air cooling is the best.

Here are my pros and cons for oil cooling

Pros:

Asthetics

Virtually unlimited lifespan of anything submerged

Very little if any maintenance necessary (so long as you take into account the fact that oil wicks up cables and do your best to prevent this)

If it's made properly, no dust

The only noise would be the fans from your radiator, which can be shut off and still maintain reasonable, though ultimately not desirable, temperatures

No hotspots

Cons:

Cost

Void all warranties

Heat can still be an issue if not designed properly

Potential mess

Mobility

Difficulty of maintenance/upgrades

No rubber aloud, so you have to spend more on cables

Cable management becomes much more important, as oil will wick up cables

Good luck selling parts-cleaning up parts that have sat in mineral oil for years is virtually impossible

Don't belief the claim that oil only influences rubber.

Oil has an impact on all plastics, more or less.

It is indeed non-reactive but it swells the structure, cable insulation materials are very suspective to this.

The fact that all is to be silicone treated should telltale

I like the idea of the evaporative system: it would be a huge inside out heat-pipe with the board on the bottom of the tank.

If your main purpose is to create something fancy

A very basic aquarium air pump would have the same effect without causing problems such as producing extra heat, or causing oil to evaporate and having greasy walls. As far as not being reactive, the puget article I referred to in my last post seems to indicate that mineral oil is nonreactive.

Are you an engineer?

An aquarium pump will generate some bubbles that will take a little bit of the heat but mostly will help the natural convection.

Those bubbles are only there for the nice effect they have in combination with the LED lights.

Evaporative cooling is the most powerful cooling you can have. (the inverse of steam heating)

No I'm not an engineer, i was referring to the bubbling effect. Although I do have aspirations to be an engineer.

Dude,

If you really want to be an engineer, and it seems like you are crazy enough in the right sort of way to pursue that path, then keeping things simple is a mantra that you will learn to march to.

Adherence to the KISS principal and not reinventing the wheel are two of the keys to success that you will discover out of the shear need to survive.

Complexity for it's own sake is something to be shunned.......unless an architect/marketer insists and is paying you to play.....if the latter applies then happy days, other wise KISS.

I'll keep that in mind.

Well some physics:

The bubbles from an aquarium pump are only making/helping the fluid to move a bit, and a little bit of air will dissolve in the fluid. The air that will be dissolved will go out on the warmer surface and form a little blanket of bubbles on that surface.

Bubbles on a surface are bad news as that is gas, preventing the cooler fluid to reach the warmer surface. When a bubble becomes to big it will go up. But due the higher viscosity and lower density of the oil the size of bubbles that can remain on a surface is bigger than compared with bubbles in water

If the temperature is high enough to make the fluid boil at the surface the boiling will take an enormous amount of heat and the vapor bubble will leave the hot area enabling new fluid to boil off. This process is so powerful that the surface will effectively be cooled to the boiling temperature in a wide range of heat output. Only when the vapor can't be evacuated sufficiently fast the temperature will go up and the boiling becomes violent.

The two phenomenons are completely different and the second is good, the first is very bad.

By combining the right inert fluid, at the right pressure the exact boiling temperature is programmable. The only issue now is that you will want to have work at low pressures and you want to see the motherboard. This limits the choices of fluids drastically. I doubt that there is a simple solution as the potential fluids I know off are prohibited due their environmental impact. (CFC's)

There are non CFC refrigerants which have similar phase change characteristics to their CFC counterparts.

Moving away from closed system refrigerant gases for a bit... Immersion in a bath of liquid nitrogen would give that nice boiling bling till it's all boiled and vented away. Maybe too cold?

I know but corrosion and flammability comes in the game when moving away from CFC's.

Liquified gasses are mostly too cold, your solder joints will break.

Unless you have a space usage certified computer.

Nitrogen is reasonably inert. Used as an inert gas charge for many applications. It won't go flash or boom.

We're in fantasy land now, gibbering in Enginese.

I'm pretty sure there is some work I should be doing....

Mineral oil has been proven to work so long as you have a sufficient radiator, which I plan to use. I'm going to stick with the tried and true.

There is one thing that potentially can cause problems. Optocouplers. I have seen where cooling liquids can seep past the lead frame on plastic encapsulated optocouplers and cause erratic operation of the optocoupler because liquid can interrupt the light from the LED to photodiode/transistor. Some signals may be optically coupled and power supplies may have an optocoupler in the feedback loop. Just something to keep in mind when you implement liquid cooling.

I don't plan on using optocouplers, I will be using copper or brass couplers.

I understand you won't be installing optocouplers in your machine, but your power supply may have an optocoupler in it as part of the control loop feedback from the isolated output side of the power supply to the line side of the power supply.

The better designed power supplies use magnetic coupling or primary side sensing but, it is possible to find optocouplers in power supply designs. Similary, there may be peripheral connections to your main board that are optically coupled. Some electronic designers will use optocouplers to 'protect' their sensitive circuits from the 'outside world'.

I'm not a PC designer so I can't really tell you what is typically used in a modern PC, but I do design power supplies so you might run across this. I just offer this up on the chance that you build this, and it works fine for a few hours or days or weeks, then suddenly for no apparent reason, it doesn't operate correctly any more.

I learned about this the hard way on one of my designs a decade ago.

Every part I selected is high end, so if optocouplers are used primarily on cheaper parts, I shouldn't have any. I'll check with evga and asrock jic though.

I found a glitsh in your system: why doing all this to improve speed and then run on Windows 7?

Ask any serious pc gamer.

Also, I'm quite familiar with navigating system settings etc, and windows 8 is quite different. I tried fixing a problem on my mom's laptop, I couldn't even find the control panel or the task manager lol.

If I;m not almost done with this project by june (or whenever windows 10 comes out) I will be getting windows 10 though.

We are not a gaming community, we are serious hard working engineers

Seeing as I'm not actually an engineer I'm not sure if thats a statement or a joke....

From what I hear it's probably a joke.

It was a joke.

We know that gamers require Windows, and most of us do "work" on windows stations too.

In my private time I use Ubuntu as this is for free and I can find the software I require for free (if I don't make any money with it.)

There is a speed difference as my professional Windows is overloaded with security measures. (P7-3520M @ 2.90GHz with 16Gb memory)

In my private time I mostly edit video or audio, and I can assure you that Ubuntu beats Windows for the same job.

I was referring to windows 7 vs windows 8, that windows 8 is not optimized for a gaming experience, its loaded with gimmicks like touch screen support, the menus i find are anti user friendly

If one of the aim of this project is better behaved gaming console, you can use a linear psu.

Gajanan Phadte

What do you mean linear psu? The psu I picked is one of the best available, but I've never heard the term linear psu.

The psu used in today's digital equipment is of the SMPS type. This makes things compact and light. This is due to the fact that the transformer is small due to the high frequency in such psu. In a linear psu, the transformer steps down(mostly) and its primary is connected to the mains directly. This makes the psu heavy as the transformer size will be considerably big and also the linear psu has to dissipate the extra voltage in the series pass element which usually is a transistor. Another disadvantsge of the linear psu is it is less efficient.

The linear psu, even if has disadvantages improves the behaviour of the load connected to it. In some cases like audio, it is the only(mostly) chioce.

Gajanan Phadte

So explain me why to use a linear PSU?

More volume and less efficient.

Yes, but better functioning pc.

Gajanan Phadte

??

really, I can't imagine that a PC would perform better with a linear power supply.

I don't see any difference when my Laptop runs on line power or batteries.

I'm sticking with what I've already picked, the evga g2 series.

http://www.coreaudiotechnology.com/products/music-servers/atx-linear-power-supply/

I found this on the net...

...last week i did see a real linear psu powered computer.
was not built for audio use, but for overclock.
actualy there is quite a gain, it does not surprise me at all.
it does indeed look quite "dumb" to have a big powerbrick next to a pc,
but on the otherhand nearly every single test shown improvement.
better overclock capacity, and even , better hdd performance to mention a few.
(probably less noise in data lines helps that hdd read/write figure)
one thing that was verry intresting is to see running temp going down with the linear psu.
this one is a bit spooky for me, i *might* have a clue why is that, but it could be verry wrong.
The local geek does plans to run as mutch of the fans and other motors as possible on an isolated supply, this is what i would not try.
The ps is verry ineffective, and uses linear regulators. without them the computer fails to start up according to him.
Even so start up procedure is complex.
The psu has dummy loads on all rails, after mains power on there is a secund of delay prior motherboard powerup, or else the ripple halts the system.
So it does not look like an easy project, but surely cool for overclocking.
I resurrected the thread to share what i seen, who knows it may be usefull for someone out there on the internetS

First, the pricepoint.... That's more than my budget for this entire project.

Second, the complex power system

Third, the fact that the oil wicks up cables, so having as few cables as possible is my goal. Yup, sticking with what I've chosen, and actually ordering it tomorrow (payday) because the psu on my current machine is going, and fast.

So, how are the moving parts of this machine that are submerged supposed to beahve in a viscous mineral oil, as compared to air, or are there any moving parts?

This has been done before and the fans are in the oil spinning around and moving the oil about in side the tub.

It's art.

What about the extra power consumed by the motor in trying to (inefficiently) pump the oil around using a designed for air impeller? Will you not burn out the motor by overheating it? I give up, use a thinner oil, and see what happens, I guess.

The motor maintains the same amount of power, and does not try to maintain the same amount of speed. Couple this with the complete lubrication and you have a fan with a virtually endless life cycle.

The oil keeps the fans well lubricated, and they have a virtually endless life cycle

The transformer immersed in the oil will have ten times the VA to that in the air. For the fan, the movement will be restricted by the liquid but now it can dissipate much more, due to better cooling.

Gajanan Phadte

What do you mean va?

V stands for voltage in volts and A stands for current in amperes. It is the power that the transformer is designed for. However, Watts is used for the load.

Gajanan Phadte