Submerge your rigs in liquid

[farlack]

Has anyone tried this?

http://www.youtube.com/watch?v=UuNkB_8fJeY&feature=related

I would have to assume it would save power not having to run fans, + air conditioner to cool down rooms.

[1715121338]

1715121338

[1715121338]

1715121338

[1715121338]

1715121338

[Frequency]

that looks great firts you think were are the fish then u see the gaer damn cool...but greassy 

[stevegee58]

It's probably fine except when you want to work with the hardware i.e. adding/replacing boards etc.
Everything will be slick with mineral oil which sounds yucky.

[Arto]

A handful of Intel servers just emerged from a yearlong bath in an oil-based coolant, and the results were remarkable. The servers ran at a PUE just above 1.0, and showed no ill effects from the oil. Is oil immersion coming to a rack near you?

The idea of immersing servers in oil to keep them cool isn't entirely new--passionate gamers have been housing their systems in vegetable oil for years. But it's time to take notice of this trend when Intel starts singing its praises as a potentially revolutionary method for slashing the price of running a data center.

The microprocessor giant just finished a yearlong test of Green Revolution Cooling's mineral-oil server-immersion technology and is very happy with the results. According to Mike Patterson, senior power and thermal architect at Intel, not only does the technology appear perfectly safe for server components, but it might also become the norm for anyone needing maximum computer power or building out data center capacity. [...]

And the technology is incredibly effective. Patterson said that whereas traditional air-cooled server racks often operate at a Power Usage Effectiveness rating of about 1.6 (meaning cooling tacks on a 60 percent increase over the power needed power the servers' computing workloads), Intel's oil-immersed servers were operating at a PUE between 1.02 and 1.03. It's possible to achieve similarly low PUE ratings with traditional air- and liquid-cooling methods, Patterson said, but getting there can require some serious engineering effort and cost a lot of money.

As for concerns over the effect of all that oil on the servers' processors, hard drives and other components, Patterson says companies probably shouldn't sweat it. When its test period ended, Intel sent its servers to its failure-analysis lab, which, he said, "came back with a thumbs up that a year in the oil bath had no ill effects on anything they can see."

[AmDD]

It still gets hot... Think of it like a deep fryer, the heat you put in the oil doesn't magically vanish, you need to cool the oil. It may not be true with a low power computer but with a mining rig, cooling the oil would be needed.

Also consider that because of the oil everything in it will heat up to whatever temp it is at. Even things that are normally cool, like USB controllers, capacitors, etc which could cause problems if heated up too high.

There was a guy who did this however, I think it was in the "pics of your mining rigs" thread.

[farlack]

It's probably fine except when you want to work with the hardware i.e. adding/replacing boards etc.
Everything will be slick with mineral oil which sounds yucky.

I've seen people on here who have 10-20+ rigs, all maxed already, so adding wont be an issue, and replacing wont either, as no dust gets to your components. I've been seeing online 3+ years with 0 problems.


The energy cost savings alone would make it worth while of having a sticky part if you needed to access it. You could get creative of ways to make eco-heatsinks to zero your heat energy bill.

[farlack]

It still gets hot... Think of it like a deep fryer, the heat you put in the oil doesn't magically vanish, you need to cool the oil. It may not be true with a low power computer but with a mining rig, cooling the oil would be needed.

Also consider that because of the oil everything in it will heat up to whatever temp it is at. Even things that are normally cool, like USB controllers, capacitors, etc which could cause problems if heated up too high.

There was a guy who did this however, I think it was in the "pics of your mining rigs" thread.

That's why you cool the oil, there are many different ways you can do it, and off the top of my head..

Hook a heatsink to frigerator coils.
Hook a heatsink that's buried into the ground.
Hook a heatsink thats connected to a pool.


Where I live, you dig 2 feet down and you're at water. A free source of heat dispersion.

[AmDD]

It still gets hot... Think of it like a deep fryer, the heat you put in the oil doesn't magically vanish, you need to cool the oil. It may not be true with a low power computer but with a mining rig, cooling the oil would be needed.

Also consider that because of the oil everything in it will heat up to whatever temp it is at. Even things that are normally cool, like USB controllers, capacitors, etc which could cause problems if heated up too high.

There was a guy who did this however, I think it was in the "pics of your mining rigs" thread.

That's why you cool the oil, there are many different ways you can do it, and off the top of my head..

Hook a heatsink to frigerator coils.
Hook a heatsink that's buried into the ground.
Hook a heatsink thats connected to a pool.


Where I live, you dig 2 feet down and you're at water. A free source of heat dispersion.

You did not mention that in your first post, it looks like you assume the heat just goes away.
A big heat sink would work well but wouldn't it just be easier, cleaner and probably cheaper to water cool the GPUs?

[HDSolar]

Yes, someone on here is running their rig in mineral oil and mining.  They made their own setup and it is working with I think five cards.  There are pics of it under the pic of my rig thread.  The only issue I can think of that would be a real pain is that the oil can get contaminated once things like dust and other forign items enter the mix.  The would then cause you to have to shutdown and drain, clean and reload as needed which could be a real pain.  I have thought about this and I would say that if your going to attempt it that you basically plan that the gear you put in will be tossed after two or three years if you get that far.  Otherwise it is really interesting idea for us heat and noise producers.

[farlack]

It still gets hot... Think of it like a deep fryer, the heat you put in the oil doesn't magically vanish, you need to cool the oil. It may not be true with a low power computer but with a mining rig, cooling the oil would be needed.

Also consider that because of the oil everything in it will heat up to whatever temp it is at. Even things that are normally cool, like USB controllers, capacitors, etc which could cause problems if heated up too high.

There was a guy who did this however, I think it was in the "pics of your mining rigs" thread.

That's why you cool the oil, there are many different ways you can do it, and off the top of my head..

Hook a heatsink to frigerator coils.
Hook a heatsink that's buried into the ground.
Hook a heatsink thats connected to a pool.


Where I live, you dig 2 feet down and you're at water. A free source of heat dispersion.

You did not mention that in your first post, it looks like you assume the heat just goes away.
A big heat sink would work well but wouldn't it just be easier, cleaner and probably cheaper to water cool the GPUs?

Well I gave a suggestion, not the solution in the first post lol. Probably would not be cheaper to water cool, I'm not to familiar with watercooling, but looking online, each is $100+, if you need 25 of them, you're spending a lot of money.

You will still have to buy the mineral oil, but at a bulk purchase, you can get a lot of oil, relatively cheap, and you can put a bunch of rigs in 1 tank with oil.

[farlack]

Yes, someone on here is running their rig in mineral oil and mining.  They made their own setup and it is working with I think five cards.  There are pics of it under the pic of my rig thread.  The only issue I can think of that would be a real pain is that the oil can get contaminated once things like dust and other forign items enter the mix.  The would then cause you to have to shutdown and drain, clean and reload as needed which could be a real pain.  I have thought about this and I would say that if your going to attempt it that you basically plan that the gear you put in will be tossed after two or three years if you get that far.  Otherwise it is really interesting idea for us heat and noise producers.

Before I posted this I was doing a bit of searching, and YT videos, the one said after 3 years, everything is still going well without switching oil, the only thing is the oil is a little darker.


Even if this did happen, I would think a siphon technique going into a filter, and just keep adding the clean oil back into the mix would take care of that. Possibly saving a few hundred a month on 10 minutes of work if needbe is a good trade.

[os2sam]

It still gets hot... Think of it like a deep fryer,

Yet, another use for GPU's, Turkey fryer.

[crazyates]

See this post here, and read the following conversation: https://bitcointalk.org/index.php?topic=7216.msg1253938#msg1253938

The end result was that while oil can work amazing for gaming, where you're not dumping a lot of heat into it 24/7, it's not the best idea for miners. It will initially (first 12 hours) keep your GPUs very cool, but once the oil heats up, it's very hard to cool the oil (and the system) back down.

[stevegee58]

Like any cooling system you have to remove the heat in some fashion.  For liquid cooling the warm liquid has to be pumped somewhere to cool, just like a car's radiator.

[farlack]

See this post here, and read the following conversation: https://bitcointalk.org/index.php?topic=7216.msg1253938#msg1253938

The end result was that while oil can work amazing for gaming, where you're not dumping a lot of heat into it 24/7, it's not the best idea for miners. It will initially (first 12 hours) keep your GPUs very cool, but once the oil heats up, it's very hard to cool the oil (and the system) back down.

That's only 1 view, with 1 person who didn't do anything different. There are a bunch of creative ways to cool it down.



Edit: A pipe going out a window, buried into the ground, mixing around with ground water. Free after you buy the copper pipe.

[Desolator]

Oil submersion isn't a good idea for anyone that knows how physics actually works.  They think "oh wow, there's this huuuuge amount of thermal capacity in that oil!  That'll work great!"  Well guess what, after a certain amount of time it heats up.  All that massive capacity changes is the amount of time it takes to get to max temperature!  So 2 weeks later, it's hot!  You still need a cooling system to cool the oil and that's much harder than air or water.  Otherwise your cooling is the surface area of the oil touching the air with no movement and no radiator of any sort.  That's not good.  You'd be better off with air at that point.

I believe thermal conduction is thermal conductivity times surface area times temperature differential.  So if you have a copper fin cooler, that has insane total surface area.  It could seriously be 1 square meter if you were to spread it all out.  Then you have a fan forcing air through it so the temp difference is always virtually the same.  Also copper conducts heat better than oil.  So with an oil pool you've got like 1-2 square feet of surface area touching the air which is not moving over the oil so the temp difference is lower as the hot air rises slowly out of the way instead of blasting by with a fan, always feeding in new air.  It's a much worse system.

[RodeoX]

I tried it with an old junk PC and a fish tank. It did work and was the quietest computer I ever had. Servicing it, however,  was as ugly as you would think.

[paraipan]

Oil submersion isn't a good idea for anyone that knows how physics actually works.  They think "oh wow, there's this huuuuge amount of thermal capacity in that oil!  That'll work great!"  Well guess what, after a certain amount of time it heats up.  All that massive capacity changes is the amount of time it takes to get to max temperature!  So 2 weeks later, it's hot!  You still need a cooling system to cool the oil and that's much harder than air or water.  Otherwise your cooling is the surface area of the oil touching the air with no movement and no radiator of any sort.  That's not good.  You'd be better off with air at that point.

I believe thermal conduction is thermal conductivity times surface area times temperature differential.  So if you have a copper fin cooler, that has insane total surface area.  It could seriously be 1 square meter if you were to spread it all out.  Then you have a fan forcing air through it so the temp difference is always virtually the same.  Also copper conducts heat better than oil.  So with an oil pool you've got like 1-2 square feet of surface area touching the air which is not moving over the oil so the temp difference is lower as the hot air rises slowly out of the way instead of blasting by with a fan, always feeding in new air.  It's a much worse system.

Here you go http://www.pugetsystems.com/mineral-oil-pc.php

It also comes with a radiator in case you need it



And another one

[Frequency]

i just can get it why.. al the effort, it looks very COOL..    or combining two hobbies in one

[farlack]

Oil submersion isn't a good idea for anyone that knows how physics actually works.  They think "oh wow, there's this huuuuge amount of thermal capacity in that oil!  That'll work great!"  Well guess what, after a certain amount of time it heats up.  All that massive capacity changes is the amount of time it takes to get to max temperature!  So 2 weeks later, it's hot!  You still need a cooling system to cool the oil and that's much harder than air or water.  Otherwise your cooling is the surface area of the oil touching the air with no movement and no radiator of any sort.  That's not good.  You'd be better off with air at that point.

I believe thermal conduction is thermal conductivity times surface area times temperature differential.  So if you have a copper fin cooler, that has insane total surface area.  It could seriously be 1 square meter if you were to spread it all out.  Then you have a fan forcing air through it so the temp difference is always virtually the same.  Also copper conducts heat better than oil.  So with an oil pool you've got like 1-2 square feet of surface area touching the air which is not moving over the oil so the temp difference is lower as the hot air rises slowly out of the way instead of blasting by with a fan, always feeding in new air.  It's a much worse system.

Did you not see my picture? You could even attach it to a refrigerator attach the coils that run thru the freezer.

[crazyates]

What's the point, if you're still gonna use NINE(!) 12cm case fans? Why not just watercool, as  any noise advantage is not thrown right out the window.

[Arto]

What's the point, if you're still gonna use NINE(!) 12cm case fans? Why not just watercool, as  any noise advantage is not thrown right out the window.

Perhaps this might be the point, at least from a data center's perspective:

And the technology is incredibly effective. Patterson said that whereas traditional air-cooled server racks often operate at a Power Usage Effectiveness rating of about 1.6 (meaning cooling tacks on a 60 percent increase over the power needed power the servers' computing workloads), Intel's oil-immersed servers were operating at a PUE between 1.02 and 1.03. It's possible to achieve similarly low PUE ratings with traditional air- and liquid-cooling methods, Patterson said, but getting there can require some serious engineering effort and cost a lot of money.

[Fcx35x10]

really cool but i agree about things getting messy

[Phraust]

Novec 7000

[freeAgent]

Novec 7000

That's very cool stuff.

[Arto]

Novec 7000

That's very cool stuff.

+1. That's a clever concept, keeping the fluid circulating without a fan or pumps.

[Phraust]

I've been messing with the idea of making a cooling setup that uses Novec 7000 for my SC Singles (Gallery here).

It's still a work in progress.  I'd shelved it after I'd heard that the new ASICs should be much more heat-tolerant, but recently there have been talks that they can be overclocked, so I am thinking about putting it together again.

[BR0KK]

that stuff is really expensive..... 

[farlack]

that stuff is really expensive.....  

I only found 1 posting online from 2 years ago saying $350/gallon
Maybe the price has changed a bit, but this would still be awesome to have for someone who has dozens of rigs going.
They would obviously be able to afford it, and the energy savings alone would be great.

[Phraust]

Yeah, it's definitely a high-ticket item, but my concerns when I was coming up with this were:

Power Consumption - I pay on average $0.36 per kwh, so with the cost of running both the singles and keeping them cool, is ridiculous.

Corrosion/Humidity - I live near the ocean, so I worry about the lifespan of the devices with all the humidity and salt in the air.

How cool it looks - If it's going on my desk, I don't want a bird nest of wires, components and junk.  I want it to look good.

[squid]

Doesn't seem worth the expensive.. now.. oil cooled asics... hrm might be an idea!

[Phraust]

When I was running all five of my singles here, the electrical costs were more than $400 a month.  If it cost $1,000 for a passive cooling setup, which would drop that to $200 or less a month, it'd pay for itself within five months.  Looking good while it did that would be a bonus.

[Cranky4u]

that stuff is really expensive..... 

+1

[farlack]

When I was running all five of my singles here, the electrical costs were more than $400 a month.  If it cost $1,000 for a passive cooling setup, which would drop that to $200 or less a month, it'd pay for itself within five months.  Looking good while it did that would be a bonus.

Awesome, thanks for posting some monthly electric costs.

[Lethos]

I've recently got my HTPC and FPGA Mining Rig stable enough I don't feel the need to constantly tweak it and went ahead and did exactly what I had always planned for it. Oil Submerged it in a thin framed glass fish tank, it looks beautiful. Still tweaking, but once it's finished I'll remember to take pictures to show it off.

[DoomDumas]

Novec 7000

Very impressive !  Thanks for sharing..

As for HDD now turning to SSD, I think the future is no moving/mechanical part at all... very interesting tech !

Imagine what tech will be availlable in the future

[b!z]

It still gets hot... Think of it like a deep fryer,

Yet, another use for GPU's, Turkey fryer.

You can also fry your hands with 'gaming' laptops 

[Serenata]

How about the mineral oil itself and the Novec 7000? Can those fluids be recycled? That's a major factor before adopting IMO.

[niko]

Novec 7000

That's very cool stuff.

+1. That's a clever concept, keeping the fluid circulating without a fan or pumps.

More than clever. It's not simple immersion cooling, it's evaporative cooling: any component reaching the boiling point of the liquid (34C in this case) will cause the liquid to boil. Heat of evaporation takes away all the heat from that point on, it's almost impossible to go above 34 C in this case. Most of vapor is then condensed at the heat exchanger, and recycled into the system. This heat exchanger is likely the most problematic part: it's got to be able to dissipate heat into the surroundings at the rate sufficient to keep the hot side below boiling point of the liquid.

[minero1]

Novec 7000

That's very cool stuff.

+1. That's a clever concept, keeping the fluid circulating without a fan or pumps.

More than clever. It's not simple immersion cooling, it's evaporative cooling: any component reaching the boiling point of the liquid (34C in this case) will cause the liquid to boil. Heat of evaporation takes away all the heat from that point on, it's almost impossible to go above 34 C in this case. Most of vapor is then condensed at the heat exchanger, and recycled into the system. This heat exchanger is likely the most problematic part: it's got to be able to dissipate heat into the surroundings at the rate sufficient to keep the hot side below boiling point of the liquid.

Good luck trying to use it in warm climate like where i live, +36ºC all year long. it would be interesting to see a Pressure-Temperature diagram of the stuff.

[niko]

Good luck trying to use it in warm climate like where i live, +36ºC all year long. it would be interesting to see a Pressure-Temperature diagram of the stuff.

There are numerous liquids available, with varying boiling points. A liquid needs to be electrically insulating, inert/non-corrosive, stable, non-toxic, and boiling point should be above your ambient temperature and less or equal to the maximum desired temperature of your components.

[farlack]

Novec 7000

That's very cool stuff.

+1. That's a clever concept, keeping the fluid circulating without a fan or pumps.

More than clever. It's not simple immersion cooling, it's evaporative cooling: any component reaching the boiling point of the liquid (34C in this case) will cause the liquid to boil. Heat of evaporation takes away all the heat from that point on, it's almost impossible to go above 34 C in this case. Most of vapor is then condensed at the heat exchanger, and recycled into the system. This heat exchanger is likely the most problematic part: it's got to be able to dissipate heat into the surroundings at the rate sufficient to keep the hot side below boiling point of the liquid.

Good luck trying to use it in warm climate like where i live, +36ºC all year long. it would be interesting to see a Pressure-Temperature diagram of the stuff.

If you're running a server or many rigs, or any rig at all, you probably have the air conditioner on?

[freeAgent]

Novec 7000

That's very cool stuff.

+1. That's a clever concept, keeping the fluid circulating without a fan or pumps.

More than clever. It's not simple immersion cooling, it's evaporative cooling: any component reaching the boiling point of the liquid (34C in this case) will cause the liquid to boil. Heat of evaporation takes away all the heat from that point on, it's almost impossible to go above 34 C in this case. Most of vapor is then condensed at the heat exchanger, and recycled into the system. This heat exchanger is likely the most problematic part: it's got to be able to dissipate heat into the surroundings at the rate sufficient to keep the hot side below boiling point of the liquid.

Yeah, that's what's most clever about this.  Obviously oil has a very high boiling point.  Even water is 100c, which is far too hot to run a computer.  This Novec 7000 stuff seems to be a perfect liquid for cooling via submersion.  As long as you can keep it from escaping when it evaporates, you're good to go.  In fact, according to this data sheet, you could refrigerate the Novec 7000 to well below freezing (for water).  It seems like something that would be fun for extreme overclockers to play around with.

EDIT: Apparently they're already starting to use this stuff in data centers: http://www.youtube.com/watch?v=N3gCavl2Y6U

[firefop]

That's why you cool the oil, there are many different ways you can do it, and off the top of my head..

Hook a heatsink to frigerator coils.
Hook a heatsink that's buried into the ground.
Hook a heatsink thats connected to a pool.


Where I live, you dig 2 feet down and you're at water. A free source of heat dispersion.

Anywhere on earth - at some distance underground you'll hit a stable region (usually around 6 to 10 feet) where you have a temp in the high 60s. No reason you couldn't passive cool an oil system by burying micro tubes. keep your pumps above ground. There are home heating/cooling systems using commercial heat exchange systems for houses... but I'm sure someone could come up with a DIY solution.

[BR0KK]

Ohhh I want to play with novec

But it's not cheap and not realy healthy ....

You have to build a competely sealed container for the computer so it doesn't evaporate.

One thing on oil rigs:

The one from puget systems ,.... If u install a GPU with a standard ati heatsink. It would pump the oil out of the aquarium. If the fan is strong enough u get a decent "oil" fountain.....:d


Quite messy setup :/

[hardcore-fs]

See this post here, and read the following conversation: https://bitcointalk.org/index.php?topic=7216.msg1253938#msg1253938

The end result was that while oil can work amazing for gaming, where you're not dumping a lot of heat into it 24/7, it's not the best idea for miners. It will initially (first 12 hours) keep your GPUs very cool, but once the oil heats up, it's very hard to cool the oil (and the system) back down.

That's only 1 view, with 1 person who didn't do anything different. There are a bunch of creative ways to cool it down.



Edit: A pipe going out a window, buried into the ground, mixing around with ground water. Free after you buy the copper pipe.

This is illegal in many countries, you are not allowed to contaminate the ground water, it needs to be a 'closed' system.
I.E, one coil for the ground water in/out  a separate coil for the cooling/heating circuit.

such a system costs several KW a day to run.

[Meatball]

Submerging your gear in oil actually works pretty well.  But there's one downside.  Forget about ever salvaging or reselling that gear.  Once you dip that stuff in mineral oil, you will never get it all off.

[Phraust]

Submerging your gear in oil actually works pretty well.  But there's one downside.  Forget about ever salvaging or reselling that gear.  Once you dip that stuff in mineral oil, you will never get it all off.

i've read that thpse novec products can also be used as cleaning solutions, so they'll get rid of most of the oil.  pricey way to clean stuff, though.

[hardcore-fs]

Arklone.

[joshv06]

Submerging your gear in oil actually works pretty well.  But there's one downside.  Forget about ever salvaging or reselling that gear.  Once you dip that stuff in mineral oil, you will never get it all off.

Thanks, I never knew that. Do you guys think mineral water could cool a card as powerful as a 7970? Or will it overheat?

[Unacceptable]

See this post here, and read the following conversation: https://bitcointalk.org/index.php?topic=7216.msg1253938#msg1253938

The end result was that while oil can work amazing for gaming, where you're not dumping a lot of heat into it 24/7, it's not the best idea for miners. It will initially (first 12 hours) keep your GPUs very cool, but once the oil heats up, it's very hard to cool the oil (and the system) back down.

That's only 1 view, with 1 person who didn't do anything different. There are a bunch of creative ways to cool it down.



Edit: A pipe going out a window, buried into the ground, mixing around with ground water. Free after you buy the copper pipe.

Funny you should bring this idea up Farlack,me & my brother have a system we're working on right now to eliminate the coil from your A/C condensor by running sealed copper tubing/pipes into the ground to dissipate the heat.Just getting BTU #'s now with a window unit,but its looking promising for existing units up to 5 tons for residential systems.

The main benefit we're pushing is coils are the main thing to fail due to corrosion in seaside homes,in ground copper will elimate that failure point & last almost forever & remove that unsightly outdoor condensor from view 

[Phraust]

wish the ground wasn't volcanic rock here, I'd love to be able to dump excess heat into it.

[firefop]

wish the ground wasn't volcanic rock here, I'd love to be able to dump excess heat into it.

Interesting locations are fun!

With volcanic rock as your starting point you're talking 300% deeper than normal to find a stable temp. So you're basically talking about digging 30-50 foot hole... putting sand (or pulverized lavarock) in the bottom over your heat exchange tubes and then filling the hole back up with rock pieces.

I don't suppose you've got a handy fissure or cave on your property? Otherwise, I think making this sort of hole happen in that location would be too expensive... maybe if you were also sinking a heat exchange system for your house at the same time...maybe.

[hardcore-fs]

This "oil" bathing is all well and good until.........

1. An Ecap explodes, spraying oil everywhere.
2. You have a "Chip-pan fire" during a critical failure.

HC

[HDSolar]

This "oil" bathing is all well and good until.........

1. An Ecap explodes, spraying oil everywhere.
2. You have a "Chip-pan fire" during a critical failure.

HC

+1

[flynn]

This "oil" bathing is all well and good until.........

1. An Ecap explodes, spraying oil everywhere.
2. You have a "Chip-pan fire" during a critical failure.

HC

+1

Maybe not. Liquids are good at avoiding things to get on fire and at absorbing exploding energy.

Fixing it stays a mess tho.

[vapourminer]

Funny you should bring this idea up Farlack,me & my brother have a system we're working on right now to eliminate the coil from your A/C condensor by running sealed copper tubing/pipes into the ground to dissipate the heat.Just getting BTU #'s now with a window unit,but its looking promising for existing units up to 5 tons for residential systems.

The main benefit we're pushing is coils are the main thing to fail due to corrosion in seaside homes,in ground copper will elimate that failure point & last almost forever & remove that unsightly outdoor condensor from view 

so basically youre talking geothermal. hope youre aware of the length of pipe youll need, my 3 ton geothermal unit (A/C, heat, preheat for DHW) needs 2000 feet of pipe buried 5 feet down, although since its for heating also it takes more pipe as it needs to pump more heat around for heating the house and DHW. heating around here takes more exchange area than straight cooling as the max temp difference is greater. IE 10 F outside to 72 inside in heat mode vs 100 F outside to 72 inside in cooling mode.

maybe 1/2 the pipe? 1000 feet call it?

[hardcore-fs]

This "oil" bathing is all well and good until.........

1. An Ecap explodes, spraying oil everywhere.
2. You have a "Chip-pan fire" during a critical failure.

HC

+1

Maybe not. Liquids are good at avoiding things to get on fire and at absorbing exploding energy.

Fixing it stays a mess tho.

SERIOUSLY Incorrect!!!!!!!!!!
1. you CANNOT compress a liquid, that is how hydraulics work!! ( & depth charges), any explosion inside a liquid transfers the energy to the edges of the liquid. (watch some old war movies where charges detonate inside a liquid)

2. With that reasoning, there would not be  "chip pan"  fires.

[flynn]

This "oil" bathing is all well and good until.........

1. An Ecap explodes, spraying oil everywhere.
2. You have a "Chip-pan fire" during a critical failure.

HC

+1

Maybe not. Liquids are good at avoiding things to get on fire and at absorbing exploding energy.

Fixing it stays a mess tho.

SERIOUSLY Incorrect!!!!!!!!!!
1. you CANNOT compress a liquid, that is how hydraulics work!! ( & depth charges), any explosion inside a liquid transfers the energy to the edges of the liquid. (watch some old war movies where charges detonate inside a liquid)

2. With that reasoning, there would not be  "chip pan"  fires.

Sorry to disagree

I used to be a scuba diver. We always reloaded our air tanks by putting them in big water containers to absorb the explosion energy if one of them had the bad idea of exploding. It also helped cooling them because compressing increase heat.

[Unacceptable]

Funny you should bring this idea up Farlack,me & my brother have a system we're working on right now to eliminate the coil from your A/C condensor by running sealed copper tubing/pipes into the ground to dissipate the heat.Just getting BTU #'s now with a window unit,but its looking promising for existing units up to 5 tons for residential systems.

The main benefit we're pushing is coils are the main thing to fail due to corrosion in seaside homes,in ground copper will elimate that failure point & last almost forever & remove that unsightly outdoor condensor from view 

so basically youre talking geothermal. hope youre aware of the length of pipe youll need, my 3 ton geothermal unit (A/C, heat, preheat for DHW) needs 2000 feet of pipe buried 5 feet down, although since its for heating also it takes more pipe as it needs to pump more heat around for heating the house and DHW. heating around here takes more exchange area than straight cooling as the max temp difference is greater. IE 10 F outside to 72 inside in heat mode vs 100 F outside to 72 inside in cooling mode.

maybe 1/2 the pipe? 1000 feet call it?

Yeah its geo thermal,but we are not designing for a heatpump app .....................yet  Just cooling for now.

Our pipes are copper (hard drawn) 1/2" outside with a 3/8" inside of it,water table is about 2-4 feet below ground level.I think it came out to about 20' per ton (2-10' sections) water jetted into the ground,a little angle seems to work better than straight down.Water temp was around 70F,not an ideal delta,but it does work very well.

[Lethn]

KABOOM!!!!

[crashoveride54902]

Submerging your gear in oil actually works pretty well.  But there's one downside.  Forget about ever salvaging or reselling that gear.  Once you dip that stuff in mineral oil, you will never get it all off.

Thanks, I never knew that. Do you guys think mineral water could cool a card as powerful as a 7970? Or will it overheat?

it wouldn't work, mineral water has minerals in it that would conduct elec. and fry the card...just like if you tryed regular water...even distilled water doesn't stay "clean" long enough...read that somewhere

One thing on oil rigs:

The one from puget systems ,.... If u install a GPU with a standard ati heatsink. It would pump the oil out of the aquarium. If the fan is strong enough u get a decent "oil" fountain.....:d

no fan is that strong, the oil is pretty thick...and really you wouldn't even need to run the fan more then 20% or even at all because the oil is suppose to do all the cooling, not the fan...

[ssateneth]

Also a great way to turn the resale value of your hardware to $0. Nobody wants a nasty greasy piece of computer hardware, plus it will alienate 99% of people who would have been interested in a non-greasy hardware. They'll think it's ruined and short circuited.

[paraipan]

Also a great way to turn the resale value of your hardware to $0. Nobody wants a nasty greasy piece of computer hardware, plus it will alienate 99% of people who would have been interested in a non-greasy hardware. They'll think it's ruined and short circuited.

You have truly spoken for yourself.

Who would resell parts? I think entire rigs only, and I would surely buy if they're nicely made and still working.

[superresistant]

Also a great way to turn the resale value of your hardware to $0. Nobody wants a nasty greasy piece of computer hardware, plus it will alienate 99% of people who would have been interested in a non-greasy hardware. They'll think it's ruined and short circuited.

Some users of mineral oil said that the hardware (like GPU) can be cleaned with alcohol and a lot of patience.
You can't shot-circuit with something non-conductive.

[blasthash]

Just look at all of y'all complaining about the oil getting hot.

Even with the lower specific heat of oil compared to water, it will still not heat up very quickly. Then with its high viscosity, the heat, as long as it is medium differential (not very hot) will either a) sit where it is (not influence other components in the bath) or b) rise to the top of the bath.

Yes, cooling is needed. But at the rate of change of temperature of that bath, your pump only has to have the speed of about a hamster wheel, and only a bit more torque.

On a technicality, you could use pure (DI) water - it is a non-conductive substance - however, your board has to be 100% chemically sterile.

Another option is a Freon tank. Freon (R-12) was the coolant used in the immersion fountains for the Cray (1 I believe) supercomputer. Run that through a HX / radiator and you have grade A cooling, plus you don't have to worry about conductive impurities in your oil.

The other way is reverse radiation. That is, instead of pumping the HMW oil through a chiller (which requires a torque-capable pump), immerse a radiator in the oil. Instead of moving the hot into the cold, you're moving the cold into the hot. Same principle.

[superresistant]

Just look at all of y'all complaining about the oil getting hot.

Even with the lower specific heat of oil compared to water, it will still not heat up very quickly. Then with its high viscosity, the heat, as long as it is medium differential (not very hot) will either a) sit where it is (not influence other components in the bath) or b) rise to the top of the bath.

Yes, cooling is needed. But at the rate of change of temperature of that bath, your pump only has to have the speed of about a hamster wheel, and only a bit more torque.

On a technicality, you could use pure (DI) water - it is a non-conductive substance - however, your board has to be 100% chemically sterile.

Another option is a Freon tank. Freon (R-12) was the coolant used in the immersion fountains for the Cray (1 I believe) supercomputer. Run that through a HX / radiator and you have grade A cooling, plus you don't have to worry about conductive impurities in your oil.

The other way is reverse radiation. That is, instead of pumping the HMW oil through a chiller (which requires a torque-capable pump), immerse a radiator in the oil. Instead of moving the hot into the cold, you're moving the cold into the hot. Same principle.

haha brilliant

[blasthash]

Submerging your gear in oil actually works pretty well.  But there's one downside.  Forget about ever salvaging or reselling that gear.  Once you dip that stuff in mineral oil, you will never get it all off.

Thanks, I never knew that. Do you guys think mineral water could cool a card as powerful as a 7970? Or will it overheat?

it wouldn't work, mineral water has minerals in it that would conduct elec. and fry the card...just like if you tryed regular water...even distilled water doesn't stay "clean" long enough...read that somewhere

You could use distilled water, provided the water is well, pure.

The only problem you would have with normal water is the fact that the water would leech ion salts off the board. Mainly the flux nasties that didn't vape off during reflow and then other unwanted gunk on the board such as fingerprint oils and what not.

I do suppose one interaction that may be chemically significant is the water and Pb (lead) in the solder itself - the local electric current on the board might cause the two to interact in the presence of other solutions, forming some weak f***'d up version of a lead acid battery.

Seal up the solder joints themselves (hot glue!), clean the board so that it's holier than Jesus, and you should be fine - not that most of that work is worth it when you can pick up mineral oil for $2.50 a quart haha

[crescendo]

I enjoy the video, Although this is very nice technique, I like this.

[niko]

About cleaning the mess: the liquid from OP is chosen as it boils at ~34C. You just need to warm up the gear above this temperature, and it all evaporates.

[BR0KK]

About cleaning the mess: the liquid from OP is chosen as it boils at ~34C. You just need to warm up the gear above this temperature, and it all evaporates.

And how do you clean oil off your hardware?

Simply put it in a dishwasher (believe me it works)

[Trillium]

About cleaning the mess: the liquid from OP is chosen as it boils at ~34C. You just need to warm up the gear above this temperature, and it all evaporates.

And how do you clean oil off your hardware?

Simply put it in a dishwasher (believe me it works)

As I explained the other day in another thread: and yes I have tried it.


As for cleaning oil off the components then here is my advice:

1. hang the card up by a piece of wire or leaning upright so most of the oil drips out / off
2. rinse off most of the oil with huge quantities of cheap dish washing liquid + hot water. this should remove the majority of the oil. And no, washing many kinds of computer components in water while they are off does not usually damage them so long as you are 100% sure it is completely dry before you power it up. Again the problem of unsealed electrolytic caps remains the only exception and problem here...
3. get yourself a few liters of your typical hardware-store-grade mineral turpentine (usually just $3/L...) and then immerse the card in a bath of it. It will dissolve the remaining traces of oil. It will also help remove any water, because they are completely immiscible liquids and you will see they will separate nicely from one another in the bath. If the oil you chose to use was not effectively removed with the mineral turpentine, try methylated spirits / denatured alcohol which is also very cheap.
4. Remove the card, hang it up near a powerful fan to dry for a long time.
5. your card is now back  to normal, ready for new thermal interface material and heatsink / fan assembly re-installation.

I highly recommend low viscosity oils.

Submerging your gear in oil actually works pretty well.  But there's one downside.  Forget about ever salvaging or reselling that gear.  Once you dip that stuff in mineral oil, you will never get it all off.

Thanks, I never knew that. Do you guys think mineral water could cool a card as powerful as a 7970? Or will it overheat?

it wouldn't work, mineral water has minerals in it that would conduct elec. and fry the card...just like if you tryed regular water...even distilled water doesn't stay "clean" long enough...read that somewhere

You could use distilled water, provided the water is well, pure.

The only problem you would have with normal water is the fact that the water would leech ion salts off the board. Mainly the flux nasties that didn't vape off during reflow and then other unwanted gunk on the board such as fingerprint oils and what not.

I do suppose one interaction that may be chemically significant is the water and Pb (lead) in the solder itself - the local electric current on the board might cause the two to interact in the presence of other solutions, forming some weak f***'d up version of a lead acid battery.

Seal up the solder joints themselves (hot glue!), clean the board so that it's holier than Jesus, and you should be fine - not that most of that work is worth it when you can pick up mineral oil for $2.50 a quart haha

Deionized water is quite corrosive. It will impress you with its ability to drag ions out of whatever it is contact with! For example if you put DI water into a stainless steel vessel, leave it a week and you'll come back to a unsighly mess! As for if it could do this with lead, I am not sure. It will do it with copper before it has a chance, just ask anyone who's put DI water into a watercooling rig with copper or brass components in the system. Even if you have it in contact with the atmosphere, it will attain an equilibrium with CO2 from the air, and hey presto, it's now no longer deionised and its conductive again. The only way to get a system to function with DI water @ 18.2 MOhm is to buy ion exchange/absorber packs that chemistry analytical labs use to get 18.2 MOhm DI from RO water. These are also used in some industrial settings. The downside? They cost a fortune. The ones I've seen in a few labs cost about $400 per canister which will last a few months in their indented mode of operation. Do yourself a favor and use something other than DI water, it will just drain your money and kill your equipment if you stuff it up the slightest bit.


As for the Novec 7000 and boiling liquid video, that is a good idea, I had considered doing that with diethyl ether or perhaps butane in a lightly pressurized container made of polycarbonate. I concluded the butane would be safer because for the same volume of ether it has much lower overall energy stored in the event of a fire. Their temps of 60+ deg C without a heatsink and under light / short term loads shows that a small heatsink would've been a good idea, there is still just too much heat to be removed from a small area (~1 inch square).

As for scaling it up, well they are going to have interesting times. The entire system in their prototype relies on extremely inefficient peltier cooling to remove most of the heat from the system. The other thing to consider is that assuming it was air tight (which would be a good idea given the cost of the liquid, you don't want it flying away...) then you would have an increase in pressure if a lot boiled in a short time, eg going from low to high load on the CPU. Increasing pressure in a sealed vessel increases the boiling point of vapourisable liquids. So if the peltier / heat pump was overwhelmed then you could find yourself in a situation where the liquid can no longer boil, and depending on the thermal conductivity of the bulk liquid itself you may find your components suddenly at catastrophic temps. Maintaining a pressurisable vessel with lots of cables going in an out and a heat exchanger as well is also non-trivial! And by pressurisable I don't mean 'held together with duct tape'!

In any case a look at the Novec products MSDS shows they are fluoro-ethers which are $$$ I would be willing to bet a few liters would set you back a couple of hundred dollars. And as for their 'non-toxic' statement in the video well I don't agree that means they are safe, since most organofluorine substances will overcome you just as well as their organic hydrocarbon precursor will (if inhaled). I would personally stick to oil for now because its simple, cheap, and works.