Different approach to Water Cooling your PSU...

Phant0m51 · 08-04-2003, 06:54 PM

Ok, what I want to do is seal my PSU completely, and fill it with Mineral Oil. I will then run some copper tubing inside the PSU and have water run through it to pick up the heat from the mineral oil.

Anybody see any problems with this (Aside from the extreme electrical shock that I will feel if the copper wire touches anything inside)?

BladeRunner · 08-05-2003, 07:40 AM

Appart from the fact sealing it all to ensure the oil doesn't leak out will be a task in itself, I'm not sure it will work. Oil has a pretty poor heat transfer ability so you will get localised component heat up, (asa the oil will act like an insulator). If the oil were circulated and cooled it may work ok, but as it will be stagnant just relying on transfer through the oil to the tubes, I doubt it would work.

Phant0m51 · 08-05-2003, 08:14 AM

Thanks for the reply. I was thinking...What if I could get the fan to rotate while it was sealed? I know it wouldn't rotate really fast, but if it did rotate it would move the oil around. I'm going to buy a little mineral oil and put a fan in it and see if it spins.

Thanks again for the reply,
Jonathan

MadHacker · 08-05-2003, 10:19 AM

if you put you copper water cooled pipes at the top of the power supply. mebe that will do it.. as the componets get hot. the minera oil will get hot.. it will rise to the top. then get cooled by the tubing then go back down... (i forget what this effect is called)
but that should be all you need.... mineral oil isn't that thick so it should work is my guess...

NOTE: only a guess.

Phant0m51 · 08-05-2003, 10:59 AM

I'm going to do some preliminary testing. I have an old PSU that I haven't used in a few months (it's only a 120 Watt PSU for a Pentium I computer). So I'm going to see if I can keep that cold with mineral oil before I attempt anything with a more expensive one.

BladeRunner · 08-05-2003, 11:03 AM

Sorry to keep sounding negative, but I just can't see it working, I'd love to be proved wrong but I very much doubt you'll get a fan to start up in oil, and if it did there would be so much drag loading it would most likely burn out.

Oil isn't like water, In my work, (motorsport), an engineer had some oil heaters made. They were basically straight versions of a standard kettle element that were inserted into the oil. The idea was they would heat the engine oil before the engines are run. The theory is fine and it would work if the heating element was much longer and coiled around the oil tank but the problem is the oil in contact with the hot area can't remove it's heat quickly enough to the rest of the oil, so it started to boil around the element. Also due to it's viscosity compared to water, (we were using a thin synthetic oil), it can't get away from the hot part either, (wont mix as well as water).

What you are suggesting is probably possible but would require a lot of development and more than just a tube running in the oil at the top, so I doubt it would work well & I'd guess the hot parts like the fets & some coils would simply over heat the oil around them, in turn over heating themselves.

Along with the sealing issues mentioned before, there are other factors to consider, like making sure the oil is inert to high voltages, just about everything will be conductive to some degree if the voltage is high enough. Another factor will be if the mineral oil is hydroscopic, (like some brake fluids)

Boli · 08-05-2003, 04:50 PM

OK I have a couple of points to make:

Firstly most PSU are full of slits and air gaps in order to allow air to circulate... this will be a task and a half to seal plus the fact it opens up and you'll have to deal with the joint between the two halves.

Secondly if you did dispurse the heat around the PSU you have to remember that parts that do not normally heat up will start to rise to a higher temperature than they were desgned to handle. In electrical terms becoming more resistive. This may lead to fuctuations in current.

I doubt you'll get better cooling and quieter than just modding your PSU with a single 5v 120mm fan mouted onto the top.

~ Boli

murray13 · 08-05-2003, 05:43 PM

You'll have to get the mineral oil moving. You wouldn't try cooling a CPU with water without the water flowing somehow. Well using mineral oil is the same in that respect. Things to take note of it does have a higher viscosity than water. It's thermal conductivity is about half that of water. The best thing is that it is electrically insulating. Maybe you could consider running a seperate pump and rad.

BladeRunner: Tho only thing your oil heater was missing is that the oil must be flowing for it to work. And it will work. I've seen them work.

redleader · 08-05-2003, 05:56 PM

Lookup the viscosity + heat capacity of your oil verses air. That should give you an idea of how well it will do.

My guess is you'll need something to stir the oil though.

08-04-2003, 06:54 PM	#1
Phant0m51 Cooling Savant Join Date: May 2003 Location: Utah Posts: 160	Different approach to Water Cooling your PSU... Ok, what I want to do is seal my PSU completely, and fill it with Mineral Oil. I will then run some copper tubing inside the PSU and have water run through it to pick up the heat from the mineral oil. Anybody see any problems with this (Aside from the extreme electrical shock that I will feel if the copper wire touches anything inside)?

08-05-2003, 07:40 AM	#2
BladeRunner Cooling Savant Join Date: Jul 2001 Location: Chesterfield Uk Posts: 459	Appart from the fact sealing it all to ensure the oil doesn't leak out will be a task in itself, I'm not sure it will work. Oil has a pretty poor heat transfer ability so you will get localised component heat up, (asa the oil will act like an insulator). If the oil were circulated and cooled it may work ok, but as it will be stagnant just relying on transfer through the oil to the tubes, I doubt it would work. __________________ Zero Fan Zone

08-05-2003, 10:19 AM	#4
MadHacker Cooling Savant Join Date: May 2003 Location: Okotoks, A.B. Canada Posts: 726	if you put you copper water cooled pipes at the top of the power supply. mebe that will do it.. as the componets get hot. the minera oil will get hot.. it will rise to the top. then get cooled by the tubing then go back down... (i forget what this effect is called) but that should be all you need.... mineral oil isn't that thick so it should work is my guess... NOTE: only a guess. __________________ "Great spirits have always encountered violent opposition from mediocre minds" - (Einstein)

08-05-2003, 11:03 AM	#6
BladeRunner Cooling Savant Join Date: Jul 2001 Location: Chesterfield Uk Posts: 459	Sorry to keep sounding negative, but I just can't see it working, I'd love to be proved wrong but I very much doubt you'll get a fan to start up in oil, and if it did there would be so much drag loading it would most likely burn out. Oil isn't like water, In my work, (motorsport), an engineer had some oil heaters made. They were basically straight versions of a standard kettle element that were inserted into the oil. The idea was they would heat the engine oil before the engines are run. The theory is fine and it would work if the heating element was much longer and coiled around the oil tank but the problem is the oil in contact with the hot area can't remove it's heat quickly enough to the rest of the oil, so it started to boil around the element. Also due to it's viscosity compared to water, (we were using a thin synthetic oil), it can't get away from the hot part either, (wont mix as well as water). What you are suggesting is probably possible but would require a lot of development and more than just a tube running in the oil at the top, so I doubt it would work well & I'd guess the hot parts like the fets & some coils would simply over heat the oil around them, in turn over heating themselves. Along with the sealing issues mentioned before, there are other factors to consider, like making sure the oil is inert to high voltages, just about everything will be conductive to some degree if the voltage is high enough. Another factor will be if the mineral oil is hydroscopic, (like some brake fluids) __________________ Zero Fan Zone

08-05-2003, 08:14 AM	#3
Phant0m51 Cooling Savant Join Date: May 2003 Location: Utah Posts: 160	Thanks for the reply. I was thinking...What if I could get the fan to rotate while it was sealed? I know it wouldn't rotate really fast, but if it did rotate it would move the oil around. I'm going to buy a little mineral oil and put a fan in it and see if it spins. Thanks again for the reply, Jonathan

08-05-2003, 10:59 AM	#5
Phant0m51 Cooling Savant Join Date: May 2003 Location: Utah Posts: 160	I'm going to do some preliminary testing. I have an old PSU that I haven't used in a few months (it's only a 120 Watt PSU for a Pentium I computer). So I'm going to see if I can keep that cold with mineral oil before I attempt anything with a more expensive one.

08-05-2003, 04:50 PM	#7
Boli Cooling Savant Join Date: Apr 2003 Location: Blackburn / Dundee Posts: 451	OK I have a couple of points to make: Firstly most PSU are full of slits and air gaps in order to allow air to circulate... this will be a task and a half to seal plus the fact it opens up and you'll have to deal with the joint between the two halves. Secondly if you did dispurse the heat around the PSU you have to remember that parts that do not normally heat up will start to rise to a higher temperature than they were desgned to handle. In electrical terms becoming more resistive. This may lead to fuctuations in current. I doubt you'll get better cooling and quieter than just modding your PSU with a single 5v 120mm fan mouted onto the top. ~ Boli

08-05-2003, 05:43 PM	#8
murray13 Cooling Savant Join Date: Sep 2002 Location: Omaha, NE USA Posts: 216	You'll have to get the mineral oil moving. You wouldn't try cooling a CPU with water without the water flowing somehow. Well using mineral oil is the same in that respect. Things to take note of it does have a higher viscosity than water. It's thermal conductivity is about half that of water. The best thing is that it is electrically insulating. Maybe you could consider running a seperate pump and rad. BladeRunner: Tho only thing your oil heater was missing is that the oil must be flowing for it to work. And it will work. I've seen them work.

08-05-2003, 05:56 PM	#9
redleader Thermophile Join Date: Jun 2001 Location: The deserts of Tucson, Az Posts: 1,264	Lookup the viscosity + heat capacity of your oil verses air. That should give you an idea of how well it will do. My guess is you'll need something to stir the oil though.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)