My mind has been wondering, as it often does, on the topic of external liquid cooling solutions. I realize having everything mounted inside the case is generally more desirable from a mobility and convenience standpoint. but i also like the idea of not being limited by case size. plus, having an external enclosure for pumps/radiators allows for a more accessible/organized case IMO. so my idea was to take one of those portable oil-filled radiator style heaters, gut the electronics and drain the oil, install a pump and fill it with water and plumb that into a CPU cooling loop. from what ive seen most of these units are able to dissipate up to 1500watts while still maintaining a safe-to-touch surface temperature, so id think theyd be able to dissipate 400-500 watts at the lower temps computer enthusiasts like to see. they have the advantage of being already designed for passive cooling, and since they are designed to sit in your home anyways, they have a certain style aspect to them (more so than a car radiator has, for example). as an added bonus, the thermostat display could be replaced with a more useful display such as pump speed or water temp. im not a mathematician, so i dont know how to calculate the heat capacity of one of these radiators, but what do you guys think? feasible or am i way off base?
Sounds like a pretty good idea. You might be better off keeping your pump inside the PC case though, that way you don't have to run an external power cable. Read up on what the Zalman Reserator did wrong and then don't do it!
Two things - one, it'll be extremely difficult to drain the oil completely. Two the temperature difference will likely mean the dissipation capacity is severely reduced.
getting the flow right might be an issue.... you will not be using convective currents so may not get the coldest liquid.... a heck of a lot of it is likely to remain static so you would only get the cooling of the route the liquid takes....
i work at an auto repair shop, we have chemicals for removing oil it would be interesting to find out at what temperatures these things run. i know the surface temp of the radiator only gives you an idea of the actual oil temp inside the unit, but if the outside is cool enough to touch, how hot could the oil really be? might have to ebay a cheap one to find out... the flow is something i hadnt considered. i assumed they used an internal pump to circulate the oil, but do they actually rely solely on convection? if thats the case, one might be able to use an old steam radiator with dedicated inlet and outlet. since they are designed for pressurized fluid to flow through them, flow should be pretty evenly distributed. and they can still be had for fairly cheap on ebay. although most of them are cast iron so any tiny bit of portability you had with the oil radiator goes out the window
This idea won't work for the simple reason they are designed to output heat through their surface area . If you look at them , although there is a large surface area it's quite aerodynamic, there is no way it'll be able to trap cold air . Look at the way ordinary rads are designed and you'll see that they are loaded with cooling fins for the air to pass over and so dissipate heat. Sorry to shoot you're idea down mate, but unless you can change the laws of psychics the best thing to do, is to use that big old rad as a heater. If you do want to get into some serious cooling have a look at motor cycle rads and oil coolers these are easily adapted to suit PC cooling needs.
they are aerodynamic because they are designed for passive operation. you dont want to trap air when your running passive. bumps and ridges and anything that sticks out horizontal disrupts convection and slows it down, making your radiator transfer heat slower. by "ordinary rads" i assume you mean standard CPU heatsinks and radiators? those are designed for active cooling. the rules change a bit in that case. however, you might be right. there may just not be enough surface area to get the job done without using a radiator the length of an entire wall. still, the idea of external radiators intrigues me...ill be keeping my eye out for other potential candidates
Maybe not Oil Radiators (guess they rely purely on convection inside) but normal water radiators are built so that the hot water goes into the top, and the cooled water is drained at the bottom. Even when both connections are at the bottom (or top, as with some bathroom radiators) there's a link inside to ensure this. Main problem is probably that they are meant to dissipate heat rather slowly and have a rather low difference between hot and cold. Check your heating system, mine has a "hot" water temperature of about 60°C and I believe the return is about 30-40°C. Don't know how this quota is in watercooled systems though.
ahh theres something i didnt know. i thought central liquid heating systems ran much hotter than that. that sounds pretty close to where youd want a computer running. thanks for the info xir. i wish i had the money to liquid cool a system and try this out. i think thats the only way to be certain.
I think you mean 60 degrees Fahrenheit, 60C is around 170F that would mean you're pumping steam through the system which is going to turn it into a very big pressure cooker . 60C on oil is still very hot, I wouldn't like to put my hand on a rad carrying that sort of heat! That said if you laid it flat you could do you're breakfast on it.
60 celsius is 140 fahrenheit, which you could touch for a moment without burning yourself. water boils at 212 fahrenheit, so even if it was 170F you wouldnt have steam, particularly if the system was pressurized. plus 60 fahrenheit wouldnt do you much good for home heating since thats a pretty chilly temp to keep your home
