House Radiator Cooling

Frozennova

Frozennova

Frozennova

Man of Honour
Joined
13 Nov 2009
Posts
11,643
Location
Northampton
Out of curiosity would it be possible to use a normal house radiator as a radiator for watercoolig.

Just a quick bit of research and a GTX295 has a TDP of 289W and an i7 920@4ghz roughly 285W and the radiator of 900mmx600mm can disperse 932W according to the Wickes website.

Would the radiator be able to disperse enough heat passively ( Say 600W allowing for discrepencies) to cool the processor and CPU effectively providing I had a large enough pump?
 
I would say its possible but since its gonna be passive, its gonna depend lagrely on ambient temperature
 
He could always use a new one. Another idea would be to use the cooling mechanism from a car. I am in posession of a cooler for a mazda MX-5 with volt modded fans and it runs totally silent.
 
neo-omega said:
He could always use a new one. Another idea would be to use the cooling mechanism from a car. I am in posession of a cooler for a mazda MX-5 with volt modded fans and it runs totally silent.
Click to expand...

AS said, a new one is perfectly possible and they are pretty damn cheap, because unlike car/computer rad's they aren't designed for massive surface area via small copper tubing and fins in a very small space, 3 things that increase price massively, copper, miniturising and intricate construction, they are massive basic and made of, actually not sure, stainless steel, Pumps not a huge issue either, it doesn't need enough pressure to fill/push water around the whole rad, as the heat will spread fine, just getting the water into and out of the rad, there are probably designs with an intake at the top of the rad so hook pump to bottom valve, through comp and dump the water in a not completely full radiator so its not working against pressure there.

But heres your issue, rad's are rated very specifically and normally with a delta T of 56, IE you'd get 900W dissipation when the internal liquid is 56C higher than room temp. Considering the water will likely be 5C or less above room temp, you won't get even remotely close to that dissipation figure, because thermal transfer increases in efficiency HUGELY the higher the temperature diff, with a very small difference you're talking about being way down an exponential scale and you'd be looking at a probably 50-100W dissipation, but with a huge resevoir that would take ages to heat up and as it heated up, you'd get increased efficiency. Without active fan cooling though you'd still not get a fantastic cooling system and its not worth the effort.

You can build a more discrete radbox if you want to go DIY on your cooling's ass, for very cheap, ultra efficiency and very quiet, but some mdf/chipboard type stuff, knock up a design for a box with a couple fans, a couple rads, a exhaust somewhere useful like a chimney or window.

Theres plenty of weird ways to cool a system very well, and wall mounted normal central heating rads isn't a great one.
 
WJA96 does this - he has 8 or 10 pc all linked to a single domestic rad in his garage.
I think he uses a heatexchanger to keep the internal loops clean.
Domestic rad are steel - not stainless. but chrome, stainless and copper ones are available.

The thing with passive watercooling is the water (and components) are at a higher temps (higher delta) - simular to stock cooling.
You can get very low temps with passive but to acheive this you need to massively over do the rads (see mine and jokester build logs)
Also domestic rad especially the one I use have a higher RADIENT heat output than watercooling rads (which are 98% convective) radient heat output/loss is not affect by ambient temp.

Have a look at skirting heaters and radient panels - they usually use small copper tube (15 or 22mm OD)
and are much easier to link to a pc loop that domestic rads or car parts.
Avoid the one with internal aluminum parts if possable, unless you have a Zalman Res and or gold/nickel plated waterblocks
These are external systems so I would recoment you get some quick connectors - makes changing hardware a breeze.

Edit: Having said all that a very small amount of air flow makes a massive differance.
I use a PA120.3 semi-passively now (with just one case fan) - the PA cools my cpu/NB/gpu without breaking a sweat.
PSU/mosfet/ram/HHD are all passively air cooled.
But I have a [email protected] and 3780 so no where near your thermal loading (temps peak at 20c above ambient)
 
Last edited:
shadowscotland said:
radient heat output/loss is not affect by ambient temp.
Click to expand...

Untrue I'm afraid, though possibly a valid approximation. Heat energy radiated per unit surface area is proportional to the forth power of temperature (stefan-boltzman x emissivity, a value from 0 to1). However this just gives gross output. The radiator is surrounded by walls which also radiate at a rate proportional to T^4. Net radiation exchange between grey bodies is

10rqro5.jpg


Can't find a page on wikipedia so I've robbed a piece from my lecture notes. Q is heat flux. A is surface area of the radiator, T1 radiator temp, T2 wall temp, sigma is stefan's constant and the backwards 3 (Greek fails me/I fail Greek) is a value from 0 to 1. Temperatures in Kelvin.

As a vaguely useful aside, radiators emit the vast majority of heat in the IR spectrum so painting them black makes no difference to radiated heat.
 
Ok radient rads are significatly less affected by ambient temp that convective rads. I stand corrected :p

Also as an aside.
modern domestic rad are about 20%/80% radient/convective (this is why emulsion paint makes little differance)
Black and white colour paint makes no differance! - it's visable colour make no differance yes.
But the paints (or material surfaces) emissivity, makes a massive differance.

Black gloss is better that white gloss - matt emulsion is better again
but white 'radiator' paint or black bbq paint is best - emissivity above .92.
Shiny metal is bad, emissivity well below .1 - chrome towel rads for example are almost 100 convective

All this relates to radient rads only - watercooling and domestic rads are convective (why a fan increase performance)
Fans have (next too) no effect on radient panels, and that's why I was going to use them.
 
Last edited:
.92 is pretty good, cheers for that. Learning lots of theory but not many examples.

I wonder if pneumonic would be interested in this thread. A side effect of peltier heat exchangers is high temperatures in the hot loop, approx 65 centigrade. This may make radiant heat exchangers a better idea than forced convection. That said, passive water cooling is normally about quiet, and his psu is not going to be quiet.
 
JonJ678 said:
.92 is pretty good, cheers for that. Learning lots of theory but not many examples.

I wonder if pneumonic would be interested in this thread. A side effect of peltier heat exchangers is high temperatures in the hot loop, approx 65 centigrade. This may make radiant heat exchangers a better idea than forced convection. That said, passive water cooling is normally about quiet, and his psu is not going to be quiet.
Click to expand...

meh... "quiet" is a state of mind :p

Interesting idea thou.. I had toyed with the possibility of a Rad Box.. I may well still need to but was going to wait and see what the water temperatures done first.. I had hoped to keep the 'hot' loop temperatures at no higher than +15c over ambient so may well not be getting hot enough to be effective
 
Konneck-o-matic maxi/ultra plus (or ultra+) is the only radient pc watercooler available I beleive.
it's the step up from the all convective maxi/ultra - less weight, higher heat output

Load of detail on underfloor and radient heating sites.

Passive watercooling is quite small as it is, so very few examples.
It might even be just me and ultra+ owners (it's a bigger deal in germany)

But domestic heating is a massive multi million market - lots of innovation.

you need to search outside computing forums for info in high emissivity paint.
DIY solar collectors/IR photography/telecopes are a good start.

Live steam model railways forums are also usefull at times -
like getting distilled water from your tumble dryer / working with copper tube
 
Last edited:
Saw some pictures on HardOCP of someone who had fitted his PC to his swimming pool for cooling - very cool indeed. Would love to have seen an update on that project.
 
My friend who is a tiler said he'll put in under floor pc watercooled heating for me, when I convert a barn to an IT office.

I'll probably run that loop into a water tank, and then drop my rad from a closed system into that tank

So the water in the floor won't be going into the pc, just the rad will be cooled by it :)
 
better to do that option but to a WC cysten - cheaper and 'tank' is frequently topped up with cold water especially in a office.

Running your cooling in a 'heated' room is not ideal - if it's an internal loop you have no choice.
But external loops can in theory be anywere - ideally same level as pc (to keep pump small) and away from a freezing risk.
Window cills, bathrooms, garages, chimneys, unheated conservatories all are good.
Ponds, swimming pools, inside underfloor heating, lofts all have problems but can work.
 
So theoretically with the radiator mounted somewhere with a nice low ambient temp tihs would work.

I hadnt thought this would be taken so seriously however again you guys amaze me again.

Thanks
 
As temi says, it would stay at ambient temperature, if the room temp is below how much heat output you processor is putting out it will stay at that temp. I wouldnt use a convector radiator beause they tend to rust after some time without any ihibitor

EG. We put furnox in at work in new systems stops them from rusting, (oxidizing) The thing is the water will eventually turn brown and mucky.

best thing to use is a copper raditor from a heat exchanger.... anythign ranging from car raditors, best thing to use is a heat matrix from a transit van they put out 2kw of heat, that should deffiently cool 150w CPU usage. :)

Stefan.
 
You must log in or register to reply here.
Back
Top Bottom