Discussion in 'Extreme and Water Cooling' started by gigs, Aug 1, 2009.
Ahh... So that's why the pix were "uploaded by Agg". I was confused.
I'm assuming no such thing. With a fixed heat load - any heat load - and zero thermal conductance, ie perfect insulation, the final temperature would be infinite. This would also be true with infinite heat load, and any finite thermal conductance. Neither of these situations exist in the real world, and are only examples used to illustrate concepts - though in your case, this illustration appears to have failed. The system will have some unknown thermal conductance, and with constant input heat and constant environmental conditions, it will approach some unknown equilibrium temperature. I'm guessing that temperature will be quite high, and that it will approach it quite quickly, but these are just guesses. Since no one has modelled it or tried it empirically yet, guessing is the same thing that the OP and everyone else in this thread is doing.
Hey Gigs i'm working for some concrete boffins at monash uni this week any questions you want me to ask shoot us a PM and i'll try them, there very freindly guys and allway keen to help.
I would have to agree that the concrete wouldn't help to cool the water. You're relying on the stretch of copper pipe to eventually cool the water before it gets back to the computer. I just think that this sort of thing would have been commercialised if it worked. For example, car radiators can use copper pipes for the water (cores) but they are attached to fins to increase the surface area so you can effectively cool it by air.
Concrete is a 'relative' insulator which means it's difficult to transfer energy (whether it be heat or electricity) across it. When something conducts, it means that the atoms in the object vibrate because it's transferring electrons to and from other atoms.
You want something attached to the copper pipe so it can freely radiate that heat. Otherwise I don't know how some people have hot showers with copper pipes under concrete running from the heater at one side of the house to the bathroom at the other.
+ As others have mentioned, if there is no insulation between the copper and the concrete you will get corrosion. Concrete is very alkaline which is why you don't touch the stuff, it will burn your skin.
Copper and concrete
Great effort for cooling but if your copper is unshielded your gona soon have problems as the two react to each other. If you get a second try there is plumbing pipe called Pex that will serve you better. Sorry to be the bearer.
depends on the mix of concrete
great work btw, Will be interesting to see the results.
Small fact you technical overthinkers are have totally missed ...THIS HAS BEEN DONE BEFORE!!! Placing lengths of copper pipe ONTOP of concrete slab works very well ...have a search on OCAU i have linked to similar setups before...FFS at least come up with something else that hasnt been bought up Al Ready in this thread it gets fucking annoying reading the same criticisms page after page guyz...You think GiGs sat up one day and thought "ill bury some pipe in my slab just for shits and giggles" he has seen others do VERY similar, noted the outstanding results and is doing it himself (kudos GiGs ;-) )...The ONLY thing that could jeopardise the cooling ability is the length of pipe in the concrete V's number of pcs on the loop.. as i bought up before.
"Stretch of copper pipe"- if concrete was such a good insulator then no heat would escape - so even your incorrect hypothesis doesn't make sense.
Yes the thermal flow rate from copper->Concrete is going to much lower than copper->air - but thats why he is putting a much longer run of copper and exposing a larger area than in air setups.
All these people claiming concrete as an insulator, obviously have never lived on a slab house oreven been barefoot in a garden shed - they are always colder than air temp - but still slightly warmer than below surface ground temps. Therefor they are being warmed by the air but more heat is being dissipated, even at the end of the day (unless its been exposed to the sun).
Geez guys, not like the figures for the thermal properties of concrete are that hard to find: http://people.bath.ac.uk/absmaw/BEnv1/properties.pdf,
1.14 W/mK for cast concrete flooring. Copper pipe on the other hand is around 400W/mK. I guess that technically makes concrete a pretty good thermal insulator. Being a good insulator also means it can store a lot of heat, large concrete blocks were used as heat banks in off peak room heaters in school rooms when I was at school (decades ago).
Not saying the idea is a good one or a bad one. Certainly some of the speculation about pipes cracking or corroding are a bit out there. Corrosion shouldn't be a problem unless the slab is damp. If it's damp there's more serious issues for the house. The pipes will not crack from thermal stress either, copper hot water pipes have lagging for insulation, even when it's not buried in concrete it's lagged.
Here's a thought though. You paid to create that heat in your electricty bill, why waste it heating a slab of concrete. An insulated header tank before the hot water heater in the house and a heat exchanger means that heat is doing something usefull.
I didn't say no heat would escape. We've already established anything can conduct given enough voltage or energy (it takes a while for concrete to warm in the sun).
But I think I missed the point. Firstly, is it unique? f'n oath! Secondly, is it better than running the pipe in ambient temp? Probably. At least, I am not building a house/shed soon to compare it.. Is it the best cooling ever? Again, missing the point
I'm not here to blast/criticise anyone. I admire these types of projects as it makes you think.
Kudos to GiGs. I respect the passion which is what keeps OCAU alive!
I'd say heat will escape, mainly because of moisture surrounding the pipe under the slab. Worst case scenario: he has an extra x litres of coolant pumped, so there will be some temp difference (the equivalent of just having a large reservoir). If it sucks, he can just saw off the copper pipes and continue with his internal WC setup.
It's people like Op that make this place interesting, trying new ideas and maintaining interest in extreme cooling.
Are you single as my wife would kill me if I did anything like that. I could put mine under the house but I'd have to put a few holes in the floor and she would spit it. Better just to keep the pease.
It's been waaaaaay too many years since I did heat transfer at uni, but some back-of-the-envelope calculations suggest that with a 150w heat load, you'd be looking at a temp rise from ground temp (= temp of the concrete slab some distance away) to the pipe of about 20ºC, neglecting temp differences along the length of the pipe (I used a length of 3m instead of 6m to balance that a bit).
So if you've got a ground temp of 15ºC, then you might be looking at a pipe wall temp of about 35ºC. Add a bit for conduction through the pipe & into the water, and you're looking at 40ºC for your water temp, give or take.
As mentioned above, I haven't taken into account the temperature gradient along the pipe, but used a length of 3m instead of 6m to offset that. I've also assumed the ground beneath the slab will conduct heat away at the same rate as concrete.
But it's late, and it's 17 years since I last opened that particular textbook (and probably 15 since the last time I had to do a heat transfer calc at work), so I cbf working out the fine details...
Thought it'd give you an idea, anyway.
I thought somebody would have mentioned this by now, but as far as I can tell.. No?
Had a look at that before I commented. It's really a different concept besides the fact that it's in the ground. Soil is a 'better' conductor, especially when moist as water is a good conductor.
Mmm, yeah looking through it, it's not the article I thought it was lol
I remember reading a buildlog not logn ago where somebody layed a bunch of pipes in the concrete when he did his garage.. Yes, I said 'a bunch' lol
Don't think he had finished it yet when I was reading the log though
You guys keep linking to this copper.org site for your arguments. You dont think a shill for the copper industry might be a bit biased? If I linked to a PEX manufacturer talking about how copper sucks and PEX is so much better would you believe that?
Hey you are published on Hackaday.com Congrats
Hey guys thanks for all the comments and lively debate that has been going on in this thread
I just wanted to set some things straight. All of these items have been dealt with in detail through this thread if you care to take the time to read it but to recap.
1. Concrete as an insulator
and will not dissipate heat for the application that I want. Please refer to this:
Thermal Conductivity - k - (W/mK)
Concrete, light 0.42
From here you can see that the thermal characteristics of water and concrete are very similar. Real world tests will answer this one very soon.
2. PEX vs Copper
Agreed PEX will last a lot longer in concrete than copper. PEX does not have the Thermal characteristics that I want ie it will not work as well as copper to transfer heat into the concrete. This is why I have not used it. I am not interested in this copper lasting decades. I am only after about ~5 years of service.
3. Corrosion of raw copper in concrete
Please take the time to read the information here:
Absolute worst case scenario. The copper may get corroded after 5-10 years and expose the coolant to the concrete. SO WHAT do you think I will care by then. Who knows what cooling solutions will be available for computers by then. I will just cut the copper off flush with the floor and move on.
4. Concrete slab cracking due to thermal expansion
Hot water pumped through copper imbedded in concrete and Heating concrete floors hydraulically isn't a new technology. In the 1930s, architect Frank Lloyd Wright piped hot water through the concrete floors of many of his buildings. No reports of concrete cracking only the copper pipe eventually failing after many decades. These days wires are laid out in the concrete and shorted to heat a slab. On average it takes 6kw + to heat a house slab just a few degrees. Ask anyone that has slab heating and most of the time they turn it off because it costs so much money to run. I am trying to absorb just 150 watts of heat into 2-3 meters of concrete, do the math guys. The water coming off the cpu block will be at the most 25deg. Concrete is elastic and expands and contracts with heat mitigating wear on the copper.
Maybe wait for me to get a few tests done before saying how much this is epic fail? If it doesn’t work I’m going to look like an right dick to the 70,000+ people that have visited this thread already. Even if it does or does not work I hope I have inspired some of you to try some crazy things too.
Trying to get hold of a pump this afternoon so I can do some experiments this weekend....
Good on you Gigs. People can calculate and argue all they want but in the end you just have to suck it and see. Even if it does fail at least you can say that you have tried it. I dont know anyone else that can say they tried to cool their pc with concrete.