articles
FORUMS!
links
contact

PC Database

Folding Team

SETI@HOME Team
RC5 Team
Genome Team

OCAU chat

Ok, firstly let me say WOW! I never anticipated such a huge response to this article. I've had over 160 emails and over 6400 visitors in the first 24 hours, I'm stunned. :) Sorry I can't reply to everyone individually - THANKS to everyone who sent info and suggestions, support and expressions of amazement. :) A lot of the responses were similar so I will address the major points raised:

• Use a fridge, with the reservoir inside, and run tubing from that to the PC
  
  Hmm, good idea, but a little expensive to leave a fridge running just for the PC. Also, not very portable.
  
  
• the clips for mounting the unit seem to be the arms from bulldog clips, used for holding lots of sheets of paper together
  
  Ahh! So they are. It's obvious when 35 people point it out to you. :)
  
  
• you can get those clips from newsagents, stationery suppliers etc
  
  Yeah ok - that wasn't what I meant.. I meant, does anyone know where to get the water-cooling blocks, not the clips. :)
  
  
• add antifreeze to the fluid
• add alcohol to the fluid to stop bacterial growth
• use "Redline Water Wetter" to make the water mix more thermally efficient
  
  Definitely considering an antifreeze mixture for the closed system. I have other plans for my alcohol. :) I think the antifreeze will have the same effects.. that Water Wetter stuff sounds good, too, anyone got any idea where to get it in Australia? (what kind of stores sell it in the US?)
  
  
• Use pure ethyl alcohol, it won't freeze even if you get it REALLY cold, it's non-conductive and very thin so good flow rates
  
  ..it's also highly flammable, produces explosive vapours and is extremely poisonous..
  
  
• Use liquid nitrogen
  
  I don't want to look like I'm saying "NO!" to all of these ideas, but.. the thermal stress would probably crack the ceramic casing of the processor. :) There's "extreme", and then theres "extreme", if you know what I mean. I'll stick with a water-based solution for now.
  
  
• Use an automobile transmission fluid radiator - small and efficient
• Use a motorcycle radiator
  
  Yep, from various people's descriptions a small oil radiator sounds like exactly what I'm looking for. I could mount it, the pump and the reservoir in a box and have fans venting through the radiator and out the top of the box. Then I could have water pipes and power for the peltier going from that box to the main machine. A definite plan is forming!
  
  
• Use an automobile fuel pump, small, in-line and very quiet
• The fishtank pump is a bad idea - it will heat the water
  
  Hmm. I agree that an inline pump may be better.. I will work around this fishtank pump for now, until I determine it to be the limiting factor. For the next big project I will definitely consider a fuel pump. If this fishtank pump is too loud I will look into one as well. :)
  
  
• Just leave the tap running into the bucket, that way it will always be cold
• Plug one end into the tap, through the heatsink and back out to the sink, turn the tap on and have cold water indefinitely
  
  Guys!!! A few people suggested this, jeez, don't you have water wastage laws where you live? :) This is Australia, we run out of water during summer - if I got in the habit of leaving a tap running all day the EPA (environmental protection authority) would come kicking in my door in no time. Plus it's expensive (ok, not _that_ expensive).. plus it's not very portable (Go to a LAN party, plug into the power, the LAN, the nearest tap..) .. :)
  
  
• Use a larger, flatter reservoir like a baking dish, you will lose heat due to evaporation
  
  Hmm, not portable, not easily built into a self-contained unit, probably wouldn't lose heat quick enough, my girlfriend's dog would drink the antifreeze and I'd get in trouble. :)
  
  
• The larger the reservoir the longer it will take to heat up
• Use a 10-gallon fishtank as the reservoir
  
  Yeah, but this is a workaround, not really a solution. I guess at some point, the reservoir would be large enough that the amount of heat it lost due to evaporation and radiation would mean it would take forever to heat up all the water.. but we run into the portability issue again, and it's impossible to build into a self-contained unit.
  
  
• it is the effective exposed surface area that is important in long term heat dissipation, not the volume of water.
  
  Yep. I will use much less fluid in the closed system. I just wanted to have a large enough reservoir to last a little while - just to prove the chip can cope.
  
  
• Use tupperware box for the reservoir inside case
• Use a little plastic project box from electronics store
• Use a cd-rom tower case, not a minitower case, they look nicer
• Drill 2 holes in a case-front faceplate (5.25" device slot), route the tubing through there, so you can put the cover back on your case. Use rubber grommets in the holes to stop the plastic cutting the tubing
  
  All good ideas. :)
  
  
• Use a peltier sandwich of 2 heatsinks, one in the water (cold), one in the air (hot) with a peltier between them.
  
  Hmm, interesting idea.. I think the radiator will be simpler, but this is a good alternative.
  
  
• Why don't you do a fully immersed system like Dr Ffreeze?
  
  Because I'm not completely barking mad. Too messy! And frankly, Dr Ffreeze's results weren't that impressive last time I checked.
  
  
• Why don't you use mineral oil?
  
  Because I don't have any. :) Actually, I think the thermal conductivity of mineral oil is not all that good (hence Dr Ffreeze's lack of outstanding results) .. the advantage is that it's not conductive and not corrosive, so if it spills inside your computer you don't care (apart from the mess). Hmm, something to consider. I think water and antifreeze or that Water Wetter stuff is the way to go. (Agg anticipates 100 emails telling him he's wrong again)
  
  
• I can't wait to see what you do with a .18 micron processor / athlon / p3
• Have you tried DEC alphas / p3 / other celerons / athlon
  
  I've only been overclocking Celerons.. because they overclock well, and they're cheap to replace if they go bang. I am always open to donations, though.. :) [especially athlons!]
  
  
• we extensively use the type of RTV you utilized at work... heres a little trick to get the RTV looking like a factory installed bead.... spit on your finger tip and smooth out the RTV... the silicone won't stick and it will look like it was meant to be there.
  
  Had to go read my article again to see what you were talking about. Oh, the sealant stuff! :) Ok, cool. Thanks for the tip!
  
  
• Check under the processor - you are probably still getting condensation underneath it, inside the socket. Remember those pins are the ones that carry the core voltage, so you really don't want to get them shorting each other!
  
  Aaarg! Ok. I'll pull the assembly apart (hopefully I can just cut that sealant stuff off) and fill the hole with more sealant. Also, I think I'll coat the back of the slocket (where the pins from the socket370 are) with sealant so they can't short either.
  
  
• Got to www.meci.com and go to the section on "coils". They have some AWESOME radiators for liquid cooling for $15-20. The inventory changes a lot, so just see what is there. I took mine and zip-tied it to a 19"" floor fan and it keeps things pretty cool. Make sure that the radiator is BEFORE the ice, or you will just warm your cold water with it. Good luck.
  
  They look great. The little one is exactly what I'm after.. shame they're in the US. Should be able to find something here, though.

Your setup is pretty cool for a 370, I have a slot 1 so the there's a little bit more water block to work with. This is about the condensation. There are two things you can do to stop it, completely. I would not use that silicone, it's messy, and your water block will still sweat when you have real cold water going through it, including your 1/4 water hoses, trust me, I'm sitting here looking at some damaged goods because of it. My father is a heating/cooling engineer and he gave me two great products that he uses. I think you have to have a license to get the stuff, but I'm not for sure. The first is a spray called "No Sweat". This shit is awesome. It sprays a fine, textured mist over all of your parts and somehow it stops condensation completely. The second is the stuff I use here. It's a black foam (very thin), and it has a sticky side that will let you apply it to any surface, copper, aluminum, you name it. What I did was totally encased my water block in this stuff and made cutouts on it for the brass hose connectors and the slug. When this stuff is "pressed" together with the chip, it provides great insulation, and it's kind of rubbery to keep anything from shorting out. I filled a whole cooler full of ice and ran it through my water block and the foam wasn't even moist.

Sean Lair

Actually, I _have_ experienced that - I put 4 trays of icecubes into the bucket while trying to hit 670, and I got condensation all along the tubing and on the copper block - even the back of the slocket was wet! Not good. That is the main reason why my current goal is to get the water constantly at room temperature, not "as cold as I can get it" .. I will definitely keep an eye out for that kind of stuff - thanks again!

Ahh.. more useful info. Thanks, looks like I will be hunting for a fuel pump and transmission fluid radiator next.

1. Radiation
2. Convection
3. Conduction
4. Evaporation

Heat will only leave the container by radiation, convection and conduction
if the temperature of the container is higher than the surroundings. This
will ofcause happen eventualy.

The only way that heat can leave a container if the temperature of the
container is lower than the surroundings is by evaporation.

I will now go through the different means of getting rid of the heat:

Radiation:

Radiation is best accomplished from black surfaces. Blank surfaces will
reflect heat. Make sure that your container is black. However paint
insulates the container so the paint layer should be very thin and it would
be best if the material itself is black. Radiation will not take place
inside the container only on the outside. The amount of heat lost depends
on the area of the surface.

Convection:

When air or water becomes hot it will start to rise and thus create
movement at the surface of the container. This movement causes heat to be
removed. from the surface of the container. You can force the air to move
with fans. Faster air or water movement equals higher removal of heat.

Conduction:

Conduction happen in solid materials if there is a temperature difference.
Counduction takes place in the CPU, heatsink and walls of the container.
The best conductoer is Copper. Another good one is Aluminum. All metals are
good conductors. Plastic, Paint, Wood, Rubber and the like are bad
conductors. Use a copper container and use copper for heatsinks.

Evaporation:

Evaporation cools the surface if the humidity is below 100 % Evaporation
can not take place at 100 % humidity. If you use an open container you will
loose heat by evaporation.

The think to remember is that if you double the surfacearea you double the
heat lost. This is why heatsinks have ribs.

Another way to cool the container is to put it inside a freezer or fridge.

Peter Sorensen

Thanks for the info Peter!

Brandon Park

Now that is a genuinely useful, original idea! Thanks Brandon!

http://hardware.externet.hu/cooling/HM1/index.htm (in hungarian, but good pics)
http://freddy-intertaiment.webjump.com/water.html - water-cooling a video card.
http://www.frostyfox.com - the master. we're not worthy.
http://www.accsdata.com/drffreeze/ - the original and best (and looniest)
Dugedug's rig - beer can for a reservoir, very cool.. but where does the heat go?!
www.benchtest.com's rig - homemade copper block!