Experiment in Water Cooling....

[alysher]

Hello all!

i am planning an experiment in water cooling that i would like some input on. Before i get into what im going to do i should let every one know that this will be my first time doing a water cooled setup. i have read most of the stickies and understand what i need for the most part, but i definitely need a couple of pointers.

now this experiment is 2 fold. first to see if it can be done this way and to assertain if there is any difference versus it being air cooled. second to teach myself the in's and out's of building a true water cooled system.

now for specs.

the loop is going to be set up on a laptop. yes, a laptop. there are several reasons as to why im using a laptop but top most(after instruction and to see if it works properly) is because of whats been going on with this laptop.

this section is me just ranting about how much of a pain this laptop has been..... you dont have to read it if you dont want to!

the laptop is a Gateway M520/7330/7426(all models are identical except for processor and graphics cards-first two are Intel p4's and the last an AMD Athlon 64) that has been giving me issues for a long time. it started out as a Gateway 7426mx(amd processor) that gave out due to power issues(stopped receiving power from the adapter). i went searching for a new motherboard and found that the intel boards were cheaper and since i didnt care bout graphics on it i figured i would go that route. i got the board and the cpu that is ment for it (ddr mem on all thank goodness) and attempted to bench test it. what happened was nothing....i had gotten a used board that had fried. i rma'ed and ended up with a second the same way. got a refund and got a NEW one. and had the strangest issue...on the internal monitor i would have white vertical bars most/if not all the way across the screen. since i happen to have a second laptop of the same type i double checked the screen. lo and behold it worked fine. when i plugged the board into an external monitor i was actually able to use it. since i had NEVER had this happen to me i researched it and found that i may have a bad board, so a second rm process and a new mobo later same thing happens.....by this point im really frustrated and put the puter on the back shelf. over the next few weeks i get 2 more boards in. one dead and a second one that repeats history.

at this point im giving up on getting this laptop up and running properly, so i decided that im going to rebuild it with out the screen and use it for a testing platform of sorts. so for the first test im going to try for water cooling.

ok now on to what im planning on doing. im going to use the original heatsink/fan combo as my waterblock/radiator. this is what is looks like...

see the copper tubes running from either side to the center area. well they are soldered to a copper block about 1/8 inch thick. these tubes are actully hollow with a copper mesh of some sort in them.

now where i need advice atm is how to attach fittings to the ends of these tubes so i can get a pump and reservoir hooked up as well.

any ideas anyone?

 

[Conumdrum]

Well, the tubes have a wicking agent and a liquid that vaporizes the liquid. Then the vapor moves to the lower pressure end of the tube that is cooler where the fins are, it turns back to liquid and is pushed back to the hot part (CPU). This is how heatpipe tech works. Turning into vapor removes heat, the vapor hold energy (heat) and is released back to the fins, which removes the energy and it returns to a liquid state.

Using the tubes for watercooling can'r be done because there is no loop possiblity unless you cut then off each tube at each end, braze a fitting on each one and run water through it. It could work.

Few issues. The pipe at the CPU part has no room to keep it thin and braze ends on. The wiking in the tubing could slow flow so bad there isn't enough flow. And the heat removal area by the CPU is probably a poor conductor for watercooling, and how regular PC waterblocks work with fins etc.

Sure, you can try it, but never seen it done. Lots n lots of questionable home engineering needs done, if your a experianced brazer and have access to make your own tubing ends, make a 3 way hose outlet, buy a pump and a radiator and res on the cheap, it could work.

You could try it.

I don't think anyone wanted to take the time to explain why and how, but I read this a few times and thought a reply would at least be the right thing to do, even tho the idea is okay, but the execution could be a nightmare.

Best of luck, good to see thinking outside the block.

 

[\dev\goat]

thinking outside the block.

 

[alysher]

Well, the tubes have a wicking agent and a liquid that vaporizes the liquid.

huh???? I had to re read this part 5 times before i realized you may have ment that its a liquid that vaporizes when it heats up.

Using the tubes for watercooling can'r be done because there is no loop possiblity unless you cut then off each tube at each end, braze a fitting on each one and run water through it. It could work.

this is kinda what i was planning, altho was going to solder on either a compression fitting or a a few inches of copper tube(1/4 inch) to make into a barb. brazing isnt something that i have done much of, but i have done lots of soldering for plumbing and electrical issues

Few issues. The pipe at the CPU part has no room to keep it thin and braze ends on. The wiking in the tubing could slow flow so bad there isn't enough flow. And the heat removal area by the CPU is probably a poor conductor for watercooling, and how regular PC waterblocks work with fins etc.

im going to take this in reverse here. the heat removal area is basicly a 1/4 inch block of copper about 2 inchs square. if it was good enough for the air cooled system wont it be fine for water cooling? if not would using a thinner base be ok? i happen to have a second heatsink from a different model of laptop that i clipped the ends on and checked the flow on. air goes through it decently but it does need a bit of pressure. how are pumps at pressure? good, bad, or ugly...or does it depend on the pump i get? as for the pipes at the cpu end, well the end of the pipes are about half an inch further along then the end of the sink, so if im careful i can remove some of the edge and keep the end strait and solder on some extra pipe or make a manifold to use all 3 pipes at the same time.

...lots of questionable home engineering needs done...

something im good at!

...the execution could be a nightmare.

sounds like most of my projects!

Best of luck, good to see thinking outside the block.

thanks for the vote of confidence!

 

[Conumdrum]

Pressure depends on the pump, it's head pressure is what matters as well as flow rates. Flow rates to move the liquid can'r be too low. In our loops a 600 GPH pump might only do 1.5 gallons per minute or less.

Sounds like you have some good ideas.

 

[alysher]

how would i check the flow rate for a single line? if i can figure this out i can multiply it by 3 and that will tell me where my pump should be correct?

what about the 1/4 inch copper block that the 3 lines are soldered to? is it going to be enough or do i need to see about finding a thinner piece of copper to replace it?