Return to

Diy Water cooling experiment "Project FLEXSEAL"


so to begin I have been a closet water cooling dude and well to learn a bit and have fun.
I am taking dual pass transmission coolers and attempting to make a passive cooled water cooling loop. I will be using a “thrift salvage” dell dimension e510 that I got for 7$. It has a 3.0GHz pentium ht (I will post some pictures). Once i am able to beat the stock cooler that is when I have achieved my goal or i have failed by breaking the computer who knows. lets find out.






That pump is absolutely garbage you’ll just end up buying another one, I know I bought one.

Very cool experiment though, I too like cludged together things




thanks for the heads up i kinda expected it to be but it was the cheapest option for this for fun, if it fails i’ll get a way better one its just to see if this works at all.


Keep in mind that you could improve temps by going with copper instead of aluminum. Not saying that you should buy new parts, I know this is a proof of concept, but when you record your results just keep in mind that’s just one way you can improve the system.


Well finding a copper transmission dual pass cooler that’s the tough part not much the rest. the main reason I kept to aluminum was mixing metals is a bad idea, as many know.


Yeah that only matters if they are contacting each other though.


after realizing nothing really comes with barbs i bought 12 just in case.


@Skelterz the pump arrived it came with the 12v power adapter even though i didn’t order it! the pump works its a bit finicky but it does the job (the rad arrived too so i tested the pump)


Nice, best of luck, looking forward to updates


testing the pump my friend had some primo fluid he let me have i don’t intend to use it for the finished build. those are some old led strips my older bro gave me.


everything has arrived time to do 24h leak test


Project Flex Tape
And leak testing done tomorrow I’ll begin testing on the p4


I’m probably going to build a test bench tray/back plane above this I feel my idea can grow a bit.


time to get the temps for idle i will be using stress to test the loop


idle temps 36.2 water temp leaving the rad


I think I ran into a Issue I didn’t expect so I didn’t think about retention on top of the cpu i should have guessed


the fix so far