Diving into cool water that's right there under us
Diving into cool water that's right there under us
I personally prefer to use heat exchangers with a closed loop and treat the cooler like a water/slush box where the radials get submerged inside. The WC loop would stay sealed, but I'd just fill it with coolant or some controlled water instead of letting ice jugs float around. Inside the cooler, any growth needs to come from condensation collecting on surfaces and then hiding among crawlies; otherwise, I'm pretty sure nothing will grow in that weird chemical mix they made using those microbial inhibitor tablets. You've said this a lot before, even when I first shared this cooling idea over 11 months ago. Have you ever tried using your own suggestion?
A few years back, with a bucket of ice and some heater core parts, I decided to test how well these things work on my P4 system when it was getting ready to be scrap'd. I just wanted to share some tips—I know I've said this before, but there are other people who might not have seen your build or config who could consider doing something like that if they were planning a project along those lines.
The heater got the water too hot and that started mold. I'm not sure if I would have given up then maybe some extra radiation before it hit the cooler? Or change the jugs more often? Just use a growth killer instead? No, I didn't give up and adding more radiators wouldn't solve it either because my GPUs don't actually need ice cooling in the first place. They work just fine with a radiator closed loop. The mold I was seeing came from condensation building up under the lid while warm, moist air filled the big space above the water when the temp hit 26C or 78F—that's hot enough for mildew to grow. I wasn't using ice when it was mostly radiators, but even then radiators had a reverse heating effect running chilled water through them so they heated the cold water faster anyway. Even with the water reaching standard loop temps, my GPU load stayed way below 580GTX stock air cooling at 40C less than before. So that's exactly the same kind of cooling I'd expect using my Rad Box to cool the GPUs too. Taking them out of the CPU loop meant the CPU went back to just ice water cooled, which brings me right back to my original goals for high multiplier overclocking versatility with my 2500K.
I wondered if that's really how it works... the outside air probably matters a lot because lots of air blows through and heats up, while the card is just throwing heat out without getting cooled down much.
Back when this thing was built years ago, it was mostly just to test if things would break or work out. It wouldn't fit today's computers easily without changing the heat exchangers and adding more water flow. The main reason I mentioned that was because the project looked similar to yours. I saw condensation on your CPU block, so I had to blow a lot of air around it while using paper towels to soak up the moisture. I don't remember there being too much water dripping down; it just seemed visible at times. I didn't think I needed any kind of temperature monitor back then—maybe even an ASUS probe, which isn't very accurate anyway. Honestly, I wasn't sure if there were many good tools for checking temps like that at the time. The main reason this design doesn't work well today is because it's just too practical to take on the road or move around a lot. Sometimes I go to LAN parties (though fewer now), or even to a friend's house just to play some games, so I don't want to keep all that water going. Plus, running a slushbox means having to constantly refill with ice from the grocery store, which I didn't really need when I was testing it out back then. I had a 5-gallon bucket and an ice bag for those early days. Sure, there are other ways to keep things cold for 24 hours without needing ice all the time, but sticking to this setup makes more sense because you don't have to worry about that part of the process. You've never doubted my cooling solution worked? No, I just wanted to offer some suggestions so other people could try building something similar too. The original thread got deleted or is lost somewhere on the forum, and even though I mentioned heat exchangers a few times, not everyone has read through the old posts. So, while the DIY results are pretty clear, it's obvious this setup works well for what you're doing right now.