In most instances you may not require cooling. However, after completing the H2O conversion of my system, I experienced lockups.
I had to add a fan only to cool the RDRAM.
A water-cooled solution would be nice. Mind you, at $40US, forget about it.
From a practical point, it would be best to use 1/4" ID barbs with a serial parallel setup. It would require 2 Y's or 2's using a 1/2"- 2x 1/4" ea
| | - 1/2"
[[[[[[[[[]]]]]]]]]
| | - 1/4" | | - 1/4"
R1 R3
| | - 1/4" | | - 1/4"
R2 R4
| | - 1/4" | | - 1/4"
[[[[[[[[[]]]]]]]]]
| | - 1/2"
With my ASUS P4T (1.7GHz 423 -115MHZ FSB'd to 1.95Ghz), I have 4 slots to fill (256MB in each for 1GB total) - they do get plenty hot! In an ANTEC 1040B case
My cooling setup consists of an Aquastealth II radiator, CPU waterjacket and chipset waterjacket with an Extreme II radiator.
Yes 2 rads with 120mm fans circulating air and an Eheim 1048 pushing the liquid. Due to the case size, I've opted to hook the rads and the pump externally.
I'm left with a PSU fan, the HD cooling fan and a fan redirected to the saturate the RAM with moving air.
I manage to maintain 4 to 7 celsius differential (average of 4).
Air cooling gave me 10 to 15 degrees difference.
Much better, but still....Cool Silence is the Holy Grail.
If I can watercool my RAM....I can get rid of one fan. Watercooling my internal hard disks will cut out another. That would leave only the PSU fan.
I will eventually replace the ATI AIW 8500DV fan with a cooling block of some kind.
I don't know if I can do this safely (ANTEC smart PSU) although suspect it can be accomplished.
So how to cool the RAM? Why wouldn't an epoxied copper block
with a straight 1/4" ID bore be sufficient?
I'll have to get my hands on a milling unit, or at least a decent drill press.
I mostly wanted to address the idea that RAM doesn't get hot. While it doesn't generate as much heat as a CPU, it does gets hot
enough to cause hang issues. (Trial & Error Proven)
|