Flow is important ofc. But if water's thermal capacity is already saturated to some lvl from taking heat away from SLI first and then getting to CPU then it will take away less heat from CPU. It does have really high thermal capacity though so it can be all fine at the end. But you don't know till you try it out. Basically ideal would be to try both ways but that way you would spend time and prolly some hose so I was directing you towards usual setup as I don't see why you would go the other way. It also depends how good the 140 rad will cool the water b4 it get's to CPU block. It is not necessarily that CPU will be cooled less efficiently but 1/2/3 C less efficient is enough if you ask me when you already spent so much on this setup. Ie. why not get the best if you can? If there's a mounting problem and it has to go the reverse way, like big difference in overall waterway length then it can be justified, maybe, so you get less impact on flow and pressure drop...
As of number two, two autonomous loops are used as far as I've been told just to separate two groups of components, like when one of them is really hot (yeah, GPUs mostly are put on their own loop) so you don't want to mess it with the others as you'll get worse temps on some of other chips in your single loop. But don't know, again how much you will gain this way as you have a lot thermal dissipation with all these rads and you would have to buy one more pump for dual loop config. Extra cost? Also that way you're avoiding flow rate problems, pressure drops if they pop out if you're having it all under water as I said. A lot of blocks adds to the restrictions. You won't have these probs with 3 blocks. Anyway, think that today's components are not that hot like I said. So why LethalHammer did two loops? Maybe just for that 1 C. Some ppl are perfectionists. I am but wouldn't go dual loop except if everything is under water and then I would rethink to use serial loop or to use dual loop - two pumps needed for both configs ofc. I would prolly go in that case with serial setup and sacrifice that 1 C only to avoid more hose clutter inside the case which you get with dual loop.
So mixing the 1. and 2. I would say you no need reverse loop, as you might lose 1 C on CPU temp that way and you already loosing 1 C more at having single loop, maybe.
Third question, not sure about it but I think ppl get very little benefit from cooling the memory.
Also think you should go the classical way like this: pump => reservoir or reservoir = > pump ( this doesn't matter) => CPU block => 140 rad => GPU blocks => 320 => pump/reservoir or reservoir pump as I've said. This way your pump will get fresh water, cooled and it would be good it gets it like this. I didn't see it at first but what you have drawn for us the pump would be getting hot water, and that's not good for it. Prolly would go without probs but can shorten it's life. Most pumps are designed to operate within the limit of 60 C water temp tops. Few degrees over it won't kill it but if it gets more on the long run it can be harmful... In case two if you follow your same drawing and use the reverse flow the pump will get fresh water but pressure drop on CPU block is increased as it is in the middle of the system, that also can affect temp for 1/2 C. All in all, my philosophy is that CPU should get the best it can. Primary component.
This is my opinion, based on what I've read, remembered and then it formed over time. I could be wrong but I'm speaking you what I know. I am a noob in wc, but do extensive research and I must say that I didn't see much of reverse loops. Hope some of the guru's will respond. If I'm wrong don't shoot me. Guru's will jump in to the rescue.
Reply With Quote
Bookmarks