Originally Posted by Rognin
The problem with parallel is that some ode the already warm water does Cleo through the blocks again after it has passes over the other blocks (ideally it wouldn't)
No, Ideally you want them in parallel to have the best possible temps. Every card will get the same temp water, and it'll go through the rad after. I really don't get where you're getting this info, but it is way off. Think of the loop as an electric circuit. The path of lowest restriction will be followed. If all paths have the same restriction, then the water will flow equally through all channels. So if it were in series , yes I'd agree to some extent that the water in the last card would be a tad bit warmer; but not enough to make a drastic change, and certainly not enough to merit a rad in between. You would also see a better flow rate due to less restriction with a parallel setup, which should equate to better overall temps (high flow normally gives better temps). I use to have 2.8LPM and had cards running at around 46 while folding with a series setup (ironically the last card was cooler than the first two). I now redid my loop with parallel and get 5LPM with cards hardly hitting the 40C with 22C ambient.
i understand how parallel cards works, it can just lead to a slightly warmer loop then serial with a rad between every card (i actually tested it a few years back, as in p4 EE days, before i did a build for a client)... the reason behind this is that there is more load being put on the rad... think about it this way... what would be more efficient... letting the freezer warm up to room temp and then cooling it down to 10F once a day, or letting it warm up to 30F and cooling it back down to 10F 4 times a day?... same applies to liquid cooling (in any application)... though your temps may only be better by 1-2C with a rad between ever card, they will still be better... i was just trying to offer the best option for lowest temps as that was what i thought the OP was asking for... plz dont blow me off as someone that doesnt know what they are talking about, as i do understand the thermo-dynamics of liquid cooling
Originally Posted by Los Hog
He is right the water is moving to fast to sit there and get hot before moving to the next card
if the water didnt change temp then there would be no heat transfer... how would how fast the water moves change the fact that the heat from the cards is being transferred to it?
Originally Posted by jjshabadoo
'Thanks for the input, I was wondering about the whole concept of the water "warming" bit it makes sense that the water is moving too quickly for this to matter.
Obviously for a four gpu loop a powerful pump is a necessity.
Would you guys go with one of those monster 2x480 rads, or two separate 480 rads?
if your talking about pure rad space they should be the same, but i would look into dual 480 rads as well as there is a better variety and you are more likely to find something that fits your needs... you could also get 140.4 rads and use either 140mm fans or 120mm fans with an adapter (what i tend to do if the case can fit 140mm rads)
what i would do (since noise doesnt seem to be an issue) is get these:
2x Phobya G-Changer 560mm
some Bitspower Low Profile 140mm To 120mm Fan Adapter
and some high powered delta fans (200+ CFM if you so desire)
one warning when you are using those adapers though is that you will have to seal the area between the fan and the adapter (and the adapter and the rad as well)... this can just be done with some cheap weather stripping or fan silencers if you are careful to how you attach them
(this is not my build, but a good example of what you should do if you use weather stripping)Edited by eskamobob1 - 12/11/11 at 10:30am