what order should i make my loop in

[rustyfender]

i have a
apogee gt cpu block, danger den 4870 x2 jacket

Pump:
swiftec mcp 665-b

resivoir:
swiftec mcres

Radiators
one swiftec mcr 320
and a single 120mm thermaltake one

they are in a cosmos s case so the swiftec rad has air from in the case pulled past and is exhausted out the top of the case and the small single fan thermochill one has air from outside the case pulled past

so i think it might be best due to the massive thermal output of my gpu that the loop goes:

res>pump>cpu>small rad>gpu>large rad>back to res

any thoughts?

[Conumdrum]

You pretty much got it right. Place the rads in any order, try to put the Pump right before the CPU. The diff in water temps will be about 1C anywhere n the loop so order for rads really don't matter. Just res before pump inlet.

[Boulard83]

+1^^

[axis01]

My personal opinion is that you should have the absolutely coldest water going directly to your CPU first. That would mean feeding the CPU straight from the rad. While it's true that the temp increase from the pump is very small, it's still adding heat to the loop. Spending $$ to get a 1C difference is up to the person building the system, but getting that 1C difference by just changing the loop order seems like a no brainer.

Axis

[Nebulous]

My order is like so:

T-line>pump>rad>cpu block>pump in a closed loop. :D

[baditude_df]

My personal opinion is that you should have the absolutely coldest water going directly to your CPU first. That would mean feeding the CPU straight from the rad. While it's true that the temp increase from the pump is very small, it's still adding heat to the loop. Spending $$ to get a 1C difference is up to the person building the system, but getting that 1C difference by just changing the loop order seems like a no brainer.

Axis

I have to agree here. It's a "why not?" type of thing. However, sometimes there are constraints(as in my case) where I had to do pump to CPU to keep the loop short as possible and avoid rediculous bends etc.

[rustyfender]

http://i729.photobucket.com/albums/ww298/rustyfender_01/PICTURE8.jpg

here's a schematic of what i think would work well i took your advice axis and badditude the flow of water isnt shown but it goes

rez>pump>single rad>gpu>triple rad>cpu>rez

any thoughts the pump and rez and both rads are already mounted from privious set up that didn't have the gpu in the loop

[thorilan]

while im of the same opinion as axis there is something you need to consider.
how much tubing will you use in each configuration

i have built loops with less than 30cm of tubing because of the design of the res and placement of the pump but it takes some skill and ingenuity and the willingness to possibly mangle a case to get things the way you want them.
BUT for all that effort you can gain 200+ mhz OC with the same noise level ( the overall goal of lowering your Dt, not just lowering it to lower it.)

overall you are deciding is the added tubing creating more resistance to the system lowering its flow enough to effect the efficiency of the rads or are you high enough on the rad curve that that extra flow is not doing much so you can turn it into more tubing , and hence reroute system to make use of the core set of rules "add all the heat at once and all the cooling at once to make a more efficient Dt"

that is where your decision lay

[rustyfender]

ya ive been thinking about that and to go this rought would require a fair about of tubing almost 10 ft. i don't like that. im currently working on a second sechematic. ill post when done.

[rustyfender]

http://i729.photobucket.com/albums/ww298/rustyfender_01/PICTURE9.jpg
there is the second schematic this one uses less tubing but it it is dependent on me being able to turn the rad around. this may be a problem as i don't belive ther is enough room for the fittings on due to the motherbord but ill find out when i get home from work

yes i know the pump does not take directly out of the reservoir but not rlly any way around that problem so its in the next best spot. there are a numbe opf inherant problem with this schematic eg. i will have to use onboard sound X( cause the pump will not fit under my creative card.

id favor the first schematic as it is far less work (due to remounting and whatnot).

[thorilan]

keep in mind there is no rule that says the pump has to be afixed to anything at all. i did one build where i had it hanging inside a pair of pantyhose (one of my more ingenius setups)
in your second drawing what if you moved the pump to take from the res by moving the pump twards the front of the case and mounting it at a slight angle so that the output can still go down to the rad and from the rad straight up to the card.
it would require you to design a suspension for it but it can be done a lot of different ways

[rustyfender]

hadn't thought of that never rlly considered that ill have to take a look at the case comp when i get home to get ideas for that im at work atm so only so much i can do from here

[thorilan]

hehe i did all my best design work at work on scratch paper =)

[jr1]

in the first post you used both thermaltake and thermochill for the single rad
if it is a thermaltake it most likely is aluminum in the rad and you should be using something to inhibit corrotion as it will get there over time
if it's a thermochill all you need is destilled water and some biocide

[rustyfender]

its a thermaltake srry about the misunderstanding. anyways i am aware of that and i am running 20% propalene glycol. i work in a heating ventilation and air conditioning engineering firm so i can get enough of it for free :)

thanks for checking though

[jr1]

np
better safe than sorry :-))

[rustyfender]

np
better safe than sorry :-))

no kidding only problem is the stuff had blue dye in it......... now the fluid has no color and 5 bucks says i can tell you exactly where it ended up

[jr1]

hehe i'm not gone bet against since i bet you know way better than me
all i'm ever useing glycol is on the rad on my machinery

[rustyfender]

i havn't actually looked to see where it wen't yet im thinking its plugged up something prolly rad or cpu block. ill look tonight when im home from work since ill be re plumbing my whole loop anyways for the new gpu block. but i taste checked the water for propylene glycol which is safe since its a food additive its the biocide that you gotta watch for. and it tasted like glycol so its only the dye that came out of sol'n. anyways ill recycle the same fluid since it is only a couple weeks old and now has no die and when i have it apart ill clean everything out. then i should be good to go. till i can get rid of this damned aluminum rad and replace with a copper one.