Routing hoses

[Nordic]

I didn't route my hoses to get the optimal temperatures, my current radiator is too weak to handle overclocking my Athlon XP anyway so it doesn't really matter how I route them (IMO). The water temp really gets up there when I overclock, I tried 1333@1750 with 2.1 Volts and the temperature was 55-60 degrees and climbing which is not healthy for the PVC tubing as the water temps get real high. 1500@1.75 Volts seems manageable (50 degrees with burnK7) so I've stopped there. Check out my routing in the pic down in my sig!

I've ordered a heater core which will change this and I will re-route my system for better cooling and replace the few PVC hose sections I have. The heater core probably goes in the back of the right compartment of my cube case so I will have to plan my routing a little different. I haven't tried adding more fans to my current radiator but I don't think that will help my cause, either way it's gonna go in favor of the heater core, perhaps I will use it in-line for additional cooling.

Compared to my old Asus GF2 GTS, the GF3Ti200 (175@220) is a really cool card. I would guess it's below 10 Watts while my GF2 was probably dissipating 15+ Watts and deperately needed active cooling when I used a Zalman GPU cooler. I don't think that my non-optimal routing of hoses influences the CPU temperature that much but it would be fun to know what gains (if signficant) I can expect if I route rad-cpu-gpu-... ?

[ECUPirate]

either way it's gonna go in favor of the heater core, perhaps I will use it in-line for additional cooling.

If you add a rad, run it in parallel w/ the old one. This effectively slows the flow rate through either rad (like using a larger hose...), while increasing total flow across the two rads (less total resistance than just one rad). Rads like low flow rates because it gives them more time to dissipate the heat.

Your temps seem very high. If your system will run at 1750mhz, then I'd suggest you find a way to cool it so you can take advantage.... you're lucky. My 1333 doesn't like to cross the 1500 mark. granted, I've not had time to try too hard..

[Brad]

use the heater core for the cpu, and use the old rad for the gpu, nb, whatever else you want.

[Nordic]

You guys have sparked my interest with these ideas

I could simply use a "T" for distributing the flow from the pump to the radiators, using the perpendicular fitting for the GPU-line as the flow along that fitting will be more restricted.
Then I add another fitting to the air-trap so I have another incoming line.. seems like little work

About my temperatures, when I run my CPU at default (1333@1.75), I idle at 14-16 degrees (Celcius) above ambient, which I don't find that bad. The problem is that I'm very close to the point where my current radiator runs out of steam - it can't dissipate much more heat without raising the water temperature significantly..

[ECUPirate]

If you are saying that you're going to run the AGP block and the CPU block in parallel with the rads, I would discourage that. You want high flow rates through the blocks, so I'd use them in series like you have them. Just put the rads in parallel. Then again, if the AGP block line has high resistance, it might not matter much...as long as the AGP card is getting sufficient flow too.

Brad: You're suggesting running the rads in series? I thought that was a no-no...?

[Nordic]

Ehm ECUPirate, I gotcha: The rads in parallell, everything else in series. Brad, what is the motivation behind your suggestion?

[Cyco-Dude]

thy this:
pump -> 'Y' fitting -> rads -> 'Y' fitting -> CPU -> GPU -> res / airtrap -> pump. make sure the res / airtrap is the highest point in the system.

what Brad was saying sounded like pump -> 'Y' fitting -> heater core goes to CPU, old rad goes to GPU, NB, and anything else on the system that you want to WC -> 'Y' fitting -> pump.
with that method, you keep the GPU and CPU on different loops.

[Nordic]

I really don't get this talk about keeping the air-trap at the highest point in the system. Sure it works like that if you have a pool of water as all the bubbles will merge to the top as they rise. But in a closed loop system all that matters is that you have good circulation so that the air bubbles are brought to the air-trap. Think of it as a bucket held upside-down in the water several meters below the surface of a big pool of water. What matters here is that all of the water in the system will pass by the "bucket", the air-trap, and it will fill with air eventually.


In places where circulation is slow, like in a large heater core, it is a good idea to keep the exit/entry fittings of that component at the highest point related to itself, topside if you know what I mean (english is not my mother tongue).

The exit/entry of that component does not necessarily have to be below the air-trap, it serves no purpose unless you have a really unique routing of your hoses where all bubbles will rise to the air-trap without circulation in the system - which is really ridiculous. But keeping the fittings topside will make it harder for bubbles to resist the flow in your lines and be dragged along until they pass the air-trap.



Ok, perhaps I've pointed out something obviously untrue here.. please say so if that's the case.

[ECUPirate]

You're on the right track, but having the airtrap at the top makes things MUCH easier. Air will rise to the highest point it can. There can be multiple 'high points' in a system. For instance, if a waterblock is positioned so that both hoses point downwards, air will accumulate in the block regardless of where the airtrap is (absent any water flow). The waterblock has formed an airtrap.

Of course, good waterflow through the system can fairly easily flush out the waterblock, but if the heater core is trapping air, that could be more difficult.

In my system, the rad is at the bottom (barbs at top), the waterblock is above that, the pump higher still, and the res is at the top. Air in the rad rises up through all of the parts and into the res, even w/out flow. Granted, small bubbles required the water flow to flush out... Bleeding this system was a breeze. I poured water in the res, and watched it flow down the system, filling up all of the parts!

[Cyco-Dude]

i guess that makes sense lol.

[Nordic]

Yes I agree, air will rise to the highest point it can. But in my opinion having the air-trap at the highest point is MUCH harder!

In my "low-rider" cube case it's very hard to get the kind of layout that you have ECUPirate. As you mentioned, good water-flow will flush the bubbles out of most systems, and that works great for me! I previously had a problem as I had no air-trap and relied on the bubbles to rise into a perpendicular fitting of a "T", filler/airbleeder setup. It simply was not effective enough, the flow in the line was so strong that it swept the bubbles past the "T"-fitting. With the air-trap in place, I fill the system in just a few minutes and the bubbles are all gone within 15-20 minutes.

I'm just waiting for my heater core now

[ECUPirate]

if it works for you, great.

what kind of hose do you have attached to the rad? it looks different from the silicon in the rest of the system. --mighty tight bend radius from rad to res... is it tygon?

[Cyco-Dude]

ack, ECU got his post in b-4 mine.
i was agreeing with nordic
even if the airtrap is the highest point, there are, as you said, other 'high points' (could be the top of the rad, WB, tubing, etc). so having the airtrap @ the top is a moot point, as long as water goes into the airtrap, air will settle in there. to get ALL air out of the components (core, WB), you'll need to play with them just the same as you would with an inline system (that is, one with no res / airtrap).
think about it...... *ouch* lol, dont think too hard thought, might hurt yourself

[Nordic]

Originally posted by ECUPirate

what kind of hose do you have attached to the rad? it looks different from the silicon in the rest of the system. --mighty tight bend radius from rad to res... is it tygon?

Actually it's regular PVC but if you look closely you can see the steel spring that I've inserted into the hose to keep it from kinking, works really well!
Here are some more pics:
http://212.32.156.24/~mattias/stuff/...eongyang_cube/
As you can see, air flow is not optimal through the radiator with the front attached, but I really don't care - the radiator is too weak either way.

[Brad]

what I meant was running a small pump, old rad, and other waterblocks as one system. and big pump, new rad, cpu block in another system, these systems aren't connected in any way at all.

thats the motivation behind my suggestion