View Full Version : Benefits of a second loop?
Right now i have one loop in a TJ07. It cools my two 7900gtx's and my cpu. When it's hot in my room my comp idles at 38 or so and goes up to 42 under load. When my rooms cool it idles at about 35 and goes up to 39. Not bad temps i think for having 3 blocks in the loop. My Video card blocks are full cover NV-78's too so my loop's really cooling alot. I'm kinda wondering how mcuh i'd benefit from a second loop just for my cpu. I have two 120mm fans in the top of my case so it's all already cut out and ready to go (love the TJ07). I could have one loop with a GT240 cooling my cpu, and have my GT360 cooling my two gpu's. But obviously if it's not going to change my temps that much then it's not something that's worth doing.
I'd just do a second loop to keep the videocards and CPU seperate. It would also add a bling affect because both loops would be different colors. At least for my setup they would.
Right now i have one loop
I could have one loop with a GT240 cooling my cpu, and have my GT360 cooling my two gpu's. Without adding a second pump there are two problems. 1.) how will you split the flow? (a wye gives no controll!) 2.) you will halve (or thereabouts) the flow rates through each loop.
There is a easy, cheep, but not conventional way to do what you want. Put a small, quality pump (eheim1048?) and a rad (2x120 or 3x120) on a "side loop. See pickey. The advantage are: There is nothing except a 1/2" rad in the new loop, so even an eheim 1048 will push almost 1gpm through a 1/2" rad. Flow splitting is not an issue. You can shut off the second rad and fans and have no effect on the exixting loop. It's cheep and it can double heat disipation! I've been using this setup fro about 7 months now and it's fantastic.
No i was planning on getting a second pump, just forgot to add that in my original post.
Without adding a second pump there are two problems. 1.) how will you split the flow? (a wye gives no controll!) 2.) you will halve (or thereabouts) the flow rates through each loop.
There is a easy, cheep, but not conventional way to do what you want. Put a small, quality pump (eheim1048?) and a rad (2x120 or 3x120) on a "side loop. See pickey. The advantage are: There is nothing except a 1/2" rad in the new loop, so even an eheim 1048 will push almost 1gpm through a 1/2" rad. Flow splitting is not an issue. You can shut off the second rad and fans and have no effect on the exixting loop. It's cheep and it can double heat disipation! I've been using this setup fro about 7 months now and it's fantastic.
That is a pretty unique idea.
I have to think about this idea for a little while because right away i think that the water would have a higher tendency to follow the second pump instead of taking the direct path to the cpu block. Since the pump 2 is trying to suck up as much water as possible. I think this would hurt the first pumps flow strength, and it would rely on the second pump to push the water through the cpu/gpu blocks. It is a very unconventional way of doing thing indeed and for max flow i think you would need to use 3 pumps in this kind of setup. I am no flow expert but logically thats what i think would happen.
It would probably be better to do something like pump1out->pump2->cpu->gpu1->gpu2->rad1->rad2->pump1in or some variation of this. This way it is a simple loop with no t-lines or y-lines breaking your pumps flow. KISS, Keep It Simple Stupid.
As for adding a second completely different loop well if your temps are pretty close to ambient (room) temp then your loop is doing a good job but if your temps are much higher than room temp then I would definately try cooling your video cards, and cpu on seperate loops. It may help lower temps a bit and definately add a "bling affect."
what pump are you using? You could just throw another rad in there.. or use 2 pumps in series with an extra rad.
Best way to check to see if work needs to be done though would be to measure your water temp (which I would imagine to be ~30C) and compare it to your ambient temp. Ideally you'd obviouslyl ike to have it as close as possible (mines about 1-2C over ambient) I've only got one CPU block with a triple rad though so it stays cool.
That is a pretty unique idea.
I have to think about this idea for a little while because right away i think that the water would have a higher tendency to follow the second pump instead of taking the direct path to the cpu block.
It would probably be better to do something like
pump1out->pump2->cpu->gpu1->gpu2->rad1->rad2->pump1in or some variation of this. TIf you think a little harder you will see that if the Eheim pump wasn't in that loop, the differential pressure between the takeoff and the return is the friction caused by about 2-3 inches of 1/2 inch tubing...essentially zilch, so very little water will flow through the secondary loop. Adding a pump to this loop does not change that fact, except for the water pumped by the Eheim. This type of setup is standard in the industrial HVAC industry for a number of systems.
"pump1out->pump2-" will not be practical for me because of the VASTLY different presures and flow rates between an Eheim 1048 and an Iwaki MD20rz(t). One of the benefits of this system is that I get the cooling of a second loop with a small, quiet, ecconomical to rum pump, since it is only pumping through a 1/2 inch rad and nothing else.
Actually, I have a check valve located between the takeoff and return, so that I can run either or both pumps depending on the particular OC and ambient temps at the time.
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.