[Daydreaming] Dual Reserators?

peterson · Post by **peterson** » Fri Sep 16, 2005 4:58 am

I reckon there are more like me who's constantly daydreaming about next configuration. Sane or insane configuration.

I have this little thought on going water cooling that won't go away. It's like breaking new ground. And i just came up with a slighly insane thought on dual Reserators.

And my question goes out to anyone with knowledge on what the extra effect would be.

If placed in series. How much more cooling could be expected?
Would it be better to have 2 loops? I.e. one for CPU and one for GPU?

It's crazy idea i know, but what the h...

acaurora · Post by **acaurora** » Fri Sep 16, 2005 6:10 am

Russ Kinder has two in series, one he bought off me. I think he used external pumps, I'm not sure. Go talk to him ;P

Post by **Rusty075** » Fri Sep 16, 2005 7:29 am

I have two reserators...but haven't gotten around to putting them to use.

For simplicity, running two independent loops would be far easier, and probably more effective. Considering that a high-end GPU is almost as hot as your CPU, it makes sense to treat them the same.

Running them in series would require using an external pump most likely. If you run them with the stock pumps you'll probably discover that the second reserator in line will overflow. There's a hole in the top, and since the second pump will have more resistance than the first pump (that's just pumping from one res. to the other), the water will enter the second tank faster than it leaves.

peterson · Post by **peterson** » Fri Sep 16, 2005 8:02 am

I couldn't imagine that. I thought the pressure would even out.
I suppose i'm no guru at watercooling yet.

autoboy · Post by **autoboy** » Fri Sep 16, 2005 8:45 am

Try running them in parallel. Having two in parallel will increase each reserators efficiency over running them in series. You also get more fluid flow with pumps in parallel. Running pumps in series increases the head (pressure). You really need to experiement with the system because the best setup depends on if your system has high head or low flow. Also, you could try running seperate loops for each and you could keep water inlet temps low for both the cpu and gpu while keeping flow relatively high. Hmm, now that i think about it i can't really decide which is better. Probably independent loops so you could get lower inlet temps and a higher overclock for your gpu.

bobo5195 · Post by **bobo5195** » Fri Sep 16, 2005 9:15 am

The restorator will in no way overflow i would not worry about that. even with a huge bad ass pump im not sure that you could get it to overflow.

Parallel i would of thought would be far better from an thermodynamics standpoint but i think seris with the two pumps will work as well. For restircitve loops seris might be better as the two pumps will double the head of the loop, this will cause the first rad to be more efficent but the second one will have less of a dt to work with so im not sure. Seris vs parallel will be very close.

Happy Hopping · Post by **Happy Hopping** » Fri Sep 16, 2005 8:55 pm

If you have 2 reservators, of course, you run it in //. It's so much easier to set up, and in the even if one loop fails for whatever reason, you still have the other one running.

The idea of running it in series is like your water heating tank at home. If you have 2 water tanks, the water in the first in somewhat boiled, and loop into the 2nd water tank to fully heat. But that idea, I can't see be remotely as efficient and powerful as // set up.

cAPSLOCK · Post by **cAPSLOCK** » Fri Sep 16, 2005 11:15 pm

As most of you know, in the human body the heart is in fact two pumps, the reason is that there are two circuits (lungs and body) in which there is a loss of pressure, so how about applying this to a computer, for example:
Reserator1 -> gfx card -> northbridge -> Reserator2 -> CPU -> Hard-Drive -> Reserator1

bobo5195 · Post by **bobo5195** » Sat Sep 17, 2005 2:06 pm

heart analogue is already accepted. The idea of mounting in seris as you desicribe is good from a heat transfer standpoint but the pressure created by a pump generates flow where ever its placed and runign cables in and out of the case is not nice.

some animals only have one ventrical as well.

SebRad · Post by **SebRad** » Sat Sep 17, 2005 3:03 pm

If I had two Reserators I'd run them in //. Probably one on CPU and second on graphics card and everything else (e.g. Northbridge, HDD). Unless was SLI/Crossfire set-up in which case First on CPU & Northbridge and second on graphics cards. I possibly/maybe read about/saw pictures of such a system somewhere.
What sort of hardware really needs 2 Reserators is open to question if this thread is anything to go by. (Top of the range dual core Prescott + SLI 6800 Ultras?)

OT: I found this and this explaining heart function. My understanding of why it’s set up this way is that the lungs require lower blood pressure than the rest of the body so the right side of the heart pushes the blood through the lungs at low pressure to the left side which then pushes the blood at high pressure round the body back to the right side of the heart.

Seb

Tibors · Post by **Tibors** » Sat Sep 17, 2005 8:58 pm

Any comparison between two reserators in series and the way the heart in mamals functions is going to give you false results. The heart is the pump in a closed loop system. With the reserator you have an open loop system.

The top of the reserator is open to atmospheric pressure. This means that inside the reservoir only the current weather and the depth below the surface determine the pressure. So, as Rusty already pointed out, with two equal strength pumps and different flow resistance in the two parts of the loop one of the reserators will overflow.

bobo5195 · Post by **bobo5195** » Sun Sep 18, 2005 12:54 pm

sorry need to have more of a think

bobo5195 · Post by **bobo5195** » Mon Sep 19, 2005 12:08 am

right the equations boil down to:

C*Q^2 = 40*h(mm)

Where c is the resistance of the loop as a constant (m and units of Q). Q is the flow rate and h is the increase in height of the resterator.

5lpm is the max flow of the resterator (max flow in real life it will be less than this but ive done enough maths already and it can easily be seen that a smalll change in Q is going to have a big change in results). C could be said to be of order maybe 5 for a worst case extremly restricitive loop. That would give a h of 3 and abit mm ie its never going to overflow. Ill post the method if anyone wants but some spell checking needs to be done.