Revision of a Volenti Concept, will it work?

[Khledar]

Well I've had this on my mind for a good year or so. But now that I'm done my term, I've got time on my hands, and water to cool.

It all started when I read this:
http://forums.procooling.com/vbb/showthread.php?t=4469

Basically my concept is very similar but done with acrylic and on a lot smaller scale. It's the scale that worries me, maybe I've gone too small to negate the cooling goodness.

Well time for pictures, they describe it best.

Apparently my rendering program thinks "Plexiglass" is green, so I switched to glass for the second two, I also moved the "pump" up further where it will actually be in the final two pictures.

The pictures are missing all the connections but basically the pump loop goes like this: Pumps(series)--> "sealed" resevoir --> Tubes --> Open resevoir -->computer --> back to pumps.


Two things I'm wondering your input on:
1) Will I have enough surface area to cool decently? (The tubes are 1/8" ID, 1/4" OD and 5" tall, there will be approximately 28 (or more if you suggest that will help) (32 are shown in renders for symmetry).

2)If I shut it off, will the pumps drain (they're the highest thing in the system)? I'm thinking with the loop setup the way I suggested the pump wont drain - but maybe I'm just not seeing it properly.

Any general suggestions are welcome too.

[EDIT: Oh, the attaching seems to have reversed the images order what so they go: 3, 2 ,1 as shown below]

[j813]

Do you mean two pumps in "series", why?
Where will the Fans be located?
Haven't made one, I'm interested on the results are.

[Khledar]

Well by series it just means the outlet of one pump goes to the inlet of the other. Parallel is the other option for connecting two pumps, in the case of parallel the inlet and outlet of both pumps are joined by a wye fitting, or such.

[j813]

Why in series, for higher flow or head?
Maybe 1 powerful pump will do instead of 2in series. Just a suggestion.

[Khledar]

Oh, yeah well I have two pumps - the system is kinda designed around using both of them, since they're just sitting around.

And if I recal correctly two pumps in series yeild better head.

[deathBOB]

Ok dumb question, whats it supposed ot do? How is everythink hooked up? I assume the water goes in the littl etubes and is cooled... What about leaving the tubes uncapped and making a fountain kinda effect...

[Khledar]

Oh maybe I didn't explain it too well.

Yeah the water is pumped through the tops of the little tubes and flows down them, cooling the water (Volenti used little cloth dealies to slow the water's descent). The water then returns to the open resevoir where it is sucked back into the computer loop, goes through the computer, then returns to the pumps, which pump the water back into the closed resevoir, and up through the tubes.


By my calculations (V1A1=V2A2 - conservation of constant density mass) the velocity coming out the the little tubes will be about 1/2 of the speed of the water through the loop. Maybe that's too fast meaning I need more tubes - which can be done. Any thoughts?

[Khledar]

Well I've got a couple pictures that hopefully help answer the questions I've had so far. I made the boxes into the actual pump dimensions and added where I want a 120mm fan to go.

[Groth]

Uhh, it looks like you're having the water drain out of the open part, circulate to the CPU, and be sucked up by the pumps?

[Khledar]

That's the idea.

[Groth]

It won't work, unless you have a pump that is specifically rated for self-priming and has several feet of suction-head. The normal little centrifigal pumps we use don't like sucking; they'll cavitate, pump very little, and possibly die.

[Khledar]

That's kinda what I was thinking (since the pumps are not self priming) - But if I set it up like that I don't see where the air would "get in" Like the top of the tubes is technically the higest point in the system as seen. So the water (in the pumps) cannot 'drain' into the sealed resevoir, as it would have to push water up. And if it tried to drain into the computer wouldn't it just suck water from the closed resevoir till...hmmm until what? Like what I'm thinking is since the pumps are between two closed bodies of water how is the air going to fill the pumps? Like the CPU loop won't get any air until the open resevoir is empty, and the pumps also wont get air unless the closed res is empty, and that water can't go nowhere. Shouldn't the levels in the two tanks equalize with a slight difference due to the increased pressure of the water at hieght in the pumps?


[Edit: Below]
Think of a siphon. If you have one big tanks and a smaller tank inside of the big one. You start a siphon fromt he full small on to the empty big one. What happens? The siphon should draw water from the small tank into the big tanks until the water level in the two are equal. If the end of the siphon tube are submerged does the siphon tube drain? I don't think so - but then I guess I should go try it :P That's kinda what I'm thinking I'd love for someone to prove me wrong BEFORE I start building though :P

[Groth]

It's not a matter of air getting into the lines (which is easily prevented by having the discharge in the closed res below the waterline of the open res), but the amount of pressure inside the pump chamber.

As a pump impeller spins, it causes high pressure on the front of the little blades/paddles and low pressure on the backside (especially the tips). If the low pressure regions are below the vapor pressure of your water, the water gets "pulled apart" and forms little bubbles of water vapor - cavitation. As those little bubbles are thrown ******d with the discharge, they get into an area with higher pressure and collapse, which is noisy and cuts your discharge pressure. The energy that goes into the creation of those bubbles is wasted as heat and sound.

As drawn, your setup would have the pump inlet pressure below atmospheric pressure, whereas our pumps are designed for an inlet pressure at (or a few of cmH2O above) atmospheric. With that reduced pressure you risk cavitation.

Net Positve Suction Head, NPSH, having the inlet pressurized: pumps love it, so should you.

Edit: Why is the forum software censoring ******d (o-u-t-w-a-r-d) ? Man, that is *****d up.

[bigben2k]

Originally Posted by Groth

Edit: Why is the forum software censoring ******d (o-u-t-w-a-r-d) ? Man, that is *****d up.

Actually, it's censoring o-u-t-w-a-r, which is one of those silly net games where you collect clicks. O-u-t-w-a-r-d should be filtered out seperately, but that's buggy... :shrug:

[Groth]

Ah, yes. I'm familiar with that nonsense. 'Twas a terrible plague of spam and redirects on usenet. I've firewall rules to avoid them and their ilk.

[Khledar]

You've got a good point there. Lol. Damn.

There won't be that much flow resistance in the loop, don't you think if I can get the flow going the risk of cavitation is low? If being a key word.

There is a reason I'm doing it like this. It's not just 'because' -- maybe I'll go look at my Fluid Mechanics notes and regret sleeping so much in that class

[Groth]

Okay, I'll bite...if it's not just 'because', it's ???

Pumps aside, I think it's workable (and would be quite stylish).

[Khledar]

The sealed bearings on the direct drive shaft of the pump have, over considerable use, begun to leak. I want to catch the drips. I guess I should have said that - there's always more than one way to solve a problem.

The pump leak situation can be found hmmm - here: http://forums.procooling.com/vbb/sho...nson+pump+leak - since then it's only gotten worse, however I haven't used either of the Johnson pumps since then - just looking to put closet clutter into use.

[Groth]

Ah, that makes sense. Well, just have them suck directly from the open res and pump to the CPU. You'll still have some negative suction head (a couple inches), but much less than sucking through the computer (a couple feet).

If you want to get fancy, come up with some sort of cantilever or floating platform so that the pumps are the same distance above the water as the water level drops due to evaporation.

[Khledar]

Originally Posted by Groth

If you want to get fancy, come up with some sort of cantilever or floating platform so that the pumps are the same distance above the water as the water level drops due to evaporation.

Now that, would be sweet!

[Heavytank2]

Simple. Styrofoam plank. Imagine shipwrecked people on flat log-boats. Just make a "log boat" out of styrofoam with a hole for the suction hose.

Neato idea, btw.

[Ls7corvete]

why not try making a res with a bunch of holes in it, shove something absobant in the holes, sponge etc, and let the water wick down. Drilling holes is easier than routing pipe IMO.