dual radiator, dual pump, setup order question

[Sokratz]

I'm going to go with a dual radiator/dual pump setup. one pump will be a little bit more powerful than the next. what would be the most effecient way to set the system up?

would it be pump-rad-pump-rad-cpu block-gpu block-res

or perhaps

pump-rad-rad-pump-cpu block-gpu block-res.

or even

pump-rad-rad- cpu block-pump-gpu block-res


thanks,
sokratz.

 

[Ackmed]

If its not going to be a seperate loop, Im not sure it would be a good idea. Just get a good pump with lots of head, like the MCP600.

 

[Jaxidian]

Originally posted by: Ackmed
If its not going to be a seperate loop, Im not sure it would be a good idea. Just get a good pump with lots of head, like the MCP600.

Yeah, you might get better performance making two seperate systems, one for CPU, one for other stuff (GPU + NB or whatever else you might wanna cool).

 

[Sokratz]

i guess that's an idea. what i'm trying to do is a put all the pumps and readiators into a seperate acrylic box so i can get all that stuff out of the computer, do better sound proofing, and make it easier to work on. i bought a second radiator because now i will have more room but i figure hte hydor l30 can't power 2 radiators and 2 waterblocks by itself. i've been having some problems iwth that pump but i've now epoxied the whole thing round nad it's working like a charm again. guess i don't want to toss 50 bux down the drain. would you put in a seperat resavoir for each block (cpu and gpu (+chipset later) or have all return lines into one larger res. That leads me to another question. Is having more water in the system better for cooling or worse?

thanks again

 

[SuperPickle]

I agree that everything probably shouldn't be inline. With that equipment, I suggest:
Loop #1: Small pump-->GPU-->Rad-->Res
Loop #2: Large pump-->CPU-->Rad-->Res

Feed both pumps from one shared reservoir. That'll give you the best flow across both waterblocks and with two radiators, good dual-cooling of the shared coolant. That'd be one hell of a system.

Question: did you forget about the northbridge?

 

[Sokratz]

I started off being a cheap schmuchk about it and decided no to the chipset. looking back... man i've ended up buying crazy stuff for this. (including my dremel that i got today *happy*) I'm trying to go for something like this (second pic down) but undernearth the computer as hot air rises. Anyhow, that's the idea. I also got sound dampening foam (under carpet foam) and I'm planning on sectioning off the pumps from the radiators inside the box. Hopefully that'll keep it from being loud. Now to figure out the dimensions and find a second pump.

Thanks,
Sokratz

 

[Sokratz]

Originally posted by: basilisk420
I agree that everything probably shouldn't be inline. With that equipment, I suggest:
Loop #1: Small pump-->GPU-->Rad-->Res
Loop #2: Large pump-->CPU-->Rad-->Res

Feed both pumps from one shared reservoir. That'll give you the best flow across both waterblocks and with two radiators, good dual-cooling of the shared coolant. That'd be one hell of a system.

Question: did you forget about the northbridge?

I've read many places that you're supposed to put the rad before the block inbetween the pump so
Pump - Rad - Block - Res because the pump introduces heat back into the water and you want to be using the water when it's at it's coldest. I originally setup my kit like you have above but am going to change it and see what happens.

Sokratz

 

[SuperPickle]

Originally posted by: SokratzIs having more water in the system better for cooling or worse?

Some will say that more is always better but I'm not convinced (though I cannot prove otherwise). I'm clearly not an expert in thermodynamics but have had built a few watercooling systems in my day. I propose that varying the volume of coolant only varies the speed at which the temperature of the coolant changes.

All other factors equal:
Heat is transferred from the block to the coolant dependant upon the temperature difference of the block and the coolant. The same can be said about the temperature transfer of the heat from the coolant to the outside air via the radiator. Eventually, this closed loop will reach equilibrium...heat in=heat out.

The coolant itself will absorb and retain heat. A larger volume of coolant can absorb a larger amount of heat so it will take longer for the system to reach equilibrium but again, eventually heat in=heat out.

That's my logic so take it for what it's worth. If there are any kind folks with some true knowledge and know that I'm blowing smoke, please set me straight.:beer:

 

[SuperPickle]

Originally posted by: Sokratz
I've read many places that you're supposed to put the rad before the block inbetween the pump so
Pump - Rad - Block - Res because the pump introduces heat back into the water and you want to be using the water when it's at it's coldest.

This makes sense but if the pump is run inline, the heat introduced is minimal. A submersible pump certainly does add heat to the coolant.

 

[Jaxidian]

Originally posted by: basilisk420

Originally posted by: SokratzIs having more water in the system better for cooling or worse?

Some will say that more is always better but I'm not convinced (though I cannot prove otherwise). I'm clearly not an expert in thermodynamics but have had built a few watercooling systems in my day. I propose that varying the volume of coolant only varies the speed at which the temperature of the coolant changes.

All other factors equal:
Heat is transferred from the block to the coolant dependant upon the temperature difference of the block and the coolant. The same can be said about the temperature transfer of the heat from the coolant to the outside air via the radiator. Eventually, this closed loop will reach equilibrium...heat in=heat out.

The coolant itself will absorb and retain heat. A larger volume of coolant can absorb a larger amount of heat so it will take longer for the system to reach equilibrium but again, eventually heat in=heat out.

That's my logic so take it for what it's worth. If there are any kind folks with some true knowledge and know that I'm blowing smoke, please set me straight.:beer:

Generally, you're correct about the equilibrium. But what you're not taking into account is that the larger the resivoir (i.e. the more the fluid), the longer it sits there passively cooling. With more fluid, basically you get free cooling as it just sits there taking longer to get back to the hot stuff. And if you've built a few water-cooling systems in your day, then you know that resivoirs can give off some decent heat (not real hot but always warm).

 

[SuperPickle]

Originally posted by: Jaxidian

Generally, you're correct about the equilibrium. But what you're not taking into account is that the larger the resivoir (i.e. the more the fluid), the longer it sits there passively cooling. With more fluid, basically you get free cooling as it just sits there taking longer to get back to the hot stuff. And if you've built a few water-cooling systems in your day, then you know that resivoirs can give off some decent heat (not real hot but always warm).

Point taken and agreed upon. I guess it just seemed to me that the amount of heat dissipated by the reservoir was negligible compared to that of the radiator unless the reservoir was designed or positioned with this benefit in mind. Perhaps the cooling capacity of a large-capacity reservoir is notable.

On a personal note, my first system only had a tiny reservoir (marine electrical box) and the subsequent systems were reservoir-free inline systems (space-conservation was a concern).

 

[Jaxidian]

Originally posted by: basilisk420
Point taken and agreed upon. I guess it just seemed to me that the amount of heat dissipated by the reservoir was negligible compared to that of the radiator unless the reservoir was designed or positioned with this benefit in mind. Perhaps the cooling capacity of a large-capacity reservoir is notable.

On a personal note, my first system only had a tiny reservoir (marine electrical box) and the subsequent systems were reservoir-free inline systems (space-conservation was a concern).

A few more comments:

The geometry of your reservoir makes a big difference as well as the internals of your reservoir. The geometry is pretty obvious - the more surface area, the more it cools. But the internals of your reservoir is very important too. Here's an exagerated example:

Consider a 10-gallon reservoir that is 5" tall, 10" wide, and however many feet long. This reservoir has excellent surface area to transfer heat. Now let's say that the your tubes are connected to this reservoir at the same end of this really long reservoir. Naturally, the water at the far end is not going to be sucked out for quite a while whereas the water at the other end is constantly being used. Because of this, you're sucking out water that is much warmer than what you could be getting.

Now the obvious solution would be to pump the water in at one end and suck it out at the other end. However, a slight improvement would be if, in the above reservoir, there was a solid wall that went down the middle of the reservoir from one end to the other end with a 1" gap at the far end. This would force the water to travel the entire length of the reservoir twice. The main improvement is not that it travels the length twice (that's not really important since it travels faster than if it went from one end to the other end) but that there are fewer dead areas due to the narrower channel.

 

[Mellman]

Yeah generally i think abigger resivoir is better...but if its designed poorly, its a moot point. If its just putting the water right back into the res, and the intake is next to the output, your suckin the same water....the idea posted above is good...if i had time...money....a duck, and two pidgeon's i'd try...but i have none of the above right now .

other suggestions people have made....the two pump situation...i've seen very few implementations done on this. most are just a backup, in case your primary fails. In my opinion, its completely un-needed to have two res and two pumps on a single system, i'd do two separate loops, since obviously mobility doesnt seem to be an issue for you considering your doing an external resivoir. if i have time i'll post more later tonight back to work now! ALL OF YOU!

 

[Ackmed]

I just use one pump, dual rad, res, CPU, and NB block. Works fine. I dont see the need for a second radiator.

I am going to do a seperate loop for my GPU in the next few months though, with its own pump, rad, res and block.

 

[Sokratz]

Well, sounds like splitting it is the best idea then. Currently I'm unable to do a chipset block (big honking HS that ain't coming off) but i'll seperate the NB and GPU from the CPU. Does anyone know how to run power supplies in series? Trying to figure out how to turn on all my fans and radiators when I turn on the main machine.

Thanks,
Sokratz

 

[0roo0roo]

from what ii've read its better to have a single large radiator.