Would 2 resevoirs lower temps??

S

steven1110

Limp Gawd
Joined
Aug 12, 2004
Messages
271
Thinking about this setup:

pump->Resevoir->radiator->Resevoir->CPU->GPU->GPU->Pump

OR maybe putting the 2nd res inbetween the CPU and GPU blocks.

Would this help out cooling at all??

ALSO...if I were to get an ehiem 1048 what cpu block should I get?? I was looking at the DD TDX but I know performance differs with the pump.
 
Unless you have a huuuuuuuge res, and even then, a little radiator would be better for cooling
 
Ok then...let me ask you this then.

Would using 2 separate WC systems (1 for CPU, 1 for GPUs) both with single 120 rad be better. Would the extra money for 2 res and 2 rad be worth it??
 
It'd be pretty messy, and from what I hear after water hits your CPU it's only a few degrees hotter, so I'm pretty sure that you could build a high end 1 loop system that will outperform a 2 loop system for less money.
 
DarkMonkey said:
It'd be pretty messy, and from what I hear after water hits your CPU it's only a few degrees hotter, so I'm pretty sure that you could build a high end 1 loop system that will outperform a 2 loop system for less money.
Click to expand...

Ok thanks for the input...there still leaves the question about which cpu block if I'm going to use a eheim 1048(this pump is pretty quiet isn't it??)
 
Go to petsmart and buy a pump, thay are cheaper and are pretty quiet (might need to do a bit of modding to get the tubing to fit). Also, If a pump gets loud I had an idea, using aluminum make a little box that can hold your pump, with 2 large holes (Fitted with rubber grommets of course) that tubes can go through, and line the box with the black sound dampening foam stuff, then clamp this box inside your case. Im probably going to do this once all my parts come in.
 
The 1048 is probably the quietest pump you can get, and in testing the difference between the mcp650 and the 1048 is only 1c difference. The 1048 is far quieter than the mcp650.
 
hehehe yah my mcp650 is damn louder than my mcp600 but for some odd reason i actually got ~2c lower with the mcp650 which everything ive read seems to contradict.
 
steven1110 said:
Ok thanks. Still the block....swiftech or Danderden??
Click to expand...

If you get the 1048 you will probably get slightly better temps with the swiftech 6000, the DD TDX (#5 nozzle) wont be far behind however.

The difference between the two blocks isn't very much...so I guess it is personal preference.
 
2 resevoirs will give you larger volume of water and give you more that wont move as much
it will give a little lower temps too
 
blindviper said:
2 resevoirs will give you larger volume of water and give you more that wont move as much
it will give a little lower temps too
Click to expand...

technically yes, in reality no. A second resevoir might add maybe one litre to the overall volume of the system. Now remember that a decent loop will be pumping the water at around 3-6lpm. That means that the water is recycling quite fast, the water doesn't really have much time at all to sit in the resevoir and passively radiate off its heat. You will not notice any performance increase from a second resevoir, or really one for that matter. The added tubing of connecting a second resevoir might even add enough extra restriction to the loop to raise temperatures.

The usefulness of a resevoir is in filling/bleeding the loop, not in the fact that it gives the water a "rest" period. (The water never really gets to rest in the resevoir anyways)
 
Erasmus354 said:
technically yes, in reality no. A second resevoir might add maybe one litre to the overall volume of the system. Now remember that a decent loop will be pumping the water at around 3-6lpm. That means that the water is recycling quite fast, the water doesn't really have much time at all to sit in the resevoir and passively radiate off its heat. You will not notice any performance increase from a second resevoir, or really one for that matter. The added tubing of connecting a second resevoir might even add enough extra restriction to the loop to raise temperatures.

The usefulness of a resevoir is in filling/bleeding the loop, not in the fact that it gives the water a "rest" period. (The water never really gets to rest in the resevoir anyways)
Click to expand...


Exactly and well put. To many poeple have the idea that a resivour will ad to the cooling of their system. But it doesnt. The only real ways to enhance the cooling is via the heat exchanger (radiator). Whether it be using one thats larger, adding more fans or more powerful fans or be it chilling the Radiator itself. But thats the only true way to make the H2o system more efective. Including combining all of the above ideas.
 
I'd go with the Danger Den TDX, it performs really well and looks cool to boot.
 
QUESTION: It was my understanding that, with an inline system, once the system is closed the amount of pressure needed to move the liquid is reduced.

HOW TRUE IS THIS??

Back on topic: I also thought, that the more water you have in a closed loop system, the more energy it will take to raise the overall temperture. So, if the water temp is at a constant 40C under load [for example] with one resevoir ........ Then adding another resevoir would decrease the overall water temp. Because of the increased volume and the more energy required to heat that additional volume of liquid.

Another example ...boiling water in a pot, on a stove, If it takes X amount of energy to boil the water over Y amount of time. then increasing the volume of water would require more energy and more time. RIGHT?

This is my understanding of heating\cooling properties of water ..... Could someone clarify or provide more information.
 
Dtech : It would take longer for the water to reach its relative equilibrium points for certain cpu load situations, but it will not change the overall temperature. Having 10% more water means that instead of taking 10minutes for the water to level out at idle temp of +3C over ambient, it might take 11minutes. It will not however reduce the temperature of the water.

The temperature of the water is dependent upon the amount of heat you are putting into the loop (from the cpu, pump, gpu and whatever else) and the amount of heat that is radiated out of the loop (mostly through the radiator). The amount of water doesn't really change that difference between heat in and heat out.
 
killernoodle said:
The 1048 is probably the quietest pump you can get, and in testing the difference between the mcp650 and the 1048 is only 1c difference. The 1048 is far quieter than the mcp650.
Click to expand...

Want to provide some linkage on that? I've seen testing that shows a bit more than that: http://forums.procooling.com/vbb/showthread.php?t=10825&page=1&pp=25

Of course, the more stuff you have in your loop (like a GPU block or NB block) the worse that's going to be.

Also, those who are recommending a TDX with the #5 restrictor plate are recommending a fairly restrictive block. If you're going to run a pump with low head pressure, a less restrictive block (like the swiftech 6000) would be a better choice.
 
I run a system with the eheim 1250 (300gph). Its dead silent, like the smaller ones. I split the output to have one line go cpu>chipset>dual 120mm rad>res>pump. the other line goes GPU>HDD>large passive external radiator>res>pump. So with one res and pump I have two loops and two radiators. Works great!
 
so heres q ?
would to radiators help drop temps
like you go res->pump->rad->cpu block->rad->res(to close loop)
 
Gun_Strife said:
so heres q ?
would to radiators help drop temps
like you go res->pump->rad->cpu block->rad->res(to close loop)
Click to expand...

You would be better of going a larger radiator and going res>pump>rad>cpu>res

OR if you were going to do 2 radiators then go res>pump>rad>rad>cpu>res

cooling before it goes into the res and after the cpu does nothing benificial for the system. One thing to remember is the pump wil induce heat into the water so cooling before the pump will defeat the purpose. I would highly recomend to always cool (run through the rad) just before going into the cpu. As it will yeild you the coolest water to that block.

The Cpt.
 
A typical resivior will not benefit your temps unless you only run your computer for 15 minutes at a time. If your computer never turns off, in the "big picture" it wont do shit.

Unless of course your resivoir is a yard sqaure and 1/16" thick and made of metal or something, but then at that point it's essentially ceased to be a mere resivoir.
 
Captin Insano said:
You would be better of going a larger radiator and going res>pump>rad>cpu>res

OR if you were going to do 2 radiators then go res>pump>rad>rad>cpu>res

cooling before it goes into the res and after the cpu does nothing benificial for the system. One thing to remember is the pump wil induce heat into the water so cooling before the pump will defeat the purpose. I would highly recomend to always cool (run through the rad) just before going into the cpu. As it will yeild you the coolest water to that block.

The Cpt.
Click to expand...

Insano, it might do you good to read the post about the configuration of your loop. It doesn't matter where the radiator is in the loop, it will still be beneficial to cooling :p

And yes, I do agree that one double 120 radiator would probably be better than two single 120 rads. But then again, there are always space constraints to consider. (also if you go with the two single radiators, consider running them in parallel so that the pressure drop from the two of them is lessened)
 
sideopen.jpg

topside.jpg

This was from my old setup that i sold, acura heater core for cpu and chevette for 6800. Two rads worked better then one 360 rad (3 120mm). Both loops went into one dual bay res and did not see a temp difference using two seperate (res). As i said this setup is gone but my dual loop aquacomputer v2000(alot more room then v1200), SLI, BLAH BLAH BLAH is waiting on its fx-53 and that will have evo 360 for 2 6800's and chevette for HDD coolers and cpu. So ill post temps and pics then.
 
Erasmus354 said:
Insano, it might do you good to read the post about the configuration of your loop. It doesn't matter where the radiator is in the loop, it will still be beneficial to cooling :p

And yes, I do agree that one double 120 radiator would probably be better than two single 120 rads. But then again, there are always space constraints to consider. (also if you go with the two single radiators, consider running them in parallel so that the pressure drop from the two of them is lessened)
Click to expand...


Read what you were talking about long ago and thats BS. If you put the cpu at the end of any loop it will recive the least cooling benifit from the system. The water pump dumps heat the Vid card dumps heat and then if I were to run it to the cpu. temp of water is allready elevated at that time. The cpu gets shit for cooling. Thats why I belive in running LARGE radaitor with Lots of CFM or chill it and run it res>pump>radiator>cpu>vid card>res. To some they say it will only yeild you maybe 1 to 5 degrees celsius but that can mean even enough for that stable oc.
 
Captin Insano said:
Read what you were talking about long ago and thats BS. If you put the cpu at the end of any loop it will recive the least cooling benifit from the system. The water pump dumps heat the Vid card dumps heat and then if I were to run it to the cpu. temp of water is allready elevated at that time. The cpu gets shit for cooling. Thats why I belive in running LARGE radaitor with Lots of CFM or chill it and run it res>pump>radiator>cpu>vid card>res. To some they say it will only yeild you maybe 1 to 5 degrees celsius but that can mean even enough for that stable oc.
Click to expand...

Would you like me to explain the physics behind my reasoning Insano?

Lets say that your processor outputs 100W (100J/s) and the typical pump will output about 10W (10J/s) into the loop. The average loop will be running at say 3 lpm, or converted for ease of calculations to come 0.05 lps. One litre of water is ~1kg. The specific heat of water is approximately 4.186J/g*K. Specific Heat is defined as the amount of energy required to raise the temperature of 1 gram of a substance by 1 degree kelvin. Now for some calculations....

The formula for calculating temperature rises used will be as follows : Q = mcΔT
m = mass, c = specific heat capacity, ΔT = Change in temperature.
Note : Because the change in temperature is the variable we want, in this case simply divide Q by mc to get the desired result.

  • Temperature increase between inlet and outlet as a result of the pump :

In one second the pump is inputting heat (Q) of Q = (10J/s)(1s) = 10J
In one second you have 0.05 litres flow through the pump so mass (m) = (0.05l/s)(1s)(1000g/l) = 50g

Now with this information we can perform our calculation :
ΔT = 10J / [(50g) * (4.186J/g*K)]
ΔT = 0.0477 K = 0.0477 C

So now we just found that the temperature increase the water experiences from the pump is about one half of one tenth of a degree centigrade...hardly enough to cry over.​

  • Temperature increase between inlet and outlet as a result of the cpu :

I am not going to bore you with these calculations. I will simply explain that because the values I assumed for CPU power output were 100W, or 10x larger than that of the pump I assumed, the resulting temperature difference will also be one order of magnitude larger.

ΔT = 0.477 K = 0.477 C

The temperature difference coming out of the cpu is less than half of a degree centigrade.​

NOTE : The flow rates that I took are for a low flow system, using 3/8" ID hosing. High flow 1/2" systems produce almost twice as much flow, meaning twice as much water traveling through in one second, meaning half as much increase in water temperature.​

I hope that I have gone into enough detail for you this time Insano. In the most extreme circumstances, a person may have a difference of two degrees centigrade from beginning to end in their water loop. You will not experience any significant gains by sending the water through a radiator right before sending it to the cpu.
 
First off Erasmus354 you say the 2 rads will help, what config would u have it. Second I say 2 120mm rads instead of one dual 120mm rad because what cause can do that if you know let me know. Because I planned on going with the shark case and big water then buying second rad to mount in front. Here is my current setup
http://www.hardforum.com/showthread.php?t=848567
which you will see I am also selling it because it won't fit in new case.
Which I had it configured wrong, now I have it res->pump->rad->cpu->res and am not using 6800 block and am seeing temps of 35~36 idle 43~46 load(varies on load) usually call of duty for 4 hours or so and thats at about 43c.

Also Aviddigi how was yours setup
 
sorry, my post was a little off the topic of the thread, it was in response to Insano holding by the fact that you should configure the loop to run through a radiator before going to the cpu. As to the fact of two radiators. As I said before, if you have room for it I believe a dual 120mm radiator would outperform 2 single 120mm radiators. If you dont have room for it, but can fit 2 120mm radiators into your case, then great go for it. The second radiator will give you some added performance.

When you configure your loop, keep in mind my proof that you dont need to worry about what goes where in the loop. Read this thread : proper loop configuration
 
Erasmus354 said:
technically yes, in reality no. A second resevoir might add maybe one litre to the overall volume of the system. Now remember that a decent loop will be pumping the water at around 3-6lpm. That means that the water is recycling quite fast, the water doesn't really have much time at all to sit in the resevoir and passively radiate off its heat. You will not notice any performance increase from a second resevoir, or really one for that matter. The added tubing of connecting a second resevoir might even add enough extra restriction to the loop to raise temperatures.

The usefulness of a resevoir is in filling/bleeding the loop, not in the fact that it gives the water a "rest" period. (The water never really gets to rest in the resevoir anyways)
Click to expand...
I didnt say how big the res would have to be either :)
 
Very interesting stuff so do you know of ne case with 2 120mm fan slots on it (together that is) because the shark has 1 in front and 1 in back which is why I would go with 2 seperate ones
 
blindviper said:
I didnt say how big the res would have to be either :)
Click to expand...

Well, the point at which a resevoir becomes beneficial to cooling is the point at which you start calling it a tank, or a bathtub full of water rather than a resevoir ;)

Gun_Strife said:
Very interesting stuff so do you know of ne case with 2 120mm fan slots on it (together that is) because the shark has 1 in front and 1 in back which is why I would go with 2 seperate ones
Click to expand...

No I dont know of any case like that, perhaps the yeong-yang cube case? Not too familiar with that one though. Generally a double 120mm radiator/heatercore requires a bit of modification to the case in oder to mount it. The most common mounting positions would be the top ceiling (sacrificing perhaps the top 2 5 1/2 drive bays, you could still use a rheobus in the slots though) and in the front of the case, which would require taking out drive cages in order to make room.
 
You must log in or register to reply here.
Back
Top