Finally going for a custom loop

Look, forget the maths. There are two types of radiators, high a low flow. If you're going for high fan speeds use low flow, if you're going for low fan speeds use high flow.

Every set up is different. The pump pressure / water flow is actually one of the most significant factors.
 
That doesn't really help my situation.

In 2 reviews the same radiator is being tested at 1.5gpm with a constant heatload. So that is the same flow rate and wattage of heat applied to the loop. at 1400rpm, one review says that it will take away 200w of heat and the other says that it will take away 450w. I don't believe that difference is because the fans have different static pressures. Not 250w different. No way.

I'm focusing on this so much because if I want to cool CPU and GPU in a loop and remain under delta t of 10. Either I will be fine with what I have picked out or I will need double what I picked out. I don't want to have bad performance for a custom loop because I didn't get enough cooling for the system.

I'm going to assume that you mean FPI rather than flow when talking about radiators. I understand that you can have higher density (in fpi) radiators with higher static pressure fans and get better performance. Or have double thick radiators with low FPI and slower fans for a quieter system.
 
Look, the best cooling will be 2 x 140mm fans so get one rad for each component. That will keep it within 10 delta.

If you want to work it out properly, tell me the C/W of the rads.
 
Last edited:
Buzz1927 said:
Responded but not answered the question. I've pulled the blocks apart and they're all fine, HOWEVER, they're not EK blocks (which your whole argument is based on).
Click to expand...

No, you made the EK comment not me. Seconldy the EK issue had nothing to do with corrosion from the coolant.

I made the comment about how silver is completely uncessary and can only do harm from a corrosion perspective. You may currently be fine, yet we're only taking your word for it, no photos?

I have also backed it up with quite a lot of evidence and counter evidence.

Just because its 'what people do' doesn't make it logical.

Almost all non-ionic coolants are also biocides thus the silver debate is pointless. It can ONLY increase the chance of galvanic corrosion and isn't required for biocidal properties.

But, i think this topic is going off course. Hes looking for info regarding the size of rads required. He can make up his own mind about the rest.
 
bigfellla said:
No, you made the EK comment not me. Seconldy the EK issue had nothing to do with corrosion from the coolant.

I made the comment about how silver is completely uncessary and can only do harm from a corrosion perspective. You may currently be fine, yet we're only taking your word for it, no photos?

I have also backed it up with quite a lot of evidence and counter evidence.

Just because its 'what people do' doesn't make it logical.

Almost all non-ionic coolants are also biocides thus the silver debate is pointless. It can ONLY increase the chance of galvanic corrosion and isn't required for biocidal properties.

But, i think this topic is going off course. Hes looking for info regarding the size of rads required. He can make up his own mind about the rest.
Click to expand...
Very funny, everything you quoted has been from EK, why don't you start a "bigfella is talking out of his arse" thread? Would be quite popular, I think..
 
bigfellla said:
Look, the best cooling will be 2 x 140mm fans so get one rad for each component. That will keep it within 10 delta.

If you want to work it out properly, tell me the C/W of the rads.
Click to expand...

I have actually heard differently to this. Speaking as someone not hugely clued up on water cooling, i thought the general consensus when considering rad sizes is number of components + 1, so for CPU + GPU a 360 rad rather than 240
 
Ok here are the graphs that I am getting confused with

First the tests from Skinnee Labs
HWLabs-SR1360_CWChart.jpg

HWLabs-SR1360_10cDelta_Chart.jpg


Here is the same radiator tested on Martin's Liquid lab

hwlabs-sr1-360-th1.png


The graphs for MLL:

[http://martinsliquidlab.files.wordpress.com/2012/04/radthermalbarcharts5.png?w=614&h=2277

Same radiator, same fan speed, same flow rate. Drastically different C/Ws.

Hopefully that will finally illustrate why this is frustrating to me and why I can't drop the math portion.
 
Last edited:
Aastii said:
I have actually heard differently to this. Speaking as someone not hugely clued up on water cooling, i thought the general consensus when considering rad sizes is number of components + 1, so for CPU + GPU a 360 rad rather than 240
Click to expand...

Yeah thats pretty true but for his case and his requirements of delta 10oC, i would suggest 2 x 360mm rads (what i meant by 2 x 140mm x 2).

At salvage-this, yes those calcs look pretty similar to my assumptions of C/W 0.07 (ideal) with a coefficient of 3 (3 - 10oC delta). So yeah, I would go with 1 360 rad per component if you want 10oC delta or less.

You can always try one rad, and if that doesn't work, the D5 Lang pump should be able to handle 2 later if you desire/need it.

If it were me, I would make one loop with CPU and GPU with 360 Rad if possible and go from there. All the maths in the world will only get you so far. You need to build it and then see. ;)
 
Last edited:
^Yes.

@Bigfella. Yeah it looks like I will have to just have to try it and see. Too bad I'll need to get a temp sensor for my res. I just want to know how it will really work out.

Originally I was going to use Swiftech's XP line because it was a good fit for the fans that I have. Would I be better off getting a thicker radiator and getting some better optimized fans for high FPI rads? Even though this case is huge it does not have the most area for watercooling. At the most could probably get away with one 360 in the top and 2 140s. One on the back and one by the HDD cages.
 
Last edited:
You wont need a temp sensor for the rad, the system will achieve equilibrium, that means eventually everything in the loop will be the same temperature. So just use the GPU diode.
 
Okedokey said:
You wont need a temp sensor for the rad, the system will achieve equilibrium, that means eventually everything in the loop will be the same temperature. So just use the GPU diode.
Click to expand...
It won't come to thermal equilibrium (which is a fairly murky term to begin with), it will reach steady state at which point each component will have a constant fluid entrance and exit temperature, assuming, of course, that the computer components are generating a constant amount of heat.
 
Last edited:
Yeti said:
It won't come to thermal equilibrium (which is a fairly murky term to begin with), it will reach steady state at which point each component will have a constant fluid entrance and exit temperature, assuming, of course, that the computer components are generating a constant amount of heat.
Click to expand...

What bloody nonsense, a closed loop will reach thermal equilibrium. Not a murky term at all.
 
If an isolated system is left long enough, it will reach a state of thermal equilibrium in itself, in which its temperature will be uniform throughout, but not necessarily a state of thermodynamic equilibrium.
 
Okedokey said:
If an isolated system is left long enough, it will reach a state of thermal equilibrium in itself, in which its temperature will be uniform throughout, but not necessarily a state of thermodynamic equilibrium.
Click to expand...
It is not an isolated system as there is heat being generated. Heat transfer deals with heat flows and temperature differences. As I already said each component will have its steady state (a much better term than equilibrium) entrance and exit temperature given a specific heat load.
 
Either way, the system will get to a 'steady state'. Its complete overkill to worry about monitoring individual system temperatures, as eventually, they will be the same.
 
Okedokey said:
Either way, the system will get to a 'steady state'. Its complete overkill to worry about monitoring individual system temperatures, as eventually, they will be the same.
Click to expand...
Yes, 'steady state' - you can look it up in any heat transfer textbook. So to be clear, by saying that "it will have a uniform temperature throughout" and "everything in the loop will be the same temperature," are you implying that the water exiting the cpu block is the same temperature as the water exiting the radiator?
 
You must log in or register to reply here.
Back
Top