water cooling setup, anything missing?

vapor matt

vapor matt

vapor matt

Associate
Joined
19 Jan 2003
Posts
2,498
Location
west sussex
here is what i want for a water cooling setup, can anyone see anything missing?

XSPC RX360 X 2 120mm Rad
Aqua Computer Aquafx gtx 680 x 2
XSPC D5 Vario with front cover x 1
Koolance CPU-380I x 1
Bitspower Z-Multi 400mm water tank tube x 1
Koolance Radiator mount x 1
Aqua ComputerTwinC connect x 1
XSPC HighFlex Hose x 4
mayhems emrald green uv dye x 1
Bitspower Premium G1/4 stubby high flow black 1/2" x 4
Bitspower Premium G1/4 high flow barb 1/2" fitting x 4
EK-PSC Adapter Rotary adapter45 deg G1/4 x 4
 
Last edited:
single triple rad for 680SLI + CPU?
you will be running some faster fan speeds to keep all of that cool (which you've not listed any fans either)

you've also listed some dye, but not fluid, so I'm guessing you are using distilled water, it would also be a good idea to add some biocide, or preferably an additive that includes a corrosion inhibitor as well
 
single triple rad for 680SLI + CPU?
you will be running some faster fan speeds to keep all of that cool (which you've not listed any fans either)

you've also listed some dye, but not fluid, so I'm guessing you are using distilled water, it would also be a good idea to add some biocide, or preferably an additive that includes a corrosion inhibitor as well
Click to expand...


thanks for that, but i have fans, and the rad will be mounted outside so should keep nice and cool. plus distilled water and anitcorrosion aditives.

I thought that rad would cool everything ok as its a top rad with good flow rates.
 
a single 360 is very tight for 3 components

rough rule of thumb is 110-120watts per 120mm on medium fans, even the best rads will struggle to do 200watts per 120mm with fast fans

you've got say 150W in CPU, plus say 200W each for graphics = 550W
more if you are wanting to push overclocks

I think you are going to be looking at 2000RPM fans to keep everything to a 10C water delta
your graphics cards will be fine as GPU blocks tend to give very good water to GPU temps, but the CPU will struggle on a warm day if you are wanting to push a big OC out of it

that wouldn't be an acceptable noise level for me, but it depends what you are trying to achieve with watercooling :)
 
a single 360 is very tight for 3 components

rough rule of thumb is 110-120watts per 120mm on medium fans, even the best rads will struggle to do 200watts per 120mm with fast fans

you've got say 150W in CPU, plus say 200W each for graphics = 550W
more if you are wanting to push overclocks

I think you are going to be looking at 2000RPM fans to keep everything to a 10C water delta
your graphics cards will be fine as GPU blocks tend to give very good water to GPU temps, but the CPU will struggle on a warm day if you are wanting to push a big OC out of it

that wouldn't be an acceptable noise level for me, but it depends what you are trying to achieve with watercooling
Click to expand...


what if i added another rad the same, would one pump still be ok?
 
price wise, I would look at a bigger single rad than running 2 smaller ones... a bigger rad will probably only be a tiny bit more expensive, where as 2 rads are double the price (obviously)

you could always switch to a DDC 18w pump (better head pressure) if flow was a concern
just get some foam or a shoggy sandwhich to mount it on

(I personally run 2 EK DCP 4.0's in series, but these are a bit trickier to mount if space is limited)

I really like Phobya rads for their price/performance, they have a quad 140mm that would be good for your purposes, or alphacool do some double / triple 180mm rads that are very good - with any of these you can then run slow (read cheap) fans as well

if you are mounting externally anyway then it opens up a lot of options
 
Last edited:
Doubling the radiator surface area will halve the air-water temperature difference. So if you're at the 10 degree point andy is (sensibly) advocating, it'll take five degrees off your cpu temperature or let you slow the fans down a fair bit. Of course, you'll also have twice as many fans.

Sticking the radiator outside is a serious game changer in winter. The heat it'll get rid of is directly in proportion to the air-water difference. If you can take the inlet temperature down to 5 centigrade, and you're aiming at a water temperature 10 degrees over a 25 centigrade room, you now have a 30 degree delta in the radiator, and it'll dump 3 times what it would in the room.

Sticking the radiator in direct sunlight in the middle of summer is unwise for similar reasons.

Watercooling is a balancing act between number of radiators (hence cost), noise of cooling (hence cost) and water temperature. That's the fun of it really. I just spent a happy hour trying to determine whether (radiator-fan-radiator) is a significant improvement on (fan-radiator) - and the answer is mostly a function of water temperature.
 
intresting stuff, well the rad will go outside, as will the pump. I will try 1 rad at first and if that dosnt do what i want cooling wise i will install a 2nd loop with another rad and pump ect for gfx cards again outside. Rads and pumps would be in complete shade all year round, the whole idea is to get a quiet as possible PC inside. fan noise is not an issue really as i wont hear them.
 
if it is going outside the house then you'll need to be wary of condensation inside your case if there is too big of a difference between in house air temp / humidity and outside / water temp
 
ubersonic said:
A quick Google would have told you it actually reduces performance
Click to expand...

A quick Google is not always better than thought. The link tells me that high fin density and low speed fans makes for a bad sandwich. That isn't exactly astonishing. Fortunately the same site has some useful information about low fin density radiators as well:

hwlabs-sr1-360-th1.png


A single radiator takes the air temperature up by 7 degrees out of the maximum possible 10. A second radiator can be reasonably expected to achieve the same assuming a constant airflow, i.e. 0.7*3=2.1 degrees. Note the efficiency going down with air velocity, the premise here is essentially that the air efficiency is significantly below 100.

At a 10 degree delta, I think the second radiator is the difference between 300W and 400W of heat. So it cannot halve the C/W as separate radiator would, but could knock 25% off it. If that's the difference between a 20 degree air/water delta and a 15 degree one, for decreased noise*, I'm interested.

*Fan speed goes up a bit to compensate for the extra radiator. I have decent 38mm fans in mind, so pressure is OK. Not much dampens fan noise quite as effectively as bolting a radiator to both sides, it's a lot of mass to vibrate up and down and the airflow is drawn through a load of channels. The only real worry is whether a high fin density radiator would be better, sadly martin's lab only tested the high density ones with 25mm fans.
 
what you are missing Jon, is that 2 rads sandwhiched mean that the fans are pulling air from a hot radiator and then pushing that hot air over the 2nd rad, which means that the air/water difference is then less meaning that the 2nd rad will be far less efficient at releasing its thermal mass

there have been multiple tests of sandwhiched rads and all of them show that adding a 2nd rad to a sandwhich improves temps by as little as 1-2C, where as moving the rads to be side by side instead of in a sandwhich improves temps by say 5-10C

if you look at the chart you posted you can see that the air from the first rad gains 7C in passing over the radiator, so the temp difference of the air to water on the 2nd rad is going to be a tiny fraction of what it was on the first rad

your argument about using it to cut down fan noise is nonsensical, because you could use far far quieter fans in the first place, 2 rads side by side and end up with much lower temps, or in fact you could just use 1 bigger rad in the first place and save even more money

2 rads side by side are always exponentially better than 2 rads sandwhiched
 
Last edited:
I may be missing something, but it wasn't that. In the example, the first radiator gets air at ten degrees below water temperature, and shifts 300W into it by raising the air temperature seven degrees. The second radiator gets air at three degrees below water temperature, and shifts 100W into it by raising the air temperature two degrees. The second radiator is not even close to as efficient as the first, but it isn't at zero either.

I'm fully in agreement that supplying both radiators with ambient air will work better than supplying one with ambient air and one with hotter-than-ambient-but-cooler-than-water air. The only reason to stack the radiators is to meet geometric constraints where no more airflow is available. I think a second radiator, sandwiched, has a reasonable chance at taking air/water delta from 20 to 15 for the application I have in mind.

edit:
Water temp: 30 degrees
Ambient: 20 degrees
Air between radiators: 27 degrees
Air after second radiator: 29 degrees

edit2:
Because the stacked arrangement does poorly with low pressure fans, it follows that the response curve from turning a fan from minimum up to maximum will be rather nonlinear. Turning the fans down to achieve quiet would be likely to push temperatures up rather faster than normal.

edit3:
We're not considering the same problem. What's your plan if one radiator isn't sufficient, but there isn't space for a longer one or multiple radiators, only for a thicker one? Re: fan noise, bolting a radiator to either side of a 38mm san ace does a pretty convincing job of quietening it. It's louder than a 1600rpm fan, but quieter than it is with one radiator removed.
 
Last edited:
Basically, a sandwich is just a bad idea, you can get better performance in the same space by using either a double thick rad or push/pull fans. Hell rad+rad+fans = worse performance than just rad+fans as you end up starving both rads of airflow in addition to blowing the heat over the second rad.

But that's been debated and solved in other threads in the past, and isn't really relevant to the OP.
 
Last edited:
We might not disagree as much as you think then ubersonic.

If you agree that going from 30mm thick to 60mm thick can be an improvement, why not from 60mm thick to 120mm thick?
 
my plan in this circumstance would be to find more space, not spend double the money on a solution that other people have already tried and shown offers no improvement

going from a 45mm rad to a 60mm (same brand, same makeup) rad you gain around 10-20W of dissipation
that is a 3% increase in dissipation from a 33% increase in thickness

assuming this were linear, which I doubt, then doubling would lead to a 9% increase, not 25% as you've summised

and in another test, going from a 30mm to an 80mm, the difference was 24% from a 166% increase, which working backwards would also mean a 9% increase from doubling the thickness
 
Last edited:
well guys my 2 rads will not be in a sandwich, they will be side by side each other doing there own loop. one good thing about seperate loops is you can always take one down if you need to.

also being outside antifreeze will be in the loops, in summer the diffrence will be good, come winter to save condense i may have to turn the fans off completely, we shall see come november!:p

looking forward to puting it all together now, will post pics when finished.
 
You're quite right matt, apologies for the derailment. Everyone is in agreement that your two shouldn't be sandwiched :). Hopefully you'll forgive a closing statement.

The missing factor is fan pressure Andy, to achieve the 25% I'm hoping for, the fans will have to be good as well as fast. I'm happy leaving this as "agree to disagree", I'll run the numbers more carefully before buying a couple of 420mm radiators.

Connecting a fan controller to a thermometer might be worthwhile as a means of turning the fans down as outside temperatures drop. Turning fans down to keep water temperature at room temperature, without running a chiller, is rather nice!

Every time winter comes round I think about hanging radiators out of a window, but I am yet to find myself living in a flat where this is a realistic option. How are you planning on routing the tubing to outside? Drilling a hole in the wall is tempting, but maybe a bit heavy-handed.
 
You must log in or register to reply here.
Back
Top Bottom