GPU Waterblock feasible?

gcaster

gcaster

gcaster

Associate
Joined
18 Jan 2011
Posts
1,187
Location
South London
Hi guys,

Another watercooling question I'm afraid. My loop is currently CPU only but I'm looking to add my GPU to the loop at some point in the near future.
Only problem is I'm not sure it's going to be feasible with my system as it is and may have to buy another 120mm rad.

I currently use a 120.2 EK coolstream rad which keeps my CPU pretty chilly but fear by adding a GPU block I'd be putting too much heat through the one radiator.
Here's a pic of my rig, I was thinking I could simply put a 120mm stealth rad behind the rear 120mm fan and have the loop going:

Res/Pump > 120.2 Rad > CPU Block > 120 Rad > GPU > Res/Pump

Is this enough radiator? My CPU temp reaches max temps of 67C so there's a bit of headroom there but on stock cooling my 6950 gets pretty damn hot (90C on full load in FurMark with the fan on auto hitting ~40%). If so should I buy the new 120mm rad alongside the GPU block or will I be able to hold out and buy it a little later?

Also, is a GPU backplate necessary/worth the cost or is it purely aesthetic? Not too sure I see the point in them! :rolleyes:

Cheers
G
 
Oh and what 120mm rads would people recommend? I could probably fit a regular one in if the stealth rads are no good but would prefer stealth.
 
A shroud between your rad and fans, (so all the air is pulled through the rad and not from around it) would make quite a lot of difference.
 
The fans above the current rad are 200mm beasts so still pull a good amount of air through. I'll no doubt make a shroud at some point but I'm a bit too busy for that at the moment.
 
Last edited:
Without knowing your cpu or its overclock/voltage, your 67C could be good (your rad has spare cooling capacity but you're limited by your block) or bad (your rad is already struggling). All water-blocks (and heatsinks) have a cooling capacity which cannot be changed. Whatever the inlet temp, the outlet will be xC higher for a given heat input in Watts. This is normally written as C/W.

My point is that you may not need another rad and indeed adding the extra resistance might negate any benefit.
 
gcaster said:
The fans above the current rad are 200mm beasts so still pull a good amount of air through. I'll no doubt make a shroud at some point but I'm a bit too busy for that at the moment.
Click to expand...



If you could focus all the power of the fans to pull air through the radiator, then you'd be removing more heat from the loop, possibly leaving you with enough cooling capacity to add your GPU block.
 
MikeTimbers said:
Without knowing your cpu or its overclock/voltage, your 67C could be good (your rad has spare cooling capacity but you're limited by your block) or bad (your rad is already struggling). All water-blocks (and heatsinks) have a cooling capacity which cannot be changed. Whatever the inlet temp, the outlet will be xC higher for a given heat input in Watts. This is normally written as C/W.

My point is that you may not need another rad and indeed adding the extra resistance might negate any benefit.
Click to expand...
My CPU and overclock are in my sig (2500k @4.8GHz).
The radiator (EK Coolstream) still runs cold so I imagine it'll be the block (EK Supreme LT) holding it back.
I'll try running it without the extra rad first and see what sort of temps I get before potentially adding the second rad. My pump isn't the best so it may be the case that the extra rad isn't an option until I upgrade the pump.

Mick-1965 said:
If you could focus all the power of the fans to pull air through the radiator, then you'd be removing more heat from the loop, possibly leaving you with enough cooling capacity to add your GPU block.
Click to expand...
As I say I will do this at some point but don't really have the time to make a custom shroud right now. What methods do people use for making shrouds and what materials are used? I was planning on knocking one up in CAD and then send the design off for it to be made in plastic.
 
Shroud designed and sent off for a production quote.
It's only about 7mm tall so not sure how much affect it'll have, just gotta hope they can make it thin enough without it being so flexible that it vibrates.
 
Many people (such as me) turn off the display of signatures, hence my question. If the rad is cool, then you've likely reached the limits of the block.

And if the rad is cool, then a shroud is unlikely to help much. rather than trying a shroud, you could just try using more or more powerful fans and see if the temps drop at all. If they don't or only marginally, then a shroud will make little difference until you can get the water hotter. A gpu block will achieve that .
 
Well, the GPU block has gone into the current loop after the CPU.
Idle temps for both now sit at around 28C for the GPU and 30C for the CPU.
Running Furmark the GPU hit 52C but stabilised at around 48C which is a whole 42C lower than with the stock cooler (and a hell of a lot quieter)! CPU did rise a bit though but obviously this is expected as it also gets loaded.
I haven't put an extra rad in yet and not sure I will. I now need a program that can stress both my CPU and GPU to 100% simultaneously then I can check max temps of both and see the effect adding the GPU block has had on my CPU temps.
 
Looking at the pic,
Just a couple of things, you've got the block on side ways. the barb closest to the edge should be at the top. ( avoids getting air trapped in the block )
and assuming that's the pump at the front with the res at the back you've got the feed to the cpu block back to front, usually water should flow into the barb closest to the centre to the outlet barb which is the furthest from the centre( just the way most are designed to be more efficient)
 
What?!
Mick-1965 said:
you've got the block on side ways. the barb closest to the edge should be at the top. ( avoids getting air trapped in the block )
Click to expand...
I haven't, see here. There's no air trapped in it, I can see that (my reason for using coloured fluid and a clear block).
Also, the etching reads left-to right as it is, I imagine it would be on it's side if the block was meant to be run top-bottom.
Mick-1965 said:
assuming that's the pump at the front with the res at the back you've got the feed to the cpu block back to front, usually water should flow into the barb closest to the centre to the outlet barb which is the furthest from the centre( just the way most are designed to be more efficient)
Click to expand...
Surely left to right makes no difference in a CPU only loop as the block is symmetrical? I am now running right-to-left but only to ensure I pass through the radiator immediately before the CPU block.


These seem pretty strange claims!
 
Tested both the CPU and GPU simultaneously now and the CPU got pretty damn hot although in real world use I doubt it'd ever see temps like this so will keep the loop it as it is at the moment. Radiator got reasonably warm so will look to see if I can get that shroud made cheap enough to be worthwhile.
Cheers

 
Am I right in thinking this is your block http://www.overclockers.co.uk/showproduct.php?prodid=WC-012-EK

The closest barb to the centre is the inlet, ie the water coming out of the pump needs to be connected to this. Good water blocks are designed to be most efficient when flowing in one certain direction.

Orientation doesn't matter performance wise, It just makes for easier bleeding if the barb furthest to the outside is at the top. Difficult to tell from the picture but it looks like there's air trapped in there. Thats why I mentioned it :)
 
Mick-1965 said:
Am I right in thinking this is your block http://www.overclockers.co.uk/showproduct.php?prodid=WC-012-EK
Click to expand...
No mine's the LT version.

Mick-1965 said:
Orientation doesn't matter performance wise, It just makes for easier bleeding if the barb furthest to the outside is at the top. Difficult to tell from the picture but it looks like there's air trapped in there. Thats why I mentioned it :)
Click to expand...
MikeTimbers said:
Unless that pic was taken before bleeding, there is definitely air trapped in there!
Click to expand...
That photo was taken about a minute after the loop was assembled so hadn't been bled. All the air is out of it now.
 
You must log in or register to reply here.
Back
Top Bottom