Need some rad advice

wolvers

wolvers

wolvers

Soldato
Joined
26 Jan 2005
Posts
5,141
Location
In........cognito
Hi all,

I'm thinking of changing the Swiftech 120mm rad in my loop for something else and I need some advice on whether I should bother and which rad i should change it for.

Current loop is; >120mm rad+res>cpu block/pump>184mm rad>gpu>

The 120mm rad is only a single row rad but the 184mm rad is a black ice two row one and I've been wondering if I would see any benefit from getting a two row 120mm rad. Thermochill 120.1 (with or without a shroud) or a Black Ice GTX 120, which is a bit cheaper, or not bother at all as it won't make much difference?

Seems to me that it should have better cooling?

Thanks in advance.
 
Marginally. I doubt the overall cooling is up to it but do you know the temperature of the gpu? I really doubt that the cpu block is correctly mounted given the temperatures. I have a single Black Ice Extreme cooling an E6600 at 1.46v and 3300 and it keeps it below 50C at load.

Can you post more details? gpu? idle/load temps of both cpu and gpu etc.?
 
MikeTimbers said:
Marginally. I doubt the overall cooling is up to it but do you know the temperature of the gpu? I really doubt that the cpu block is correctly mounted given the temperatures. I have a single Black Ice Extreme cooling an E6600 at 1.46v and 3300 and it keeps it below 50C at load.

Can you post more details? gpu? idle/load temps of both cpu and gpu etc.?
Click to expand...

Spec in sig.

CPU is 38c idle, 70c load. Block is the Apogee Drive (block and pump combo).

GPU is about 53c full load. Can't remember the GPU idle temps. I'm sure that's good for an 8800GTS@660.

Gonna lap the CPU tonight.

Was just wondering if the Noctua fan on the 120.1 rad was pants for rads as the 2x92mm rad has Xilence red wings on and seems to be performing better. Maybe they're not comparable though. Got another 120mm fan, might try that tonight too.
 
wolvers69 said:
Spec in sig.

CPU is 38c idle, 70c load. Block is the Apogee Drive (block and pump combo).

GPU is about 53c full load. Can't remember the GPU idle temps. I'm sure that's good for an 8800GTS@660.

Gonna lap the CPU tonight.

Was just wondering if the Noctua fan on the 120.1 rad was pants for rads as the 2x92mm rad has Xilence red wings on and seems to be performing better. Maybe they're not comparable though. Got another 120mm fan, might try that tonight too.
Click to expand...

"Spec in Sig". You might be surprised to know that I have signature display turned off partly because I browse these forums from work sometimes and many of the sigs border on being work-unsafe and mostly because they're a waste of bandwidth!

So the cpu is 70C; the exiting water is then cooled by a 92.2 rad after which it can cool the gpu to 53C before going through a 120.1 and then "cooling" the cpu. I don't know the CW rating of the Apogee Drive but I would be very surprised if it couldn't manage better than a 27C delta so...

...I think the apogee drive is not making proper contact with the cpu.

What happens if the cpu is at load and the gpu isn't? Run 32M SuperPi or something which will only stress the cpu. If the temps are not lower than when you run 3DMark (for example) then the problem is not the rads but the cpu-Apogee contact. If the cpu load temp is much lower than better radiators might help.
 
MikeTimbers said:
...I think the apogee drive is not making proper contact with the cpu.
Click to expand...

Thanks for your reply.

E2160 (@3.2), P5K-E, 8800GTS (@660/951/1528) is the spec in question.

I've just finished lapping the IHS and checking that the Apogee Drive is seated properly, which I believe that it is, and the load temps in orthos (small ffts) have not improved at all.

I like your theory 'What happens if the cpu is at load and the gpu isn't? Run 32M SuperPi or something which will only stress the cpu. If the temps are not lower than when you run 3DMark (for example) then the problem is not the rads but the cpu-Apogee contact. If the cpu load temp is much lower than better radiators might help'.

What if the temp of the cpu is the same in both scenarios ie, with or without gpu load?
 
If the temps are the same in both scenarios it wouldn't make any sense at all. Is that what actually happened?

1) gpu at load not affecting temps would imply the 120.1 can remove all the gpu's heat

2) gpu at idle not affecting temps would suggest that the 120.1 is doing nothing to aid cooling.

They can't both be true.

OK. so let's assume that both radiators are 100% efficient. At idle, the cpu is 38C which is quite high. Take off the case sides and let ambient air hit the rads. Does the idle temperature go down? (Or are you in Australia?) If it doesn't, are you accurately measuring the temps?

It's entirely possible that the water temperature hitting the cpu combined with the efficiency of the Apogee minus the added heat of the pump bit results in a delta which seems high by conventional standards. But if turning off the gpu doesn't lower the cpu temperature when being cooled by both the 92.2 and the 120.1 then a bigger rad to replace the 120.1 won't help.
 
With the 3d view in ati tool running as well as orthos I get 54c gpu and 69c cpu (it actually fluctuates between 62c and 69c) so I don't get it.

It's 71c cpu and 43c gpu with orthos running on it's own! :confused:

I can, however, prove your theory about the 120.1 doing nothing to aid cooling. I've got another fan than runs at a much higher speed and should lower the cpu temp when running flat out, or not if the rad is not aiding cooling.
 
You must log in or register to reply here.
Back
Top Bottom