Is this right?
 

I have been reading around quite a few different forums to find out if my temps are inline with others. I am now sure that mine are high based on previous posts and other forums.

I ran across this schema for proper water flow and mine is nothing like it. Mine is pretty much the exact reverse of what this shows. Could this be causing my higher than expected temps?

Here is the schema I found on amdmb.com's forums:

http://www.aaclan.op4ctf.com/files/pics/misc/schema.jpg

My temps are typically 18-25 degreez over ambiant.

Degrees C or F? If F, then they aren't that bad. If C, then....

As far as hookup... I side with hook the output of the radiator to the CPU block. That way the coolest water enters the CPU block. Also, all heat adding components (blocks and pump) are before the rad, meaning water entering the rad will be at its highest. In both cases this gives the largest DeltaT, meaning better cooling (basic thermodynamics, the larger the DeltaT the greater the heat flow, coolest water going into CPU block, hottest water going into the rad).

Yeah it's F on the ambiant temps. I will reconfig to see if this improves.

Right now i got Pump>GPU>CPU>rad>resov. and back

mine will be bleeder/pump/rad/cpu/ --> back

How are you measuring temps? If it is through MBM or something like that, the probe is inaccurate anyways.
-Kev

having the rad just before the cpu is the best, but I honestly doubt it makes that much of a differance, maybe one or two degrees

I am monitoring the temps via a DigiDoc5 and MBM. As you can see from the pic below, the probe is very close to the core and should be pretty accurate.

Should I try and lap the block? This has not been done yet.


http://www.aaclan.op4ctf.com/files/pics/misc/probe.jpg

Do it if you think its imperfect. And it probably is.

I should be able to tell by the imprint left on the block once it is removed. With my SK-6 you would see a perfect replica of the core with the text when removed. I am also going to remove the shim since that might be causing it not to seat completely.

I'll wait for the next Laker game and so I can lap while watching the game!

I sorta got side tracked because of this damn HTPC I'm building. Fun stuff for sure. I want to be the first one to water cool a HTPC!

Whats a HTPC?

The shim might be your problem, if the shim does not ahve a slot cut out for your temp probe.


If youa re stacking your probe on the shim, or vice versa, then you are prolly lifing that edge of your block off of the core.

shims are just a problem in general, they just suck

No, tell me how you really feel? I am going to lap it and remove the shim... This should take care of it!

I really really really hate shims

LOL.. I get the point.

Shim=bad

I love shims, they insure a good contact without crushing the core,

U can send it to me if ya want!!!!

they are good assurance they will go onto the core themselves meaning you will cook the cpu, or they will slide onto a trace cooking the cpu, or they will lessen the pressure so you will get worse temps, or they won't be thick enough so you will have a false sense of security, or they can be too thick so your cpu doesn't contact the heatsink at all.

they are training wheels for the 5 year old Kylie's, and Tom's of this world, get rid of them all :mad:

* Sends Brad a gross of shiny new copper shims to play with * :evilaugh:

Now I was refering to a good quality shim. that is siliconed to the chip to insure its placement.

Granted there are alot of crappy shims out there. But it doesnt take a rocket scientist to check over everything to make sure non of what you described occures. Trust me, I have a heavy hand and love to crank down bolts. Im used to it, working in an auto shop and machine shop, its second nature. Its really easy to crush a core. (POOR POOR 1.2 T-bird)

also if you are swapping out blocks alot for testing, then they are second to none in speeding things up and insuring accuracy.