Apogee GT + Conroe owners Look inside.

Well I got thrown another wrench in my testing (thanks for that Gabe) and if you run conroe and an apogee GT you will want to pay attention. I conveyed to swiftech that the GTX is performing MUCH better than the GT on my bench, was told this should not be, not by as much as I'm seeing (more that 2c fwiw). Anyway, the key is to rotate the block! Duh, why did I never think of this? Of course this only works for socket 775 as the mount points are perfectly square. It seems the apogee GT performs best when the flow goes over the WIDTH of the die and not flowing the LENGTH. So far 2 mounts have confirmed this by me - better on the order of 1C or so. Now I will need to see what other blocks will show some improvement. So thanks for the wrench. Oh, incase you have no idea what im talking about, perhaps these pictures will explain.

This is the die orientation of conroe.


This is how I tested the 5 mounts of apogee GT - this is how I would normally mount every block as would most people me thinks.


Now here is the vertical mount which performs ~1C better.

Do you think this is because of the die's shape or its location?

It has to be the location. With the new orientation, there is a much more direct flow path of the water to the actual CPU die.

Hmmm, next time I flush my system maybe I'll redo my tubing to try it out.

OniKoroshi said:

Hmmm, next time I flush my system maybe I'll redo my tubing to try it out.

Click to expand...

Your loop is so tight you can't turn it 90degrees?

--pak

was the water block seated better with a thinner amout of AS5 the second time around?

in the first pic with the red squar does this squar indicate the orientation of the tow dies? Are they really THAT big?

Mycobacteria said:

was the water block seated better with a thinner amout of AS5 the second time around?

Click to expand...

Nope same same. In my testing of all these water blocks I perform 5 mounts of each block... the apogee was tested 5 times in horizontal position. First 2 runs of vertical position are ~1c better.

Not surprising at all TBH... and if the water is coming in through the top(where the processing core is) and out the bottom(where the L2 is), you should be getting the best possible results...

Now, mabe try a better thermal compound(Ga,Sn alloy is the best IMO) and strip that Conroe naked!

You should then see a good drop in tempatures... depending on how thick the copper to the Apogee is... since it'll be the IHS.

Shell said:

Not surprising at all TBH... and if the water is coming in through the top(where the processing core is) and out the bottom(where the L2 is), you should be getting the best possible results...

Now, mabe try a better thermal compound(Ga,Sn alloy is the best IMO) and strip that Conroe naked!

You should then see a good drop in tempatures... depending on how thick the copper to the Apogee is... since it'll be the IHS.

Click to expand...

Huh? I think you got the other way around. The first pic indicates the cores. Hence the second pic show normally what ppl would mount the block. However, nik found that if mount the block the UNUSUAL way which the horizontal way, it "gives" better temp.

Altho, i think it has more to do with the core temp sensor location than anything.

BTW, Conroe can be naked? I dont think so as i believe the IHS is soldered on the cores.

Very interesting Scott. With the four port outlet of the FuZion I wonder how the orientation impacts performance?

SiGfever said:

Very interesting Scott. With the four port outlet of the FuZion I wonder how the orientation impacts performance?

Click to expand...

You know I'm wondering the same thing Although, I don't see how it could.

MameXP said:

Huh? I think you got the other way around. The first pic indicates the cores. Hence the second pic show normally what ppl would mount the block. However, nik found that if mount the block the UNUSUAL way which the horizontal way, it "gives" better temp.

Altho, i think it has more to do with the core temp sensor location than anything.

BTW, Conroe can be naked? I dont think so as i believe the IHS is soldered on the cores.

Click to expand...

They can be lidded, but its a VERY risky operation unless you're real careful, probably WAY too risky for your average OCer. Basically you lid it like anything else, but the finale is taking a torch to the cap center to melt the solder enough for it to pop off. I've seen a couple succeed, but there's been a few attempts failed posted too.