Quote:
Originally posted by EMC2
Look closely at that pic that Rick posted. Notice how the water isn't even flowing around the sides of the block except directly under the inlet and outlet.
Also with the "crosshatch" on top of the slug such as it is, I suspect the flow through the valleys of the hatching is very low. Good flow probably only extends at most a third of the way into the valleys. The only good side is the flow on top of the block would be quite turbulent.
icel0rd -
What processor, what multiplier, what FSB, what memory settings, running what program do you get your 9C over ambient. And what specifically is at that temp of and how is the measurement taken?
|
Athlon XP 1800+ (actual speed is something like 1.53 GHz
FSB speed = 133 (266 effective since both ends of the wave are used)
Memory settings are set in performance mode on my mobo (not sure what it sets it to, but I can verify).
The temperature was taken on the processor alone via the probe on the board using motherboard monitor and the gigabyte util that comes with the board. I know people are going to lecture about the inaccuracy of the probes on board, but what matters to me is that the last setup I put together using cheaper compound and a less efficient setup with the same components produced poorer results compared to how things run with the way I have my layout now. I never removed the processor, so the probe never got touched.
Before I changed my system layout and corrected hastey mistakes and such, at the previous temperature readings, I got 45 to 48C at full load under 31 to 32c ambient. That was with an inefficient radiator setup under the chassis without a fan blowing directly through the fins or pulling directly through the fins. I had a little desk clip on fan blowing under the chassis until I found a way to cool the rad correctly (ghetto, I know). I also had no chassis fan on the mobo side of my YY cube case to draw or expel air.
What I changed specifically:
Removed the rails that provided support for the very long AGP PCI cards since I didnt need it.
Mounted Rad in the bottom left of the chassis face (inside) against the chassis wall.
Cut a hole in the chassis for my 120mm fan to suck the air through the rad fins. Cut a matching hole through the plastic face for the fan to expel air from.
Installed a rubber gasket around the fan hole to make sure that air was being drawn 100% from inside and that the seal was good.
Fan walls (not blades) were rubber coated and mounted between the plastic face chassis cover and the chassis over the fan hole with the gasket.
Eliminated the chassis fan on the drive bay side of the case (rear) for silence and since my 500W ATX PSU has 2 fans in it and it expels air from that side of the chassis anyway.
Hose lengths were reduced because the radiator was now much closer to the res/pump. New hose was installed since I noticed that the stupid pipe joint compound I used was the stuff that doesnt harden, and it got into the hose and coated it with a white goopy film. Bad move last time. I didnt pay attention because I did a rush job.
Remounted the processor after thoroughly cleaning the block and the core with rubbing alcohol. I noticed that the edge of the block was shifted a bit and was barely sitting on the lip of the processor mounting socket (not an optimal fit and I used cheap compound to top it off). I applied Antec reference silver compound carefully and correctly this time and mounted the block.
Well, that covers it. BTW, I ran SETI@home for an hour to peg the machine at 100% when I took the reading.
I also did some 3d gaming (anarchy online, Q3A) for a few hours and noted the temp immediately after end game. In Anarchy Online, I just run in window mode and check the temps in the system tray.
-Ice-