Pro/Forums - View Single Post

Cathar · 10-06-2003, 01:30 AM

Well, I was being a little silly in that other thread, but just to point out how silly some motherboards thermal sensors are. You can't really make comparisons.

I also report actual voltages, not BIOS settings. My CPU is actually set to 2.15v in the BIOS, and a 450MHz (~20%) higher frequency. All up, my CPU is probably putting off around 25% more heat than yours. If you're CPU was putting out the same heat as mine, the difference would probably be more like 38C.

Let's run with that 38C difference and try to account for it shall we?

The BIOS I was using was Ver 1.4, which reports CPU temperatures at about 10C lower than any other NF7-S Rev 2.0 BIOS.

So we have 28C left to explain.

The Iwaki pump isn't doing an awful lot with respect to the final temperature. Maybe gaining me ~1C over using something like an Eheim 1250.

27C left to explain.

The dual-radiators with car heater-core blower, now there's the trick. Radiator based water-cooling, much like regular air-cooling, is all about surface area and air-flow. I have around 3x the radiator surface area of a DTek Pro core setup, and probably around 3x the air-flow of a regular 12cm fan setup as well. My fans/radiators keeps my water very cool, down to near on 3C above radiator air-intake temperature. I'd estimate that this gives me about 4-6C on most other setups. Let's call it 5C.

So 22C left to account for.

My radiator air-intakes on right near the floor-level. Despite room temperature typically being ~22C the air-intake temperature is down around 18-19C. Let's call it a 4C difference.

So 18C left to account for.

You're at 28C ambient too, I'm at 22C ambient.

So, 12C left to account for.

This CPU does seem to be quite cool running by itself, giving about 3C lower temps than my other Barton XP2800+. I guess it's part of the reason why it is able to hit >2.8GHz due to lower internal resistances for whatever inexplicable reason that is.

So that's about 9C left to account for.

That 9C then is probably what the Cascade SS is offering. I don't know what block you have, but that 9C includes stuff like proper mounting of the block on the CPU, good thermal paste, good lapping of the block, and just the efficiency of the block itself.

So you see, it can all be a big sum of many small factors that can easily add up to ~30-40C differences.

10-06-2003, 01:30 AM	#2
Cathar Thermophile Join Date: Sep 2002 Location: Melbourne, Australia Posts: 2,538	Well, I was being a little silly in that other thread, but just to point out how silly some motherboards thermal sensors are. You can't really make comparisons. I also report actual voltages, not BIOS settings. My CPU is actually set to 2.15v in the BIOS, and a 450MHz (~20%) higher frequency. All up, my CPU is probably putting off around 25% more heat than yours. If you're CPU was putting out the same heat as mine, the difference would probably be more like 38C. Let's run with that 38C difference and try to account for it shall we? The BIOS I was using was Ver 1.4, which reports CPU temperatures at about 10C lower than any other NF7-S Rev 2.0 BIOS. So we have 28C left to explain. The Iwaki pump isn't doing an awful lot with respect to the final temperature. Maybe gaining me ~1C over using something like an Eheim 1250. 27C left to explain. The dual-radiators with car heater-core blower, now there's the trick. Radiator based water-cooling, much like regular air-cooling, is all about surface area and air-flow. I have around 3x the radiator surface area of a DTek Pro core setup, and probably around 3x the air-flow of a regular 12cm fan setup as well. My fans/radiators keeps my water very cool, down to near on 3C above radiator air-intake temperature. I'd estimate that this gives me about 4-6C on most other setups. Let's call it 5C. So 22C left to account for. My radiator air-intakes on right near the floor-level. Despite room temperature typically being ~22C the air-intake temperature is down around 18-19C. Let's call it a 4C difference. So 18C left to account for. You're at 28C ambient too, I'm at 22C ambient. So, 12C left to account for. This CPU does seem to be quite cool running by itself, giving about 3C lower temps than my other Barton XP2800+. I guess it's part of the reason why it is able to hit >2.8GHz due to lower internal resistances for whatever inexplicable reason that is. So that's about 9C left to account for. That 9C then is probably what the Cascade SS is offering. I don't know what block you have, but that 9C includes stuff like proper mounting of the block on the CPU, good thermal paste, good lapping of the block, and just the efficiency of the block itself. So you see, it can all be a big sum of many small factors that can easily add up to ~30-40C differences.