Thermal resistance?

[Harold7]

I'm doing some resaerch on finding the best CPU cooler for my T-Bird 1.2 GHZ and was wondering... when a value of thermal resistance is given in degrees C/W, is a lower number better for efficient heat transfer... is 0.45 C/W better than 0.64 C/W?

I'm leaning toward the Volcano II to replace the Silver Orb that I got originally with it... my CPU running at 55 C at idle is just too high for me. http://www.sysopt.com/forum/smile.gif

[Fingers]

Lower numbers represent more efficient heatsinks.

There are a few calculators at Benchtest.com (http://www.benchtest.com/calc.html) for figuring out wattage and C/W.

[Szech]

Yep, lower is better. If you get a heatsink that has low .30's, it's pretty good. If you get one that is in the .20's, it's really good. Usually this requires more unorthodox means such as delta black label fans that I personally cannot stand.

[Harold7]

Thanks for the info. After doing some more research, the ThermoEngine from ThermoSonic looks like a very good cooling solution for my processor... just haven't decided whether to go with the version using the Delta fan or to opt for less noise.

The advantage of the Volcano II (which can also use the Delta fan) is that the fan can be replaced with out having to remove the heatsink from the processor... a clear advantage to limiting the physical stress on the body of the processor. http://www.sysopt.com/forum/smile.gif



[This message has been edited by Harold7 (edited 04-30-2001).]

[lazerbeam]

If you haven't bought one yet you need to go to AMD's website and check out their list of recommended heatsink/fan combo's. I did a lot of research in this area and discovered that some heatsinks, although recommended by AMD for their CPUs, will not fit on certain brand/models of motherboards. Also, make sure the heatsink you pick has thermal mounting tape.. If not you must buy and use thermal grease!! If you use grease, put it on evenly, equal to 2 or 3 sheets of paper thick.. More is not better.

[club_med]

thermal mouting tape or grease, whats better ?.
cm.

[Szech]

Thermal grease is much better than the PCMs (Phase changing materials). Plus, it's a lot easier to remove. Arctic silver is a high quality thermal grease, but it will cost you. Also... AMD's recommended heatsink/fan combos aren't necessarily the best coolers. I didn't see the FOP38 on the list, for instance.

[Harold7]

Well, I replaced my Chrome Orb with a Volcano II using Artic Silver II paste and my idle temp dropped from 55 C to 50 C using the stock fan.

I got a 60mm 30 CFM Delta (5700 rpm)to replace the stock fan and my idle temp dropped to 45 C, plus the Delta fan is much quieter than the fan that comes with the Volcano II.

As soon as my 7000 rpm 60mm Delta comes in, I'm going to use it to replace the generic 60mm case exhaust fan I mounted about 3" from the heatsink... the fan I'm using now runs at 4500 rpm, so I should be able to suck heat out of my case and directly from the base of my heatsink much more effectively. http://www.sysopt.com/forum/smile.gif

[paul233]

Harold if your tbird is really idling it should see the low 30s depending on the heat probe location. How does your system temp compare to room temp? With good case cooling you can get system = to room which helps a lot.

[Harold7]

Paul233

With room temp at 27 C right now, my motherboard temp is 31 C and my CPU temp is 48 C and my computer has been running for about 3 hours already.

By idle I mean my system configuration at boot up... Norton AV, DirectCd and other programs loaded... System Monitor load shows 18 %, if I run my system to 100 % load, temps go up to 51 C on the CPU , but drops back to 48 C quickly after I reduce load back to normal.

I can put 3D Mark 2000 in a loop and let it run for an hour and I get the same results, so it appears that the Volcano II and Delta fan are working adequately... I'm suspicious of the ASUS A7V KT133A temp probe readings anyway, since it appears to relate directly to its physical proximity to the CPU core.

I think that the stable temps I'm reading at normal CPU load and the relatively small temperature rise I get when running the CPU at full load are a better indication of the quality of my heatsink and fan combo...someone showing a temp probe reading in the 30's I believe is more an indication of the physical location of the probe in relation to the CPU and the heatsink than any consistent temperature standard.

Without knowing that we have exactly the same physical placement of the probe on our boards, a 30 C temp reading for you may equate to a 45 C reading on mine if my probe is closer to the CPU than yours.

Once I put in the 7000 rpm case fan, I should be able to remove heat from the case more efficiently and drop temps a couple of degrees C, so as long as the CPU temp remains stable at around 45-48 C I'll be satisfied. http://www.sysopt.com/forum/smile.gif