Well, I finally got around to building my new system with its Conroe processor (you can see my write-up of it at http://www.silentpcreview.com/forums/vi ... hp?t=37017
), and my temperatures are odd. After all the discussion earlier in the thread, I decided to go with the Thermalright SI-120 as a decent compromise between cooling ability and weight. Using the Speedfan 4.32 beta under Vista Ultimate RTM, my idle temperatures (with the machine having been on all day) are pretty stable as follows
Temp1: 39 degrees C
Temp2: 40 degrees C
Temp3 (Case Temp, I believe): 29 degrees C
Core0 Temp: 54 degrees C
Core1 Temp: 52 degrees C
This is pretty much what I've been seeing for the last several days. Unfortunately, I have no idea what Temp1 and Temp2 are. I'm wondering if they're what used to be what Speedfan reported as the CPU temperature. They're very stable and seem to stay about 10 degrees over the case temperature. The Core temperatures are also very stable and, at idle, stay about 25 degrees over the case temperature. The really odd thing is that when I ran the Prime95 CPU-area Torture Test for a while (only for about 5 minutes as a start), the only thing that happened was a couple of degrees increase in case temperature (and Temp1/2) and about a 5 degree increase in the Core temperatures.
Regardless, as I've seen in other posts about Core temperatures on Conroe, they don't seem to tie in well with those thermal constants arrived at here. For instance, the Thermalright SI-120 has a thermal constant of 0.47 degrees C/W. At a maximum load of 65W, I should see
65W x .47 degrees C/W = 30.55 or about 31 degrees C
increase over case temperature at load (which is about what I saw with my short test). But, at idle, and assuming a 20.4W load (per the article at http://www.lostcircuits.com/cpu/low_e/5.shtml
), I should see
20.4W x .47 degrees C/W =9.59 or about 10 degrees C
increase over case temperature (again, at idle). I don't. I consistently see an increase of 25 degrees. If my wild guess at Temp1 and/or Temp2 being the old CPU temperature, then at idle, I do get that 10 degree increase. But, my load temperature is way off.
I have to play with this some more to try to figure things out. But, I thought I'd just add my results to this thread I started all that time ago.