[pHaestus]

Take a stab at the "goodness" of your results by calculating a W for the CPU using the dT across wb. It should be mostly flat because the wattage doesnt change. It wont be perfectly flat because as you lower the C/W then more of the CPU's heat goes into the wb and less into the traces of the board. You'll also find with a "below CPU" sensor that there's another problem. There's a baseline temperature that is difficult to go below because the air near the mobo is heated up by the mosfets and other bits all around it. This tends to make good wbs all perform even closer to one another than you'd expect. It's basically the same problem as with using an insocket thermistor, it's just slightly less pronounced. I had all kinds of trouble trying to use an undercore probe with the MCWCHILL tests if you'll recall.

If you want Jaydee I'll solder you up a spare 6657 and CPU. I'm not sure exactly what chips I still have around the house as spares (maybe a 2500+ Barton and a 1200MHz Morgan Duron or something like that?)