Well, I have some preliminary temp data. I took my trusty indoor/outdoor probe thermometer and removed the plastic housing that covered the actual probe to get the best thermal connection possible. I hot melt glued (how ghetto is that?) the probe diode or thermistor or whatever that little blue bulbous thing is onto the area of the PC board that correlates with the back side of the GPU core.
Preliminary conclusion:
My ghetto heatsink leaves the Thermaltake heatsink in the dust.
Semi-scientific approach:
I haven't measured any temperatures with the Thermaltake or the factory heatsink yet, but I found an online review for the Thermaltake heatsink that gives temperatures measured using the same probe placement on another GF4 TI4200, so they should at the very least give an idea of the ballpark temperatures we're dealing with.
Temps immediately after Windows login:
Case air (MBM): 78.8F 26.0C
GPU: 102F 39C
Temps after running Milkdrop (openGL winamp vis plugin) for about 20 minutes:
Case air: 80.6F 27.0C
GPU: 106.9F 41.6C
The review data I found for the Thermaltake showed these temps (I was running my card at stock clock speeds for my tests):
As you can see, my highest recorded temperature was lower than this person's lowest recorded idle temperature. This could mean that our tests were different (or flawed) enough to get different results, or it could mean that my cooler kicks ass. I'll have to mount my thermaltake back on there one more time to find out the truth.