Sorry for all the posts. The extra vNB is looking promising (currently 1.20v). I just ran Pi 32M and Aquamark 3, both of which crashed the machine before, but now they run great. Ahmad, you were probably right -- the NB voltage was too low before. At least now I know it can do at least 430 FSB on 1.14v. I'll Prime this and hopefully return with good results from what I thought was a dog of a chip.