Stability at 504
At 464 MHz and 2.1 volts, this system is rock solid. Win95, 98, NT4, and 5 run with no unusual problems and the few lock-ups I have experienced I attribute to software, rather than hardware problems.  Unfortunately, even at 2.5 or 2.6 volts the same can't be said for my Celeron at 504.  I have been running at 504 on and off for a couple of weeks now and while the lock-ups are not that often, they are there in greater numbers than I feel is acceptable.  At this point I can't justify the added speed if it means that I must give up stability.

Peltiers?
One thing I have noticed is that with 2.5 and 2.6 volts, there is a large difference between the temperature of the slug and that of the heat sink when starting any CPU intensive operation.  The CPU is heating up very quickly with the high voltage.  At 2.1 through 2.3 volts, the processor gains heat slowly enough that the heat sink is able to absorb it as it is produced.  This has me thinking that the circuitry inside the slug is running much hotter than the temps I am observing by measuring the slug's temperature.  This, in turn, has me thinking  that it may be possible to coax a little more stability out of the celery at 504 with some active cooling.  Yep,  you guessed it, I'm thinking about peltiers again.

I have placed an order for a couple of peltier elements and have rounded up all of the hardware necessary to assemble them into a heat sink or two to test.  Unfortunately my understanding of "two day air freight" and that of Mecci Surplus seem to be different, and I am still waiting for the TEC's to arrive.  Soon, I hope.