Overclocking 3800 X2 and burning in at high voltage

--------------------------------------------------------------------------------

Ok, I think there may just be a possiblility that my x2 just may overclock to a decent amount. However the voltage required makes me worried. Now that the colder winter Florida months are setting in my CPU boots up with an average 32c. I was able to boot up at 2750 with the "oh so cold" weather but at high voltage: 1.65v on the cpu. Everything loaded fine, despite not trying to run 3d apps. My problem is I don't want to exceed the voltage so much that I burn the cpu up. I have a weak core, it's core 1 and it requires at least 0.5 volts more than core 0 to run equally. right now I'm running a stress test using sp2004 on core 1 only:

run 7 minutes:

Core 1 @ 2590 @ 1.60 volts
CPU temp is 42-43c while burning only that core.

at the same 2590mhz rating I can burn core 0 at ~1.525v and get ~ 40-42c temps

To me, given my underground drug chemistry knowledge of days long past, I know that silicon can't be messed up easily very much at those temps. However, I know this is about the borderline temp that I'm at high risk. Moreover, the 1.6v I'm running at right now is about close to the 20% overvoltage limit that I'm willing to take. I know I booted up at 1.65v with 2750 before, but this I think is way too much....plus sp2004 wouldn't last more than 15-30 seconds before having an error.
I just want to burn the weak core in first at night while temps are low and go from there.
I can not lower my voltage anything below 1.35 (stock volts) as my bios doesn't permit it. So, I've taken to the opposite side and decided to overclock as high as possible with the absolute mininum voltage I can be relatively safe with. My problem is I know that there is a correlation between cpu temps and voltage. I have no idea what those extremes are. I'm assuming anything over 1.625 at above 35c is going to burn it up. I feel if I stay at 1.6 and below and my temps are below 42c then I should be within limit. At the moment, I am burning the weak core (1) at 41c for a total of 17 minutes without error. I've "feel" that there has been some progress by burning the cpu in, but it may be all subjective an relative to the fact that I may have just fined tuned my ability to get the parameters in the "sweet zone."
Could some of you comment on the voltage limits/temp limits of these processors and what "should" be safe to burn in. Keep in mine that I'm trying to get the highest possible frequency at the minimum voltage and temps to burn overnight in the colder weather.

Any opinions, comments, flames are appreciated.
Thanks.

I'd be cautious running at 1.65 unless you have a badass water cooled setup.

Get your X2 stable at a more resonable clock first, like 2.5Ghz @ 1.45V (that was my first stability check run). Run Prime95 for 12+ hours, for good burn-in and stability check. You might want to up it to 1.5V just to make sure you don't error, or if you just want to burn use a burn-in utility.

Just for clarification. When I stress one core my temps stay well below 40C, even on hot days. You may peak over 50C running both cores, so watch yourself.

After awhile your core won't respond to over volting, and you'll just be carelessly over heating your CPU to get those last couple of Mhz. From what I've seen, about 2.5-2.7Ghz is a resonable over clock for an X2 on air, depending on how good your chip is and how much gusto you have with the voltage knob.