First a little teaser:
The attachment SSTIMG09.PNG is no longer available
Then a full story:
About a month ago, a batch of three AMD K6-2+ 570 MHz (570ACZ) from a well known seller finally arrived. I prepared my trusty Gigabyte GA-5AA (cache stable at 115 MHz FSB), nice cooler with a copper core and began. First I quickly confirmed that each is capable of 600 MHz @ 2.0V - this was no problem. Then I set 2.2V to check maximal OC potential. One of the CPUs (#3) was able to load Windows at 690 MHz and seemed stable:
The attachment 690_l3.png is no longer available
The attachment 690_l3_bench.png is no longer available
Unfortunately Prime95 quickly showed an instability. At that point I said "OK, let's see, what 2.4V can do!". With this voltage I didn't want to torture this CPU in Prime, so I just tried how far can I go. At 720 MHz (6x120 with disabled L3), computer went through POST, but even DOS applications were crashing. As a final resort I tried combination 700 MHz (5x140 with disabled L3) which seemed stable in DOS and this is where I took this:
The attachment SSTIMG07.PNG is no longer available
But Windows was a no-go. Funny thing is, I confirmed later that 140 MHz FSB is unstable even without cache, so for 700 MHz Windows "experience" I would need a better board.
Finally I wanted to check how far can I go at a standard 2.0V. To my great surprise it was 660 MHz (6x110):
The attachment 660_l3_20v_stable_forum.png is no longer available