SPONSORED LINKS
> > "Theoretically, you *should* be able to replace the cpu with a DX4-100 chip > > > > be careful, this means changing the fundamental external clock speed from > > 25MHz to 33 MHz, which of course is aquite independent of socketed CPU's. > > > > However, the original announcement of the 750 said that it would be > > available with 100 MHz chips, so this may be possible. > > WAITAMINNUT! I keep getting told that the 750's (Not the 755's or > anything--speaking only of the 750's here) are *not* upgradable in their > processing speed whatsoever. > > When did this suddenly change, or was I misled in the first place? Sorry, I inadvertently added to this confusion by talking about the 750's having socketed CPU's. The 755's have socketed CPU's and come in two incarnations: a 50Mhz version and a 75Mhz version. The 50's are actively advertised as upgradable to 75's. I was told by IBM that they were using a bizarre version of the DX4 which came in a 50Mhz version, but this is probably wrong since a clock tripler would require a 16.667 Mhz system bus to generate 50Mhz. On the other hand, Gary Udstrand seems to think that the 50Mhz version uses a DX2 -- which would explain the ability to upgrade: you still have a 25Mhz system bus. The only problem with this theory is that the DX2 is a 5 volt chip, the DX4's are 3.3 volt chips... To add more confusion to the subject, someone (sorry, I lost the e-mail) mentioned that the chip itself is not socketed, but there is a daughterboard that is socketed. Now this would explain a DX4-50, since a daughterboard would imply that you have a seprate oscillator as well. Does anyone have the technical specs on their 755x-50???? BTW, John Kim correctly noted that the 750's have an IBM proprietary chip called the 486SLC running at 33Mhz. This chip contains a math co-processor and should _not_ be confused with the Intel 486SL chip which does _not_ have a math coprocessor.