First trial I did was without extra capacitors on the back of the board and i couldn't get stable work if frequency was higher than 233MHz...
So after changing voltage from 1,8V to 1,97V without any improvement in stability) I decided to put low-ESR capacitors on 1,9V line and add 2 extra MOSFET transistors (they are not really necessary; I did it just to be 100% sure).
Only after all these operations CSPPC started to work stable.
Interesting. How are you measuring the CPU supply voltage? If you're using a multimeter or similar instrument, then I suspect you're not seeing the whole story. It's quite possible that because the faster CPU draws significantly more current, the ripple voltage will have increased dramatically. Your multimeter would not necassarily pick up this ripple voltage, and this would explain why stability was not achieved simply by modifying the power supply for an increased output voltage. The supply could still be dipping well below the required supply voltage for fractions of a second, and these problems would not be seen on a multimeter. If you have access to a scope, it'd be interesting to have a look at it.
Of course, this would also explain why adding larger capacitance (I assume we're talking about the smoothing caps on the supply rail here) would improve the situation, since these would deal with the ripple voltage problem.
I'm not quite sure what you're getting at with the additional MOSFETs though. Are you saying you've changed the MOSFETs in the PSU for higher spec components or that you've added some additional ones? If the latter I can't really comment on that unless I saw the mod.
Happy Hacking!
Apple Hammer
