PPC 604E processors

[Lunarnut]

Apple used these processors for their Macintoshes. Would say a 300 MHz processor work in place of a 233 Mhz one for the Amiga Accelerators? Has anyone tried this?

[downix]

Don't think the crystal is setup for it.

[Lunarnut]

The crystal? Can you elaborate?

[Framiga]

IIRC the max multiplier is 4.5 in a CSPPC. With a 66.666 Mhz, you'll achieve about that.

Will it works? try and let we know! :-D (with a huge cooling system, of course)

[Lunarnut]

I would imagine it would work. Back in the day, Pentiums (if I have my info right) were basically batch tested at certain speeds/temps. The faster speed/temp then the higher rating the processor got. I'm not sure if this applies to PPC processors and from reading around there does seem to be slight differences between versions.

So it might work...

[Homer]

Hmm, but would the 300Mhz cpu work at 233Mhz   :-?

[Framiga]

yes ok Lunarnut but pentiums were a different things.

Here in Amigaland the biggest problem would be the ram controller buswidht (and probably the project/routing itself also).

Someone tryed to install a 300 Mhz 604e, with problems about stability. Now we don't know if these issues were generated by a suboptimal BGA soldering or else.

Look at this page, where this guy tryed about everything to achieve the better performance/stability ratio. (look also at the temperature achieved)

DaveM overclocking page

[AJCopland]

Homer wrote:
Hmm, but would the 300Mhz cpu work at 233Mhz   :-?

Yes it would. The question being asked is, could you remove the 233MHz 604e CPU and solder in place a 300Mhz 604e CPU.

The answers are both yes, but the 300MHz cpu would only be running at 233MHz because the crystal which generates the clock frequency at which the CPU runs is only running at a speed AND multiplier that produces a 233MHz speed.

For this to work as the original poster wished they would have to replace the CPU with the 300MHZ _capable_ one, and then change the crystal and onboard multiplier to generate a 300Mhz cpu clock.

Andy

[Framiga]

Homer wrote:
Hmm, but would the 300Mhz cpu work at 233Mhz   :-?

yes of course :-)

[Framiga]

AJCopland wrote:

Homer wrote:
Hmm, but would the 300Mhz cpu work at 233Mhz   :-?

Yes it would. The question being asked is, could you remove the 233MHz 604e CPU and solder in place a 300Mhz 604e CPU.

The answers are both yes, but the 300MHz cpu would only be running at 233MHz because the crystal which generates the clock frequency at which the CPU runs is only running at a speed AND multiplier that produces a 233MHz speed.

For this to work as the original poster wished they would have to replace the CPU with the 300MHZ _capable_ one, and then change the crystal and onboard multiplier to generate a 300Mhz cpu clock.

Andy

nope, the multiplier goes to 5x (look at that page)

[AJCopland]

Framiga wrote:
nope, the multiplier goes to 5x (look at that page)

What's that got to do with the question that I answered? :-P

Andy

[Framiga]

ops sorry! i've overlooked one. Apology! :-)

[Lunarnut]

Well my main point of asking was if I couldn't find a PPC 233 processor then could a find a faster Mac and scavange its CPU for an Amiga project. I would only run it at 233 Mhz.

[AJCopland]

Lunarnut wrote:
Well my main point of asking was if I couldn't find a PPC 233 processor then could a find a faster Mac and scavange its CPU for an Amiga project. I would only run it at 233 Mhz.

Ah I see, well yes it will just run at whatever speed you set it to run at, I once booted a 300Mhz at 600Mhz due to poorly documented jumper settings. It will run like that... just not for very long :-o The clockspeed of a CPU is what it's supposed to be run at but you can feed it almost any actual clockspeed with varying side effects (like instant cpu death if its too fast).

Andy

[Framiga]

ah ok, if you find someone who can do BGA reworking/reballing then yes, no problems.

edit- IMHO the worse problem, seems to be reworking an already populated PCB (very populated in the CSPPC) and then sometimes the fault is not the CPU itself but some broken traces in some internal layer.

A guy here in Italy, tryed to send the CSPPC to an italian BGA specialist (he fixed/X-rayed some traces also).