Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: redrumloa on August 06, 2004, 03:33:48 PM
-
As the topic states. I'm thinking about the possibility of finding a professional shop that does extensive SMD reworking. Why not take the 604e 233 chip off and put say a 300 or 400 in it's place? The board may not be able to handle it you say, maybe yes maybe no. I'd also have the tiny jumpers removed and dip switches put in it's place.
So whatcha say techno gear-heads?
-
Just from what I recall, there might be a 250Mhz version
that is compatible. I believe the faster 604e that were
used in upper-end powermacs like the 8600 and 9600 use
different voltages and bus speeds. They also used a different
motherboard vs the models that use the 604e in the 180-233
range. My gut feeling is these will not work with the cyberstorm.
Again, I am no expert on this.
C Snyder
-
It has been done, but not easily ... the 300Mhz 604e's run at a different voltage, also getting the existing CPU off is not easy to do without damaging the board.
I'm trying to remember when it was done and by who ... have a trawl through the newsgroups archives on google think the details are in there.
-
Well,
Wasn't a few of the early G3's pin compatible with the 604e?
Then the change would be like putting a 060 on a 040-card (with the voltage difference).
Also, wasn't the 603 availible in speeds up to 300 Mhz?
-
Even if the early G3s are pin compatible, I don't think
they are actually faster than the 604e. The G3 was more of
a replacement for the 603 if memory serves. I do know
that the 604e stayed in service for a few years after the
introduction of the G3 in server and workstation
machines. I think the 604e has a better FPU than the
early G3s. I need to research this some more.
C Snyder
-
Red
My advice would be to contact Bplan and Gerald Carta and TK. Gerald designed the board and knows everything there is to know. I know they are busy atm but maybe try to call...
magnetic
http://www.bplan-gmbh.de/
-
I don't think they are actually faster than the 604e.
Yeah, I have heard too that the early g3s were slower than the 604s at the same clockspeed. And the g3s were based on the 603 and the 604 design was dropped although it was superior. I might be wrong though..
-
Some 604e information (nice chart on page 2):
http://search.freescale.com/cs.html?url=http%3A//www.freescale.com/files/32bit/doc/fact_sheet/PPC604EFACT.pdf&qt=powerpc&col=product&n=2
604e "user manual" (engineering heave stuff):
http://search.freescale.com/cs.html?url=http%3A//www.freescale.com/files/32bit/doc/ref_manual/MPC604EUM.pdf&qt=604e&col=product&n=7
Addendum to "user manual":
http://search.freescale.com/cs.html?url=http%3A//www.freescale.com/files/archives/doc/ref_manual/MPC604EUMAD.pdf&qt=604e&col=product&n=6
The user manual doesn't seem to talk very much about voltage stuff though.
604e technical summary, has brief discussion about seperating the 2.5V core coltage supply from the 3.3V IO buffer supply:
http://search.freescale.com/cs.html?url=http%3A//www.freescale.com/files/32bit/doc/data_sheet/MPC604E.pdf&qt=604e&col=product&n=10
This may indicate the possibility to change a 2.5V regulator on the CSPPC to a 1.8V regulator for the faster versions. This assumes nothing else onthe board runs off the 2.5V supply, if anything else does then it coule be more problematic. Probe around witha meter to see.
I went to www.freescale.com and did a keyword search for 604e. If you find a couple part numbers for exactly what chips you'd like to use you can search by part number as well and possibly get more detailed docs for that.
But check the voltages used by other things, you might be lucky enough to just have to change a regulator and the PPC chip, as well as probably put a bigger heatsink on and have things go well. But changing the PPC chip is going to be some work, and may be costly... I look forward to hearing if you have any success in this!
-
Hey. 8-) That's something I'd be very interested in knowing the results too...
-
Assuming you got a 400MHz part and all the voltages correct, what is the limit for the bus multiplier? 5x?