faster, Faster, FASTER

[Dave Hanna]

In article <493 at icus.islp.ny.us> lenny at icus.islp.ny.us (Lenny Tropiano) writes:
>In article <1185 at umbc3.UMD.EDU> alex at umbc3.UMD.EDU (Alex S. Crain) writes:
>...
>|>
>|>	Has anyone played with the clock speed on the 3b1? The hardware seems
>|>to be of excellent quality, and it seems like 12Mc shouldn't be unreasonable.
>|>I think that the modem is on its own clock, and the power supply should be 
>|>able to handle the extra load on a machine without expansion cards. 
>|>I have an AC fan, so heat shouldn't be a problem...
>|>
>
>I'm not an electronic's guru, but the UNIX PC's have a MC68010L10 CPU chip
 [ Discussion of need to swap out for 68010L12 12.5Mhz chip. ]
 Note also other chips that are tied to the clock - e.g., I assume
 the 3b1 uses a MC68451L10 - that would also have to be upgraded.
>I wonder if this is feasible?  Comments from any hardware gurus?  :-)
>
>-Lenny

I am far from a hardware guru either, and I have no idea what's inside
the 3b1, never having had it open or seen a schematic, but in most
system designs I have worked with, the CPU and crystal changes are the
least of your worries.  The main one is memory.  And it's not just a matter
of substituting faster chips (e.g. 100ns for 120's, or 80ns for 100's),
because there is a fairly complex train of pulses that control the
addressing of theRAMS.  The delays between those pulses are, in some
part at least, fixed by propagation delay of buffers and/or delay lines,
and don't speed up when you speed up the clock.
    Also, you can't tell just by changing the clock/CPU and trying
it, because there may be enough margin in the design that it will
ALMOST work, meaning that it will work most of the time, but
get flaky every once and awhile, or it will work at room temperature
but get flaky when it heats up, or anyone of a number of situations.
And we all know that if it's going to fail, it's a whole lot better
to fail consistently!  Debugging timing problems gives designers
ulcers.
    If you're going to try it, you need to get the schematics and work
out a timing diagram to see what happens to the margins when you speed
the clock up.
    Hope this helps.
-Dave

-- 
Dave Hanna,  Daltech MicroSystems    |  "Do or do not -- There is no try"
P.O. Box 584, Bedford, TX 76095      |                        - Yoda
(214) 358-4534   (817) 540-1524      |
UUCP:  ...!killer!gtmvax!dave        |