Re: parity memory

[roman songdog eskimo com (Bill Roman)]

Dale Jackson wrote:

> I recently purchased two non-Digital 64 mb true parity 60ns FPM SIMMs
> (128 mb) from someone who had them working in a multia udb. [...]
> I purchased them for an alphastation 200 4/233. The as200 had only two
> Digital 8 mb 70ns parity SIMMs in it and it runs fine with the two 8 mb
> simms....
> 
> But the as200 will not boot up at all with the two 64's in it (with or
> without the two 8's). It stalls at the initial point of checking memory,
> I guess. [...]
> 
> Are there any jumpers on the mb that need to be changed, or are these
> just bad simms for the as200?

Avanti-class machines (including AlphaStation 200) are picky about
memory: 70 nsec. exactly, please, no slower, no faster.  This is
not a rumor: it is The Word from the sustaining engineering group.
This doesn't mean 60 nsec. memory will never work; it means there's
no guarantee it will.

Why would this be so?  There could be a variety of reasons: faster
chips could be sensitive to glitches on signal lines that slower ones
ignore; data hold time out of the chip might be shorter, shaving the
margins; faster chips draw higher-frequency current spikes when
accessed, and bypass capacitors on the board might not be able to
absorb these transients.

> Where can I pick up some good inexpensive simms for the as200?

This is the real question.

-- 
Bill Roman  (roman@songdog.eskimo.com / roman@songdog.uucp)   running linux