Re: Zip GS questions

[Barry Rees <Barry.Rees@prodigy.net>]

pubpc1@library.ucla.edu wrote:
> Barry, I have the latest catalog from JDR Microdevices and it seems that
> they are again carrying the 32k x 8, 28 pin DIP, 600mil chips.  These
> SRAMs have been reported to be good to the vacinity of 10MHz.
> 
> To use the 300mil SRAMs without major surgery is pretty easy.  Just build
> yourself a 300mil to 600mil socket adapter like I did.  Get a 600mil 28pin
> DIP component carrier and a 300mil, 28pin wire-wrap style DIP socket.
> Bend the wire-wrap socket's contacts side ways and solder these to the top
> of the DIP component carrier's corresponding contacts.  These custom
> adapters work well for my 13.75MHz upgraded board (they are used in the TAG
> sockets).
> 
> Paul Creager said that the adapters from Digi-Key didn't work.  The reason
> that these custom ones are better than the Digi-Key ones may be that
> I used high-quality parts with gold contacts, thus no oxidation and less
> noise.
> 
> If Nate wants to, he is free to update his Zip/TWGS FAQ with this info
> (also should state than 15ns is too fast for both TAG and CACHE and should
> use slower ones in CACHE - the FAQ should note this).
> 
> :)
> 
> -Scott G.

Scott,

I too have JDR's latest catalog. The "not available" SRAMs that I was
referencing are the 600 mil SRAMs in the 35 to 45ns range -not- the 300
mil ones. ALso, I am aware of the adapter setup. I have gold plated
adapters that I bought from Ironwood Electronics (California). I did
some serious testing about nine months ago and had better results than I
had hoped for, but not stable enough to continue using. While I was able
to reach 16 MHz speed for sustained periods of time with 15ns SRAMs
(from JDR) in the TAG sockets and 45ns Sony SRAMs in the data sockets,
my system would randomly crash with $0201 memory error. With additional
testing, I came to the conclusion that (possibly) the 15ns SRAMs were
actually too fast to use with the rest of my hardware setup. I had spent
several days testing, didn't have any 20 or 25ns SRAMs (only 12ns ones
that made the situation worse) to test my theory, and by that time had
"had enough". So for the time being, I went back to the 12/64
configuration. 

After the first of the year I hope to get some 20 & 25ns
chips and do more testing. If that doesn't pan out, I am going to ask
Doug Pendleton to update a sprare ZipGSX from it's current state of 9/64
to 16/64. AT that point it will be cheaper than me spending more and
more money and -maybe- reaching my goal. I'd like to do it myself (for
the satisfaction), but the upgrade needs to also be cost effective.

Barry Rees