Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!mips!ptimtc!nntp-server.caltech.edu!toddpw
From: toddpw@nntp-server.caltech.edu (Todd P. Whitesel)
Newsgroups: comp.sys.apple2
Subject: Re: Extra Ensoniq RAM
Message-ID: <1991Jun17.110401.7382@nntp-server.caltech.edu>
Date: 17 Jun 91 11:04:01 GMT
References: <1991Jun15.031457.9681@micor.ocunix.on.ca> <1991Jun15.031838.9787@micor.ocunix.on.ca> <16425@smoke.brl.mil>
Organization: California Institute of Technology, Pasadena
Lines: 45

gwyn@smoke.brl.mil (Doug Gwyn) writes:

>In article <1991Jun15.031838.9787@micor.ocunix.on.ca> taob@micor.ocunix.on.ca (Brian Tao) writes:
>>I just want to know if it is technically feasible.

>Nope.

Gee, at least tell him why. A raw 'nope' is really unhelpful.

It's technically feasible, although it is not something to try unless you
know what you're doing. If no one does it over the summer then I might try it
when I get back (I'll want my roommate's oscilloscope) and then I can give
people cookbook instructions. In any case, it requires some permanent warranty
killing modifications to the motherboard unless you want to invite disaster,
so caveat usor.

According to the schematics, all the signals are available (but you'll have to
figure out which of the two unconnected DOC pins to use, and what -- if any --
timing requirements it has). Practically, however, it'd be a bitch unless you
want to desolder the DRAMs and replace them with sockets (get nice machined
ones that can be soldered to the motherboard on both sides). You'd also need
a test clip or two to get the bank-select signals from the DOC and elsewhere.

To figure out which DOC pin is being used as the bank select (I suspect that
one of the unconnected DOC pins, probably "BS", is the bank select bit and
that you can pick from two pairs of DRAMs with it), find an oscilloscope and
write some test programs with the DOC (not for the squeamish) and watch the
outputs while comparing them to known ones like the DOC's RAS and the GLU's
CAS line.

To hook up the real test setup, replace both DRAMs with machined sockets and
make a test board that plugs into them and has room for four DRAMs (wire a
pair of DRAMs to each socket in parallel, except for the CAS and OE lines
which will probably need to be decoded according to which bank is being
accessed). Use test clips to get the necessary signals from the DOC and
elsewhere (you may have to use something cheesy like an announciator output
if the Sound GLU bank bit is not available in the hardware (and it is a PLCC
so attaching a jumper to it could be nasty in any case). It would be nice if
the extra bank were enabled by the reserved bit in the Sound GLU, but don't
count on it. The sound tools don't have any way to specify the extra bit anyway
so you'd have to set it between DOC RAM loading calls or write your own sound
software anyway.

Todd Whitesel
toddpw @ tybalt.caltech.edu