Path: utzoo!utgpu!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!mailrus!bbn!oberon!caesar.usc.edu!mlinar
From: mlinar@caesar.usc.edu (Mitch Mlinar)
Newsgroups: comp.os.cpm
Subject: Re: MEX Overlays and SCSI Implementations
Message-ID: <13522@oberon.USC.EDU>
Date: 18 Nov 88 20:09:41 GMT
References: <8811140554.AA25836@ucbvax.Berkeley.EDU> <3193@mipos3.intel.com>
Sender: news@oberon.USC.EDU
Reply-To: mlinar@caesar.usc.edu (Mitch Mlinar)
Organization: University of Southern California, Los Angeles, CA
Lines: 20

In article <3193@mipos3.intel.com> dbraun@cadavr.UUCP (Doug Braun ~) writes:
>
>I did just this for my one-of-a-kind CPM system, and it has worked out
>very well.  The SCSI Host Adapter is indeed a glorified parallel port,
>and I built one from 6 or so TTL chips.  If you used one of those SCSI
>controller chips, the design wuld probably be even simpler.
>

I did the same thing using my Xerox 820-II parallel port to talk to
a SCSI board (Shugart SA1610).  This one did not require any chips, just
cabling.  Unfortunately, the cabling is point-to-point to transform Xerox
parallel ports into SCSI; also, you can only operate one SCSI card.  However,
that is enough to run two hard drives.

It was a bear to get the timing right.  The -I was too slow for writing, but
fine for reading which screwed me up for awhile.  I also tried an Omti card
with the same results.  -II, or -I with speedup, is fast enough to keep up
on these two cards at least.

-Mitch