Path: utzoo!attcan!uunet!samsung!uakari.primate.wisc.edu!uflorida!news
From: news@uflorida.cis.ufl.EDU (USENET Master)
Newsgroups: comp.os.os9
Subject: Re: Building a Multipak for OS9 on the CC3
Message-ID: <22392@uflorida.cis.ufl.EDU>
Date: 22 Feb 90 05:54:59 GMT
References: <1261@unsvax.NEVADA.EDU> <2280@ultb.isc.rit.edu>
Reply-To: se@beach.cis.ufl.edu ()
Organization: UF CIS Department
Lines: 31

From: se@beach.cis.ufl.edu (Sina Eatemadi)
Path: beach.cis.ufl.edu!se

   1) Build your own MPI:  This would be more trouble than it is 
   probably worth.  You could, however, build a Y-cable with
datadate/address line buffering and slot selection.  Get the service 
   manual for Tandy part # 26-3024  
   
   2) Buy a third party MPI.  Two are currently advertised in the 
   Rainbow.

   3) Build a double or triple Y-cable:  These cables are fairly easy
   to make but you must keep a few things in mind.  
   Keep the cable short (under 3 inches), otherwise you'll get bus 
   timing problems.
   Only have one active ROM on any device plugged into this cable.  
   pull any RS-232 pak ROM's, hard drive autoboot Rom's, etc.
   WATCH FOR HARDWARE ADDRESS CONFLICTS.  These are fun to deal with.
   Most (all?) CoCo hardware is not fully addressed decoded.  The Tandy
   disk drive controller is about the worst.  The RS-232 pak is decoded down 
   to 4 bytes and will work with just about everything.
   Lastly, contrary to what someone posted earlier, the B&B interface will
   NOT plug into a Y-cable unless it is modified.  Even if the autoboot
   ROM is pulled, the CoCo still senses 'something' at the standard ROM 
   addresses ($C000+) from the B&B.  This conflicts with the disk controller
   rom and will cause the CoCo to lock-up.  There is a hardware mod for
   this problem available from Chris Burke.

   Hope this is of help to everyone...

   latr    ChrisM    se@beach.cis.ufl.edu