Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!zaphod.mps.ohio-state.edu!unix.cis.pitt.edu!dsinc!bagate!cbmvax!daveh
From: daveh@cbmvax.commodore.com (Dave Haynie)
Newsgroups: comp.sys.amiga.hardware
Subject: Re: GVP Series II/A2000 problems
Keywords: GVP, A2000, 4.3
Message-ID: <15672@cbmvax.commodore.com>
Date: 7 Nov 90 17:37:49 GMT
References: <1484@sumax.UUCP>
Reply-To: daveh@cbmvax.commodore.com (Dave Haynie)
Distribution: na
Organization: Commodore, West Chester, PA
Lines: 56

In article <1484@sumax.UUCP> davidy@sumax.UUCP (David L. Yee) writes:

>	 I have a rev 4.3 motherboard. Gary said that Commodore made a
>	   number of 4.3 motherboards without a certain resistor. This
>	  resistor is a 470 ohm, 1/4 watt resistor that (I believe) he said
>	  was supposed to be connected to the DTACK line of the 68000 (My
>	  memory on the DTACK portion is a little hazy, the rest is not.)

That's reasonably correct, but not quite.  During the 4.3 run, we switched
from a Gary gate array made for us by Toshiba to a full custom Gary made at
MOS.  The change's significance was mainly in the DTACK* signal created by 
Gary.  On the Toshiba part, this was an active output that went 3-state when 
taken over (like when OVR* is asserted by an expansion bus device).  On the
MOS part, this was an open drain output.  The original A2000 design had a 
4.7K pullup on DTACK*, basically just to make sure that DTACK* was kept high
when no device was driving it.  With the open-drain or open-collector parts,
a pullup resistor is necessary, and is essentially responsible for determining
how long a particular signal will take to go high.  So, for production, we
required that any machine built with MOS Gary also have a 470 Ohm pullup added
to it, so that DTACK* would rise fast enough.

>	GVP claimed that they had been in touch with Commodore, and that 
>earlier this Monday they had concluded that this was why the Series II RAM's 
<were failing. He said that this was Commodore's fault, and that it was a 
>random occurance as to which rev 4.3 boards were missing this resistor. 

Well, the existence or non-existence of that pullup is certainly Commodore's
fault.  The failure of GVP's controller is based on GVP's failure to take into
account this difference.  Of all the hard disk controllers on the market, this
new GVP controller is apparently the only one that has a sensitivity to this.
Also, it's not guaranteed that the Toshiba part can properly sink DTACK* when
you have the effective 427 ohm pullup on DTACK* resulting from the addition
of the extra 470 ohm resistor.  So if you're adding the resistor and the
system gets flakey, go out and get a MOS/CSG Gary while you're at it.

This difference, by the way, has been documented at the last few DevCons, as
well as on the net here and on bix.  It's not like we're trying to hide it
or anything.  

>   *****Is their really such a problem with the rev 4.3 motherboards? I 
>hope someone from Commodore can help out here, because I don't know who to
>call or email to to find out the scoop from Commodore's side. I do not want
>to unfairly malign Commodore if GVP is trying to pass the buck. (GVP was
>very courteous and helpful, though.)

There is a difference, as I have described.  GVP thinks it's a problem, but
they seem to be the only ones.  Obviously, if you can avoid having ANY 
electrical difference between revisions of a system, that's a good thing.
But it's not always possible to have 20/20 foresight.  It is always possible
to have 20/20 hindsight, and it sure looks like GVP may need some glasses to
avoid this kind of thing in the future.

-- 
Dave Haynie Commodore-Amiga (Amiga 3000) "The Crew That Never Rests"
   {uunet|pyramid|rutgers}!cbmvax!daveh      PLINK: hazy     BIX: hazy
	Standing on the shoulders of giants leaves me cold	-REM