[Coco] Un-upgraded 3124 Multipak/Coco3 Question

[Richard Atkinson]

On the subject of bus conflicts, I'd like to share a few thoughts.

First, a bus conflict when reading from a conflicted address only
occurs for a nanosecond or so each time. If you read from this address
often, in a loop for example, it might cause problems, but an
occasional read should not matter too much.

Second, the seriousness of the conflict depends on the underlying
technology used in the two bus agents. If they're both NMOS, such as
in the Commodore 64, there is no problem. This is because NMOS
technology uses passive pull-up resistors for logic 1 (high) and
active pull-down transistors for logic 0 (low). The situation where
one bus agent is trying to assert logic 1, and the other is trying to
assert logic 0, will result in a logic 0 on the bus. In effect the
result of the operation is the logic AND of the two results asserted
by the two bus agents.

If both bus agents are CMOS, such as in later PCs, there is a problem
because CMOS technology uses active pull-up transistors and active
pull-down transistors. The above situation causes one bus agent to
turn on a pull-up transistor and the other bus agent to turn on a
pull-down transistor, resulting in an undefined logic level on the
bus. If one of the transistors has a lower on resistance than the
other (possibly the NMOS pull-down transistor is lower than the PMOS
pull-up transistor), then this transistor will 'overpower' the other
and the result will be better defined.

It sounds like the GIME, which is possibly CMOS, would be conflicting
with TTL buffers in the MPI. In that case, the worst situation would
be when the GIME tries to assert logic 1 and the MPI tries to assert
logic 0.

Richard


On 8/1/07, Mark Marlette <mark at cloud9tech.com> wrote:
> Mike,
>
> I'm not even going to try to do the agree/disagree repeat.... :)
>
> I stand corrected, you are correct.
>
> Thanks for keeping me honest. :)
>
> Mark
>
> Quoting Mike Pepe <lamune at doki-doki.net>:
>
> > Mark,
> >
> > I hate to disagree with you after you disagreed with my agreeing with
> > you... ;)
> >
> > The 244's buffer the data bus from the RAM banks.
> >
> > The CPU is buffered from the data bus via a 245...
> >
> > ...but the general data bus goes right out the cart port. The GIME and
> > MPI would indeed be duking it out if there's an addressing conflict.
> >
> > -Mike
> >
> > Mark Marlette wrote:
> >> Mike,
> >>
> >> I hate to disagree with you after you agreed with me but.....
> >>
> >> That is not exactly true. The GIME is buffered by the 244's. The
> >> MPI is buffered by the 245, so those two would fight for the data
> >> bus.
> >>
> >> The GIME wouldn't be damaged the buffers possibly would be.
> >>
> >> Mark
> >> Cloud-9
> >> http://wwww.cloud9tech.com
> >>
> >> Quoting Mike Pepe <lamune at doki-doki.net>:
> >>
> >>> I second that. Essentially the MPI and the GIME chip will respond to
> >>> the same addresses. Reading is probably the worse of the two in terms
> >>> of potential damage, since the GIME and MPI will be driving the data
> >>> bus at the same time. It could potentially kill the GIME. Not likely,
> >>> but it is possible.
> >>>
> >>> -Mike
> >>>
> >>> Mark Marlette wrote:
> >>>> David,
> >>>>
> >>>> The MPI and CoCo3's memory map conflict. Depending on what the
> >>>> program is accessing for addresses, depends on the damage or
> >>>> possible damage.
> >>>>
> >>>> It is a real issue and should be a concern to non upgraded CoCo3
> >>>> MPI owners.
> >>>>
> >>>> Mark
> >>>>
> >>>> Quoting David Grimmel <navydave1 at hotmail.com>:
> >>>>
> >>>>> I know I need to upgrade my white 3124 multi-pak to be 100% compatible
> >>>>> with my Coco 3, but havent gotten around to it yet. While reading a
> >>>>> Marty Goodman Q&A in the July 88 iss of Rainbow, he says, "even if your
> >>>>> Coco 3 {appears} to work fine with your Multi-Pak, the upgrade is
> >>>>> required. The upgrade may prevent slow, subtle damage to your Coco 3".
> >>>>> I knew that using it un-upgraded my cause small glitches and software
> >>>>> problems, but I didnt think it would cause any physical permanent
> >>>>> damage. What damage is the good DR. referring to? Have I already
> >>>>> damaged my Coco 3 by using it with an un-upgraded 3124 multi-pak?
> >>>>>
> >>>>> Dave
> >>>>>
> >>>>>
> >>>>>
> >>>>> --
> >>>>> Coco mailing list
> >>>>> Coco at maltedmedia.com
> >>>>> http://five.pairlist.net/mailman/listinfo/coco
> >>>>
> >>>>
> >>>>
> >>>> --
> >>>> Coco mailing list
> >>>> Coco at maltedmedia.com
> >>>> http://five.pairlist.net/mailman/listinfo/coco
> >>>
> >>>
> >>> --
> >>> Coco mailing list
> >>> Coco at maltedmedia.com
> >>> http://five.pairlist.net/mailman/listinfo/coco
> >>
> >>
> >>
> >> --
> >> Coco mailing list
> >> Coco at maltedmedia.com
> >> http://five.pairlist.net/mailman/listinfo/coco
> >
> >
> > --
> > Coco mailing list
> > Coco at maltedmedia.com
> > http://five.pairlist.net/mailman/listinfo/coco
>
>
>
> --
> Coco mailing list
> Coco at maltedmedia.com
> http://five.pairlist.net/mailman/listinfo/coco
>