[Coco] Grr!!! Bad MPI! No biscuit!

Sat May 5 16:57:02 EDT 2007

I believe Marty Goodman once recommended that the ls245 in the CoCo AND
the MPI be replaced with a faster version (the ''F'' version, I think
he mentioned). -r

> -------- Original Message --------
> Subject: Re: [Coco] Grr!!!  Bad MPI!  No biscuit!
> From: Gene Heskett <gene.heskett at verizon.net>
> Date: Sat, May 05, 2007 1:48 pm
> To: CoCoList for Color Computer Enthusiasts <coco at maltedmedia.com>
> 
> On Saturday 05 May 2007, Joel Ewy wrote:
> >Gene Heskett wrote:
> >> On Saturday 05 May 2007, Joel Ewy wrote:
> >>> Jim Cox wrote:
> >>>> I have a MPI board that someone gave me.  I am not sure if it
> works or
> >>>> not, but it's yours.
> >>>>
> >>>> Jim
> >>>> PS: We need a new MPI for the CoCo
> >>>
> >>> Here's a humble suggestion for anyone contemplating such a thing:
> >>> Orient the board perpendicular to the desktop so that any cartridges
> >>> plugged into it will be parallel with the desktop.  Having the disk
> >>> controller stick up 8.5" from the surface of the desk is just a bit
> >>> silly.
> >>>
> >>> As much as I loathe trying to find places to plug them all in,
> maybe an
> >>> external transformer would be a good idea here...
> >>>
> >>> Would it be feasible to put bi-directional bus drivers in a short
> >>> cartridge and connect the CoCo to the MPI via a short length of
> ribbon
> >>> cable?  That would provide much greater flexibility in placement, and
> >>> probably reduce the chances of bumping things out of place.
> >>>
> >>> JCE
> >>
> >> This idea falls over due to a bus contention in the mpi's internal
> timing,
> >> which in turn causes a ground bounce that is only barely tolerable
> if the
> >> grounding ear clips in the side port have a good clean grip on the
> mpi's
> >> ground plane.  Additional jumpers of heavy wire to tie the mpi's
> rather
> >> surpentine ground plane together also help, as does additional
> bypassing
> >> near the ls245.  You can make it stable, as shipped, its a house of
> cards.
> >
> >Are you saying that this problem is intrinsic to any CoCo multi-pak
> >design, or just Rad Shak's?  FWIW, I'm musing about a possible future
> >product, not modifications to the existing design.
> 
> I'd say it would be pretty well builtin, Joel, unless someone actually
> went to 
> the trouble of fixing the timing problems, which are in actual fact
> quite 
> narrow, existing for not more than 10ns, but during that time its an
> absolute 
> crowbar across the ls245 in the mpi, and whatever is driving the buss
> on the 
> coco's side of the edge connector.  The ground bounce, measured from the 
> coco's ground plane, say at the contact ears, over to a ground plane 
> connection around the ls245 and its environs, can exceed a volt for 8
> to 10 
> ns in the shacks first version.  The mods I made to mine pull it down to 
> around 150mv, but its still there.  They also lowered the operating 
> temperature of that ls245 by a noticeable with the back of your finger 
> amount.
> 
> I think, but won't swear, that its the turn around delay in the ls245
> itself, 
> and that one of the faster chip families version of it might alleviate
> it at 
> least as well as the extra grounds and bypassing I added to mine did.  I 
> should have taken the time to socket it and experiment but didn't. :(
> 
> There needs to be a few ns of dead time, where the ls245 is truly open
> in both 
> directions on a switch in directions, but I'm not sure how one would
> go about 
> separating the read and write functions in order to achieve a buss
> that's 
> tri-stated on both sides of it.  That is whats needed to eliminate it
> though.  
> I'd guess that 5ns of dead time would be sufficient, based on what my
> measly 
> 100mhz dual trace scope says.  5ns is of course faster than my scope,
> but it 
> is useable to around 200 mhz as I've used it to look at a channel 8 tv 
> transmitter (175mhz), just not anywhere near full height traces when
> stuff 
> that fast goes by. 10ns should cover it all, and that still leaves the
> coco 
> with a couple of eons to do stable reads and writes through it.
> 
> >JCE
> >
> >>>> On Sat, 5 May 2007 03:12:01 -0600
> >>>>
> >>>>  Willard Goosey <goosey at virgo.sdc.org> wrote:
> >>>>> My old-style MPI is apparently toast.  My CoCo has therefore lost
> >>>>> access to its hard drive.  :-(
> >>>>>
> >>>>> I suspect power supply problems.  Either that or the upgrade PAL is
> >>>>> loosing its mind.
> >>>>>
> >>>>> Just what I need, something else broken.
> >>>>> Willard
> >>>>> --
> >>>>> Willard Goosey  goosey at sdc.org
> >>>>> Socorro, New Mexico, USA
> >>>>> "I've never been to Contempt!  Isn't that somewhere in New Mexico?"
> >>>>>   --- Yacko
> >>>>>
> >>>>> --
> >>>>> 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
> 
> 
> 
> -- 
> Cheers, Gene
> "There are four boxes to be used in defense of liberty:
>  soap, ballot, jury, and ammo. Please use in that order."
> -Ed Howdershelt (Author)
> Inara: "It wasn't entirely a disaster." 
> 
> Mal: "I got stabbed! Right here!" 
> 
> Inara: "You also lined up exciting new crime."
> 				--Episode #4, "Shindig"
> 
> --
> Coco mailing list
> Coco at maltedmedia.com
> http://five.pairlist.net/mailman/listinfo/coco