Date   : Tue, 26 Jul 2005 23:22:55 +0200 (BST)
From   : Johan Heuseveldt <johan@...>
Subject: Re: ARM copros, speech cartridges, real timeclocks, etc

Hi Jules,

On Tue 26 Jul, Jules Richardson wrote:
> On Tue, 2005-07-26 at 13:54 +0200, Johan Heuseveldt wrote:

> >   Technical: Indeed this is a bus: all connector pins are connected to
> >               each other, except the ULA Select line, which comes from,
> >               lets say, an (74LS)138, allowing 8 2ndProc's. A latch (as
> >               the physical paging register) and address decoding
> >               completes the hardware, perhaps added with an extra power
> >               input.
> 
> So what's the issue with bus loading on the Tube? Does it just not have
> any buffers at all or something (which could be added) - or is there
> more to it than that?

They are certainly an issue. It must have. Well, that seems very logical
to me, as the current cable length is quite short, and must have been a
design issue. And only a single CoPro is expected. That's clear enough.

For the moment I deliberately didn't touch this, as there are some
practicalities involved which need further thoughts, and it looks like
this day just haven't enough hours to write about it, here on the list.
Perhaps another day. If there is an interest I would like to know.

> > A note to Sprow: Would this be of interest of you, to create such
> >                   a 'Tube Selector' board, and make it part of your
> >                   current list of hardware projects?
> >                  More sophisticated approaches are possible. like
> >                   prioriterized interrupts ('like' on the SCSI bus),
> >                   so some brain storming on the list seems valid. :-)
> 
> I thought there were no interrupts at present 

Correct. I didn't say any differently. Did I?

> and it was just done by the beeb polling the Tube every so often - or
> did I misread that?

It's polling constantly. It devotes all its processing time on this
polling, only being interrupted by its own interrupt system. If of
interest I dig out the hand made disassembly, and show you. It's just
a few lines. From that, it can also be seen that VDU has a certain
priority over the rest of the Tube protocols.

> Adding a prioritized interrupt system to something that has no interrupt
> system at all might get messy?

Ah, I see that my writing wasn't clear enough:

It was exactly the point that no interrupt system was in use, ever. But
the Tube ULA /has/ it on board. In the light of more than one CoPro, this
is, or could be, a valueble addition, as it is present in the hardware,
ready to be used, though needing new software.

Of course this software has to be born yet, and is incompatible with the
current use of CoPro's. But that was something I also was trying to make
clear: If we/someone goes for the multiple CoPro's approach, normal use is
out of the question, as it makes no sense allowing two, or even more,
CoPro's writing to the screen.

The idea is to create something totaly different, and with which the
technical user will come up with a total different system. 

> Now if I could get a Z80, 6502, 32016, 80186, 80286 (from the ABC310),
> ARM, Torch Z80, Torch 68000 and Torch Graduate all on the same beeb it'd
> be something of a monster! :-) 
> (I'd be inclined to make the Dragon 64 copro then too)

I have something similar for the last; a Tandy (Radio Shack) CoCo.
(Color Computer)

> Of course there'd probably be all sorts of horrible software conflicts
> having support for that lot in the same beeb at once...

That was part of my story elsewhere. I can't see any use of multiple
CoPro's connected /and/ working, with the software currently used on
them. The point is to go further, where Acorn had stopped.

No doubt they had plans too, otherwise the interrupt from Data Register 4
/in/ the Beeb, caused by writing into Data Register 4 /from/ the CoPro
side, which is /never/ used, wasn't designed to be useless for ever.

If this was not present, then it would make no sense to think about it,
as it means changing the hardware of the Tube ULA. I wouldn't even
think about it, as nothing would have triggered that. But it /is/ there,
creating opportunities which can give so much added elegance to any new
design on our debated Tube/multi CoPro's issue. The actual needed
hardware is basicly simple, but can have some advanced issues as well,
which is a different cup. 

I hope it is more clearly now?


greetings,
Johan

-- 
Johan Heuseveldt <johan@...              >
  aka  waarland

  The best place is a Riscy place
 
"There is much Obi-Wan did not tell you."