please see comments embedded below
regards,
Dick
-----Original Message-----
From: Allison J Parent <allisonp_at_world.std.com>
To: Discussion re-collecting of classic computers
<classiccmp_at_u.washington.edu>
Date: Saturday, August 07, 1999 7:54 AM
Subject: Re: Z-80 hack
><I've got a little job to do which would be easy with an AMPRO Little-Board
><like the REV-0 boards I have got, except that it's too slow. I was
thinkin
><that the easiest way to do this thing is to take a Z-80H from the selected
><12MHz capable ones I have, team it up with a 64Kx8 SRAM, and a set of the
>
>The number of mods to get a LB to run at 12mhz would be better invested in
>building a new board from scratch.
>
Since this is something that's only got to work once, I thought I'd put the
SRAM on a small scrap of perf, along with a PAL to fiddle with the timing.
The reason for not fiddling with the nWAIT line is that it doesn't time
right with the instruction fetches, such that the peripherals will see the
RETI opcode. If this gets too involved, I'll simply have to suffer through
doing it from the parallel port on the PC, but I already have some code
written many years ago (in BASIC) which does the job nicely, and I'll
probably never use it again.
I'm working on restoring a dozen or so 8" floppy drives which malfinction in
one sense or another, to the extent that I can't verify/adjust their
alignment until these repairs are effected. The code was written for an
S-100 box I used back in the '80's which ran a 12MHz (hand-picked) CPU
together with some fast SRAMS. I don't want to fire up the S-100 stuff
right now, and the AMPRO provides sufficient packaging to allow me to
install a mezzanine board on the processor socket. My HDC adapter sits
there anyway, and the code is on that drive. It steps through the process
of troubleshooting the logic on Shugart 800/801 boards, Siemens 100-8's, and
a few others I don't have to deal with today. The timing probably isn't
even that critical, but I prefer to handle the job this way since I've given
away the hardware I previously used. The thing I have to build is just a
port in and a port out, to monitor and control the signals to/from the drive
without having to read and write. This also allows monitoring of the
signals at various test points in the circuit without having to use the
'scope, which only does half the job.
>
><'A' type peripherals, provided I switch the clock to 4MHz each time I see
a
><IORQ* or an (M1* & I/O*) or a RETI on the data bus during M1*, resetting
th
><higher speed on the next M1*.
>
>there is a wait line for this kind of stuff. I'd upgrade to faster
>peripherals 6 or 8 mhz ones are still available.
>
I found that the wait line doesn't work for this sort of thing. It's got to
do with the way the peripherals work. Switching the clock rate does work,
though.
>
>The FDC, SIO, CTC, RAM(and timing chain) all have Tacc that are way to slow
>for even 6mhz.
>
>I hacked one to 8mhz and then put it back, it was never as solid in
>operation.
>
Their RAM timing was stupid.
>
><Have you ever tried something like that? What am I overlooking here?
>
>Basic design. Many of the utilities are timing sensitive. Catch TCJ and
>order some back issues for the 8mhz mod. The back issues are still
>available.
>
What's a TCJ?
>
>Also get issue 79, the P112 board may really be what you want.
>
><I used to do things like that with the 65xx and 68xx parts all the time,
bu
><never with a Z-80 because of the peripherals and the mode-2 interrupt,
which
><the AMPRO board probably doesn't even use.
>
>The ampro does use mode 2, otherwise there is little point to using z80
>peripherals.
>
I never saw much advantage in using Z-80 with mode-2 interrupts anyway. I
guess it's just a matter of style. I like simple interrupts, and if it gets
too complicated, use another system to handle the I/O that makes it
complicated. It's much easier to use three computers than to figure out how
to use one with three interrupts. I guess I'm lazy . . .
>
>allison
>
>
Received on Sat Aug 07 1999 - 12:24:37 BST