68K ISA project
ethan.dicks at gmail.com
Tue Jul 13 08:56:11 CDT 2010
On Tue, Jul 13, 2010 at 3:30 AM, Ben <bfranchuk at jetnet.ab.ca> wrote:
> The MMU bit reminded about DACK GROUNDED. I think that was the news letter I
> saw on the web.
Close - DTACK Grounded. DTACK is DaTa ACknowledge, a signal pin that
allows for async memory operations... different memory mapped I/O
chips or RAM or ROM, etc., can each have different access times, and
when their subsystem "knows" that data contents are valid, that
subsystem grounds DTACK to tell the processor to continue with the
memory cycle (this is in contrast to 1970s 8-bit microprocessors which
had fixed-length memory cycles, so you had to have memory and I/O
pretty much all the same speed or resort to stall cycles. The trick
about grounding DTACK is that when the right phase of a machine cycle
comes up, it sees DTACK (/DTACK, really) asserted and keeps on going -
as long as your memory and I/O are fast enough, the machine runs at
full speed with a non-complex memory control circuit (ISTR the
COMBOARD design (c. 1982) uses two bi-polar PALs and maybe one or two
TTL parts as well to generate /DTACK at the "right" time, earlier for
DRAM, later for EPROM, etc).
The author of DTACK Grounded was rather fanatical about the issue (you
don't have to read very much to pick up his opinions).
> Guess what ! All about hooking a 68000 to a apple II for numeric processing.
> The news letter may be easier to find than a 68020.
If you mean paper copies vs a real chip, maybe not, but for PDFs, the
second google hit reveals all...
In my own pile of parts, I have tubes of real Synertek and
Motorola-sourced 68000L8 chips, tubes of 68000P8s, and a smattering of
oddballs like 68000L12s, 68000Z8s, and maybe a 68000FN or two, but no
loose 68020s. OTOH, I do have two ancient Mac 68020 boards that I
bought long, long ago (back when one might care to stuff an
accelerator into a machine vs buying a faster machine) but never got
working. ISTR they are stuffed with MC68020RC16s or something close.
I never did much with the '020 in the Amiga world; I jumped right from
the straight 68000 (with the occasional 68010 "upgrade") to the 68030
(A1000 -> A3000)
At one point, a 68000 motherboard with ISA slots would have fascinated
me, but then I started making GG2 Bus+ cards, so any Amiga I run into
that has slots (A2000, A3000, A4000...) can have ISA cards.
Ultimately, I did very little with ISA parallel ports, a bit with ISA
serial cards, enhanced an IDE driver for 1995-era drives, and did a
lot with network interfaces. I always wanted to fiddle with unusual
cards, but the lack of DMA on the GG2 Bus+ curtailed some options.
Something that would still fascinate me is something I saw once - a
68000 CPU (not peripheral card - I have those) for Qbus. More so now
than 20 years ago it'd be a curiosity, but that's the sort of machine
I'd want to try to bring up from bare metal (but it'd be better with
an on-board SCSI controller since real Qbus SCSI controllers are
desired by many, including myself, for MicroVAXen).
More information about the cctalk