1 Oct
2015
1 Oct
'15
5:30 p.m.
@Chris / All
Unfortunately, I've never eyeballed the logo of the locally-based MCT corp
so I can't comment on it, but again, it's an intriguing thought. And yes,
it's quite possible that the PO knew or ever purchased this Altair from an
employee there. He (PO) indicated that he bought the system second-hand, in
the early 1980s, from an engineering colleague who worked elsewhere in the
locality.
And the PO is the kind of guy who would have hacked-out the cheapest and
most expedient solution to whatever was presented. Don't get me wrong, the
guy is quite sharp and well-educated, but just a real practical 'bailing
wire and duct tape' kinda guy.. half EE, half farmer or something. ;-)
Oh - about it being a 'kit'.. I gave a good look, and it's hard to say for
sure, but the board +is+ totally hand-soldered. But the work is excellent.
So it's either a well-built kit or a smaller-run factory board - but it
wasn't wave / batch soldered.
Also, I'm curious about the 50-pin headers.. why were they never installed?
All of the other work was done, all the expensive parts (chips) are
in-place - so why not a couple of cheap headers? Yes, a few resistor SIPs
are gone, as is one chip missing, but still, you get the point.
On Thu, Oct 1, 2015 at 6:56 PM, Chris Elmquist <chrise at pobox.com> wrote:
On Thursday (10/01/2015 at 06:27PM -0500), drlegendre
. wrote:
Now a question..
Can someone give me a quick rundown on how the CPU communicates with this
board? Does the board show up as a few bytes in the memory map, like on
page zero? Does it connect directly to some registers in the CPU? How
does
data move from the CPU / buss into & out of
the board?
In short, how does the computer know where to "find" the board - and how
do
they converse? I'm only concerend with the
serial portion, the rest is
still a mystery - the 50 pin headers might be anything from parallel
ports
to (proprietary?) controller interfaces.
This is where I'm holding onto my theory that it is a custom design for
a specific purpose-- as I do not see the usual jumpers or DIP switches
to set I/O or memory addresses. If it were a generic card built for
general purpose use, it would almost certainly have DIP switches to set
an I/O or memory address decode.
On S-100 systems, you have both I/O and MEMORY space. Things like serial
port cards were almost always in I/O space and were decoded within a 256
byte block. Different S-100 systems had their console or other I/O at
different addresses and with different chips (8251, 6850, 2661 and other
UARTs with internal register sets as well as TR1602, AY-3-1015 and similar
"dumb" parts).
Without any jumpers or switches to set the decode, you will have to
reverse the design to figure out how they did it. It might be I/O
mapped but it might be memory mapped since it was likely purpose built
for a specific application. Along with this, you will probably not have
operating system support that understands how to talk to this setup so
you will need to modify a CP/M BIOS or other OS I/O support to understand
how to talk to it.
The AY-3-1015 UARTs are "dumb". Their framing format is decided by
strapping inputs to the chip and then they present a byte-wide input
register and output register and strobes to read or write those registers.
This would likely be handled by buffers with enables to gate these paths
onto the S-100 bus.
I concur with Chuck that there were and are a lot of "MCT" companies.
I just have a high confidence that the logo on Bill's board matches the
logos I am familiar with for the local MCT here in town.
Chris
--
Chris Elmquist