17 Jan
2017
17 Jan
'17
8:53 a.m.
On Tue, Jan 17, 2017 at 1:15 AM, Adrian Graham
<witchy at binarydinosaurs.co.uk> wrote:
On 15/01/2017 14:38, "Tony Duell"
<ard.p850ug1 at gmail.com> wrote:
Basically, memory mapped I/O means having devices addressed as memory that
perform I/O functions. You access them with the same load/store
instructions that
you use on real memory. On some processors (6502, 6800, 68000. PDP11, etc)
that's all you have, there are no special I/O instructions. On others
(8080, 8085,
Z80, PDP8, P850, etc) you have special I/O instructions accessing I/O devices.
The address spaces are totally separate, I/O location 0 has nothing to do with
memory location 0. On the 8085, an I/O instructon (IN or OUT) will cause
IO/M to be asserted (other state from when the CPU is accessing memory).
Note that on a machine with I/O instructions (like the 8085) there is
(a) nothing
to stop you having memory mapped I/O (that processor can access memory),
and (b) nothing to stop you having a mix of memory mapped and I/O mapped
I/O. You might have simple devices mapped as I/O ports, but video memory
(which is a sort-of I/O device in that storing something there causes it to
appear on the screen) memory-mapped. As an aside, the TRS-80 model 1
had almost everything (video, keyboard, printer port, etc) _memory mapped_,
the only standard I/O mapped device was the cassette unit.
But do you know it''s not doing I/O. OK
IO/M is never going into the
right state for
I/O, but what that _really_ means is that the 8085 is never executing
any IN or OUT
type instructions. But of course memory mapped I/O is possible
(storing or loading
at particular locations that happen to be I/O devices) on any processor that
can
access memory (including the 8085). I've seen small 8085 and Z80 control
systems
with only memory-mapped I/O.
I pondered that too but the reference says IN and OUT are used for
non-memory mapped I/O and there's a few of those instructions in the code.
Whether they're being executed at this point in time is moot. I've now traced all of them and its associated
pair of supporting chips
(LS04 and an MM74C906) and it's a tape controller, it's only using port 2
and all the lines go to the tape drive header.
They're the big green rectangles visible in this picture -
Ah.... I can see what appear to be thick-film resistors on them
(the black rectangles). Are there more conventional components
on the underside?
There are
also 3 modules on the phone side which I can't find anything
about, marked "NKT NMC1515", NMC1516 and NMC1517.
Are these potted blocks, or can you see the components on them? http://www.binarydinosaurs.co.uk/STCexecutelboard.jpg
The D8741A is above them and the SAA5070 LUCY chip is to the right.
Ah, I'd forgotten there was an SAA5070 on this board...
Could this be part of the serial data transfer?
There will be incoming data
at 1200 baud. There should be some kind of demodulator (maybe one of the
modules) and a serial-to-parallel converter You've not mentioned a serial chip
(is there one), if not then I would expect it to be simulated in software.
Maybe on the 8085, maybe on the 8741.
LUCY does that, it's also where the keyboard connector's lines split off so
the whole data bus goes up to the keyboard module too. I now need to check Now, having just typed that it's making me think
of what Allison said about
lightning or ESD, I know the previous owner of this machine powered it up
before putting it on eb*y and 'the smoke came out' which I thought initially
was just the RIFA mains filter popping (it had), but look at this picture:
http://www.binarydinosaurs.co.uk/STCExecutelBlownCaps.jpg
These are on the tape drive controller board and I thought they'd rotted
through exposure to moisture for several years but could they have exploded
instead? The damage looks old so I don't think that power up is responsible.
What is the tape drive? That board has a distinct look of Philips about it. What
tapes does it use? If I were a gambling man I would guess at Phlips
minicassettes
(not microcassettes). I think I know that drive...
If it is the drive I am thinking of, I have one somewhere, meaning I can look up
the capacitors.
But I would expect the thing to produce video without it.
I don't suppose you have a logic analyser? This is the sort of problem
that would
have me using said instrument to see what the processor is executing.
-tony
