29 Jan
2010
29 Jan
'10
7 a.m.
Having learned assembly language programming on the beautifully simple architecture and
instruction set of the 6800, the Byte magazine article linked to below that I read when it
was originally published really impressed me. In the 6809 they made one of the earliest
efforts I know of to really tweak an already great uP instruction set based upon an
analysis of existing software:
http://tlindner.macmess.org/?page_id=119
29 Jan
29 Jan
11:26 a.m.
Having learned assembly language programming on the
beautifully
simple architecture and instruction set of the 6800, the Byte magazine
article linked to below that I read when it was originally published
really impressed me. In the 6809 they made one of the earliest efforts
I know of to really tweak an already great uP instruction set based upon
an analysis of existing software:
I found the 6809 to be by far the nicest 8-bit CPU I ever worked with.
The instruction set was simple and very orthogonal, the fact that you had
various relative addressing modes meant you could write truely
position-independant code, there were 2 stack points, and so on. Unlike
certain chips I could name, there were no major misfeatures that I came
across.
Of course the problem (as we all know) is that it came out too late. By
that tine everybody was using the Z80 or 6502. Oh well.
It always suprised me that hre BBC micro used the 6502 rather than the
6809. By the time the Beeb was designed, Acorn had made a 6809 processor
board for their System machines, so they must have had experience with
the chip. THe Beeb is nice, but a Beeb with a 6809 processor would have
been something else :-)
-tony
11:46 a.m.
Tony Duell wrote:
Around here, it was Apple ]['s and C-64's. The Coco was the only cost
6809 machine I can think of, but RS designed for BASIC rather than
business machine. RS did have a 68000 machine, but I think the next year
they switched to in house PC clones.
Having learned
assembly language programming on the beautifully
simple architecture and instruction set of the 6800, the Byte magazine
article linked to below that I read when it was originally published
really impressed me. In the 6809 they made one of the earliest efforts
I know of to really tweak an already great uP instruction set based upon
an analysis of existing software:
I found the 6809 to be by far the nicest 8-bit CPU I ever worked with.
The instruction set was simple and very orthogonal, the fact that you had
various relative addressing modes meant you could write truely
position-independant code, there were 2 stack points, and so on. Unlike
certain chips I could name, there were no major misfeatures that I came
across.
Of course the problem (as we all know) is that it came out too late. By
that tine everybody was using the Z80 or 6502. Oh well. It always suprised me that hre BBC micro used the 6502
rather than the
6809. By the time the Beeb was designed, Acorn had made a 6809 processor
board for their System machines, so they must have had experience with
the chip. THe Beeb is nice, but a Beeb with a 6809 processor would have
been something else :-)
I thought the BBC micro was designed, just before the 6809 came out.
I still think the 6502 was cheaper than the 6800. I think the 6502
was ~$50 when it came out compared to ~$250 of the 8080A and 6800.
-tony
Ben.
12:49 p.m.
----- Ben wrote:
Around here, it was Apple ]['s and C-64's. The
Coco was the only cost
6809 machine I can think of, but RS designed for BASIC rather than
business machine. RS did have a 68000 machine, but I think the next year
they switched to in house PC clones.
The Coco was also capable of running OS-9, I had this.
They also had one for the SWTPC-6809 and other SS-50 bus machines
It was even ported to the 68K
And it was a quite powerful OS for it's time on an 8bitter- even more than CP/M.
=Dan
--
http://www.vintagecomputer.net/ragooman/
12:50 p.m.
Tony Duell wrote:
Around here, it was Apple ]['s and C-64's. The Coco was the only cost
6809 machine I can think of, but RS designed for BASIC rather than
business machine. RS did have a 68000 machine, but I think the next year
they switched to in house PC clones.
The only common 6809 machine in the UK was the Dragon, which was based
on the same Motorola application note as the CoCo. Of cource the CoCo was
also sold here (althohgh AFAIK the CoCo3 never was).
Acorn made a 6809 CPU board for their Eurocard based System machines, it
normally ran Flex09. It's not easy to find.
The curious Tiger used a 68B09 for I/O (along with a Z80A for the main
processor and a 7220 graphics chip).
There were, of course, various special-purpose embeeded 6809 controller
boards.
Having
learned assembly language programming on the beautifully
simple architecture and instruction set of the 6800, the Byte magazine
article linked to below that I read when it was originally published
really impressed me. In the 6809 they made one of the earliest efforts
I know of to really tweak an already great uP instruction set based upon
an analysis of existing software:
I found the 6809 to be by far the nicest 8-bit CPU I ever worked with.
The instruction set was simple and very orthogonal, the fact that you had
various relative addressing modes meant you could write truely
position-independant code, there were 2 stack points, and so on. Unlike
certain chips I could name, there were no major misfeatures that I came
across.
Of course the problem (as we all know) is that it came out too late. By
that tine everybody was using the Z80 or 6502. Oh well. It always suprised me that hre BBC micro used the
6502 rather than the
6809. By the time the Beeb was designed, Acorn had made a 6809 processor
board for their System machines, so they must have had experience with
the chip. THe Beeb is nice, but a Beeb with a 6809 processor would have
been something else :-)
I thought the BBC micro was designed, just before the 6809 came out. I still think the 6502 was cheaper than the 6800. I
think the 6502
was ~$50 when it came out compared to ~$250 of the 8080A and 6800.
That's almost certainly true. The 6502 was a cheap processor when it
first came out, But by about 1980 the other chips had really dropped in
price.
-tony
1:34 p.m.
On 29/01/2010 20:50, Tony Duell wrote:
I'm not sure. The Beeb was 1982 or so.
1981. The 6809 came out a little before that; I'd guess 1979-80 because
although it's listed in a couple of 1980 data books I have, the specs
are incomplete. Acorn's 6809 board came out in late 1980, so the 6809
MPU can't be any later than that.
When the Beeb was designed, Acorn had a lot of expertise with the 6502,
and since they hadn't much time, I suspect the 6502 was an easier and
more obvious choice. Probably cheaper, too, at that time.
--
Pete Peter Turnbull
Network Manager
University of York
1 Feb
1 Feb
5:32 a.m.
Tony Duell wrote:
No... I spent many a year searching, too. I have the manual, which I think
includes the schematic (if not, I got that separately), but I never did come
across the physical board.
I'm not sure. The Beeb was 1982 or so. I thought the 6809 was out by
then. If not, it implies Acorn were still working on their System
machines, which seems a little curious.
I'm not sure quite when Acorn sold the rights to the System range to Control
Universal, but 1984 sticks in my head for some reason. If so, it implies that
in 1982 they still thought there was some mileage left in the System hardware.
I suppose that at the very least there were still quite a few System-based
fileservers around.
Remember that there was a *lot* of in-house expertise in 6502 development; the
Atom was almost a "System rack in a home computer box", just as the BBC was an
"Atom on steroids". I'm sure they considered other CPUs, but it made sense
to
stick with what they knew.
cheers
Jules
Tony Duell
wrote:
Around here, it was Apple
]['s and C-64's. The Coco was the only cost
6809 machine I can think of, but RS designed for BASIC rather than
business machine. RS did have a 68000 machine, but I think the next year
they switched to in house PC clones.
The only common 6809 machine in the UK was the Dragon, which was based
on the same Motorola application note as the CoCo. Of cource the CoCo was
also sold here (althohgh AFAIK the CoCo3 never was).
Acorn made a 6809 CPU board for their Eurocard based System machines, it
normally ran Flex09. It's not easy to find. Having
learned assembly language programming on the beautifully
simple architecture and instruction set of the 6800, the Byte magazine
article linked to below that I read when it was originally published
really impressed me. In the 6809 they made one of the earliest efforts
I know of to really tweak an already great uP instruction set based upon
an analysis of existing software:
I found the 6809 to be by far the nicest 8-bit
CPU I ever worked with.
The instruction set was simple and very orthogonal, the fact that you had
various relative addressing modes meant you could write truely
position-independant code, there were 2 stack points, and so on. Unlike
certain chips I could name, there were no major misfeatures that I came
across.
Of course the problem (as we all know) is that it came out too late. By
that tine everybody was using the Z80 or 6502. Oh well. The curious Tiger used a 68B09 for I/O (along with a
Z80A for the main
processor and a 7220 graphics chip).
The HH machine? I've never seen a real one, but it sounded like an interesting
beast.
There were, of course, various special-purpose
embeeded 6809 controller
boards.
I do have a Control Universal 6809 board (I don't recall if it's System bus or
STE now), so making that 'do' something is a possibility one day.
I pulled a whole stack of 6809s off some scrap telephone exchange boards many
years ago. I'm not sure whose product it was - AFAIK BT never used an exchange
which had 6809s, so it must have been some kind of private branch exchange.
It always suprised me that hre BBC micro used the 6502
rather than the
6809. By the time the Beeb was designed, Acorn had made a 6809 processor
board for their System machines, so they must have had experience with
the chip. THe Beeb is nice, but a Beeb with a 6809 processor would have
been something else :-)
I thought the BBC micro was designed, just before the 6809
came out.
12:02 p.m.
No... I spent many a year searching, too. I have the
manual, which I think
includes the schematic (if not, I got that separately), but I never did come
across the physical board.
You're going to hate me. I've got _2_ of them. One with the normal ROM
for the 40 column VDU card, the other with a ROM for the rare 80 column
VDU (and yes, I have that card too).
The curious Tiger used a 68B09 for I/O (along
with a Z80A for the main
processor and a 7220 graphics chip).
The HH machine? I've never seen a real one, but it sounded like an interesting
beast. Remember that there was a *lot* of in-house expertise
in 6502 development; the
Atom was almost a "System rack in a home computer box", just as the BBC was an
To the extent that the expansion bus was a System bus connector. You
could fit one System eurocard inside an Atom -- at least one of my Atoms
has the econet interface doen that way (another one of my Atoms has the
'proper' econet interface PCB plugged into the solder side of the mainboard).
The video ciorcuitry was very differnt, though. THe Atom used the 6847
VDG chip. The System VDU cards used 6845s, and in the case of the 40
column one, an SAA5050 teletext charactger geenrator ROM.
-tony
2:36 p.m.
Tony Duell wrote:
Git ;) Was the 80-col VDU just the stock System one (Acorn p/n 200,019), or
something else? I've got at least one of the System ones (along with many of
the other System boards that they made). The 6809 CPU board has always proved
elusive, though...
I've never heard of another 80x25 board than the '019 one, but I think my
System-era document collection's only about 80% complete so I might be missing
something!
There's one under my bed (seriously). Alas I don't have the disk unit,
which makes it farily useless (the ROMs in mine are simply a disk
bootstrap).
Yes, I remember you telling me about it... I recall the museum being offered
one quite a few years ago, but then the donor went all quiet (as sometimes
happens!) so I've never seen a 'live' one.
Was there something unusual (local processor or strange interface etc.) about
the disk unit that makes it hard to just couple some drives up?
Indeed - I've seen a couple of Atoms with System lab interface boards
shoe-horned into the case like that (and my Atom with the disk unit just uses
a System FDC, as I'm sure you know). I've seen two different board layouts for
System Econet cards, incidentally.
No... I spent
many a year searching, too. I have the manual, which I think
includes the schematic (if not, I got that separately), but I never did come
across the physical board.
You're going to hate me. I've got _2_ of them. One with the normal ROM
for the 40 column VDU card, the other with a ROM for the rare 80 column
VDU (and yes, I have that card too). The curious Tiger used a 68B09 for I/O (along with a
Z80A for the main
processor and a 7220 graphics chip).
The HH machine? I've never seen a real
one, but it sounded like an interesting
beast. Remember that
there was a *lot* of in-house expertise in 6502 development; the
Atom was almost a "System rack in a home computer box", just as the BBC was an
To the extent that the expansion bus was a System bus connector. You
could fit one System eurocard inside an Atom -- at least one of my Atoms
has the econet interface doen that way (another one of my Atoms has the
'proper' econet interface PCB plugged into the solder side of the mainboard).
The video ciorcuitry was very differnt, though. THe
Atom used the 6847
VDG chip. The System VDU cards used 6845s, and in the case of the 40
column one, an SAA5050 teletext charactger geenrator ROM.
Yep. I always wanted to build a colour encoder board for the Atom (I've got
the documentation, but never had the physical board) but maybe that's
irrelevant now I'm in a non-PAL country :-)
Oh, I've got a couple of Control Universal "CU Graph" graphics boards with a
System bus. These use the EF9366 graphics display processor for the "grunt
work". There are two boards in each set, one containing the I/O decoding
logic, 9366 IC, 16KB of DRAM and video outputs, the other containing 32KB of
DRAM and a mysterious* 26-way IDC connector brought out to the card edge.
According to the EF9366 docs, the boards should do 256x512 in 8 basic colours.
* I don't have proper documentation for them and have never traced a schematic
out. I wondered about light-pen connectors (and the 9366 chip does support a
light pen), but 26 pins seems a bit excessive. Maybe they just bring some
other useful signals out to that connector.
cheers
Jules
2:51 p.m.
Tony Duell wrote:
Git ;) Was the 80-col VDU just the stock System one (Acorn p/n 200,019), or
something else? I've got at least one of the System ones (along with many of
As far as I know it'sjust hte standard System one. The 6809 and 6502
buses are very similar, so there'd be no need to have special cards for
the 6809. Certainly the memory and FDC cards are the normal ones.
No... I
spent many a year searching, too. I have the manual, which I think
includes the schematic (if not, I got that separately), but I never did come
across the physical board.
You're going to hate me. I've got _2_ of them. One with the normal ROM
for the 40 column VDU card, the other with a ROM for the rare 80 column
VDU (and yes, I have that card too). the other System boards that they made). The 6809 CPU
board has always proved
elusive, though...
I've never heard of another 80x25 board than the '019 one, but I think my
System-era document collection's only about 80% complete so I might be missing
something!
There's one under my bed (seriously). Alas I don't have the disk unit,
which makes it farily useless (the ROMs in mine are simply a disk
bootstrap).
Yes, I remember you telling me about it... I recall the museum being offered
one quite a few years ago, but then the donor went all quiet (as sometimes
happens!) so I've never seen a 'live' one.
Was there something unusual (local processor or strange interface etc.) about
the disk unit that makes it hard to just couple some drives up?
The disk cotnroller was in the drive unit. The connectors on the bottom
of the Tiger are essentially system buses (and all 3 are differnt -- I
think 2 of them carry differen subsets of the Z80 bus, the other one the
6809 bus, but I might be mis-remembering it).And I don't have schematics
of the disk controller board. I suppose I could disassmeble the ROMs and
see what sort of FDCchip it's probalby looking for and at what address,
but I've got a lot of other projects to complete first.
The curious Tiger used a 68B09 for I/O (along with a
Z80A for the main
processor and a 7220 graphics chip).
The HH machine? I've never seen a real
one, but it sounded like an interesting
beast. Remember
that there was a *lot* of in-house expertise in 6502 development; the
Atom was almost a "System rack in a home computer box", just as the BBC was an
To the extent that the expansion bus was a System bus connector. You
could fit one System eurocard inside an Atom -- at least one of my Atoms
has the econet interface doen that way (another one of my Atoms has the
'proper' econet interface PCB plugged into the solder side of the mainboard).
a System FDC, as I'm sure you know). I've seen
two different board layouts for
System Econet cards, incidentally.
My guess is that it's the connector for a parallel-interfaced keyboard,
thus putting human input and output on the same PCB.
-tony
The video ciorcuitry was very differnt, though.
THe Atom used the 6847
VDG chip. The System VDU cards used 6845s, and in the case of the 40
column one, an SAA5050 teletext charactger geenrator ROM.
Yep. I always wanted to build a colour encoder board for the Atom (I've got
the documentation, but never had the physical board) but maybe that's
irrelevant now I'm in a non-PAL country :-)
Oh, I've got a couple of Control Universal "CU Graph" graphics boards with
a
System bus. These use the EF9366 graphics display processor for the "grunt
work". There are two boards in each set, one containing the I/O decoding
logic, 9366 IC, 16KB of DRAM and video outputs, the other containing 32KB of
DRAM and a mysterious* 26-way IDC connector brought out to the card edge.
According to the EF9366 docs, the boards should do 256x512 in 8 basic colours.
* I don't have proper documentation for them and have never traced a schematic
out. I wondered about light-pen connectors (and the 9366 chip does support a
light pen), but 26 pins seems a bit excessive. Maybe they just bring some
other useful signals out to that connector. 
29 Jan
29 Jan
7:20 p.m.
Tony Duell wrote:
When I moved from the 6800 to the 6809 (in assembly language - *many*
years ago) I was sort of astounded and at the same time very pleased by
the way many of the little subroutines I had written for the 6800 became
one instruction in the 6809. I think it will always be my favorite
8-bit CPU. My only annoyance at the time was the fact that there was no
way for the software to reset the companion UART chip, whose number I've
completely forgotten by now. 6821 maybe???
Later,
Charlie Carothers
Having learned
assembly language programming on the beautifully
simple architecture and instruction set of the 6800, the Byte magazine
article linked to below that I read when it was originally published
really impressed me. In the 6809 they made one of the earliest efforts
I know of to really tweak an already great uP instruction set based upon
an analysis of existing software:
I found the 6809 to be by far the nicest 8-bit CPU I ever worked with.
The instruction set was simple and very orthogonal, the fact that you had
various relative addressing modes meant you could write truely
position-independant code, there were 2 stack points, and so on. Unlike
certain chips I could name, there were no major misfeatures that I came
across.
Of course the problem (as we all know) is that it came out too late. By
that tine everybody was using the Z80 or 6502. Oh well.
It always suprised me that hre BBC micro used the 6502 rather than the
6809. By the time the Beeb was designed, Acorn had made a 6809 processor
board for their System machines, so they must have had experience with
the chip. THe Beeb is nice, but a Beeb with a 6809 processor would have
been something else :-)
-tony
30 Jan
30 Jan
10:28 a.m.
When I moved from the 6800 to the 6809 (in assembly
language - *many*
years ago) I was sort of astounded and at the same time very pleased by
the way many of the little subroutines I had written for the 6800 became
one instruction in the 6809. I think it will always be my favorite
8-bit CPU. My only annoyance at the time was the fact that there was no
I was a confirmed Z80 hacker until I bought a CoCo 2 in a clearout sale.
And after reading the 6809 data sheet I fell in love with that chip.
way for the software to reset the companion UART chip,
whose number I've
completely forgotten by now. 6821 maybe???
The 6821 is the parallel chip (2 16 bit perts). The serial chip, commonly
called te ACIA is the 6850. It needs an external baud rate generator.
There was also the 6852 synchronous serial chip and the 6854 which did
some other flavour of synchonous comms and was used in the Acorn Econet
system
Of course you could also use 6502-series I/O chips on a 6809 (and vice
versa), the buses are essentially the sane,. The 6522 VIA is a nicer
parallel chip (with counter/timers, a shift register, etc). The 6551 is a
serial chip with a built-in baud rate generator.
-tony
8:19 p.m.
Tony Duell wrote:
Many thanks to you and Henk for reminding me what the Motorola "UART"
chips were called/numbered. All this discussion is making me want a
CoCo unfortunately. Never had one when they were "new" but always
thought they would be a lot of fun - maybe a garage sale will turn one
up some day. Just what I don't need of course - yet another project or
piece of hardware.
Later,
Charlie C.
When I moved
from the 6800 to the 6809 (in assembly language - *many*
years ago) I was sort of astounded and at the same time very pleased by
the way many of the little subroutines I had written for the 6800 became
one instruction in the 6809. I think it will always be my favorite
8-bit CPU. My only annoyance at the time was the fact that there was no
I was a confirmed Z80 hacker until I bought a CoCo 2 in a clearout sale.
And after reading the 6809 data sheet I fell in love with that chip.
way for the software to reset the companion UART
chip, whose number I've
completely forgotten by now. 6821 maybe???
The 6821 is the parallel chip (2 16 bit perts). The serial chip, commonly
called te ACIA is the 6850. It needs an external baud rate generator.
There was also the 6852 synchronous serial chip and the 6854 which did
some other flavour of synchonous comms and was used in the Acorn Econet
system
Of course you could also use 6502-series I/O chips on a 6809 (and vice
versa), the buses are essentially the sane,. The 6522 VIA is a nicer
parallel chip (with counter/timers, a shift register, etc). The 6551 is a
serial chip with a built-in baud rate generator.
-tony
1:32 a.m.
From: "CSquared" <csquared3 at tx.rr.com>
Sent: Saturday, January 30, 2010 4:20 AM
To: "General Discussion: On-Topic Posts Only" <cctech at classiccmp.org>
Subject: Re: 6809 SBC
When I moved from the 6800 to the 6809 (in assembly
language - *many*
years ago) I was sort of astounded and at the same time very pleased by
the way many of the little subroutines I had written for the 6800 became
one instruction in the 6809. I think it will always be my favorite
8-bit CPU. My only annoyance at the time was the fact that there was no
way for the software to reset the companion UART chip, whose number I've
completely forgotten by now. 6821 maybe???
Later,
Charlie Carothers
I am somewhat a Motorola-guy, so a bit biased :-) Of the 8 bit CPUs the
6809 indeed would be my favorite. The 6821 PIA (parallel I/O) has the RST*
pin, the 6850 ACIA (serial I/O) does not.
- Henk.
7:57 a.m.
Henk Gooijen wrote:
I am somewhat a Motorola-guy, so a bit biased :-) Of
the 8 bit CPUs the
Umm so why do all your projects end up with PDP 11's or
8's :)
6809 indeed would be my favorite. The 6821 PIA
(parallel I/O) has the RST*
pin, the 6850 ACIA (serial I/O) does not.
As I am currently working a CPLD/2901 homebrew computer design I will
have to say that is my favorite at the moment. A 12 bit op code does
make for a nice 8 bit micro style of computer, and I can manage to
sqeaze in a full 18 bit address space. +-15 bits branches/18 bits calls.
Ben.
- Henk.
8:09 p.m.
Henk Gooijen wrote:
From: "CSquared" <csquared3 at
tx.rr.com>
Sent: Saturday, January 30, 2010 4:20 AM
To: "General Discussion: On-Topic Posts Only" <cctech at classiccmp.org>
Subject: Re: 6809 SBC
Ah yes, 6850 and ACIA, thanks for the memory jog. I had completely
forgotten both of those. I can't remember now precisely why I felt the
need to reset the ACIA with a software controlled line, but ISTR I
*really* needed to occasionally, and a power cycle was the only way. I
do remember the little system had several of the ACIA's attached to one
6809 and I controlled a number of unique hardware devices that way. I
got to do some 68000 stuff a few years later and liked it quite well
too, though that was using C - my very first C as a matter of fact,
along about 1986 or 1987.
Later,
Charlie C.
When I moved from the 6800 to the 6809 (in
assembly language - *many*
years ago) I was sort of astounded and at the same time very pleased
by the way many of the little subroutines I had written for the 6800
became one instruction in the 6809. I think it will always be my
favorite 8-bit CPU. My only annoyance at the time was the fact that
there was no way for the software to reset the companion UART chip,
whose number I've completely forgotten by now. 6821 maybe???
Later,
Charlie Carothers
I am somewhat a Motorola-guy, so a bit biased :-) Of the 8 bit CPUs the
6809 indeed would be my favorite. The 6821 PIA (parallel I/O) has the RST*
pin, the 6850 ACIA (serial I/O) does not.
- Henk. 
31 Jan
31 Jan
11:49 a.m.
On 1/30/10 8:09 PM, "CSquared" <csquared3 at tx.rr.com> wrote:
Ah yes, 6850 and ACIA, thanks for the memory jog. I
had completely
forgotten both of those. I can't remember now precisely why I felt the
need to reset the ACIA with a software controlled line, but ISTR I
*really* needed to occasionally, and a power cycle was the only way. I
do remember the little system had several of the ACIA's attached to one
6809 and I controlled a number of unique hardware devices that way. I
got to do some 68000 stuff a few years later and liked it quite well
too, though that was using C - my very first C as a matter of fact,
along about 1986 or 1987.
Later,
Charlie C.
Hmmm, possibly It is because I am more of a hardware guy anymore than a
software guy, but couldn't you make an adapter pcb for the ACIA chip, that
had a relay and a few other componants on it, like a 555 set as a timer, and
when the 555 is triggered, it would power off your acia for a preset amount
of time by tripping the relay so the VCC pin no longer gets voltage, then at
the end of the timing cycle it returns the relay to normal position and the
acia comes back on and you set baud, etc....
12:46 p.m.
On 31/01/2010 19:49, Geoffrey Reed wrote:
On 1/30/10 8:09 PM, "CSquared" <csquared3
at tx.rr.com> wrote:
> Ah yes, 6850 and ACIA, thanks for the memory jog. I had completely
> forgotten both of those. I can't remember now precisely why I felt the
> need to reset the ACIA with a software controlled line, but ISTR I
> *really* needed to occasionally, and a power cycle was the only way. I
> do remember the little system had several of the ACIA's attached to one
> 6809 and I controlled a number of unique hardware devices that way.
Hmmm, possibly It is because I am more of a hardware
guy anymore than a
software guy, but couldn't you make an adapter pcb for the ACIA chip, that
had a relay and a few other componants on it, like a 555 set as a timer, and
when the 555 is triggered, it would power off your acia for a preset amount
of time by tripping the relay so the VCC pin no longer gets voltage, then at
the end of the timing cycle it returns the relay to normal position and the
acia comes back on and you set baud, etc....
I am really puzzled by this. The ACIA has no hardware reset line, but
it's very easy to send it a master reset command in software. In fact,
you have to do that to clear the hardware reset condition after a power
up, because the ACIA stays in the inactive/reset condition until you
program it.
Charlie, you weren't using GTE 68C50s, were you? They did have a few
weird bugs.
--
Pete Peter Turnbull
Network Manager
University of York
9:37 p.m.
Pete Turnbull wrote:
On 31/01/2010 19:49, Geoffrey Reed wrote:
I'm pretty sure they were Motorola parts. As I told Geoffrey, it has
been a very long time - like probably 40 years - and I've forgotten a
lot of the details, and just sort of remember the pain. I think the
ACIA would somehow get into a state where it would not pay attention to
the master reset command.
Over the years, because of this and other similar experiences, I sort of
developed a philosophy that every peripheral of an embedded controller
should have a software controlled reset line that could unconditionally
reset the device independent of the state of the device. Without this
capability, I don't believe it is possible to create a truly robust
system. By robust I mean a system that can continue to run for months
with no intervention needed to clear an error condition. Another
example of the sort of experience that led to this philosophy are the
modems I mentioned to Geoffrey. A reset command to the modem has no
effect if the firmware in the modem is off in the weeds somewhere.
Later,
Charlie C.
On 1/30/10 8:09 PM, "CSquared"
<csquared3 at tx.rr.com> wrote:
> Ah yes, 6850 and ACIA, thanks for the memory jog. I had completely
> forgotten both of those. I can't remember now precisely why I felt the
> need to reset the ACIA with a software controlled line, but ISTR I
> *really* needed to occasionally, and a power cycle was the only way. I
> do remember the little system had several of the ACIA's attached to one
> 6809 and I controlled a number of unique hardware devices that way.
Hmmm, possibly It is because I am more of a
hardware guy anymore than a
software guy, but couldn't you make an adapter pcb for the ACIA chip,
that
had a relay and a few other componants on it, like a 555 set as a
timer, and
when the 555 is triggered, it would power off your acia for a preset
amount
of time by tripping the relay so the VCC pin no longer gets voltage,
then at
the end of the timing cycle it returns the relay to normal position
and the
acia comes back on and you set baud, etc....
I am really puzzled by this. The ACIA has no hardware reset line, but
it's very easy to send it a master reset command in software. In fact,
you have to do that to clear the hardware reset condition after a power
up, because the ACIA stays in the inactive/reset condition until you
program it.
Charlie, you weren't using GTE 68C50s, were you? They did have a few
weird bugs.
1 Feb
1 Feb
12:32 a.m.
On 01/02/2010 05:37, CSquared wrote:
Pete Turnbull wrote:
> I am really puzzled by this. The ACIA has no
hardware reset line, but
> it's very easy to send it a master reset command in software.
I'm pretty sure they were Motorola parts. As I
told Geoffrey, it has
been a very long time - like probably 40 years - and I've forgotten a
lot of the details, and just sort of remember the pain. I think the
ACIA would somehow get into a state where it would not pay attention to
the master reset command.
I don't recall that ever happening to me, but I can well believe it :-)
--
Pete Peter Turnbull
Network Manager
University of York
7:46 a.m.
Pete Turnbull wrote:
On 01/02/2010 05:37, CSquared wrote:
Intel parks I can see, Motorola parts no!
Pete Turnbull wrote:
> I am really puzzled by this. The ACIA has no
hardware reset line, but
> it's very easy to send it a master reset command in software.
I'm pretty sure they were Motorola parts. As
I told Geoffrey, it has
been a very long time - like probably 40 years - and I've forgotten a
lot of the details, and just sort of remember the pain. I think the
ACIA would somehow get into a state where it would not pay attention
to the master reset command.
I don't recall that ever happening to me, but I can well believe it :-)
7:37 p.m.
Ben wrote:
Pete Turnbull wrote:
I really like Motorola stuff too, but ISTR the 6850 package was pin
limited and there was no room for a discrete reset line. Y'all are
going to force me to see what I can dig up on that old project yet. :)
It's been so long though, and since I think the project was scuttled, I
really doubt I have any documentation of any kind left. Maybe some of
my old REI friends remember more about it than I seem to, though I think
I've lost touch with the other folks who worked on that system.
Of course I could get myself a 6809 system, connect a 6850 and figure
out how to lock it up. Hey, what a great idea...
Later,
Charlie C.
On 01/02/2010 05:37, CSquared wrote:
Intel parks I can see, Motorola parts no!
Pete Turnbull wrote:
> I am really puzzled by this. The ACIA has no
hardware reset line, but
> it's very easy to send it a master reset command in software.
I'm pretty sure they were Motorola parts. As
I told Geoffrey, it has
been a very long time - like probably 40 years - and I've forgotten a
lot of the details, and just sort of remember the pain. I think the
ACIA would somehow get into a state where it would not pay attention
to the master reset command.
I don't recall that ever happening to me, but I can well believe it :-)
31 Jan
31 Jan
9:13 p.m.
Geoffrey Reed wrote:
On 1/30/10 8:09 PM, "CSquared" <csquared3 at tx.rr.com> wrote:
I'm sure that would have worked. We did something similar a few years
ago so the software could reset external (non-plug-in) modems where an
I/O line controlled a small relay whose contacts were in series with the
low voltage modem power. That worked quite well.
Back when I was coping with the ACIA's I was in a working environment
where the hardware guys were sometimes not too eager to listen to the
software guys input, to be ultra polite about it. :) This was a new
board design, and they were not about to respin the board to solve what
they probably considered to be a software problem. So, we just sort of
lived with it. Once the system was pretty well debugged and tested I
think it was not so much of an issue. I really don't recall by now how
the ACIA's would get into a state where only a reset could get them
unstuck. It's just been way too many years and way too many other
projects between then and now, but I do remember it being a bit of a
pain. ISTR that particular project was scuttled before it ever made
into production, so we never found out whether it would be a problem in
the field or not.
Later,
Charlie C.
Ah yes, 6850 and ACIA, thanks for the memory jog.
I had completely
forgotten both of those. I can't remember now precisely why I felt the
need to reset the ACIA with a software controlled line, but ISTR I
*really* needed to occasionally, and a power cycle was the only way. I
do remember the little system had several of the ACIA's attached to one
6809 and I controlled a number of unique hardware devices that way. I
got to do some 68000 stuff a few years later and liked it quite well
too, though that was using C - my very first C as a matter of fact,
along about 1986 or 1987.
Later,
Charlie C.
Hmmm, possibly It is because I am more of a hardware guy anymore than a
software guy, but couldn't you make an adapter pcb for the ACIA chip, that
had a relay and a few other componants on it, like a 555 set as a timer, and
when the 555 is triggered, it would power off your acia for a preset amount
of time by tripping the relay so the VCC pin no longer gets voltage, then at
the end of the timing cycle it returns the relay to normal position and the
acia comes back on and you set baud, etc....
30 Jan
30 Jan
11:27 a.m.
From: "Ben" <bfranchuk at jetnet.ab.ca>
Sent: Saturday, January 30, 2010 4:57 PM
To: "General Discussion: On-Topic and Off-Topic Posts"
<cctalk at classiccmp.org>
Subject: Re: 6809 SBC
Henk Gooijen wrote:
True! Just trying to combine hobbies. However, for something completely
different, the main screen software in my StarShip uses the 6809 :-)
to plot lines (Bresenham algorithm) and arcs/circles, but that would go OT.
- Henk.
I am somewhat a Motorola-guy, so a bit biased :-)
Of the 8 bit CPUs the
Umm so why do all your projects end up with PDP 11's or
8's :)
...snip...
Ben. 
8:44 p.m.
On 30 Jan 2010 at 20:27, Henk Gooijen wrote:
True! Just trying to combine hobbies. However, for
something
completely different, the main screen software in my StarShip uses the
6809 :-) to plot lines (Bresenham algorithm) and arcs/circles, but
that would go OT.
Was the 6809 all that wonderful for the time? Consider that the
NS32016 was just starting to show up around 1980. Far more advanced
architecture. If it wasn't for NS's attention-deficit-disorder, it
might have found its way into many more systems.
I've got an NS32CG16 (32016 with graphics extensions) that's waiting
for a "sometime" SBC project.
--Chuck
31 Jan
31 Jan
2:55 a.m.
From: "Alexandre Souza - Listas" <pu1bzz.listas at gmail.com>
Sent: Sunday, January 31, 2010 2:04 AM
To: "General Discussion: On-Topic and Off-Topic Posts"
<cctalk at classiccmp.org>
Subject: Re: 6809 SBC
Yup. See the bottom of my webpage. There are two small "gearboxes".
One of them links to my StarShip. It is a project which did not get much
attention the last 5 years or so. I hope to start on it this summer. That
will
also give a lot of new material on that site.
- Henk (www.pdp-11.nl)
True! Just
trying to combine hobbies. However, for something completely
different, the main screen software in my StarShip uses the 6809 :-)
to plot lines (Bresenham algorithm) and arcs/circles, but that would go
OT.
StarShip? 
5:55 a.m.
Yup. See the bottom of my webpage. There are two small
"gearboxes".
One of them links to my StarShip. It is a project which did not get much
attention the last 5 years or so. I hope to start on it this summer. That
will
also give a lot of new material on that site.
Nice control console :oD Will it blen...eh...fly to mars? :o)
- Henk (www.pdp-11.nl)
HAHAHAHA PA8PDP!!!
UHAUAUHAUHA!!! This one was GREAT! :oD
73' de PU1BZZ Alexandre :)
8:59 a.m.
On Jan 31, 2010, at 5:55 AM, Henk Gooijen wrote:
Yup. See the bottom of my webpage. There are two small "gearboxes".
One of them links to my StarShip. It is a project which did not get
much
attention the last 5 years or so. I hope to start on it this
summer. That will
also give a lot of new material on that site.
Neat! Very nice work!
-Dave
--
Dave McGuire
Port Charlotte, FL
True! Just trying to combine hobbies. However, for
something
completely
different, the main screen software in my StarShip uses the 6809 :-)
to plot lines (Bresenham algorithm) and arcs/circles, but that
would go OT.
StarShip?
2:59 a.m.
From: "Chuck Guzis" <cclist at sydex.com>
Sent: Sunday, January 31, 2010 5:44 AM
To: "General Discussion: On-Topic and Off-Topic Posts"
<cctalk at classiccmp.org>
Subject: Re: 6809 SBC
On 30 Jan 2010 at 20:27, Henk Gooijen wrote:
If the National CPU had received more attention back then, it might have
been a lot more successful. Perhaps the absence of software and developers
for it contributed to that. It would have helped if the popular magazines
like
Byte etc. had written more interesting articles. I know that some Oc?
copiers
had some National chips (32xxx, really can't remember) working as pixel
processors for the generation of the bitmap. Never worked on that part of
the machines though, always worked on the real-time control software.
- Henk.
True! Just trying to combine hobbies. However,
for something
completely different, the main screen software in my StarShip uses the
6809 :-) to plot lines (Bresenham algorithm) and arcs/circles, but
that would go OT.
Was the 6809 all that wonderful for the time? Consider that the
NS32016 was just starting to show up around 1980. Far more advanced
architecture. If it wasn't for NS's attention-deficit-disorder, it
might have found its way into many more systems.
I've got an NS32CG16 (32016 with graphics extensions) that's waiting
for a "sometime" SBC project. 
7 Feb
7 Feb
3:47 p.m.
William Blair wrote:
Having learned assembly language programming on the
beautifully simple architecture and instruction set of the 6800, the Byte magazine article
linked to below that I read when it was originally published really impressed me. In the
6809 they made one of the earliest efforts I know of to really tweak an already great uP
instruction set based upon an analysis of existing software:
http://tlindner.macmess.org/?page_id=119
I was introduced to the 6809 when I was at university and it just seemed
right. I came home and bought a CoCo2 and an EDTASM+ cartridge just to
play with it. I still have it.
Is there a source for 6829 MMUs? I would love to build a 6809 SBC with
an MMU that could run OS-9.
-chuck
8 Feb
8 Feb
8 a.m.
Charles H Dickman wrote:
William Blair wrote:
Well you can get the data sheet here: Get it now, to
the link works.
Who knows in 24 hours.
http://www.macmess.org/downloads/mc6829.pdf
Having learned assembly language programming on
the beautifully simple
architecture and instruction set of the 6800, the Byte magazine
article linked to below that I read when it was originally published
really impressed me. In the 6809 they made one of the earliest efforts
I know of to really tweak an already great uP instruction set based
upon an analysis of existing software:
http://tlindner.macmess.org/?page_id=119
I was introduced to the 6809 when I was
at university and it just seemed
right. I came home and bought a CoCo2 and an EDTASM+ cartridge just to
play with it. I still have it.
Is there a source for 6829 MMUs? I would love to build a 6809 SBC with
an MMU that could run OS-9. -chuck
Ben.
1:31 p.m.
-----Original Message-----
From: chd at chdickman.com
Sent: Sun, 07 Feb 2010 18:47:55 -0500
To:
Subject: Re: 6809 SBC
William Blair wrote:
Having learned assembly language programming on
the beautifully simple
architecture and instruction set of the 6800, the Byte magazine article
linked to below that I read when it was originally published really
impressed me. In the 6809 they made one of the earliest efforts I know
of to really tweak an already great uP instruction set based upon an
analysis of existing software:
http://tlindner.macmess.org/?page_id=119
I was introduced to the 6809 when I was at university and it just seemed
right. I came home and bought a CoCo2 and an EDTASM+ cartridge just to
play with it. I still have it.
Is there a source for 6829 MMUs? I would love to build a 6809 SBC with
an MMU that could run OS-9.
-chuck From what little I've managed to gather, the 6829
had some limitations that made it either difficult to use, or otherwise impractical for
many (if not most) applications. I've read someplace, for instance, that you
were restricted to four tasks. I suspect for that reason,
they're rarely seen. They were rare even in their own time.
A better approach, may be to create some hardware to emulate either
a Gimix or SwTPc style DAT (dynamic address translator). The key
element here, is that both of these are supported by some version
of OS/9.
-Jeff
1:41 p.m.
On Mon, Feb 8, 2010 at 4:31 PM, N0body H0me <n0body.h0me at inbox.com> wrote:
From what little I've managed to gather, the 6829
had some limitations that made it either difficult to use, or otherwise impractical for
many (if not most) applications. ?I've read someplace, for instance, that you were
restricted to four tasks. ?I suspect for that reason,
they're rarely seen. ?They were rare even in their own time.
I'd never heard of it before today (I was into the 6502 then as now).
Glancing at the datasheet, yes, there are registers to handle 4 tasks,
but you can apparently stack up to 8 6829s in one system for 32 tasks.
With 20 bits of memory addressing, it seems a reasonable balance (32
tasks in 1MB), but finding room on a board to stuff eight 40-pin DIPs
between the CPU and memory seems the harder job.
-ethan
1:59 p.m.
Ethan Dicks wrote:
I'd never heard of it before today (I was into the
6502 then as now).
Glancing at the datasheet, yes, there are registers to handle 4 tasks,
but you can apparently stack up to 8 6829s in one system for 32 tasks.
With 20 bits of memory addressing, it seems a reasonable balance (32
tasks in 1MB), but finding room on a board to stuff eight 40-pin DIPs
between the CPU and memory seems the harder job.
Just stack them ;-)
most signals are connected parallel ...
3:32 p.m.
-----Original Message-----
From: emu at e-bbes.com
Sent: Mon, 08 Feb 2010 14:59:28 -0700
To:
Subject: Re: 6809 SBC
Ethan Dicks wrote:
I'd like to see that! :^)
Seriously, though, I don't think Moto had general-purpose operating
system memory management in mind when they designed that part. The
feeling I get is that it was seriously slanted towards embedded
applications, where the size and number of tasks would be more fixed.
While it's true that OS/9 can be (and has been) embedded, I suspect
that the interest here would be to use it as a 'general purpose'
OS for different applications.
-Jeff
____________________________________________________________
FREE ONLINE PHOTOSHARING - Share your photos online with your friends and family!
Visit http://www.inbox.com/photosharing to find out more!
I'd never heard of it before today (I was
into the 6502 then as now).
Glancing at the datasheet, yes, there are registers to handle 4 tasks,
but you can apparently stack up to 8 6829s in one system for 32 tasks.
With 20 bits of memory addressing, it seems a reasonable balance (32
tasks in 1MB), but finding room on a board to stuff eight 40-pin DIPs
between the CPU and memory seems the harder job.
Just stack them ;-)
most signals are connected parallel ... 
3:21 p.m.
New subject: 6829 was 6809 SBC
N0body H0me wrote:
From what little I've managed to gather, the 6829
had some limitations that made it either difficult to use, or otherwise impractical for
many (if not most) applications. I've read someplace, for instance, that you were
restricted to four tasks. I suspect for that reason,
they're rarely seen. They were rare even in their own time.
A better approach, may be to create some hardware to emulate either
a Gimix or SwTPc style DAT (dynamic address translator). The key
element here, is that both of these are supported by some version
of OS/9.
The Dragon Beta/128 prototype also has a DAT implemented with a couple
of small SRAMS and a bunch of LS/PAL logic, used the output of one of
it's PIAs as a task select, and supported 16 tasks, and again was
supported by OS-9.
For those who don't know the Dragon Beta/128 was a prototype machine
that Dragon data where working on when they went bust, it was pretty
impressive specs for a machine of it's sort, dual 68B09s, up to 768K of
RAM etc. Details and pics here : http://www.dragondata.co.uk/
Cheers,
Phill.
--
Phill Harvey-Smith, Programmer, Hardware hacker, and general eccentric !
"You can twist perceptions, but reality won't budge" -- Rush.
5197
days inactive
5207
days old
37 comments
17 participants
participants (17)
-
afra＠aurigae.demon.co.uk -
ard＠p850ug1.demon.co.uk -
bfranchuk＠jetnet.ab.ca -
cclist＠sydex.com -
chd＠chdickman.com -
csquared3＠tx.rr.com -
emu＠e-bbes.com -
ethan.dicks＠gmail.com -
geoffr＠zipcon.net -
henk.gooijen＠hotmail.com -
jules.richardson99＠gmail.com -
mcguire＠neurotica.com -
n0body.h0me＠inbox.com -
pete＠dunnington.plus.com -
pu1bzz.listas＠gmail.com -
ragooman＠comcast.net -
wbblair3＠yahoo.com