6809 Single Board Computer project

[coppice]

Nice, I like the 6809.

Another rare, or invisible, chip was the 6840 floppy controller. Did anyone ever get one?

The floppy controller was the MC6843.  It was not very popular, but it was definitely produced (I have at least one).  The MC6840 was a triple timer.

I also used the MC6843. I implemented my own driver so that it emulated the discrete floppy controller card in the original MC6800 development systems, and allowed the Motorola development tools to be run on a wider range of hardware.

[ron.owens]

Nice, I like the 6809.

Another rare, or invisible, chip was the 6840 floppy controller. Did anyone ever get one?

The floppy controller was the MC6843.  It was not very popular, but it was definitely produced (I have at least one).  The MC6840 was a triple timer.

I also used the MC6843. I implemented my own driver so that it emulated the discrete floppy controller card in the original MC6800 development systems, and allowed the Motorola development tools to be run on a wider range of hardware.

I didn't even know the 6843 existed until I saw it posted here. And I actually worked at Motorola back in the day! By the way... they are going for $40US on eBay.

[basinstreetdesign]

Wow.  Good on ya for resurrecting the venerable 6809.  I used one in a couple similar projects in 1981-84 to develop a specialized computer for a film processing lab.  With two stack pointers and a very symmetric and concise instruction set, around where I worked it was acknowledged as the most elegant 8-bit machine there was.  Imagine having not just one but TWO branch never instructions, a short branch never (2 machine cycles) and a long branch never (3 instructions) as well as the usual bunch of other branch instructions.  But with the ability for position-independent or re-entrant code (think Towers of Hanoi), anything was possible.

Too bad it was so slow or else it could have done some real damage in the early PC market.

[NEALC]

...However, the  mmu capability of the gime *is* well documented and my github repository has a functional equivalent coded in vhdl.

Neal

[ron.owens]

...However, the  mmu capability of the gime *is* well documented and my github repository has a functional equivalent coded in vhdl.

Neal

That is very cool. Do you have a link? I really wish I could get a FPGA in through-hole, but I would settle for a surface mount part to get a great MMU.

[NEALC]

https://github.com/nealcrook/multicomp6809/blob/master/multicomp/Components/MMAPPER2/mem_mapper2.vhd

Note that this code is not register compatible with the gime mmu, but is (afaik) functionally compatible.

Neal

[bingo600]

https://github.com/nealcrook/multicomp6809/blob/master/multicomp/Components/MMAPPER2/mem_mapper2.vhd

Note that this code is not register compatible with the gime mmu, but is (afaik) functionally compatible.

Neal

Just tried if it would fit one of the cheap EPM240 boards , and it's a tight fit but does 

http://www.ebay.com/itm/1PCS-NEW-EPM240-Altera-MAX-II-CPLD-Development-Board-/272086159418

Code: [Select]

Fitter Status Successful - Wed Aug 24 21:32:55 2016
Quartus II 64-Bit Version 13.0.1 Build 232 06/12/2013 SP 1 SJ Web Edition
Revision Name mem_mapper2
Top-level Entity Name mem_mapper2
Family MAX II
Device EPM240T100C4
Timing Models Final
Total logic elements 235 / 240 ( 98 % )
Total pins 43 / 80 ( 54 % )
Total virtual pins 0
UFM blocks 0 / 1 ( 0 % )


/Bingo

Edit: Note a MaxII EPM240 is NOT 5v tolerant

[ale500]

Cool project !!!, I'd have used a couple CPLDs instead of GALs, just because I have no GAL programmer.

A while ago, I got a MC6847 and a 74LS783 to recreate a CoCo1 , I also picked up a MC68B09E, still on the drawing board...

[obiwanjacobi]

Edit: Note a MaxII EPM240 is NOT 5v tolerant

It can be done (8-3):
https://www.altera.com/content/dam/altera-www/global/en_US/pdfs/literature/hb/max2/max2_mii51009.pdf

[bingo600]

Yesss - My retro 6809 is alive.

Thanx Neil , for selling me a PCB.
And answer my kazillion questions on startup.

Code: [Select]

6809 CamelForth v1.1  20 Mar 16

OK HEX
OK 1000 1FFF PASS
OK HEX
OK 1000 1FFF 10 PASSES ................
OK HEX
OK 0800 CFFF 10 PASSES ................
OK


I got the PCB assembled , but uses serial as console instead of VGA/KBD.
Got the 512K Ram going , and thats what is tested in above listing.


Now to connect the SD adapter & make a SD card 

/Bingo

[bingo600]

And it's booting FLEX 

6809 CamelForth v1.1  20 Mar 16

OK
FLEX FLEX......................
FLEX 9.0

DATE (MM,DD,YY)? 09,04,16

+++
+++dir

DIRECTORY OF DRIVE NUMBER 0
DISK: FLEX09BT  #0    CREATED: 27-JUN-97

FILE#   NAME   TYPE  R  BEGIN   END   SIZE    DATE     PRT


    1  FLEX    .COR     25-41  26-0E    22  27-JUN-97
    2  SD      .BIN     29-17  29-17     1  17-JUL-97
    3  CONS    .BIN     29-18  29-18     1  17-JUL-97
    4  WINDRV  .CMD     01-1F  01-20     2  27-JUN-97
    5  ASN     .CMD     01-21  01-21     1  27-JUN-97
    6  TIMERON .CMD     01-22  01-22     1  27-JUN-97
    7  TTYSET  .CMD     01-23  01-24     2  27-JUN-97
    8  DATE    .CMD     01-25  01-26     2  27-JUN-97
    9  ERRORS  .SYS  R  01-27  01-47    33  27-JUN-97
    ..
    ..
    ..

I feel like 20 again 

/Bingo

[ale500]

The 6809 is largely an improved 6502, adopting the 6502 improvements and adding more on top.

Can you elaborate ? In which regard is the'09 an improved 6502 instead of an improved 6800 ?

[Fungus]

The 6809 is largely an improved 6502, adopting the 6502 improvements and adding more on top.

Can you elaborate ? In which regard is the'09 an improved 6502 instead of an improved 6800 ?

Nah, they're totally different beasts.

The only real similarity is they both start with a '6' and have four digits in their name. Apart from that? Nada. 

(...and they both come in 40-pin DIP packages)

[bingo600]

The 6502 was designed by former Motorola designers , and some of the 6502 instructions could be considered an improvememt over the 6800.

If you can program a 6800 in assembler you won't have many problems , switching to a 6502.

A bit like the 8080 vs Z80 , although the 6502 isn't opcode compatible with the 6800..

The 6809 ... Well simply the "best" 8 bit instruction set ever invented , but as a CISC some of the opcodes takes a lot of cycles.
The 6309 even added more features to the MCU , while maintaining backwards compatibility.

/Bingo

[woodchips]

Ah, yes, you are correct, the floppy controller was the 6843, how silly.

The 6502/6800? Different instructions, but better? An overflow flag is a wonderous thing, but a PUSHX and PULX instruction really would have been useful.

[Fungus]

If you can program a 6800 in assembler you won't have many problems , switching to a 6502.

All the addressing modes, etc., are completely different.

Two accumulators, 16-bit addressing registers (LEA instructions), two stacks...

A 6502 would feel really restricted after using a 6809.

You could switch over, sure, but I don't think knowing 6809 would give you any special advantages over knowing any other assembly language of the era.

[legacy]

interesting topic
any news? 

[vladk]

I used a SN74LS610 MMU from Texas Instruments. ... I was thinking, since the 74LS610 and the MC6809 chips are very available on eBay, at least my design could be built by a few people.

Could you please share here the contents of GALs used in the MC6809 project?
For example, how exactly do you form the "Write STROBE" signal for the 74LS610 MMU ?

Regards,
Vladislav

[bingo600]

If you can program a 6800 in assembler you won't have many problems , switching to a 6502.

All the addressing modes, etc., are completely different.

Two accumulators, 16-bit addressing registers (LEA instructions), two stacks...

A 6502 would feel really restricted after using a 6809.

You could switch over, sure, but I don't think knowing 6809 would give you any special advantages over knowing any other assembly language of the era.

I never compared the 09' to a 6502 , i wrote 6800

/Bingo

[NEALC]

I have done some work with an fpga-based 6809 system and ported a bunch of pre-existing software, some of which might  be of interest to you:
"buggy" a debug monitor with line assembler/disassembler
"FLEX" the swtp OS
"NitrOS9" a open source OS9 rewrite
"FUZIX" an open-source unix for small systems

Huh? that's what I wrote in Reply #9 on: August 21, 2016. Are you a spambot?

[David Hess]

I think it was mainly because of Motorola's lack of focus. They were more diversified company (radios thru all sorts of semiconductors) while MOS Technologies, Zilog and Intel were mostly microprocessor companies.

This was my experience at the time when evaluating the 68HC11 via the 68HC811 and 64HC24 port replacement unit.  Availability was so poor that we abandoned and anything from Motorola which was not second sourced.  As someone mentioned, the 68008 should have been available with the 68000 but Motorola just did not seem interested in having IBM as a customer.

The 6809 stood out the rest by its neat and extensive instruction set that was a pleasure to use. There was even a kind of instruction that acted as kind of a pointer to a pointer. Very handy for lookup tables.

Double indirect addressing proved to be a real millstone around the neck of the 68000 microarchetecture when superscalar operation and pipelining were implemented.  Linus even uses it as an example of not thinking ahead.

[coppice]

I think it was mainly because of Motorola's lack of focus. They were more diversified company (radios thru all sorts of semiconductors) while MOS Technologies, Zilog and Intel were mostly microprocessor companies.

This was my experience at the time when evaluating the 68HC11 via the 68HC811 and 64HC24 port replacement unit.  Availability was so poor that we abandoned and anything from Motorola which was not second sourced.  As someone mentioned, the 68008 should have been available with the 68000 but Motorola just did not seem interested in having IBM as a customer.

Was that around 1994, when Motorola was cracking up from its own success? They did so well with design ins for HC11 and HC05 devices in the previous couple of years that they lacked the production capacity to service those customers when it came to mass production. The repercussions of that made many designers stay away from Motorola parts for a long time.

[David Hess]

I think it was mainly because of Motorola's lack of focus. They were more diversified company (radios thru all sorts of semiconductors) while MOS Technologies, Zilog and Intel were mostly microprocessor companies.

This was my experience at the time when evaluating the 68HC11 via the 68HC811 and 64HC24 port replacement unit.  Availability was so poor that we abandoned and anything from Motorola which was not second sourced.  As someone mentioned, the 68008 should have been available with the 68000 but Motorola just did not seem interested in having IBM as a customer.

Was that around 1994, when Motorola was cracking up from its own success? They did so well with design ins for HC11 and HC05 devices in the previous couple of years that they lacked the production capacity to service those customers when it came to mass production. The repercussions of that made many designers stay away from Motorola parts for a long time.

It was around 1994 but it was a continuing problem for years before and after and I did get the idea that they had some major customers on allocation.  The result was dropping 68HC11 for PIC which while not completely satisfactory was at least available.

I finally received the 64HC24s that I needed months later and I still have them somewhere.

[coppice]

Nice, I like the 6809.

Another rare, or invisible, chip was the 6840 floppy controller. Did anyone ever get one?

The floppy controller was the MC6843.  It was not very popular, but it was definitely produced (I have at least one).  The MC6840 was a triple timer.

I also used the MC6843. I implemented my own driver so that it emulated the discrete floppy controller card in the original MC6800 development systems, and allowed the Motorola development tools to be run on a wider range of hardware.

While sorting through some old papers today I came across the schematic I produced for that MC6843 based floppy disc controller in about 1978 or 79. A very nostalgic moment.    The circuit has an NE564 PLL, and some counters and flip-flops to recover the clock from the disc data. It also uses a whole 8 bit port of an MC6821 parallel port interface for drive select, side select, door lock, and so on. Later floppy disc controller chips were more complete than that.

[jbeeksma]

Hi Ron,

In 1976 me and some frined started building our own Motorola 6800-based computer. Some later, we built a Mark II based on the 6809. It is the best microprocessor for assembly programming EVER!

We built the 'OS' ourselves, and reverse engineered an assembler and editor and a BASIC interpreter from hex dumps. Those were the days.

I've been looking for a single-board 6809 based microcomputer to relive the old days, and would love to work with you on the software, if you're still looking for help in that department.

Just drop me a line if you're interested.