6809 Single Board Computer project

[ron.owens]

I've been working on a 6809 Single Board Computer design and I wanted to share it with the community. I've attached the schematic as a pdf. I would be interested in hearing from anyone interested is helping out, or just has an opinion to share. I think I've gotten it right, but that usually means there are a couple of nasty bugs I'm overlooking.

This design includes a Memory Management Unit (MMU) so the CPU has an address space of 8MB. Although it's not a flat linear memory map, it's a lot to work with. There will definitely be software required to manage the MMU. I was hoping that this board will be able to run OS/9. I'm sure that will require some modifications to the OS to handle the MMU.

Also there is a USB interface so a flash memory stick can be used for mass storage. This will also require modifications to the OS, if OS/9 is used. There is some interesting I/O on board - a stereo sound generator circuit and an Arithmetic Processing Unit (APU), along with two serial ports and two parallel ports. There is an expansion bus so more stuff can be added later. I've also added the USB File Manager module schematic. I'll post more on this later.

I'm primarily a hardware guy. I'm looking for a software person to help with a ROM monitor and work on the OS, If you think that is something you would like to try, please post.

Best regards,
Ron

[PuterGeek]

OS/9 and 6809! Now that triggered a flashback from college! Great chip and OS for the time!

You shouldn't have a problem if you used the Motorola MMU. If memory serves me it was supported by at least some versions of OS/9.

But I think you posted the wrong schematic...

[PuterGeek]

Wow! Made me go Google some 6809 history!

Did you find stock of a 6809 somewhere?

Are you using the 6829 MMU and where did you find one?

I taught a class to Georgia Tech grad students on the 6829 probably about 1982. It was the first exposure most of us had to a MMU.

[ron.owens]

I used a SN74LS610 MMU from Texas Instruments. MC6829's are virtually nonexistent. I don't think they were ever manufactured in quantity. I searched for the part a lot and came up empty handed, I wish I could have gotten them, the 6829 is a better part than the 74LS610. But I used what I could get. 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. It doesn't seem fair to post a circuit that can't be built.

I'm currently routing the PCB and should be finished with it soon. I see that the main schematic didn't post. I'll include it here. The other schematic is the USB interface.

[PuterGeek]

Nicely done with an interesting mix of old and new! 

I like seeing an old favorite get some love but an unusual choice. Why the 6809?

[ron.owens]

Once upon a time, in a career long ago, I worked at Motorola in the late 70's. Never got around to building a 6809 board. A Radio Shack CO-Co was as close as I got.

Besides, it's one of the best 8-bitters ever made.

[Fungus]

Besides, it's one of the best 8-bitters ever made.

Definitely better than Z80 and 6502.

I'm surprised it wasn't more popular, unless "price".

[PuterGeek]

Besides, it's one of the best 8-bitters ever made.

I have to agree on that point! And it was one of the last of the separate processors.

The microprocessor professor at Georgia Tech, Dr. John Peatman, was a big Motorola supporter. Probably helped that Motorola donated to the college. Dr. Peatman worked at HP during the summer so we had a lab full of HP development systems tied together on a IEEE-488 'network'.

One of the fun projects I did was create a 6809 structured assembler for the development system. HP had a universal assembler tool that managed all the common tasks (symbol recognition, hash table, etc.) so all I had to do was handle the stuff unique to the structured assembler.

And in case you're wondering, a structured assembler has macro like instructions for common programming structures such as IF THEN ELSE and DO WHILE. A step short of a C compiler (which we didn't have) but made coding a little cleaner.

The programming we did was mostly BASIC09 and a little assembly. I don't think we had a native assembler and BASIC09 ran under OS/9.

BASIC09 compiled into "p-code" which was interpreted at runtime. Made for fairly compact code and reasonable execution speed.

Thanks for the prompt for a trip down memory lane! Good times when I was just getting started!

[PuterGeek]

I'm surprised it wasn't more popular, unless "price".

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.

Much like the Z80 was an improved 8080, the 6502 was an improved 6800. The 6809 is largely an improved 6502, adopting the 6502 improvements and adding more on top.

But it was the end of the 8-bit standalone era. Motorola came out with the 68000 (in the giant 64 pin DIP) and Intel brought out the 8086. Motorola moved the 8-bitters to microcontrollers like the 6801 and later the 68HC11.

[NEALC]

Hi Ron,
how is your board design coming along? 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

Details on all of these on github https://github.com/nealcrook/multicomp6809

Regards
Neal

[bingo600]

Ohh

Brings back memories of my old Flex system , my first computer.

I even have that fpga board , and some ram ...... hmmm .....

/Bingo

[Pjotr]

6809... Brings back memories. Build many precision servo controllers with it and even a sophisticated sequencer for controlling a Moog analogue music synthesizer those day's. But better than a Z80? Depends on the state of the moon at your birth IMHO Real programming those days was assembly coding. 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. Must have still somewhere a floppy with an assembler that ran on a Atari1040.

[T3sl4co1l]

Funny, I recently came into some old chips, including 6809s (but none of the aforementioned MMU parts, oh well).  List: http://seventransistorlabs.com/Images/Chip_Inventory.txt

Tim

[coppice]

Besides, it's one of the best 8-bitters ever made.

Definitely better than Z80 and 6502.

I'm surprised it wasn't more popular, unless "price".

Price favoured the MC6809, as it was a small chip, but it was a little too late. 16 bit devices were already getting traction, and the 8088 was available for people needing a low pin count device with an 8 bit external bus. If the MC68008 had been part of the initial launch of the 68k, rather than an afterthought, the 68k family might have been a lot more successful.

[ron.owens]

The aforementioned MMU parts, 6829, were never manufactured. Motorola only produced 'engineering' samples'. It's too bad because that is actually a very cool part. I have heard of it being implemented in a FPGA and works quite well. My weapon of choice for MMU is the 75LS610. It works fairly well in 6809 and Z80 designs, but was originally designed by Texas Instruments to work with the TMS9900, a 16-bit processor. As such, it requires a few logic tricks to get it to play nice with 8-bitters. However, if successfully implemented it can allow 20-bit addressing. I would think if someone was going to go to trouble of designing a 6829 MMU in a FPGA, they might as well make a 6809 with a 24-bit address bus and register support to make it work. Now that would be a very cool part.

[Pjotr]

An 8-bitter with 24 bit address bus (or MMU)? What are you up to? Apart from embedded control those 8-bitters were used for administration tasks for a while (under CPM etc.). Only for the latter a MMU was profitable to handle larger chunks of data. Most embedded control tasks didn't need it. I think more than emulate original parts and running software of those days has no use.

[ron.owens]

You are absolutely right, it's of no use. It's based on vintage CPUs and inherently has no use by obsolescence. Most people designing and building vintage computers do it for enjoyment. Much the same way someone restores an old automobile when a much improved version can be purchased from a dealership. I imagine very few people build 8-bit computers with microprocessor from the 1970's and 1980's to make money.

I recently watched a video of a guy building a fully functional artificial hand from a brand new coffee maker. And I thought I had too much time on my hands! I think he did it because he could. Given his talent, I would think he could make a fantastic hand from more standard parts. I think hobbies are seldom done because of their 'use'. But that's just my opinion.

[bingo600]

Nice info

http://www.baltissen.org/newhtm/74ls612.htm

http://www.6502.org/users/andre/icapos/mmu65.html

/Bingo

[woodchips]

Nice, I like the 6809.

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

[edavid]

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.

[xygor]

The Coco 3 used the TI MMU and ran OS/9.  That might make a good starting point.  The memory limit may be hard-coded into OS/9 in such a way that will make if difficult to increase.  Also applications such as BASIC and the assembler might have to be modified.  Or maybe not.

[edavid]

The Coco 3 used the TI MMU and ran OS/9.

It did not use a TI MMU - the MMU was in the "GIME" ASIC.

[xygor]

The Coco 3 used the TI MMU and ran OS/9.

It did not use a TI MMU - the MMU was in the "GIME" ASIC.

Who made the ASIC?

[edavid]

The Coco 3 used the TI MMU and ran OS/9.

It did not use a TI MMU - the MMU was in the "GIME" ASIC.

Who made the ASIC?

I think VLSI.

[ron.owens]

Good luck getting any useful information on the GIME. It's even harder to get a chip, unless you dissect a CoCo III. In that case, you might as well use the CoCo III! It was very specific to that design.