MECB 6802 CPU Card

[PaulShipley]

Hi,

I am just sending my first version of a MECB 6802 CPU Card off for PCB manufacture. It is based on the 6502 design, as these chips are very similar. Indeed, there was a legal dispute around this with MOS having to redesign their first version.

I will share the details once I have tested it.

[djrm]

Greetings Paul, I have an interest in 6800 computers but never having used one. The nearest was a 6303 dev board I used for a project. I'm looking forward to exploring the SWTP operating systems used with 6800 and 6809 processors. I see the 6802 is software compatible with the 6800. I did buy some chips a while back, including a 6800, but sadly not a 6802. Best regards, David.

[PaulShipley]

Hi David, I was using the 6800 and 6802 chips as part of my university course back in the late 70s. The 6802 is functionally equivalent, though not totally pin compatible, with the 6800. The major difference is that the 6800 requires a very precise dual phase clock generated by the 6875 chip, which even in the 70s was often very hard to obtain. The 6802 included an internal clock which could be driven either by a crystal, or a suitable external clock signal. I believe that the function of the 6875 could be reproduced by a modern PLD (eg ATF16V8B), but I have not tried this.
I didn't get to use SWTP at the time, and I'm hoping to revisit that and some of my old projects from that era.
HTH

[epaell]

Hi Paul,

Nice, I look forward to hearing more about this! I grew up with my brother's SWTPC 6800 in the late 1970's with a whopping 20 KB of static RAM (5 x 4 KB RAM boards using 2102 memory chips). It had a HUGE linear transformer to power the thing and it weighed a tonne! We never had a disk drive so it was all terminal and cassette operation but I had fond memories of it. I was really excited when I saw a company (Corsham Technology) selling SWTPC-like SS-50 kits with 6800/6809 and 6502 boards but by the time I discovered it the owner of the company was diagnosed with a terminal illness and ultimately passed away. Thankfully he managed to get much of the information regarding these pushed online (so his legacy could live on) but there are still snippets that are missing (making it difficult to reproduce the boards). I did manage to make a version of his SD-based disk drive for interfacing with the MECB-6809 card and MECB-Motorola I/O card (I have used it to run FLEX-09 and nitrOS9) - it should also work with the 6800/2. I was also thinking about modifying the 6502 board to accommodate the 6800 (as I do have one on hand) but was put off a bit by the clock circuitry requirements (and didn't want to risk spending large sums on potentially dodgy 6875 devices). However, recently I came across this implementation:

http://www.amy.hi-ho.ne.jp/officetetsu/ ... chdata.pdf

Which uses a simple and inexpensive 8-pin PIC to generate the clock. I've recently programmed one of these for this use and it "appears" to work but I haven't tried to use it to drive the 6800 yet - it is something I'd like to try though. I did previously created an emulated environment on a cheap ESP32 board and that was able to boot up FLEX and run programs but it would be nice to run on an original processor - somehow it makes it feel more real I'll definitely have to put some more thought and testing into it once work commitments slow down a bit.

Cheers,

Emil.

[PaulShipley]

Hi Emil,

That's interesting! I see that they are using a PIC MCU to emulate the 6875 signals - which is mind blowing when you think about it. We were trying to use the regular 74xx TTL parts at the time (late 70's) and it was nearly impossible due to the speeds required to achieve the timing. Now an 8 bit MCU (which would have to be the great, great, great descendant of the 6800) is fast. powerful, and cheap enough to be used to emulate the 6875 clock for the 6800.

I only had a home-designed 6800 built on a protoboard with 2K RAM (the 6116 was amazing at the time). Which more or less matched the Motorola D2 Kits we use at Uni.

I have the PCBs back, now I just need to do all the soldering and test it.

[epaell]

Hi Paul,

Ah, the good old 6116 - the 6809 system I inherited from my brother (after the SWTPC 6800) had those. They did seem amazing at the time!

It sometimes feels like obscene overkill to use modern devices to emulate something relatively simple like a clock or reset circuit or TTL gate but I think I've become a bit more pragmatic of late. Given that that the PIC MCU only costs a couple of dollars compared to tens of dollars for a "potential" 6875 I think it is an acceptable way to go (just a bit more of a hassle because it needs to be programmed ... but even that is quite easy these days).

Good luck with the testing - I look forward to hearing your updates.

Cheers,

Emil.

[Editor]

The major difference is that the 6800 requires a very precise dual phase clock generated by the 6875 chip, which even in the 70s was often very hard to obtain.

I just noticed this interesting topic thread!

Related to my just completed DREAM 6800 re-creation project, I wondered if you knew that "Electronics Australia" published a MC6875 substitute circuit in 1979, during the then MC6875 shortage.

I haven't tried it, but it's probably a good place to start for a more retro friendly MC6875 substitute (given it obviously worked at the time, as an official Michael.J.Bauer designed DREAM 6800 replacement circuit for the MC6875).

Here it is:

[epaell]

Nice, I searched high and low for something like this a few months back and failed miserably. Thanks for posting. When I get a chance I'd really like to try the PIC version to see how well that works. I just want to clear up some 68008-related things so that it I don't feel quite so embarrassed by the state of what I have in GitHub so far.

[PaulShipley]

Thanks for posting Greg! I remember there being several designs (of varying quality) at the time, but I was unable to find any when I searched recently.
I'm sure this could be easily replicated by a GAL (or PIC, ATTINY, etc), but care would have to be taken as the timing of both the rising and falling edges of the three clocks (phi 1 mpu, phi 2 mpu, phi 2 bus) is critical. It's a highly choreographed dance, which was the downfall of many designs.
This was one of the reasons why the 6800 disappeared almost as soon as the 6802 was available, as that only needed a crystal or a single external clock.

[djrm]

I'd be interested to see how the clock signals could be implemented using a old, I imagine implementing the clock delays introduced by the RC networks in the circuit Greg posted would be impossible. I have very little experience in pld programming though. Maybe a hybrid with pld and RC networks could be made to work somehow.
Btw, can an oscillator be implemented in pld?
Best regards, David.