So 1980 - that as the year I discovered Unix and C. I don’t think C has really changed much since then - compilers have gotten better at bleating at you about errors, and producing better code, but that’s more a function of bigger memory, faster cpus, so they can do more. The main change might be K&R style function definitions to ANSI style. longjump is still a thing.
1980 was also the year I first used Primos and wished I hadn’t.
But by 1980 I’d already been using the Apple II (and other micros in a small way) for a couple of years. I missed CP/M early on, but used it in the mid-80’s for a lot of stuff.
The other thing - PL/1 - and PL/M. Bits of CP/M are written in PL/M but I don’t recall ever seeing a compiler for it - but the past few days a video of SPL/M has been published - see
You’re right about modern stuff being too complex though - I have a “bare metal framework” for the Pi 1 and there are bits of it that I could not write, so took the code from elsewhere (the EMMC driver and USB) But using a CPU from the 1980s or making your own shouldn’t be that hard. I’d strongly suggest avoiding the 65816 though.
The BBC Micro does provide an excellent platform for theoretical retro cpu/computer design though - assuming you can write an emulator for your new CPU in either ARM assembler, C or BBC Basic then you can run it on the PiTube co-processor with the Beeb providing the rest in the form of screen, keyboard, filing system, network, etc. That’s my plan for my project when I get some time.
I don’t know about 8080 PL/M, but I believe I have the disks and manuals for Intel PL/M 86 with 286 extensions (possibly for OS/2?) in my basement. If this is a thing of interest, I can move discovery up on my timetable, I assumed this was pretty well-trod ground.
No interest to me - it was specifically the older 8080 stuff.
But having recently re-read the pl/m manual (all 136 pages of it) I wonder now what I might have been thinking of. It has some interesting features but also some obscure ideas. They may have seemed like a good thing/choice at the time though. Or maybe I’ve just become too used to the ideas of C/BCPL, etc.
The BBC Micro does provide an excellent platform for theoretical retro cpu/computer design though - assuming you can write an emulator for your new CPU in either ARM assembler, C or BBC Basic then you can run it on the PiTube co-processor with the Beeb providing the rest in the form of screen, keyboard, filing system, network, etc.
As far as I know, they don’t seem to have made a North American model.
Speaking of weird hacks, I wonder if anyone has put a 6809 in a Apple II.
There was (is) a version of the BBC Micro for the US. I have one at home. It has additional RF shielding and a BASIC that supported COLOR as well as COLOUR.
They’re not that hard to get though - plenty on ebay, etc.
I was checking on the Canadian ebay, nothing there.
As for the 6809 mod for the apple II, I was thinking of 6809
pcb replacing the 6502 on the apple II mother board.
Its data path is only 8-bit wide so it is fairly slow. If you are comfortable with programmable logic, a compact 32-bit data path design can be done using CPLD.
I like 22V10’s as my generic programable logic part.A 68020 needs a 4 layer
PCB while a 68HC000 can use a 2 layer PCB. I will check the 68020 prices
if it is cheap, I can pick one up off ebay.
22V10 makes design of 68020 as well as 68000 much easier. 68020 should work with 2-layer PCB, but 4-layer PCB are so cheap now, I’d go with 4-layer pcb for most designs.
Bill
Edit, my most recent 68000 design. It uses 68SEC000 which can run at 3.3V. It is a mezzanine board on top of a cheap EPM240 CPLD development board from eBay. EPM240 provides the boot ROM, serial port, interface to mass storage. Still under development.
Yep, a 22V10 can save lots of standard TTL and speed up the access. I have a 68020 designed in 1990’s in all TTL logics, a couple 22V10 probably can replace all that TTL logics.
They removed the E clock on the 6820 for the 6850.
Why are all the great designs from the UK?, I got the $100.
Oh wait, I have to convert from $100 Canadian to US $,
From US $ to Euro’s $. From Euro’s to Pounds $.
If you ever a design done for 114 PGA 68020’s
PCB gerber files (zipped) works best for me.
I can get the PCB’d made in China for about $100, and shipped here in a week.
E clock is not absolutely needed to use 6850. You can always decode the 68020 control signals to drive E. 6850 is pretty simple, however. You may want 68K peripheral like 68681 or 68901.
I’m in USA, so I mean US$100. Which should be plenty for 68020. As comparison, I still sell assembled and tested CB030(68030) for $85 and Tiny68K (68000) for $60 as posted on retrobrewcomputers’s board inventory. So it can be cheaper if you build it from scratch. I do have a 68020 design published with gerber photo plots, but it uses SMT CPLD, so you probably don’t care for that. I’ll dig up the 68020 design I did for Highland University in 1990’s and see if the standard TTL can be replaced with a couple 22V10.
Bill
Edit, the 68SEC000 I posted yesterday is $6 68SEC000, two $6 512K RAM, cheap 2-layer pcb, few dollars of hardware and $9 EPM240 dev board from eBay. So it is like $35 total, so yes, $100 68020 is entirely realistic.
Still Counting pennies. The prices are right, but I have to order from china
because shipping to Canada is $15 to $30. Found RetroBrew Computers Wiki
Using a 2681 and a 68H000 at 16 mhz. Finding KICAD footprints is the delay with the PGA 68020. Since this board design is not to run 68K code but rather emulate something else decoding may be rather sloppy.
