For a vintage computer project, which processor is easier to use?

[technix]

I have some vintage computer projects on the sketch boards for a while, but I am still undecided which to pursue first. Which of the two would be easier to use in your opinion: Intel 8088 + Intel 8087, or Motorola MC68HC000? Aspect to be consider includes the simplicity of the system design, and availability of compilers, tools, libraries and existing programs.

I don't have a lot of discrete 74HC logic at my disposal, thus I have to resort to the use of CPLD devices (EPM7032 and EPM7128) for glue logic. I am not very good at HDL, and the PLD devices aren't very big, thus the glue logic should be minimal. (Does anyone know anything about 5V FPGA preferably by Altera/Intel? That may allow me to build more complex glue logic down the road, like a programmable peripheral address decoder or a memory management unit)

I am not that keen on coding everything in assembler, so availability of some kind of ISO C compiler is almost a must, and the ability to write startup code in as little assembler as possible is a bonus point. I trust the optimizer of a modern compiler much more than the one of an older compiler. Also the toolchain should be free and preferably open source. My software work platform is Linux and macOS so Windows-only tools are out of consideration.

[DaJMasta]

What about something like a Z80 based one?  I remember a single board computer with probably less than a dozen chips in a pretty small form factor and it's of similar vintage/speed.

If you're willing to delve into FPGAs, there's also the option of making a soft core entirely inside one - not the same if you're looking for a nice hardware project, but it could be a pretty good alternative for complete emulation.

[C]

Have you studied ether of the two?

I see the 8088 as an 8-bit computer stretched to a 16-bit instruction set and a 1 meg address space, Segmented memory.
From a software side try to create an array with more then 64k of data

68000 is 32-bit with a 16-bit external bus. The big cripple was that the 68 pin chip needed more pins to get 8 more address lines out of chip.
The 68000 was also built to have many 68000's connected to a bus
Better hardware & better software.

Think of Why the Z80 & 68000 are still taught in collage classes.
 

[Benta]

Due to its asynchronous external bus, the 68000 needs more glue than the others. On the other hand, it is far more flexible when it comes to mixing peripherals and memories of different speeds.

[Elf]

I have a similar project on my list, with more extensive use of an FPGA for the rest of the system. These are not on your list of chips, but perhaps worth looking at:

• WDC 65C02 (same arch. as Apple II, NES, etc.) or 65C816 (Apple IIgs, derivative used in SNES)
• Zilog Z84C00 (same arch as Kaypro, TRS-80, etc. Sort of similar to Gameboy)

The same old architectures from the mid-70s, all suitably "vintage," but are still in production and are available in modern surface mount packages. The WDC products work at 3.3V, which solves your FPGA voltage issue. Zilog's Z80 successors, the Z180 and eZ80, also have 3.3V parts available and support the Z80 instruction set with extensions.

All have free C compilers from the vendor (WDC Tools, ZDS) and plenty of learning material available. I found the following books useful:

• Programming the 65816: Including the 6502, 65C02, and 65802
• 6502 Assembly Language Programming
• Z80 Assembly Language Programming

[technix]

Have you studied ether of the two?

I see the 8088 as an 8-bit computer stretched to a 16-bit instruction set and a 1 meg address space, Segmented memory.
From a software side try to create an array with more then 64k of data

68000 is 32-bit with a 16-bit external bus. The big cripple was that the 68 pin chip needed more pins to get 8 more address lines out of chip.
The 68000 was also built to have many 68000's connected to a bus
Better hardware & better software.

Think of Why the Z80 & 68000 are still taught in collage classes.

I have studied both of them. My university used 8088/8087 as the platform though...

8088/8087 have the benefit of more software available - I can create some environment that runs vanilla MS-DOS and get a lot of software running there, since the IBM PC XT architecture is fairly open and can be replicated on a DIY system, and I can even add some PC AT features like the DS12C887 RTC. Some PC XT compatible graphics and sound system would be difficult to design though, but I think I can make it work using some dual port SRAM and a Cortex-M4F GPU. ISO C targeting 8088 are older though - I don't think there is a compiler supporting full C99.

68000 have the benefit of better software support - for example the full GNU toolchain is still available. However I have trouble gluing the memory and other peripherals to the CPU. And I don't think there are ready made software for a custom 68000 system. Is 2MB SRAM + 64kB EEPROM enough for a 68000 system?

[technix]

What about something like a Z80 based one?  I remember a single board computer with probably less than a dozen chips in a pretty small form factor and it's of similar vintage/speed.

If you're willing to delve into FPGAs, there's also the option of making a soft core entirely inside one - not the same if you're looking for a nice hardware project, but it could be a pretty good alternative for complete emulation.

I have considered Z80 but is there any ISO C compiler for it?

For FPGAs, I am not good in HDL. I do have a FPGA dev kit but I have very little clue how to get a soft core on it.

[dexters_lab]

always had a soft spot for the 68k after my time with the atari st & tt

there are plenty of low ic count single board designs for the 68k like the "Flight 68k Training System", i recently sold one on ebay for £43 so they don't cost much and could be a good starting point for your own design.

[C]

68000 have the benefit of better software support - for example the full GNU toolchain is still available. However I have trouble gluing the memory and other peripherals to the CPU. And I don't think there are ready made software for a custom 68000 system. Is 2MB SRAM + 64kB EEPROM enough for a 68000 system?

The memory is very easy
 If you think about an 8-bit memory board. You are just using two with a common address bus. Separate data bus & control bus for each memory board
Very easy to do. You can easily connect a Z80 to this memory also. A Z80's A0 selects which board of the two.

The HP Pascal compiler that HP used internally to create operating systems, compilers & assemblers, If I remember correct was 200k
The system I used back in early 1980's worked fine with 512k
HP9826 HP9826
Hp had a lot of systems based on 68000, some single user like 9826 some multi=user.

The custom in-house system had many CPU boards on VME bus with local bus on cpu board and local expansion connector.

As I said, The 68000 was designed to run many CPU's on a bus with very powerful IO

You could learn a lot pushing one to limits unlike other CPU's listed in thread.

[technix]

always had a soft spot for the 68k after my time with the atari st & tt

there are plenty of low ic count single board designs for the 68k like the "Flight 68k Training System", i recently sold one on ebay for £43 so they don't cost much and could be a good starting point for your own design.

What is the 68681 and 68230 doing there? Are CMOS variants available for those if they are important for SRAM-based system? For the glue logic it is doable in EPM7032.

[colorado.rob]

Also, take a look at http://sdcc.sourceforge.net/

SDCC is a retargettable, optimizing Standard C (ANSI C89, ISO C99, ISO C11) compiler suite that targets the Intel MCS51 based microprocessors (8031, 8032, 8051, 8052, etc.), Maxim (formerly Dallas) DS80C390 variants, Freescale (formerly Motorola) HC08 based (hc08, s08), Zilog Z80 based MCUs (z80, z180, gbz80, Rabbit 2000/3000, Rabbit 3000A, TLCS-90) and STMicroelectronics STM8. Work is in progress on supporting the Microchip PIC16 and PIC18 targets. It can be retargeted for other microprocessors.

[dexters_lab]

always had a soft spot for the 68k after my time with the atari st & tt

there are plenty of low ic count single board designs for the 68k like the "Flight 68k Training System", i recently sold one on ebay for £43 so they don't cost much and could be a good starting point for your own design.

What is the 68681 and 68230 doing there? Are CMOS variants available for those if they are important for SRAM-based system? For the glue logic it is doable in EPM7032.

they are a UART, Parallel IO and timers, there was a training program with that board so i expect small projects to do with serial, parallel and general purpose i/o using them as part of the training

[technix]

always had a soft spot for the 68k after my time with the atari st & tt

there are plenty of low ic count single board designs for the 68k like the "Flight 68k Training System", i recently sold one on ebay for £43 so they don't cost much and could be a good starting point for your own design.

What is the 68681 and 68230 doing there? Are CMOS variants available for those if they are important for SRAM-based system? For the glue logic it is doable in EPM7032.

they are a UART, Parallel IO and timers, there was a training program with that board so i expect small projects to do with serial, parallel and general purpose i/o using them as part of the training

I have 16C550 for UART. The parallel I/O can be problematic though (and that is one thing I don't want to waste CPLD logic elements on.)

[technix]

Also, take a look at http://sdcc.sourceforge.net/

SDCC is a retargettable, optimizing Standard C (ANSI C89, ISO C99, ISO C11) compiler suite that targets the Intel MCS51 based microprocessors (8031, 8032, 8051, 8052, etc.), Maxim (formerly Dallas) DS80C390 variants, Freescale (formerly Motorola) HC08 based (hc08, s08), Zilog Z80 based MCUs (z80, z180, gbz80, Rabbit 2000/3000, Rabbit 3000A, TLCS-90) and STMicroelectronics STM8. Work is in progress on supporting the Microchip PIC16 and PIC18 targets. It can be retargeted for other microprocessors.

I don't have the best experiences with SDCC though...

[C]

As I stated MEMORY is EASY
Start with two static rams 8-bits wide.

the CPU works with
8-bit value 
 Which byte is selected by A0 You have controls to read/write to only one memory chip.

16-bit Value
  You have two possible ways in memory,
Data is aligned in which read/write 16-bits, use both chips at same time.
Data is UN-aligned hardware does two 8-bit reads or writes.

32-bit value
  You have two 16-bit operations or 8,16,8 operation.

A0 is in UDS, LDS and with these two signals allow all of the above where one chip or both is used.

You have an address valid signal(AS) and a data valid(DS)

The 68000 works from a NO memory state.
Where Z80 and other chips like this assume all memory is fast and exists, you need a slow down signal. Most have no way of signaling NO Memory.
68000 works from no memory and times out when no memory exists. The signal is now a speed up(DTACK).
This is not a new Idea, The PDP-11 and others used it. Makes having a very long bus very easy. Makes connecting two backplanes with a cable easy. Less distance from CPU is faster.

[colorado.rob]

I don't have the best experiences with SDCC though...

I've never used it myself (not exactly true, but it has been a *very* long time).  What sort of problems did you run into?

[technix]

It appeared to me that a lot of you prefers MC68HC000 over Intel 8088. Please keep in mind that *existing software* is also a metric in my consideration. If I built a retro computer using 8088 I can build it in the style of an IBM PC XT compatible with some AT features, complete with a compatible BIOS. This will allow vanilla MS-DOS as well as a good selection of existing DOS programs to work, including games. Heck my system even have a FPU for the most demanding titles.

I am not familiar with the 68000-based systems (they are virtually nonexistent in China) so is there any open standards comparable to IBM PC?

[technix]

The 68000 works from a NO memory state.
Where Z80 and other chips like this assume all memory is fast and exists, you need a slow down signal. Most have no way of signaling NO Memory.
68000 works from no memory and times out when no memory exists. The signal is now a speed up(DTACK).
This is not a new Idea, The PDP-11 and others used it. Makes having a very long bus very easy. Makes connecting two backplanes with a cable easy. Less distance from CPU is faster.

For the 8088 since I am generating the clock from the chipset CPLD anyway I can clock scale as I see fit. A clock prescaler with a read/write configuration register isn't that hard to implement.

[legacy]

If I built a retro computer using 8088 I can build it in the style of an IBM PC XT compatible with some AT features, complete with a compatible BIOS. This will allow vanilla MS-DOS as well as a good selection of existing DOS programs to work, including games. Heck my system even have a FPU for the most demanding titles.

[C]

In my book the 8088/8086 was not a good system. I think it's a waste of time to learn poor hardware. For software you could use an IBM PC XT emulator. Again Software foundation is bad.

And you are comparing the 8088/8086 to what could be a multi-user and multi-CPU system.
If you build your boards correctly. the extra boards could all be used.
You would have a multi-CPU system.

Want to play or work with fancy then replace the 68000 with a 68010. This lets you run many operating systems at same time on one CPU

So a lot of old poor XT code VS learning great hardware. If you want copycat then use the VME bus.
 

[legacy]

about x86, a friend some time ago wrote

Within a couple of months, the license of the original i486 drops, and you are free to re-implement it without a license from Intel, so what I am going to do aims to recreate it as an open platform, with an aim for making it viable for embedded work.

I still wonder WHY? x86 is an ugly architecture, and we have Quark and Edison to play with x86-binaries, so ... I can't really understand.

A dude on hackday made a "GameBoyAdv"-like console (well, it's the "shell-like version"), using one of those 486-embedded boards, an Oled LCD, and a battery, and he said that it makes sense since he wants to use x86 binaries (gameconsole emulators?) which don't happen to have a decent opensource equivalent.

I can't find the hackaday topic atm, but it's there

edit:
found! here it is!

[free_electron]

4004  or, for a real challenge : mc14500 ( a 1 bit cpu)

[legacy]

4004  or, for a real challenge : mc14500 ( a 1 bit cpu)

LOL, 8008 basic 
(
edit:
a dude actually made it
the code size is 722 bytes)

[Bruce Abbott]

Please keep in mind that *existing software* is also a metric in my consideration.

What *existing software* do you want to run? If it's pre-compiled MSDOS applications then you will need to reproduce an entire IBM PC, either at the hardware level or using a custom BIOS to interface to your non-standard hardware (with reduced compatibility). If you are happy to recompile source code then you have more flexibility, but then the CPU choice also becomes less of an issue.

If I built a retro computer using 8088 I can build it in the style of an IBM PC XT compatible with some AT features, complete with a compatible BIOS. This will allow vanilla MS-DOS as well as a good selection of existing DOS programs to work, including games.

I think you may be underestimating the amount of work required to produce a sufficiently XT compatible machine, particularly for games.

My first PC was an actual IBM, but it was not compatible with a lot of PC software because it was an IBM JX. For me it was an excellent choice because I had the technical reference manuals and a desire to learn the PC architecture inside and out. The incompatibilities taught me a lot because I had find out why they occurred and fix them - including hacking into games, writing drivers and making my own hardware. However if you don't want to do that it could be a nightmare.

The 8088 seems attractive due to its low pin count and compatibility with Intel peripherals, but needs a lot of specific support hardware to become a fully functional XT. And in the end all you have is another crappy XT clone. What's the fun in that?

An interesting bit of history - Why the IBM PC Used an Intel 8088

"For IBM it was extremely different because this was a project where they let a supplier -- a partner, whatever you call us -- shape the definition of the machine and provide fundamental elements of the machine. When they first came to us, their concept was to do an 8-bit computer. And the project was more notable because they were going to do it so quickly and use an outside company....The novel thing was: could you work with outsiders, which in this case was mostly ourselves but also Intel, and do it quickly? And the key engineer on the project, Lou Eggebrecht, was fast-moving. Once we convinced IBM to go 16-bit (and we looked at 68000 which unfortunately wasn't debugged at the time so decided to go 8086), he cranked out that motherboard in about 40 days." - Bill Gates.

Basically, IBM decided to use the 8088 because it was cheap, and they just wanted something to compete with other 8 bitters. If they had picked the much more powerful 68000 they might still have a hold on the PC market today. As it was they realized their mistake (and tried to correct it with the PS/2 line) too late.
 

I am not familiar with the 68000-based systems (they are virtually nonexistent in China) so is there any open standards comparable to IBM PC?

Compared to the PC not as much, but if you want to run existing PC software (including games) you need full PC compatibility anyway so any other standards are irrelevant.

OTOH why limit yourself to recreating a boring PC? Be adventurous, learn new stuff, try doing something everybody else hasn't done before!

I am not that keen on coding everything in assembler, so availability of some kind of ISO C compiler is almost a must, and the ability to write startup code in as little assembler as possible is a bonus point.

Few of us code everything in assembler (I am one of the exceptions) but 68000 machine code is much easier to work with than 8086. Orthogonal instruction set, no segment registers, few instructions to learn, powerful easy to remember opcodes and neat looking code. Compared to Intel it's a dream. Why don't you try it?

My software work platform is Linux and macOS so Windows-only tools are out of consideration.

MacOS was originally 68000, and Linux is based on Unix which the 68k instruction set is particularly well suited to. It may be more familiar than you think...

Motorola 68000

"the 68000 became the dominant CPU for Unix-based workstations including Sun workstations and Apollo/Domain workstations, and also was used for mass-market computers such as the Apple Lisa, Macintosh, Amiga, and Atari ST. The 68000 was used in Microsoft Xenix systems, as well as an early NetWare Unix-based Server."

Notable systems
The original Apple Macintosh and early successors use the 68000 processor as their CPU.
The original Apollo Computer workstations, DN100, DN400 and DN600 use two 68000 processors as main CPU.
The 1981 SUN workstation and its subsequent commercial spinoff the 1982 Sun-1 workstation used the 68000 as their CPU.
The Sega Genesis game console uses a 68000 processor (clocked at 7.67 MHz, 15/7 of the NTSC video colorburst frequency) as its main CPU and the Sega CD attachment for it uses another 68000 (clocked at 12.5 MHz).
SNK's Neo Geo.
The Commodore Amiga 1000, 2000, 500, 600 and CDTV use the 68000 processor as their CPU.[24]
The Atari ST computers use the 68000 processor, initially with a clock speed of 8 MHz, and later switchable to 16 MHz in the Mega STe.
CDTV, the world's first compact disc based multimedia platform, uses the 68000 processor as its CPU.[25]
The TI-89 Graphing Calculator uses the 68000 processor at 10, 12, or 16 MHz, depending on the calculator's hardware version.
The CP System uses the 68000 processor at 10 MHz;

 
 

[Bruce Abbott]

68000 works from no memory and times out when no memory exists.

Unless you ground DTACK.

DTACK Grounded

[C]

68000 works from no memory and times out when no memory exists.

Unless you ground DTACK.

DTACK Grounded

And if you do so you have a lot less capable of system for a very small savings.
Easy to replace hardware with software at a cost of a slower system.
Easy to leave out hardware and get a less capable system.
Also not hard to save when you have more than one CPU in a system.

[eugenenine]

IIRC the VME bus was a standard built around the 68k.  You make a backplane then plug in processor boards, memory boards, etc.

[Bruce Abbott]

And if you do so you have a lot less capable of system for a very small savings.
Easy to replace hardware with software at a cost of a slower system.
Easy to leave out hardware and get a less capable system.
Also not hard to save when you have more than one CPU in a system.

OP asked which processor is easier to use, not which one can go faster or have more CPUs in a system.

It has been suggested that the 68000 is harder to use than the 8088 because it has a 16 bit bus and requires generation of a DTACK signal. But if ease of use is the goal then you could use a 68008 (8 bit data bus) and ground DTACK. Of course all your memory and peripherals would then have to run at 2MHz bus speed, but that's not a problem for modern chips. Less expandable than it could be? Sure. But easier.

[eugenenine]

OP asked which processor is easier to use, not which one can go faster or have more CPUs in a system.

6502, has the ease of hardware interfacing of the 8088 and the easier coding of the 68k.

[westfw]

keep in mind that *existing software* is also a metric in my consideration.

Oh.  Well in that case, I think you're screwed.  Most of the 68k systems of that era (or at least most of the ones that have had any software preserved) has highly proprietary operating systems that were heavily dependent on "elegant" but proprietary custom hardware.  (I'm thinking of Amiga, Macintosh, and the early unix workstations (SUN-1 !))

Go with the 8088 (or get a NEC V-20, for some near additional capabilities and nearly identical hardware.)  No need for an 8087 (I'm not sure why you included that, since the 68k doesn't have an FPU.)  MSDOS and the PC BIOS are at least designed for portability across "similar" hardware, and there has been continued development of things like FreeDOS.

Other possibilities for retro-computing with some hope of running old software include a Z80 running CP/M or a 6809 running OS-9. (I'm particularly fond of this recent 4-chip Z80 hardware: https://hackaday.io/project/19000-a-4-4ics-z80-homemade-computer-on-breadboard - it's been made to work by several people in several formats, and is currently running CP/M, and Forth, and "ROM" BASIC.  The basic design might work for other CPUs, too...)  (Hmm.  There was an OS-9 for 68k; I'm not sure if it ever got much traction...)

6502 has similar problems to 68k  - popular software (C64, Apple ][, Atari) was dependent on custom hardware.  In general, when you look at classic personal computers, there was a LOT of custom hardware, and most of the "really clever" programming came from manipulating that hardware, rather than being generic xx-cpu software.  (It's a bit depressing.)

(Keeping in mind that while most of that "personal computer revolution" was happening, I was off programming mainframes and not paying THAT much attention.  But that's how I remember things...)

[technix]

Also if I want to build a PC XT compatible there are a lot of reference material and test code out there. The system design can be checked against the source code of DOSBox. The BIOS code can be checked against the released original IBM BIOS as well as code from SeaBIOS and coreboot projects. And a test boot of vanilla MS-DOS 6.22 can be used to exercise various components.

[Bruce Abbott]

if I want to build a PC XT compatible

Then use an 8088. Why even consider 68000 if that's what you want to do?

[amyk]

Why not a 486/AT clone, then you can actually run more interesting software: http://www.s100computers.com/My%20System%20Pages/80486%20Board/80486%20CPU%20Board.htm

[technix]

if I want to build a PC XT compatible

Then use an 8088. Why even consider 68000 if that's what you want to do?

I have the 8088 already, but got a bit stumped at how to hook up 8088, 8087, 8288 and 8284 into the actual CPU complex, how to hook the CPU complex to the platform controller CPLD, how to interpret the signals from the CPU complex, and how to make the CPLD-based chipset work.

[legacy]

[Nusa]

You're mostly asking questions that are easily answered with basic search skills.

Assemblers for all of the above? Easily found. Including plenty of cross-assemblers written in portable C, since these chips predated the PC and binaries were often compiled with whatever computer was available.

How to build a 8088 based XT clone? Easily found, including schematics for single-board computers. Ditto for the other CPU's mentioned.

If the real goal is simply to have a working XT clone, the cheapest route is still to just buy one. Plenty of working antiques out there for not that much money. If you want original glue logic IC's, look for XT/XT clone motherboards, as is/for parts ones will be cheaper. Often the chips are even socketed.

[amyk]

You're mostly asking questions that are easily answered with basic search skills.

Assemblers for all of the above? Easily found. Including plenty of cross-assemblers written in portable C, since these chips predated the PC and binaries were often compiled with whatever computer was available.

How to build a 8088 based XT clone? Easily found, including schematics for single-board computers. Ditto for the other CPU's mentioned.

If the real goal is simply to have a working XT clone, the cheapest route is still to just buy one. Plenty of working antiques out there for not that much money. If you want original glue logic IC's, look for XT/XT clone motherboards, as is/for parts ones will be cheaper. Often the chips are even socketed.

You can even get the schematics of the real thing: http://www.minuszerodegrees.net/manuals.htm#IBM

[westfw]

(It's one of the things that annoyed me about that TV series "Halt and Catch Fire."  The trope that the IBM PC was this secret bit of high technology that it took some super-hackers to "reverse engineer."  When in fact the IBM was a horribly BORING hardware design, most of which was published.  ("published" but not at all "open source."  IIRC, the hard part was finding engineers who hadn't been "contaminated" by looking at any of the published material.)

There were even x86-based personal computers that ran MSDOS, but were NOT IBM-PC compatibles (eg the HP-150; the first PC I owned.  (Not counting the CP/M system that never got more than half-assembled.))

[NivagSwerdna]

I have some vintage computer projects on the sketch boards for a while, but I am still undecided which to pursue first. Which of the two would be easier to use in your opinion: Intel 8088 + Intel 8087, or Motorola MC68HC000? Aspect to be consider includes the simplicity of the system design, and availability of compilers, tools, libraries and existing programs.

68000

What is your end goal?

[rstofer]

In my view, the EASIEST chip to use is the Intel 8085.  It is a superset of the 8080 but is static CMOS so there is no minimum clock rate.  Back in the early '80s, I built several projects with this chip and its brethren.  Given the availability of static CMOS RAM in 64kx8, the memory system is quite simple.

You can run CP/M on such a system with a little effort.  Interfacing a Compact Flash device is quite easy.

CPU + AddressLatch + CMOS RAM + EEPROM + Memory Map Logic (map EEPROM into memory space at startup, map it out when some condition occurs) + Address Select Logic for IO devices + a couple of UARTS and you're good to go.

Or, just buy one of the Zilog eZ80F91 boards - they run at 50 MHz and have all the required memory.  Add a daughter card containing the Compact Flash and a couple of UART->USB devices (or one of the newer quad UART->USB) and you're up and running on CP/M with 16 logical drives.  CP/M RIPS at 50 MHz.

[Elf]

[...] but is static CMOS so there is no minimum clock rate.

WDC's 65C02 also has this helpful property.

Or, just buy one of the Zilog eZ80F91 boards - they run at 50 MHz and have all the required memory.

Plus Ethernet with a BSD sockets implementation. Definitely nice little boards, although I wish the modules were a bit cheaper.

I didn't realize it was easy to get CP/M going on them. Interesting!

[technix]

A lot of you are suggesting 68000 for this vintage computer project. I don't have the required peripheral chips yet (and I doubt whether my 8-bit USB and Ethernet chips would work on a 16-bit bus) so please suggest chips for UART, display output stack, Ethernet and USB-OTG. Or is there any CMOS variant of 68008 (which have a native 8-bit bus)

[Nusa]

A lot of you are suggesting 68000 for this vintage computer project. I don't have the required peripheral chips yet (and I doubt whether my 8-bit USB and Ethernet chips would work on a 16-bit bus) so please suggest chips for UART, display output stack, Ethernet and USB-OTG. Or is there any CMOS variant of 68008 (which have a native 8-bit bus)

The 68008 in any varient hasn't been manufactured for over 20 years, nor was it ever a big seller. I don't know how easy it is to find old stock.

Using a 16-bit data bus on an 8-bit device is trivial. You just use the lower 8 bits. You do, of course, still need the full address decoder for the chip select line.

[legacy]

[legacy]

the EASIEST chip to use is the Intel 8085

yup, 6800 : 68hc11 = 8080 : 8085
I believe 

[technix]

Rounding up again.

* Among the two options I mentioned, 68HC000 is easier to use.
* It is okay to use 8-bit peripherals with a 16-bit bus, by hooking the chips to just the low byte. Drivers have to be made flexible enough to accommodate this.
* 2MB of SRAM is enough for a 68000 retrocomputing platform.
* When using 68000, forget existing code.

[NivagSwerdna]

I have some vintage computer projects on the sketch boards for a while, but I am still undecided which to pursue first.

This thread seems to be oscillating wildly.  What do you actually aim to achieve?
If you want cool and vintage... Cosmac 1802 (if it was good enough for the moon landing), 6502 (PET) or Z80 (Jupiter ACE?).  1802 and 6502 are very simple especially with SRAM.  Z80 slightly more complex if you use DRAM. 
For the time 6800 and the 68000 were cool, the 68000 being used in quite a few communications systems.  The 6800 was surplanted by the 6502 which was much much cheaper. TBH they are just a bit dull IMHO. Also 8080, 8085... yawn... what came after was far cooler.
Or... at a tangent... how about re-creating a classic arcade game (they often had multiple processors)... something along the lines of... http://www.fpgaarcade.com/replay/

[Fungus]

Have you studied ether of the two?

I see the 8088 as an 8-bit computer stretched to a 16-bit instruction set and a 1 meg address space, Segmented memory.
From a software side try to create an array with more then 64k of data

68000 is 32-bit with a 16-bit external bus. The big cripple was that the 68 pin chip needed more pins to get 8 more address lines out of chip.

Not if it's a 68008. 

PS: Z80 is probably easiest.

If the real goal is simply to have a working XT clone, the cheapest route is still to just buy one.

Building an XT clone is a LOT of work.

I wouldn't build anything with the goal of "running old software". Nobody will be impressed and there's emulators for that.

If I was going to build a retro computer I'd dream up a new one.

[westfw]

Quote

is static CMOS so there is no minimum clock rate.

This is also true of the current CMOS Z80 chips (Z84c00) (and I think of a lot of the more recent CMOS versions of most "classic" chips, like the 80c88 you can get from Intersil (sticker shock!), the 80c18x (might be an interesting choice; has some built-in features that make systems easier to build, but will still run MSDOS), etc)

[rstofer]

It's pretty tough to stop this thread from oscillating when the original topic seemed to be 'vintage' and then Berkeley Sockets comes up.

It comes down to a question of "how vintage?".  Mid '70s or mid '80s (perhaps mid '90s).  To me, the Altair 8800 ('75) defines 'vintage'.  Everything else is a "Johnny come lately".  OK, I stretch it to maybe include the Apple II and certainly the 8085 (6502, 1802, Z80).  The introduction of the IBM PC in 1980 pretty much wiped out the idea of 'personal computing' and started 'just commercial computing'.  Anybody could run an IBM PC.  With the Altair, it took a certain amount of understanding.  The Apple II was a lot easier to understand.

So, which vintage are we talking about?

[colorado.rob]

Vintage to me is anything that was built before you were born or maybe existed when you were a young child.  If the OP is in their 20s, vintage isn't that long ago.

If you're doing "vintage", why not do something interesting like a PDP-11?

https://opencores.org/project,w11

[rstofer]

Vintage to me is anything that was built before you were born or maybe existed when you were a young child.  If the OP is in their 20s, vintage isn't that long ago.

If you're doing "vintage", why not do something interesting like a PDP-11?

https://opencores.org/project,w11

But I predate the Eniac!

ETA:

That PDP-11 project seems like fun.  I have the Digilent board...
The ONLY thing that makes these projects worth the effort is if there is software to run.  In the PDP-11 case Unix V5 runs.  You can't expect better than that!

The IBM1130 software is available and it's pretty easy to build a system using the information provided in the Functional Characteristics manual.  It won't be cycle accurate but it will run at 50 MHz whereas the original that I used ran at 400 kHz.  Pretty good speedup.

http://ibm1130.org/

Way back in '70, this was the first computer I ever used or even saw!  It helped immensely getting through undergrad using the IBM Electronic Circuit Analysis Program (similar to LTspice but with punched cards and no built-in components).  For school work, it was all about the plotter.  There's nothing like an accurate Bode' Plot to attach to homework.

I still use my FPGA implementation whenever I feel the need for Fortran IV and a plotter!

[colorado.rob]

But I predate the Eniac!

https://en.wikipedia.org/wiki/Jacquard_loom
 

[NivagSwerdna]

It's pretty tough to stop this thread from oscillating when the original topic seemed to be 'vintage' and then Berkeley Sockets comes up.

I think you will find this thread was 100% socket free until you mentioned it?

https://www.eevblog.com/forum/microcontrollers/converting-struct-in-to-array-changing-byte-order/msg1248367/#msg1248367

[rstofer]

It's pretty tough to stop this thread from oscillating when the original topic seemed to be 'vintage' and then Berkeley Sockets comes up.

I think you will find this thread was 100% socket free until you mentioned it?

Reply #40:

https://www.eevblog.com/forum/microcontrollers/for-a-vintage-computer-project-which-processor-is-easier-to-use/msg1253379/#msg1253379


True, it was tangential to my describing the eZ80F91 CP/M project...

I'm starting to like that PDP-11/70 project...

[NivagSwerdna]

I'm starting to like that PDP-11/70 project...

Now you are talking.  Just need to add a VT100 with smooth scroll turned on. Software wise you just need DECUS Adventure and for fun TECO.

[Bruce Abbott]

The introduction of the IBM PC in 1980 pretty much wiped out the idea of 'personal computing' and started 'just commercial computing'.  Anybody could run an IBM PC

Anything from the 1980's is now 'vintage'.

When the IBM PC was released in 1980 it had a keyboard, 64k RAM, and a tape interface. Everything else was an option. You could have one or two 360k floppy drives, but the only graphics card available was monochrome text. Compared to other 'personal' computers of the day it sucked, but it had one sure-fire selling point - the letters I B M.

The other reason the PC became so popular is that it was incredibly easy to reproduce. Clone manufacturers quickly switched from making Apples to PCs, which they were able to sell for half the price of a real IBM. Since almost everything was an option, 3rd party manufacturers had an enormous potential market to exploit. But it wasn't until the 1990's that it took over from 'personal' computers, because until then it really was just a 'business' machine.
 
I think the main reason old PCs are often not considered 'vintage' is that a modern PC can still (for the most part) run the same software. And the early IBM wasn't really a computer anyway - it was little more than a motherboard with CPU and RAM, stuffed full of 3rd party cards. With so many upgrade options and the continually advancing specifications, there isn't any point at which can you say this PC is vintage, but that one isn't.

It also doesn't help that so many of them are still around. I have 5 working PCs of various vintage, a few more that just need a new power supply or hard drive, 4 laptops, and a cupboard full of parts. Time for a cull!

[rstofer]

I walked past a PDP-11 one time and that's as close as I ever got.  Been in the same room with a couple of Vaxen...

I'm just not a PDP-11 guy but since I already have the board(s), it might be fun to install the Xilinx toolchain under Debian Linux and see how it works out.  I rather like the author's idea of using a peripheral server under Linux.  I could see converting my IBM1130 to do the same thing.

I don't have the Nexys3 board which will give much faster USB transfers and also has more onboard SRAM but if I can get the system running on a Spartan 3 Starter Board, I will know if I like it enough to part with $270 on an obsolete board (if they still have them).

[rstofer]

The introduction of the IBM PC in 1980 pretty much wiped out the idea of 'personal computing' and started 'just commercial computing'.  Anybody could run an IBM PC

Anything from the 1980's is now 'vintage'.

See, I wouldn't consider an IBM PC (beyond, maybe, the original) to be "vintage".  By the time IBM got in the business, there were dozens of S-100 machine types already running (Altair, CompuPro, Cromemco, etc).  The 8080 machines had UCSD Pascal two years before the IBM PC was introduced and this finally allowed these machines to do useful business computation.

Furthermore, current machines are simply evolutionary, not revolutionary.  The Altair 8800 was revolutionary and I suppose the Apple II fits that concept as well.  Perhaps even the Commodore Pet...  Maybe, if I was pushed, I would include the original IBM PC because it introduced a new bus structure that was a vast improvement over S-100.  But nothing beyond that...

I still think the PC-AT keyboard is the best keyboard ever invented.

[Elf]

It's pretty tough to stop this thread from oscillating when the original topic seemed to be 'vintage' and then Berkeley Sockets comes up.

Wow, I did not realize that offhandedly mentioning the socket library was so disruptive. BSD sockets came out in 1983 with 4.2BSD by the way. As in, run on systems like the VAX 11/780, of similar vintage to the PDPs you are discussing. I never had a PC older than an Apple ][e, but I do still have (and use) a lot of old UNIX systems like VAX, Alpha, various generations of Sun, SGI, etc. The concept of networking came around a lot earlier for multiuser systems than it did for personal computers.

As you point out, vintage has a lot of personal meaning. I'm sure to a lot of people even a Pentium II with a 56k modem is "vintage" now. It is pretty hard to decide the dividing line between what is vintage and what is not, since so many things that were created in the 70s-90s are still in use, virtually unchanged. I think of vintage as anything on display on the 2nd floor of the Living Computer Museum, which by the way is worth a visit, unless mentioning this is also too off topic and is going to send someone into a tizzy.

As far as these projects go, I generally just think of it as trying to construct your own computer around an older architecture (e.g. 68k, 6502, Z80, et. al.). Some people want to recreate the functionality of an entire original machine, some people want to use an original ceramic chip and discrete logic with a rat's nest of breadboard wires, some people (like me) just want to use the old CPU and the rest can go. I have a 6502 game system project queued up where I want to use a modern FPGA as an MMU to do paged memory, use either an FPGA or a DSP for a tiling and sprite engine, have coprocessors for USB mass storage, and output to HDMI at 720p. I enjoy the mix of old and new. I don't feel the need to go back to cassette tape program storage or to try and faithfully rebuild an 8" floppy drive, but I can still enjoy writing the "controlling software" in 6502 assembly, having to be creative with use of memory and cycles.

For technix, the original poster, it looks like he already has some older chips on hand that he wants to find a use for but also wants to incorporate modern components; e.g. the mention of USB OTG and Ethernet.

[colorado.rob]

If you want to do something legacy from the 80's, how about a Cray?  I mean, those things just *looked* cool.

http://www.chrisfenton.com/homebrew-cray-1a/

[NivagSwerdna]

The other reason the PC became so popular is that it was incredibly easy to reproduce.

In my first job after university I worked for a company that amongst other things developed (token ring) network cards for PCs... our design was coming along nicely and then IBM moved to the IBM PS/2 and patented/licenced the new card bus (MCA).... we received our first real PS/2 and within minutes of arriving we had completely dissembled it... to our dismay it didn't work when we reassembled it.... didn't like the battery backup being removed or something like that.  But it wasn't long before either the bus was licenced or legally challenged... can't remember the details.  This thread is making me feel very old!

[rstofer]

There is also a Computer History Museum in Mountain View (Silicon Valley) that is well worth the trip.

[technix]

For me the vintage part is the CPU core complex design and maybe some code (if the hardware is built to spec with some existing standards like IBM PC XT.) I am fully okay with using modern parts for peripherals and system bus (the Intel EPM7128 CPLD system bus controller, WCH CH375 USB OTG controller, and WCH CH395 Ethernet controller.) And I am born in 1993 so anything in the 80s are vintage to me, not just the original IBM PC XT (or AT)

[colorado.rob]

How many of you geezers feel old now?!? 

[technix]

How many of you geezers feel old now?!? 

For my experiment environment it is almost entirely legacy free (No PS/2, no VGA, no RS232) so I cannot even use true vintage designs (or I won’t be able to even write the disks for or communicate with the old machine from my usual workstation.)

I cannot even reuse the display of my computer with the adapter chips as there is no way a vintage computer can drive a 4K 2160p60 monitor.

[cdev]

The number one thing I remember from that era was that the older x86 family of chips had an unbelievable number of workarounds for this or that that needed to be incorporated into code (and of course the tools that exist today were absent)

the kludges made the early intel x86 platform quite painful to use compared to say the motorola 68000.

[NorthGuy]

For my experiment environment it is almost entirely legacy free (No PS/2, no VGA, no RS232) so I cannot even use true vintage designs (or I won’t be able to even write the disks for or communicate with the old machine from my usual workstation.)

I cannot even reuse the display of my computer with the adapter chips as there is no way a vintage computer can drive a 4K 2160p60 monitor.

Why vintage? Wouldn't it better to design your own?

[rstofer]

The number one thing I remember from that era was that the older x86 family of chips had an unbelievable number of workarounds for this or that that needed to be incorporated into code (and of course the tools that exist today were absent)

the kludges made the early intel x86 platform quite painful to use compared to say the motorola 68000.

Maybe have the video come out as R-G-B and use a TV set for a display.  Are you sure your monitor doesn't have VESA VGA capabilities?  There are a number of monitors around that can handle VGA or Composite Video.  I have a little 7" monitor from AdaFruit that might be workable:
https://www.adafruit.com/product/1667

5-1/4" floppies are kind of a PITA.  It is dead simple to add a Compact Flash disk to just about any bus oriented system.  If you have a disk image, you can use the Linux dd command to write the image.  If you want to create 16 drives like I did for the Z80 system, you can write some code to build empty drive images in a large file, copy in the real image for the first drive and then write the entire file to the Compact Flash with one command.  Somehow I got the CP/M image plus Kermit loaded on the first drive and I transferred everything else from the 8" floppies on my CompuPro Z80 machine.

[Nusa]

How many of you geezers feel old now?!? 

For my experiment environment it is almost entirely legacy free (No PS/2, no VGA, no RS232) so I cannot even use true vintage designs (or I won’t be able to even write the disks for or communicate with the old machine from my usual workstation.)

I cannot even reuse the display of my computer with the adapter chips as there is no way a vintage computer can drive a 4K 2160p60 monitor.

I don't feel like a geezer yet, but I am getting up there. I've experienced most of these CPU's the first time around as a Computer Science student and professional, so I don't feel any great urge to recreate them myself.

PS/2 and VGA didn't exist in the vintage era you're trying to replicate, so why would you need them now? RS232 was, and it's easily adapted to USB with a cheap adapter cable. Which means you can use any modern computer as a serial terminal, which includes most smartphones. If you want emulations of VT100 terminals or the like, those exist. In any case, that gives you the sort of interface you ACTUALLY HAD back in the day. (Well, it might have been an even older paper terminal TTY with paper tape reader/writer setup, but unless you actually own that antique you can forget that.)

If you're truly vintage, all the computer itself will have is idiot lights, toggle switches, and UART or RS232 port. Part of the experience, if you aren't old enough to have lived it, is being able to stop the computer clock, single-step programs, and modify memory locations using nothing but the lights and toggle switchs on the computer panel.

[technix]

How many of you geezers feel old now?!? 

For my experiment environment it is almost entirely legacy free (No PS/2, no VGA, no RS232) so I cannot even use true vintage designs (or I won’t be able to even write the disks for or communicate with the old machine from my usual workstation.)

I cannot even reuse the display of my computer with the adapter chips as there is no way a vintage computer can drive a 4K 2160p60 monitor.

I don't feel like a geezer yet, but I am getting up there. I've experienced most of these CPU's the first time around as a Computer Science student and professional, so I don't feel any great urge to recreate them myself.

PS/2 and VGA didn't exist in the vintage era you're trying to replicate, so why would you need them now? RS232 was, and it's easily adapted to USB with a cheap adapter cable. Which means you can use any modern computer as a serial terminal, which includes most smartphones. If you want emulations of VT100 terminals or the like, those exist. In any case, that gives you the sort of interface you ACTUALLY HAD back in the day. (Well, it might have been an even older paper terminal TTY with paper tape reader/writer setup, but unless you actually own that antique you can forget that.)

If you're truly vintage, all the computer itself will have is idiot lights, toggle switches, and UART or RS232 port. Part of the experience, if you aren't old enough to have lived it, is being able to stop the computer clock, single-step programs, and modify memory locations using nothing but the lights and toggle switchs on the computer panel.

To me PS/2 and VGA are already vintage - in contrast my home workstation is built around USB and DisplayPort for peripherals and display, and a further upgrade to Thunderbolt 3 is planned. I know that RS232 can be easily adapted (and for homebrew boards I would just build the USB to TTL UART chip right in there.)

For me since there is a CPLD with JTAG boundary scanning in the system bus and generates the processor clock, I can skip the toggle switches and idiot lights and use the boundary scan features of the CPLD as a makeshift JTAG debugger.

[Muxr]

I always found 6502/6510 to be pretty easy, in fact in college they used 6502 to teach us assembly. I was doing x86 assembly by that point.

[technix]

Quote

is static CMOS so there is no minimum clock rate.

This is also true of the current CMOS Z80 chips (Z84c00) (and I think of a lot of the more recent CMOS versions of most "classic" chips, like the 80c88 you can get from Intersil (sticker shock!), the 80c18x (might be an interesting choice; has some built-in features that make systems easier to build, but will still run MSDOS), etc)

I found some 80286 and 80287 though. Are those any good?

And I also found someone selling Am486DX4-75 in QFP package...

[technix]

I always found 6502/6510 to be pretty easy, in fact in college they used 6502 to teach us assembly. I was doing x86 assembly by that point.

My school had 8086 real mode assembly and later ARM9.

Is ARM9 anywhere near vintage?

[Muxr]

I always found 6502/6510 to be pretty easy, in fact in college they used 6502 to teach us assembly. I was doing x86 assembly by that point.

My school had 8086 real mode assembly and later ARM9.

Is ARM9 anywhere near vintage?

it's decently old I suppose but it's a far cry from the simplicity of the 8bit CPUs from the 80s.

[Elf]

I found some 80286 and 80287 though. Are those any good?

80286 is a bit of an oddball in the x86 lineup. If I recall correctly, it is the first processor with pmode and a page descriptor table, but it is also only a 24-bit address space. I think it was also missing a lot of features compared to 386 32-bit pmode.

If you like x86 assembly and want a quirky x86 chip I guess you could use it? But if not, I'm not sure what other redeeming features it would have.

[westfw]

I found some 80286

Ugh.  The 286 was annoying.  Just when all the mainframes, minicomputers, and 68k unix systems had decided that "paging" was The Way To Do Memory Managment, Intel released the 286 with "segmentation", implemented in such a way as to make linear addressing of large amounts of memory (more than 1M?) impossible (nearly impossible?)   IIRC, it didn't do much more beyond "faster 8086" and didn't last very long before the (much better) 386 came out.  Most of the 286 systems had 1M of RAM, but it was slightly easier to use than trying to have that much on an 8088 (once you considered that you also need BIOS, and IO cards, and such.)

[technix]

I always found 6502/6510 to be pretty easy, in fact in college they used 6502 to teach us assembly. I was doing x86 assembly by that point.

My school had 8086 real mode assembly and later ARM9.

Is ARM9 anywhere near vintage?

it's decently old I suppose but it's a far cry from the simplicity of the 8bit CPUs from the 80s.

There are reasonably integrated ARM9 chips though, like AT91SAM9260 or IMX233 or those near ubiquitous S3C2440 modules. Although I don’t like S3C2440 to a passion for its lack of mainline Linux support (in contrast Allwinner has some mainline Linux support and Microchip have near perfect mainline Linux support)

[NivagSwerdna]

Is ARM9 anywhere near vintage?

Nope.  The ARM 1 co-processor attached via 'tube' to a BBC Micro B could be considered vintage.

Early Acorn Archimedes also definitely vintage. (ARM3?)

[legacy]

Early Acorn Archimedes also definitely vintage. (ARM3?)

&& ARM RISCPC/600, with StrongARM CPU board 

At least it also supports a PC-guest card able to to run MSDOS

[legacy]

[NivagSwerdna]

That looks far too modern.

http://www.sunrise-ev.com//membershipcard.htm

[technix]

Early Acorn Archimedes also definitely vintage. (ARM3?)

&& ARM RISCPC/600, with StrongARM CPU board 

At least it also supports a PC-guest card able to to run MSDOS

I don’t think I can find those chips, but coupling an ARM to an Am486DX4-75 is an interesting idea. That is not vintage enough though.

Also since both the 486 and ARM9 are supported by Linux and mainline Linux have a heterogeneous core scheduling the support can be interesting.

[legacy]

Well ... 

[rstofer]

Well ... 

So, "vintage" is now "last Thursday"...

What about a Zynq-7000 with dual ARM9 cores and perhaps enough FPGA fabric to implement a Cray-I as the scheduler.

[legacy]

So, "vintage" is now "last Thursday"...

I was commenting the idea of putting a 486 chip coupled with an ARM chip, oh, with the benefit of linux, of course 

Considering the complexity of the whole .. well, it's hundred light years far from the concept of fun.


About fpga, well ... they save PCB problems, but I would suggest neither complex nor big projects since they usually are time consuming, with a progressive probability of being frustrating.

[legacy]

That looks far too modern

Usually vintage is triggered on something that is at least 10 years old, but you miss that computers don't measure the time in human  years

Human-years and computer-years are in log2 relation, and since "vintage" is a human concept ... well you can measure the equivalent "vintage factor" applied to computers.

1 human-year, is 2 computer-years
2 human-years, are 4 computer-years
3 human-years, are 8 computer-years
4 human-years, are 16 computer-years
5 human-years, are 32 computer-years
6 human-years, are 64 computer-years
7 human-years, are 128 computer-years
8 human-years, are 256 computer-years
9 human-years, are 512 computer-years
10 human-years, are 1024 computer-years
11 human-years, are 2048 computer-years
12 human-years, are 4096 computer-years
13 human-years, are 8192 computer-years
...
20 human-years, are 1M computer-years
21 human-years, are 2M computer-years
22 human-years, are 4M computer-years
23 human-years, are 8M computer-years
24 human-years, are 16M computer-years
25 human-years, are 32M computer-years
26 human-years, are 64M computer-years
27 human-years, are 128M computer-years
28 human-years, are 256M computer-years
29 human-years, are 512M computer-years
30 human-years, are 1G computer-years
31 human-years, are 2G computer-years
32 human-years, are 4G computer-years


Therefore everything that is older than four human-years is "computer-vintage"

Corollary

Mesozoic era, 66 million human-years ago

Everything in computer science that is older than 26 human-years is ... "computer-jurassic" as well as a fossil of the Mesozoic era 


Note
how to calculate log2(x)
log2(x) = log10(x)/log10(2) = ln(x)/ln(2)

[brucehoult]

I think you have it the wrong way about. That would imply computer progress was incredible in the distant past.

More like: there's as much difference between a computer from today and a computer from 2007 as there is between a computer from 2007 and a computer from 1997, or between a computer from 1997 and a computer from 1987, or between a computer from 1987 and a computer from 1977.

This is logarithmic.

[colorado.rob]

Growth in raw computing power over the past decade, save for some niche areas, has been rather stagnant.  The difference in computing efficiency is where things have really changed over the past decade.

[legacy]

Moore's law, the number of transistors in a dense integrated circuit doubles approximately every two years

More transistors, more complexity, more features ... more cores. I was kidding on computer shapes of things older than 26 years: don't they look like fossil reptiles of the Mesozoic era, often reaching an enormous size?

Those called "big iron boxes", they are big size, heavy, and with a little brain. Oh, and a modern cell phone is thousand times faster with billion more ram.

A SGI machine like Onyx eats in order ot Kilowatt, it's big, bloated full heavy metal, and .... it's not able to play doom.

From Lord Crimson to a modern iPhone I see progress like from a trex reptil to a ape able to move his first step into the art of math

Computers are a pure math application which is able to walk on its legs. Next step: the singularity, which we call A.I. as the first computer aware that we, human beings, wanted to collect computers for vintage

I say "wanted" because in a short time A.I. will succeed and overcome our biological limits. I believe they will consider us like ... apes eating bananas, so they will probably want to collect our fossilized bones for vintage.

100 years in the future ... - those 3D models, acquired on fossils bones by a laser scanner and rebuilt on 3D models, were human beings around 0x0000FF312A4B23DE (2054 DC) ... look at their skulls,  there was no plug since they were unable to connect to the Matrix without the use of external devices which they called "smartphones" and "laptops" -

Ohhhh, all machines are now surprised - so primitive, those human beings -

[technix]

Well ... 

So, "vintage" is now "last Thursday"...

What about a Zynq-7000 with dual ARM9 cores and perhaps enough FPGA fabric to implement a Cray-I as the scheduler.

Nope. For me the definition of “vintage” here is still rigidly “released or announced in 1992 or earlier” since I am born Q1 1993. And only the processor core complex counts towards this limitation.

[Bruce Abbott]

For me the definition of “vintage” here is still rigidly “released or announced in 1992 or earlier” since I am born Q1 1993. And only the processor core complex counts towards this limitation.

I got my last new 'vintage' computer - an Amiga 3000 - in 1991. It cost NZ$7200, about the same price as a Compaq 386DX with similar specs. I eventually upgraded it to a 50MHz 68060 and 16 bit graphics card. However by that time Windows 95 and Pentium PCs had arrived, and with the demise of Commodore the Amiga could no longer keep up. I stupidly sold it for a few hundred dollars - now A3000's are truly vintage and worth a mint!

[legacy]

The above theory, computers evolution will be faster since they aren't limited by biological limits, is perfectly shown in movies like: Automata (2014)!



yup, no nice pinups there(1), but ... you have a female sexybot in the casting
a few humanoids robots by Honda, and Antonio Banderas is a good actor

(1)
Melanie Griffith  ...
Birgitte Hjort ... maybe 


if you are looking for a few hours of good entertainment on computers, I recommend this movie 

[brucehoult]

Well ... 

So, "vintage" is now "last Thursday"...

What about a Zynq-7000 with dual ARM9 cores and perhaps enough FPGA fabric to implement a Cray-I as the scheduler.

Nope. For me the definition of “vintage” here is still rigidly “released or announced in 1992 or earlier” since I am born Q1 1993. And only the processor core complex counts towards this limitation.

Wow, recent perspective!

Linux was already out. NeXT (which became OS X and iOS) had been out several years. Windows NT was in public beta and on the verge of release. C++ was already widespread. Perl and Python were already out. Java was in development. The Apple Newton handheld with ARM CPU had been demoed at CES.

Very little in our computer landscape has changed other the speed, capacity and price since 1992. Web sites. Mobiles with radio connectivity to internet. The move from 32 to 64 bit.

I think for me "vintage" will always refer to pre IBM PC and Macintosh.

[colorado.rob]

I think for me "vintage" will always refer to pre IBM PC and Macintosh.

I think it means something more than that.

Vintage: 1. the year or place in which wine, especially wine of high quality, was produced. 2. relating to or denoting wine of high quality.

Not all wine is vintage.  No bottle of Boone's Farm is "vintage".  Same goes for computers.  And, as with the Boone's Farm example, popular does not equate to vintage.

A vintage computer has to be historically significant.  The first IBM PC is vintage, most other PCs are not.  The first Macintosh is vintage, most others are not.  The NeXTcube is certainly vintage.  I'd argue that the first iPhone and HTC Dream are both vintage -- and they're less then 10 years old.

[legacy]

Amiga 3000

Amiga is crazy.

I mean things like a "towerized"(1) A1200 with CyberStorm CPU accelerator, RAM expansion to 128MByte, mediator PCI to plug a PCI-video card, a PCI-sound-card, a SCSI expansion, and a couple PS/2 mouse and keyboard adapters.

At the end it costs something like 2K euro and ... it's not Amiga, it's a strange beast 



The above photo shows a "PowerPC" accelerator that mounts two CPUs, a classic 68k and a PPC.

Does it make sense  ?




(1) they want to put the Amiga A1200's motherboard inside a tower PC-like case, since the motherboard is neither AT-form factor nor ATX, they need to made a custom case, usually adapting a PC case.

Of course there also kit, but they are not cheap.

[Bruce Abbott]

Vintage: 1. the year or place in which wine, especially wine of high quality, was produced. 2. relating to or denoting wine of high quality.

Not all wine is vintage.  No bottle of Boone's Farm is "vintage".  Same goes for computers.  And, as with the Boone's Farm example, popular does not equate to vintage.

The meaning of a word is determined by usage, not the dictionary definition. To gauge what most people consider to be a 'vintage' computer, check out the listings on eBay and websites such as the Vintage Computer Federation and oldcomputers.net.

A vintage computer has to be historically significant.. The first IBM PC is vintage, most other PCs are not.  The first Macintosh is vintage, most others are not.

But pretty much any home computer made in the 80s and 90's is considered 'vintage'. Why? Because that was the time when real innovation was occurring and each machine had its own unique characteristics - good or bad - that made it significant.

Modern PCs will never be 'vintage' because they are just a platform to run mainstream software on, and none has any unique and interesting feature that separates it from the others. this is the real reason that PCs and Macs lose out in the 'vintage' stakes - not due to popularity or quality, but because they are boring - by design. 
 
     

[Bruce Abbott]

Amiga is crazy.

In the 1990's I ran a computer shop that sold Amigas and PCs, and I also helped a friend with his electronic repair business. He had an A1200 which he preferred, but needed a PC for some stuff. We put his A1200 motherboard in a tower case along with a PC motherboard, and connected the two together via serial ports so they could share files and use the same keyboard, mouse and monitor. It worked brilliantly.

Without that he would have needed two separate setups which was impracticable, and I'm guessing the PC would have won out by necessity. 'Towerizing' an A1200 may seem crazy, but is actually a good way to extend its life without losing most of its unique characteristics.

I mean things like a "towerized"(1) A1200 with CyberStorm CPU accelerator, RAM expansion to 128MByte, mediator PCI to plug a PCI-video card, a PCI-sound-card, a SCSI expansion, and a couple PS/2 mouse and keyboard adapters...

Does it make sense  ?

The reason for having all those extras bolted onto an A1200 is to run the latest AmigaOS and new software titles that require it. Does it makes sense? Depends on what you want to do. Certainly it doesn't make sense financially - but that's not what 'vintage' computing is about.

My own A1200 is languishing in the corner right now because I can't be bothered spending that kind of money on it. Yet I have spent thousands on older machines that I didn't have back in the 90s. My current project involves the crappiest home computer ever from that era - and I am having a ball getting it to do stuff that was never imagined.

[eugenenine]

I did the same with an Amiga 500, I had it in the left of a case and a 286 then later a 386 in the right.  I was working on a backplane for the bottom that would have Amiga 2000 slots, the pinouts were the same between the 2000 expansion slots and the one on the side of the 500 but the 500 just had a few less power and ground pins.

[Bruce Abbott]

Other crazy stuff I did:-

Put a ZX81 inside the metal case of an amber-screen monitor. Had a backplane with 4 slots, and a proper keyboard made from individual key switches. Built an interface card to connect the ZX81 to an IBM 5444 removable cartridge hard drive. 2.4MB of storage per 14" disc! The drive weighed over 25kg, and needed a 24V 6A power supply.
 
My Amstrad CPC664 came with 64k of dynamic RAM. Then a few months later the CPC6128 was released, with double the memory.  So I unsoldered all the chips off the board and replaced them with 256k DRAMs. Built a board that bank switched the RAM in 16k blocks (emulating the 6128 banking switching scheme and more). Used a bipolar fuse-link PROM to do the mapping, programmed one bit at a time by flipping dip switches and pressing a button to blow each fuse.

Made a floppy drive expansion board for the Amiga CD32, and installed a 3.5" floppy drive inside the CD32 case. The interface emulated all the essential functions of two 8520 CIA chips using plain TTL logic.   

[dnotq]

I skimmed the thread before posting, but missed if "easier to use" was clarified?  Easy in terms of interfacing the CPU, programming, existing software availability and tools?

The 8-bit CPUs like the Z80, 6502, etc. are probably the easiest to interface due to the limited data and address bus.  There is also a lot of information and books available for building systems based on the classic 8-bit CPUs.

The 8088/8086 are fine as long as you promise to *not* make a PC or XT clone. ;-)  Make a unique system, it will be way more fun.  You have to demux the address / data bus though, so extra glue logic right off the bat just to start interfacing the CPU.  Well documented though, and lots of software tools.

The Z80 is fast and the 6809 is probably the most advanced 8-bit CPU ever.  I find both require getting your head around the architecture to program them with any kind of efficiency.  But once you master their assembly language, they are very powerful.  Interfacing is straight forward, and there are tons of example systems out there for reference, books, info, etc.

The 6502 is nice to program, so I have been told.  Probably decent tools available and used in a lot of classic computers.

The TMS-9900, TMS-9995, and TMS-99105 seem to borrow heavily from the PDP-11 and have a nice assembly language.  There are compilers available, albeit limited.  Those CPUs also have a memory-based register file, which is interesting and good for fast task-switching (if that is important).  Interfacing is straight forward, but they have a strange I/O implementation called the "CRU".  TI seemed to be experimenting with a lot of ideas with that CPU family.

The 68000 is a nice CPU with a nice assembly language, like the TMS-9900 family it is very orthogonal (IIRC).  Certainly there are tools and software.  What is really nice about the 68000 is the large flat address space, with a decent size physical address bus.

[brucehoult]

I skimmed the thread before posting, but missed if "easier to use" was clarified?  Easy in terms of interfacing the CPU, programming, existing software availability and tools?

I agree this was never really clarified.

The Z80 is fast and the 6809 is probably the most advanced 8-bit CPU ever.  I find both require getting your head around the architecture to program them with any kind of efficiency.  But once you master their assembly language, they are very powerful.  Interfacing is straight forward, and there are tons of example systems out there for reference, books, info, etc.

The Z80 was much easier to interface than the 8080, but the 8085 was easier again.

As for the 6809, yes easy to work with electrically and the nicest "8 bit" CPU. In fact, once you get past documentation differences and look at the actual opcodes and their encodings, you find that there is not a lot of difference between the 6809 and the 8088. Both have eight registers, with a (slightly different) mix of 8 and 16 bit. Both have opcodes to select secondary pages with more opcodes. Even the 6809 indexed addressing mode byte is very similar to the 8088 mod/rm byte. Perhaps the biggest difference is that the 6809 allows only a register as the destination of arithmetic, not  the memory operand. (except inc/dec/inv/neg/clr).  The 6809 is a good fit for high level programming languages and an easy target for compilers.

The 6502 is nice to program, so I have been told.  Probably decent tools available and used in a lot of classic computers.

No, it's awful. By far its best feature was that it was ten times cheaper than anything else at the time it was introduced. And it was a little bit pipelined, so it ran as fast as the memory would permit. It's extremely lacking in registers, so the easy way to program it is to treat Zero Page as 256 registers (or 128 16 bit registers, as you'll want quite a few as pointers, at least). In many programs that lets you put all the local variables of your various functions in more or less dedicated "registers" so you don't need to worry about stack. Which is fortunate as the support for a C-style stack is non-existent -- there's 256 bytes for function return addresses, and you can stash a register there temporarily (push/pop) within a function, but that's it. Recursive, reentrant, or position independent code is incredibly painful (all of them are trivial on the 6809). Because you're constantly shuffling things between Zero Page and the actual registers you need a lot of instructions to do anything and the code density is awful by modern standards -- even for 8 bit calculations, but (more than) doubly so for 16 bit calculations. Something like X->a = Y->b + Z->c needs seven instructions and thirteen bytes of code and 26 clock cycles for an 8 bit field (clc; ldy #b; lda ($Y),y; ldy #c; adc ($Z),y; ldy #a; sta ($X),y). Just awful. X->a = Y->a + Z->a is better at five instructions and nine bytes and 22 clock cycles. But not much.

[fourtytwo42]

I am very sorry to go off at a slight tangent OP but when I was a lad "vintage computers" didn't have a processor chip as that didn't exist The earliest machines I was involved with were built using IC gates BUT a friend of mine worked on those built with transistors (no IC's at all). So the modern day appreciation of "vintage computer" often makes me laugh.........sorry..

[brucehoult]

I am very sorry to go off at a slight tangent OP but when I was a lad "vintage computers" didn't have a processor chip as that didn't exist The earliest machines I was involved with were built using IC gates BUT a friend of mine worked on those built with transistors (no IC's at all). So the modern day appreciation of "vintage computer" often makes me laugh.........sorry..

Those aren't something you can buy from a catalogue, obviously, which seemed to be what the OP was asking about.

You could emulate them in an FPGA. And rather a small one. And at 100x their original clock speed.

[technix]

I am very sorry to go off at a slight tangent OP but when I was a lad "vintage computers" didn't have a processor chip as that didn't exist The earliest machines I was involved with were built using IC gates BUT a friend of mine worked on those built with transistors (no IC's at all). So the modern day appreciation of "vintage computer" often makes me laugh.........sorry..

Those aren't something you can buy from a catalogue, obviously, which seemed to be what the OP was asking about.

You could emulate them in an FPGA. And rather a small one. And at 100x their original clock speed.

If I am interested in this type of computers I would try write one in VHDL and put it into a FPGA instead of cooking up a massive circuit board for it. (Hey I am even trying to consolidate the glue logic on those older computers into a few CPLD chips...) Although for a long term project it may be fun (and expensive) to implement a transistor-only computer (probably using a lot of MOSFETs in CMOS configuration instead of the more common BJTs and diodes, just to keep the power draw in check...)

[fourtytwo42]

There is a free project somewhere with an entire spark written in VHDL, and it runs, I tried it many years ago while evaluating different FPGA's There is one here http://vlsicad.eecs.umich.edu/BK/Slots/cache/www.gaisler.com/products/leon2/leon.html

[technix]

Project update.

For this project I am going to stick to 80C88 and CPLD. However I will not build the system with a lot of MS-DOS compatibility (so I can place stuff in more convenient places. And a lot of modern stuff I want to implement really does not play nice with DOS.)

Here are the features I have decided to give the system:

* Processor complex: 80C88 CPU + 8087 FPU
* RAM: AS7C4096 = 512KB SRAM
* ROM: AT28C256 = 32kB EEPROM
* Bus complex: MAX 7000S series CPLD + 82C59 PIC + 82C53 PIT + 82C37 DMA
* Display: HD44780-based character display, mapped to I/O port.
* Input: USB keyboard and mouse.
* Sound: Direct DAC output with DMA (similar to Covox Speech Thing or Disney Sound Source)
* Connectivity: 1x CH375 USB OTG on I/O port, 1x CH376 USB Host (4 ports via USB Hub) on I/O port, 1x W5100 Ethernet in memory map, 1x 16C550 UART on I/O port, 1x 82C51 UART on I/O port.
* Storage: 1x microSD card slot over CH376.