Best 1970s Microprocessor

[hgm]

The matchbox contained a 3-digit 7-segment LED display taken out of a pocket calculator. Only two digits were used, alternating the from- and to-square of the move. It was driven by an 8-bit latch that served as output port: 7 lines for the segments, and one to select the 1st or 2nd digit. The moves were entered by a touch contact made from two staples. The amplified touch current was connected to the interrupt input of the CPU, which was timing the delay between interrupts. By tapping the contact you could step through the alphabet, and pause when you reached the desired symbol (which was displayed). A long press set it thinking.

Later I added a piezo disc for beeping, connected to the highest-order address line, which was not needed for addressing the RAM. (Which was only 8KB, an connected to the 12 lowest address lines, ignoring the other 4, so that it appeared in the 64KB memory space 8 times.) By making the program jump between the low and high half of the memory space in the right frequency the piezo would beep.

The program was in the RAM. This was a CMOS RAM, kept under power by a watch battery. The program got loaded into it by driving the 6502 clock and reset line from the output port of a 'real' computer, as well as the data bus (through an IC socket in the matchbox, into which you could plug a flatcable). The 6502 was manipulated this way to generate memory addresses, and when the proper addres appeared the data that needed to be at that address was written through the cable in that memory location before the 6502 was made to complete the cycle by reading that same data back.

[towforce]

The matchbox contained a 3-digit 7-segment LED display taken out of a pocket calculator. Only two digits were used, alternating the from- and to-square of the move. It was driven by an 8-bit latch that served as output port: 7 lines for the segments, and one to select the 1st or 2nd digit. The moves were entered by a touch contact made from two staples. The amplified touch current was connected to the interrupt input of the CPU, which was timing the delay between interrupts. By tapping the contact you could step through the alphabet, and pause when you reached the desired symbol (which was displayed). A long press set it thinking.

Later I added a piezo disc for beeping, connected to the highest-order address line, which was not needed for addressing the RAM. (Which was only 8KB, an connected to the 12 lowest address lines, ignoring the other 4, so that it appeared in the 64KB memory space 8 times.) By making the program jump between the low and high half of the memory space in the right frequency the piezo would beep.

The program was in the RAM. This was a CMOS RAM, kept under power by a watch battery. The program got loaded into it by driving the 6502 clock and reset line from the output port of a 'real' computer, as well as the data bus (through an IC socket in the matchbox, into which you could plug a flatcable). The 6502 was manipulated this way to generate memory addresses, and when the proper addres appeared the data that needed to be at that address was written through the cable in that memory location before the 6502 was made to complete the cycle by reading that same data back.

Coming up with that idea and patiently getting it to work was an impressive feat. At the time, it would have undoubtedly have been the smallest chess computer ever built!

[hgm]

It was partly born from necessity: I was inviteded to participate in an anniversary edition of the CSVN Dutch championship at a time I had already given up chess programming, while I did no longer have the computer that could run my old chess program. (I had switched to 6809 for my other computing needs.) So I decided to build one, and chip technology had advanced so much that what originally had required two 10" x 10" PCBs (an AIM'65 plus memory card) would just require a hand full of chips. First I wanted to build it inside the box for a compact cassette tape, including 4 AA batteries to power it. But then I realized that more than half of the volume was batteries. So I decided to cheat, and use a DC power adapter that would be plugged into the wall socket, so no one would see it. The remaining components then could be made to fit into a smallest-size matchbox. Although I had to saw off part of the large 40-pin CPU chip to make that fit. (There were only unused address pins on that end anyway.)

[Dann Corbit]

I can't believe nobody mentioned the Motorola 68000 (1979).
Like the Amiga, it was miles above the competition.

Honorable mention goes to the 6510, because it was the heart of the Commodore 64.
The 8086 + 8087 barely missed (1980).

[towforce]

I can't believe nobody mentioned the Motorola 68000 (1979). Like the Amiga, it was miles above the competition.

Looks good - link. To be honest, I wasn't aware of it - not sure why.

Honorable mention goes to the 6510, because it was the heart of the Commodore 64.

I clearly remember how impressed I was that an affordable home computer had 64 Kb of RAM! Different times.

btw - back in the day, they weren't called "8-bit microprocessors" - they were just "microprocessors" (just as today we call personal rolling vehicles with petrol-burning engines "cars" - but they might be known as ICEVs in 20 years or so).

[JoAnnP38]

I've programmed in assembly language on them all and I prefer the Z80 because it seemed like the information was a little more abundant at that time. My favorite 8 bit processor from around that era was the Motorola 6809. It's ability to do mixed 8/bit - 16/bit addressing & operations give it a little more power. The OS OS-9 was also a joy to use compared to CP/M. Oh the memories.

[Dann Corbit]

I can't believe nobody mentioned the Motorola 68000 (1979). Like the Amiga, it was miles above the competition.

Looks good - link. To be honest, I wasn't aware of it - not sure why.

Honorable mention goes to the 6510, because it was the heart of the Commodore 64.

I clearly remember how impressed I was that an affordable home computer had 64 Kb of RAM! Different times.

btw - back in the day, they weren't called "8-bit microprocessors" - they were just "microprocessors" (just as today we call personal rolling vehicles with petrol-burning engines "cars" - but they might be known as ICEVs in 20 years or so).

The 68000 was absurdly better than the 8088 chosen for the IBM PC.
The 8088 did not even have unique addresses because you had to page. So the only way to know if two addresses were the same was to multiply.
Cringe, cringe, cringe -- why oh why did IBM choose that horrid thing?

[Dann Corbit]

I've programmed in assembly language on them all and I prefer the Z80 because it seemed like the information was a little more abundant at that time. My favorite 8 bit processor from around that era was the Motorola 6809. It's ability to do mixed 8/bit - 16/bit addressing & operations give it a little more power. The OS OS-9 was also a joy to use compared to CP/M. Oh the memories.

The best assembly set ever was the VAX.
They had an instruction for everything you can imagine and some that you can't.

[Bo Persson]

Cringe, cringe, cringe -- why oh why did IBM choose that horrid thing?

I'm sure you know, and that the question was rhetorical. But anyway...

The 68000 used a 64 pin package, while the x86 package only had 40 pins. So lower cost.

IBM even used the el-cheapo 8088 with an 8-bit memory bus, because that made the motherboard narrower. And they could then get away with only 8 memory chips, instead of 16.

They even chose the clock frequency of 4.77 MHz, as that could share the reference crystal with the US TV-frequency for the video board. It could have run at 5 MHz, but that would have cost several dollars extra!

[Ras]

Cringe, cringe, cringe -- why oh why did IBM choose that horrid thing?

Because the 68k was launched, but not ready yet, and IBM already had peripherals around the 8085, so they were familiar with Intel chips.