That poll widget is terrible — I can’t tell whether I was allowed to vote in the poll or not, or whether I did vote — I am a subscriber, but the poll is grayed out. I touched it (I’m on an iPad), and it immediately went to the results, and I have no idea whether my vote was registered or ignored. It seems to be indicating I voted for the Intel 8-bits, which was not what I wanted to vote for, but I can’t really tell.
For what it’s worth: IMHO, enough has been written (in incredible detail) elsewhere about the Intel 8008 and derivatives; I’d prefer to see a deep dive on any of the others. The 6809, F8, SC/MP all seem interesting. Z80 a bit less so, just because again it’s a very widely covered/emulated CPU.
I agree, the poll widget is pretty average. From what I can tell the small tick indicates your selection. Substack really needs some usability work... for example, reading old posts and keeping track of what you’ve read, it seems to mark things as read fairly randomly and you can’t filter by read/unread
I also agree re 8080/Z80, and 6502... enough written already. I like the more odd CPU’s, 2650 for example
Hi Chris, Thanks so much for subscribing and sorry you've had a bad experience with the poll - I agree that the widget isn't great. I've just tried it on an iPad and it gave me a tick in a circle, next to the option name, when I voted for one of the options.
I agree that there has been a lot written about the Z80 and the Intel CPUs so one of the others would be an interesting option. If you let me have a preference I'll change your vote manually - otherwise I'll put you down as a 6809 by default! Thanks again.
There was also the RCA-1802 which had many fans at the time.
A particularly long-lasting line started with the Intel 8048 aka MCS-48 which was a microcontroller. It was different from the 8080, and its direct descendant MCS-51 was an architecture still showing up in designs decades later. If you wanted to control a washing machine that was how you did it. It was dirt cheap and ubiquitous, until ARM made a real effort to go low cost with their M-series around 20 years ago. Even then it took a while to win.
Hi! Thanks for bringing up these great CPUs / microcontrollers. The 1802 featured (very briefly) in today's post, and we had a look at the 8048 in a lot more depth a few weeks ago [1]. Hoping to do a write up on the 8051 in a couple of weeks. I really enjoyed writing about the 8048. As you say the 8051 has lasted for decades, possibly as all the IP protections have expired now I think.
Fantastic site! Just as a suggestion for a future article, it would be nice to know why Nintendo chose to implement a CPU in its Gameboy that is a hybrid of the 8080 and the Z80. Why the extra effort instead of just sticking with the Z80? How they managed to come with it with Sharp, licensing wise?
Hi Alvaro, Thanks so much for your kind words. That's a very interesting question on the Gameboy CPU. I think Nintendo used a modified 6502 with no BCD in the NES to avoid the only 6502 patent.
I got the Sony msx hb201 with the z80 in 1986, I stopped by the Sony service center and they just gave me the service manual with full schematics and a year later I built my own z80 based single board computer using an eeprom programmed with an open source eeprom programmer cartridge on the Sony
It's a great design. I think the Acorn / BBC Micro designers reckoned that a 4MHz 6502 was as fast as any of the new 16 bit CPUs coming onto the market in the early eighties.
At least your 255 didn't go down to 0 in Zelda when you added one!
I am a part of the NetBSD project, and NetBSD is well known for its ability to support a wide array of old and new hardware architectures. Although, I've never used or programmed anything outside of modern Intel x86, but I've hung out with many developers who still run and maintain some of these hardware. I think NetBSD still runs on the Motorola 68k, so it would be interesting to learn more about it.
Fantastic that you're part of NetBSD. That's a really great project.
The 6800 was a few years before the 68000 and, of course, much, much less powerful. I don't think they had much in common other than the name!
There have been attempts to put Unix like OS's on the 6800 but they're all limited by the obvious restrictions in memory (to 68k) and the lack of any memory protection.
It is insufficient to characterize the 6809 as a "slightly better 6800". It implemented two under-appreciated game-changers: separate user-mode and system-mode stacks, and position-independent code.
Hi Frank, Thanks for commenting. It wasn't my intention to imply the 6809 was a slightly better 6800 - as you say there were lots of important upgrades. I'll be writing more about this in an upcoming post on the 6800 and 6809. Best wishes.
That poll widget is terrible — I can’t tell whether I was allowed to vote in the poll or not, or whether I did vote — I am a subscriber, but the poll is grayed out. I touched it (I’m on an iPad), and it immediately went to the results, and I have no idea whether my vote was registered or ignored. It seems to be indicating I voted for the Intel 8-bits, which was not what I wanted to vote for, but I can’t really tell.
For what it’s worth: IMHO, enough has been written (in incredible detail) elsewhere about the Intel 8008 and derivatives; I’d prefer to see a deep dive on any of the others. The 6809, F8, SC/MP all seem interesting. Z80 a bit less so, just because again it’s a very widely covered/emulated CPU.
I agree, the poll widget is pretty average. From what I can tell the small tick indicates your selection. Substack really needs some usability work... for example, reading old posts and keeping track of what you’ve read, it seems to mark things as read fairly randomly and you can’t filter by read/unread
I also agree re 8080/Z80, and 6502... enough written already. I like the more odd CPU’s, 2650 for example
Hi Paul, Hmmm, the Signetics 2650 - now that would be an interesting option for a post .....
Hi Chris, Thanks so much for subscribing and sorry you've had a bad experience with the poll - I agree that the widget isn't great. I've just tried it on an iPad and it gave me a tick in a circle, next to the option name, when I voted for one of the options.
I agree that there has been a lot written about the Z80 and the Intel CPUs so one of the others would be an interesting option. If you let me have a preference I'll change your vote manually - otherwise I'll put you down as a 6809 by default! Thanks again.
There was also the RCA-1802 which had many fans at the time.
A particularly long-lasting line started with the Intel 8048 aka MCS-48 which was a microcontroller. It was different from the 8080, and its direct descendant MCS-51 was an architecture still showing up in designs decades later. If you wanted to control a washing machine that was how you did it. It was dirt cheap and ubiquitous, until ARM made a real effort to go low cost with their M-series around 20 years ago. Even then it took a while to win.
Hi! Thanks for bringing up these great CPUs / microcontrollers. The 1802 featured (very briefly) in today's post, and we had a look at the 8048 in a lot more depth a few weeks ago [1]. Hoping to do a write up on the 8051 in a couple of weeks. I really enjoyed writing about the 8048. As you say the 8051 has lasted for decades, possibly as all the IP protections have expired now I think.
[1] https://thechipletter.substack.com/p/intels-inside
Fantastic site! Just as a suggestion for a future article, it would be nice to know why Nintendo chose to implement a CPU in its Gameboy that is a hybrid of the 8080 and the Z80. Why the extra effort instead of just sticking with the Z80? How they managed to come with it with Sharp, licensing wise?
Hi Alvaro, Thanks so much for your kind words. That's a very interesting question on the Gameboy CPU. I think Nintendo used a modified 6502 with no BCD in the NES to avoid the only 6502 patent.
Sad to see the RCA 1802 of Elf II fame missing from the list
Sorry about that! If it's any compensation a post will feature another '1802' chip very soon :)
1985 was a simpler time :-)
I got the Sony msx hb201 with the z80 in 1986, I stopped by the Sony service center and they just gave me the service manual with full schematics and a year later I built my own z80 based single board computer using an eeprom programmed with an open source eeprom programmer cartridge on the Sony
I had to look it up http://www.msxarchive.nl/pub/msx/mirrors/hanso/service_manuals/sonyhb101sm.pdf
it details every single 74xx series chip with pin-outs and TTL logic inside. crazy right ?
That's amazing. My sense is that documentation was on another level at this time.
Maybe the craziest part .. I was a 12 year old walking into the service center and they were just like .. here you go buddy, have fun with that :-)
The Z80 was also big in at least the GDR (East Germany), not sure how many were exported to other East Bloc countries.
The GDR produced an unlicensed clone, the U880: https://en.wikipedia.org/wiki/U880
It was the CPU for GDR-made computers like the KC 85 (https://en.wikipedia.org/wiki/KC_85) or the (office PC) PC1715 (https://en.wikipedia.org/wiki/PC_1715).
+1 for the 6502, which was used in the NES. I still remember being limited to 255 rupees in Zelda.
It's a great design. I think the Acorn / BBC Micro designers reckoned that a 4MHz 6502 was as fast as any of the new 16 bit CPUs coming onto the market in the early eighties.
At least your 255 didn't go down to 0 in Zelda when you added one!
I am a part of the NetBSD project, and NetBSD is well known for its ability to support a wide array of old and new hardware architectures. Although, I've never used or programmed anything outside of modern Intel x86, but I've hung out with many developers who still run and maintain some of these hardware. I think NetBSD still runs on the Motorola 68k, so it would be interesting to learn more about it.
Fantastic that you're part of NetBSD. That's a really great project.
The 6800 was a few years before the 68000 and, of course, much, much less powerful. I don't think they had much in common other than the name!
There have been attempts to put Unix like OS's on the 6800 but they're all limited by the obvious restrictions in memory (to 68k) and the lack of any memory protection.
Yes, I can imagine. Even the 68k support is barely working with modern software getting more and more demanding.
3E 80
C9
It is insufficient to characterize the 6809 as a "slightly better 6800". It implemented two under-appreciated game-changers: separate user-mode and system-mode stacks, and position-independent code.
Hi Frank, Thanks for commenting. It wasn't my intention to imply the 6809 was a slightly better 6800 - as you say there were lots of important upgrades. I'll be writing more about this in an upcoming post on the 6800 and 6809. Best wishes.