Is his last name pronounced fah-Jeen? There's an vintage Italian bicycle frame maker from my youth with that name. Besides being a bicycle nerd, I got to hack on Z80s on several platforms, mainly the Kaypro II under CP/M using Microsoft's M80 assembler.
We hobbyists won't let the ol' Z80 die that easily. I'm working on a Z80-based computer with 512K of paged RAM, full bells and whistles like Ethernet, MIDI, OPL2, SD card, 800x480 LCD, serial, PS/2 and more to come. It's just executing NOPs at the moment while I am waiting for the Flash programmer to show up, and will not stop until CP/M and Zork are running! :)
@@RetroBytesUK Sure! It might be a month or two until things are ready, but it's also a good thing since I've found and fixed two bugs on my PCBs (and will probably find more). I'll be more than happy to release everything on GitHub!
I must say that I have come to the conclusion that your channel is my top retro tech YT channels. Your topics, commentary, detail and humor are spot on. I find myself rewatching your videos more and more other as I pick up on details I missed the first time.
It's quite crazy how so many retro videos feature the Zilogs heavy hitters but the company itself is like ignored completely. Thank you for making this!
Nice overview. I used the Z80 from 1979 until about 1987. By then you could get 16 bit memory for the S-100 bus, and 24 bit addressing. I used a homemade bank selector to access the extended memory, and had a 256kb card for the machine. When the Z280 came out, I moved to that. It was a 16 bit Z80 compatible CPU with a highly extended instruction set, and I made (and sold) an adapter that plugged that into a standard Z80 socket. The Z280 also had a memory pager built in, so that addressing 256kb in 4kb pages was possible. Turned out the Z280 was buggy, and I ended up working for Zilog to try to fix some of the bugs. We spun the chip with a series of fixes, but it turned out the fundamental bugs were masking bigger and more serious bugs in the chip, and the CEO cancelled the chip. It now lives on as the EZ80, but I learned my lesson and headed on to the 80386. What you are referring to as the Z800 became the Z280. The original design had a lot of problems, so you can't really say they ignored it as much as it had fundamental design issues and they canceled the project. Then it came back as the Z280, and got cancelled again, then came back yet again as the EZ80. However the EZ80 was not compatible with the Z280, so there you go. You can still get the Z80 in the form of more advanced chips, and the "death of the Z80" headline was somewhat inaccurate. Yes, they discontinued an obsolete form of the chip, but there are many other instruction compatible versions. I think the most sensible design is to use one of several FPGA versions of the Z80 instruction set. Then you can use a relatively cheap FPGA with a Z80, several peripherals, and a chip with 100's of I/O's to boot.
@@christopheroliver148 4kb page lookup. So the top 4 bits selected one of 16 pages in the lookup, and each of those pages had a 12 bit page address, so 12 bits page+12 bits page address = 24 bit total address.
@@christopheroliver148 via a 16 page lookup, indexed by the top 4 bits of the address. Each page had a 12 bit page address, so 12 bit page and 12 bit page address = 24 bits. Do a web search and you can still find the manual for it if interested.
I still can remember watching my brother on the family summer vacation writing down z80 assembly by hand and then writing down the binary in hex and finally "linking’ it by counting code length.. at age 12-13! Back home it was tested on an zx81..
Assembler instructions on paper, then hand translated to hex machine code, then entered in data statements in Basic. This was the way I wrote lots of Z80 assembler for MSX.
The only think I've done comparable was hand disassembling the ROM from a Kaypro II that along with the video page was a bank swap with the lowest addressed 4K. That little spiral bound reference for Z80 machine code got rather worn.
Z-80 + CP/M will always have a place in my heart. I loved my Kaypro II Z80 at 2.5Mhz 64K of RAM and two 191K Single Sided Double Density full height 5 1/4 inch disk drives, CP/M 2.2. Oh, those were the days. Thanks for making this video.
I used a Sanyo MBC-1000 with a Z80A and CP/M well into the DOS era. Didn’t see a reason to change until lured by a deal for a PC with a 20 MB hard drive.
New RetroBytes video? This just made my weekend much more relaxing and I get to learn more about a CPU I knew little about growing up, but had used without knowing.
Love your videos. I hope you keep making them for a long while. I started out with in the C64/Amiga/286 era but didn't really get into things until the 286 so it is nice to see some of this older stuff in detail.
I find it interesting that at that time, all the CPU manufacturers thought creating new products meant new architectures, so the Z800 and Z8000 were completely new designs nothing like the Z80. It turned out that the way forward was evolutionary with extending present architectures, as with the x86 line. Code compatibility and developer familiarity were way more important that radical new architectures (as Intel themselves discovered with iAPX432 and Itanium).
@@ArneChristianRosenfeldt We can't blame them really. The technology was developing very fast and nobody thought sticking to a simple architecture was the way forward, the designers all wanted to create new chips that were more advanced architectures, a "mainframe on a chip" etc. We have the benefits of hindsight that they didn't have.
@@ian_b I read that Zilog did not want to use microcode. As we know (hindsight) this means that we have to use RISC and microcode come back through the back door as cache logic / pipeline stalls/ instruction queue. RISC as a chip is complex. Zilog could have employed perfectly orthogonal addressing mode and instruction or even load-store. Though what I don’t get: reg-reg 3 stage pipeline is so simple. ARM uses this. Why did Zilog not emphasis reg-reg instructions? Intel 8008 had a hardware stack. Maybe push and pop so could effectively be reg-reg instructions. Overflow would call a trap to RISC code which swaps out values to a larger stack in RAM. All known at the time.
@@ArneChristianRosenfeldt Not sure I understand all your points. The Z80 stack is very flexible (I write for PIC MCUs these days and the small hardware stacks are a pain). They could indeed have made the Z80 more orthogonal or done many other things but having an accumulator was quite normal at the time and the 16 bit semi-accumulator hl was nice. Also pipelines were a long way in the future. They were I think focussed on creating an 8080 superset and did a very good job of that. If I had a hindsight criticism I would say they should have pushed for higher speed earlier. I still love the Z80 instruction set and sometimes write 8 bit PIC code as pseudo-Z80 with file registers called a, h, l etc. 😁
Side-note: like the 4004, the 8008 was also a contracted design. This time for the Datapoint minicomputer company. They wanted a single-chip implementation of their 2200 minicomputer, and asked both Intel and Texas Instruments to have a shot at it, but turned down both designs for being too slow compared to what they'd already built out of discrete chips. TI couldn't find a buyer for their implementation, but Intel managed to sell the 8008 as a general-purpose microprocessor. The 8080 was derived from it based on customer feedback, and the 8086 as a 16-bit design derived from that.
I still, somewhere, have my Captain Zilog t-shirt from when I took their Z-8000 course. Great architecture, shame the Z-8001 had such problems in the first batch. And the Z-80 was such a pretty chip. As an example, even 45 years later I still remember the LDIR instruction (LoaD Increment Repeat - EDB0) and the registers you needed to load before calling it. Having to do that in a loop on processors that came out years later still seems primitive.
@@lovemadeinjapanI don't think so. No matter what kind of trickery you do with the stack, you would still have to run a loop with individual instructions for each byte. LDIR is the fastest way to move data on the Z80. Only reason some people don't use it is for compatibility with the 8080.
They probably could've made the Z8000 work if they did the Sega Mega Drive/Genesis approach (~10 years later) and integrate the older part as a subcomponent of the new system, for example the Z80 could've done memory management or IO handling for the Z8000 when the Z8000 was running, or it could've run in legacy mode and run Z80 code directly. It could have been as simple as switching from real mode to protected mode on a 286, or even just use interrupts to pass off instructions to the Z80, or it could have been like switching endianness on a PPC in that you'd need a reboot to switch between environments. Either way it would have been a useful stopgap to wean people off the earlier ISA onto the new one. As was the fate of so many architectures, though, it probably still wouldn't have saved it without a killer app for the Z8000. If only Zilog could have ditched Exxon and gotten the IBM PC deal, they could've made MSDOS a Z8000 exclusive.
It's a neat little CPU! The first time I learned about the Z80 microprocessor was my father talking about it in the ColecoVision we had back when I was little.
IMHO, their mistake is not making Z80 based microcontroller early, like competing intel 8051 controller. 8051 is not code compatible with massively popular 8085. Considering popularity of Z80 then, such microcontroller would be instant hit.
In 2001, they did just that: The Zilog eZ80 is an 8-bit microprocessor designed by Zilog as an updated version of the company's first product, the highly-successful Zilog Z80. The eZ80 is binary compatible with the Z80, but operating almost three times faster at the same clock frequency.
I love this background style... lots of slow "beauty pans" over some of our favourite old machines... with that nice wooden table for contrast. Very nice, very gentle. So much better than the rather hectic computer graphics that I think I've moaned about in the past.
Love all your videos, I often take myself to a cafe at Meadowhall and watch your videos while drinking my coffee and eating my eggs Benedict. Thank you another brilliant video.
29:40 Cue the infamous, possibly apocryphal story about Gary Kildall (the boss of CP/M) bailing on meeting IBM rep's because he wanted to fly his plane that day, & IBM hitting up Microsoft for an OS instead of CP/M, & ending up with MSDOS instead.
This video will definitely help me in my research on the Z80. I, too, hope to create my own homebrew Z80-ish based computer someday. Perhaps the eZ80? This video is saved in my "watch later" list!
Every time you say burglar alarm I hear burger alarm. Great video as always I just finished up watching one of your older videos and I was like I wonder if there's anything new and this is the newest one I hadn't watched
Sure was a lot of mentions of the intel 4004 being used in a single chip calculator... Only problem being it doesn't really work without the 4001, 4002, and 4003 support chips. If I recall, there were nine chips in the final product.
Early calculators all ran on chips like the Hitachi HD38800 and the NEC µPD7502, 4 bit microcontrollers with built in RAM, ROM and display driving circuits for controlling VFD tubes.
I still like the idea to have ALU, registers and microcode on one chip. Clock generator only needs a few connectors -> different chip . For the system bus a complicated driver is needed anyway.
@@lovemadeinjapan The HD38800 was introduced in 1978, a long time after the Intel 4004. Many other chips were used in calculators before 1978, including many single-chip designs, although the 4004 was obviously not a single chip design.
@@GodmanchesterGoblin The TMS 1000 is one of the earlier all-in-one designs, from 74. The nice thing of the NEC and Hitachi was the ability to drive 30V for the VFD tubes without external circuits.
While the biggest slice of the pie in the retro computing days goes to the z80 and 6502, the 6809 also had a look-in too... The z80 is also used in loads of early digitally controlled synths too (as was to a lesser degree the 6809, but never the 6502, weirdly)
I haven't watched the episode *yet* - but when I heard a few weeks ago that the Z80 was about to be discontinued the question what the content of your next video would be was a no-brainer 😉
A fascinating history indeed! Different architectures, different ways to achieve the same essential result, continually evolving, building on pillars of previous knowledge into the devices of today. ( sorry that was a bit poetic, I know) Thank you for sharing! :)
@@legiran9564 I wonder why consoles left Power . IBM sold a lot of fabs. Do they still have critical mass? I thought diversification was a good thing in the past? Higher margins for main frames? Vendor lock-in? That would be evil . Big blue as it was before Microsoft and Intel came along. Phillips and Siemens concentrate on medical because sick people will pay any price. But now they go abroad for medical services.
@@carguyuk7525 the first computer I ever touched was our school's cassette 380Z... hand assembling Z80 code with a pencil and paper. We later got another, running CP/M and that had an actual Z80 assembler. Not as much of a delight as the BBC micro's assembler (the CP/M text editor was quite hard work) but lovely all the same.
My first experience of computers was at school in 1978, computer studies - the classroom had a commodore pet 16k - 6502 processor I think and four challenger IP-8 with 8k and needed a monitor screen - any portable tv would do and they were! We were learning flow charts and BASIC programming - eeh a long time ago now!
1:30 and 12:40 Loved the shot of the Amstrad CPC 6128plus! A machine if it came a few years earlier would have been great. (Working on some hardware stuff for it atm.)
The ATR8000 for the Atari 8-bit was an add on that happened to include a Z80 and could run CP/M. These days, the Fujinet has (or had at one point) a virtual CP/M you can connect to. (Full disclosure: I've never had an ATR8000, nor have I played with CP/M. I did start on a Timex Sinclair 1000 (basically, US version of ZX81), though soon after got my beloved Atari 1200XL.)
One mention of either the Z80 or the 6502 and for some reason I'm reminded of the huge ad for Computer Adventure World above the Wirral Line escalators at Liverpool Central Station. I seem to remember it was still there long after the shop had closed.
31:30 - we did have Unix for home computers in 1985, SCO Xenix. Just no home user was going to pay for it, or set it up - Concurrent had the same issue. MS had planned in '81 to evolve MSDOS into single user Xenix, which they developed on the VAX, and sell Xenix for multitasking. When AT&T started selling SystemV (making MS think they couldnt compete with an older Unix of their own) and MS made an with IBM agreement to develop OS/2, all the devs were put on OS/2 and Xenix was sold off via a sort of merger to SCO, who finished the port in ~84.
The Commodore 900 was not going to be a home system. It was squarely aimed at re-entering the business machine market that CBM, naturally, had started out in, experiencing success with the PET, but had subsequently abandoned or deprioritised by focusing on home computers. The workstation model probably wouldn't have done very well against contemporary workstations based on the 68000 and other processors with 32-bit architectures. Faggin admitted that the Z8000, with its 16-bit architecture, was unsuited to dealing with the increasing demands of image processing and other workstation-related tasks. Like all low-cost workstations, by the time they would have been launched, the price wouldn't have been substantially lower than the competition. Commodore experienced that later on with the Amiga 3000UX. One can argue whether the server or minicomputer model might have done slightly better as a lower-cost replacement for something like a PDP-11, although it would have been competing with systems based on the 68000 and other 32-bit processors, running genuine Unix, these aiming to nibble at the lower parts of the VAX market. It would probably have found itself boxed into a shrinking area of the market, with potential customers paying a decreasing premium to get those 32-bit systems instead.
Good job, but I was into Motorola and their 6809 microprocessor. Just built a 6809 emulator based on an Arduino Due. Brings back a lot of good memories. :)
@scottlarson1548 The problem with these kinds of discussions is that the comparisons typically involve overly simple non-optimized operations that don't really reflect real-world scenarios. For example in a single shift operation a 1mhz 6502 can potentially complete the task faster than a 4.7mhz 8088 however when running a series of shift operations, as would happen in real software, the 8088 is actually about 2.6 times faster which means it outperforms even a 2mhz 6502. When running code optimized for each cpu a 2 mhz 6502 is roughly equal to a 4mhz Z80, however, for complex operations the Z80 actually pulls slightly ahead. That's part of why vector and 3D games were so much better on the PC and even the spectrum than the C64. Games of the Apple II are a better indication of the 6502s real performance as it lacks the custom chips of the c64.
@@dgmt1 The problem with your discussion is that I was talking about a 6502 running at *2 MHz* not 1 MHz. You're comparing a 4 MHz Z80 with a C64 and an Apple II both running at 1.023 MHz, half the speed I'm talking about. It doesn't sound like you're familiar with the internals of these two machines.
It seams to have been a thing over the years ,of people thinking that comparing clockspeeds across different CPU designs has any great meaning. In 90s I remember a few press release from Apple complaining about people using clock speed to compare PPC and x86
@@RetroBytesUKthat remains an issue to this day. In the 90s, we got a lot of CPUs with a number in the name that wouldn't reflect the actual clock, but the comparable clock a Pentium would run at to get the same performance. In the 2000s, that continued with the naming of the AMD Athlon series of CPUs.
After listening to the bit about traps I remembered Ordinance Survey did the same thing with their maps. They would features on the maps that did not exist in real life to catch out companies copying their maps. Someone told me the AA got caught copying by this method.
The original Zilog plan was to start out with the Z8 (a microcontroller), then the Z80 general purpose processor and go from there. None were planned to be compatible with the others. They were forced to change the order and release the Z80 first, but the Z8 eventually did come out in 1979. Note that without the MMU the Z8000 could only address 64KB compared to 1MB for the 8086/8088 and 16MB for the 68000. It was the 1986 Z80000 that was Zilog's first 32 bit processor.
Ralph Ungermann went on to found Ungermanm-Bass, a network hardware and software maker. UB's Ethernet cards and XNS software were in general use at Microsoft when I was hired into the LAN Manager development group in 1990. They were also the first makers of twisted pair Ethernet interfaces I saw, circa 1992.
Back many years ago I built a Z-80 based system from scratch. It had a Z-80B running at the full 6MHz, 64K of 150nSec static ram, 32K ROM, 24 bits of parallel I/O, and a Z8530 SCC that supported 4 channels of serial I/O. That was a great little machine in its day.
SPOILER: The 4004 was very much NOT a "single chip" solution for calculators! It was certainly more integrated than previous calculator solutions, but it required a lot of support chips to make a calculator! Eventually, single chip calculators were developed in the 80s, but they were highly integrated solutions that were the 80s equivalent of modern SOCs.
I would love to find an old trainer or better build one like this I think. I don't learn from books but hands on following along with good documentation. I don't know if they teach that way but at the time I was in school computers were in their boring phase at least from a learning perspective. You might get a typing or a basic programming class but that was about it. How to use but nothing explaining how anything really worked.
If you hit the Computerphile YT channel, where they have explanations from CompSci professors, the majority of them had a ZX-80 as their first computer. Multiple of them talked about how they got to know the entire machine, hardware and software, and that contributed to their interest that led to them becoming PhDs and professors.
Those of use who got to cut our teeth on z80 and 6502 based systems I think had it really lucky. It can just become cheap enough that normal people could try it out, and the systems where not so complex that you could not learn the whole system .
I got a timex sinclair 1000 a while back basicly new in the nice sinclar foam clam box... thing. I guess its time to give it a go. (Amazing what you can find in the Free section on crageslist)
@ 31:18 - Coherent was absolutely superb: I used it on PCs from the day it was released for x86 to the bitter end. Never knew it was going to be on Commodore - I'd have bought one instantly if it had appeared. Topic for another video, perhaps?
My first contact with the Z80 was the NASCOM 1 in 1978. I got to like the Z80 although I did also use 6502 based Apple 2 at work. For some reason I preferred the Z80 over the 6502, but that may be because I encountered it first. The first ever microprocessor I encountered was the Nat Semi INS8060 SC/MP development system with a whole 256 bytes of RAM.
Nice piece of history. The short clips of stock footage on endless repeat are distracting, though. Some shots of logic block diagrams would look better than a fake stock market clip played dozens of times.
In the early 2000s i was working at a computer refurbishing/retailer and would often see z80 driving the OSD on computer monitors and TV PCBs. My first computer was a sinclair clone so I knew the chip label and was surprised of how many I've seen labelled as zilog insteand of the japanese equivalents from NEC and Hitachi.
how come they couldnt see the backward compatibility being important, i believe ic z800/8000 were b.compatible, zilog would be still around and doing well....
It’s a little crazy how much similarity there is between the z80 and the 6502 Both effectively copies of an existing design, but better and cheaper. And both basically the only chip the companies that made them found success with.
I think that latter part makes a lot of sense. Since they tried to make successor designs by themselves, rather than following the methodology that led to their initial success... For instance, if the Z800 had copied/taken inspiration from the 8086 and the Z8000 a Pentium...
Lets see. A zx82, a Spectra video 328 and a 738. Both running cp/m. Did the Epson px-,8 run on z80? Also had dome other msx machines. Yups z80 ead a big part ofy early computing 😊
My first computer was a Nascom 1 with 1K of video RAM and 1K of user RAM, leaving 960 bytes that the monitor ROM didn't use. This was clearly not enough for BASIC or even an assembler, so I did my coding directly in Z80 hex, looking up each code in the fat Mostek book, gradually getting quicker as my brain cached the frequently used instructions. Then I started learning the instruction group patterns and became a lot quicker and eventually memorized them all! I can still code Z80 now without looking up the instruction codes!
IBM did ship with CP/M. They shipped with three OS choices. PC-DOS (Microsoft MS-DOS), CP/M-86 (Digital Research) and UCSD p-System (University of California at San Diego). The problem was that CP/M-86 was late and missed the launch date. It was also vastly more expensive. It also wasn't backward compatible with CP/M for the Z-80 and 8080. While the instructions were similar, it still meant tweaking and recompiling existing software. As Zilog found out with the Z800, almost compatible isn't compatible enough and if customers have to start from scratch anyway... When the IBM 5150 Personal Computer shipped you had the choice of p-System (really only useful for cross-platform academic projects written in UCSD Pascal) or paying $240 for the promise of CP/M-86 in a few months if it actually did ship (Digital Research was already very late with it), or PC-DOS which was available at ship time for $40. As a result of these problems, 96% of PC orders included PC-DOS, 3% included CP/M-86 and less than 1% included p-System.
Its that odd case of they shipped cp/m 86 with the PC, but simultaneously basically did'nt for all practical intentions. I get why Kildal was upset, I think he had every right to be, especially when it turned out DOS contained some of his code.
@@borisgalos6967 That's not what the court found. The code MS purchased that formed the basis of MS DOS, did contain sections of code lifted from CP/M. MS at the time put forward the argument they where not to know and purchased the code in good faith, no one could really prove that statement either way, so the court accepted that argument, but none the less the first version of MS shipped with some CP/M code. So MS paid out.
Even though I know the history of Zilog and their processors almost as well as my own what I want to see is the RetroBytes take on it. Cue 20's music and a PCB way advert (both most welcome by the way- at last a quality sponsor not offering to look after my balls just maybe my reball)
I loved the Z80 and 6502....how about a video on the ESP32? (I have a webcam and home automation that uses this chip)...yeah, it's not a CPU, it's a MC, but I think there's a fine line between the silicon?
Since we are taliking about that "era" (yeah I was there.. I AM 65) and the 8 bit CPUs available there were 2 more popular archetectures. The 8048/51 that was used in most keyboards and (if you were good) embedded systems. The 6812/68xx series used mostly for embedded systems.
Part of the fun of having a retro-style micro is, well, building and maintaining it. At least, in my opinion. But, I can understand why some would see it as a "why bother" time sink.
Whlie the 6800 was popular, the Z80 was my go-to. Used a TI machine with an awful keyboard. Years later, as our Unix minicomputer used z80s as the serial line processor, we taped up a Timex Sinclar 1000 next to the console.
While listening to this, I'm writing 6502 assembler code for an operating system I'm creating 🙂 Intended for a computer with both a 6502 and a z80 (it also runs CP/M), original 1980s era hardware. My new OS also runs on hardware which was available in 1986.
@@RetroBytesUK Commodore 128. There is no public git repository for now, eventually there will be one, but first want to have enough of the core functionality in a somewhat usable state.
It’d be kind of funny to put the Video Beast into a Commander X16. I’m guessing it’d be pretty easy with X16’s expansion architecture. It’d be driving a different monitor than the X16’s built in Vera video controller, but it’d be an interesting experiment.
3:00 So Fagin went to work for Oliver Twist, I mean, Faggin worked for Olivetti! I'm now going to examine Dickens to see if he wasn't really the Nostradamus of computer tech! ;)
While the biggest slice of the pie in the retro computing days goes to the z80 and 6502, the 6809 also had a look-in too... The z80 is also used in loads of early digitally controlled synths too...
It did power the dragon 32 the most welsh of all the 8bit micros. Quiet of few in the past have said the price point is what stopped it getting more widly used.
@@RetroBytesUK Yep, I've got a dragon 32! The dragon 64 was interesting in that it could run OS-9 on its 6809 which had multitasking etc (needed more than 32k to run)
@@leeselectronicwidgets My friend Tony has that up an running on his, he has quiet the Drangon 32/64 collection. With some interesting hardware, including a replacelment gfx card (well its not a card, but you know what I mean).
@@RetroBytesUK The 6502 came about because Motorola were not interested in what the customers actually wanted, so the 6502 designers left Motorola and the 6800 to create the 6500 - cut down to what customers wanted and cheaper; a lawsuit later and it evolved into the 6502.
Yes, Z180 (and Hitachi HD64180) took the architecture into many industrial applications due to its memory reach and its rich set of embedded peripherals. Nothing unusual today, but around 1990 when I designed with it, it was a game changer.
I really like this video; thanks for making it. However, I noticed one thing: Federico Faggin's last name is not pronounced correctly. Since the video emphasizes that he isn't as recognized as he should be, I think getting his name right is important. You can hear Federico introducing himself in many videos on UA-cam.
I was firmly more in the 6502 camp (I'm still are) but had many z80 based devices over the years, for example several of my car radios were managed by a z80
You keep saying S-BUS, it's not, it's S100. Another error CP/M was NEVER implemented on the 6502! Most CP/M 80 was implemented in z80 code. The Intel 8085 became very popular as a microcontroller. There were lots of cheaper 2nd sources for the z80 from NEC, sharp, Mitsubishi etc
@@RetroBytesUK It was on ported in the last couple of years as a technical challenge and performed not useful function, and only a small number of developer and demonstration tools were written for it.
My dad was the CFO at Zilog for a while and we used to play racketball with Federico on occasion. Super nice and very smart man. OG engineer
Is his last name pronounced fah-Jeen? There's an vintage Italian bicycle frame maker from my youth with that name. Besides being a bicycle nerd, I got to hack on Z80s on several platforms, mainly the Kaypro II under CP/M using Microsoft's M80 assembler.
We hobbyists won't let the ol' Z80 die that easily. I'm working on a Z80-based computer with 512K of paged RAM, full bells and whistles like Ethernet, MIDI, OPL2, SD card, 800x480 LCD, serial, PS/2 and more to come. It's just executing NOPs at the moment while I am waiting for the Flash programmer to show up, and will not stop until CP/M and Zork are running! :)
Nice to here, if you've got any information available pulbicly for it send me a link. I would be interested to have a nose.
hope you can make a low chipcount video controller that runs 5V as well, no FPGA or 32 bit Pico as sidekick please.
Are you using RomWBW?
I've loved the Z80 since my ZX81. I wished it had evolved into our time like x86.
@@RetroBytesUK Sure! It might be a month or two until things are ready, but it's also a good thing since I've found and fixed two bugs on my PCBs (and will probably find more). I'll be more than happy to release everything on GitHub!
I must say that I have come to the conclusion that your channel is my top retro tech YT channels. Your topics, commentary, detail and humor are spot on. I find myself rewatching your videos more and more other as I pick up on details I missed the first time.
Thanks Donnie, thats nice of you to say.
@@RetroBytesUK Your very welcome.
It's quite crazy how so many retro videos feature the Zilogs heavy hitters but the company itself is like ignored completely. Thank you for making this!
Nice overview.
I used the Z80 from 1979 until about 1987. By then you could get 16 bit memory for the S-100 bus, and 24 bit addressing. I used a homemade bank selector to access the extended memory, and had a 256kb card for the machine. When the Z280 came out, I moved to that. It was a 16 bit Z80 compatible CPU with a highly extended instruction set, and I made (and sold) an adapter that plugged that into a standard Z80 socket. The Z280 also had a memory pager built in, so that addressing 256kb in 4kb pages was possible.
Turned out the Z280 was buggy, and I ended up working for Zilog to try to fix some of the bugs. We spun the chip with a series of fixes, but it turned out the fundamental bugs were masking bigger and more serious bugs in the chip, and the CEO cancelled the chip. It now lives on as the EZ80, but I learned my lesson and headed on to the 80386.
What you are referring to as the Z800 became the Z280. The original design had a lot of problems, so you can't really say they ignored it as much as it had fundamental design issues and they canceled the project. Then it came back as the Z280, and got cancelled again, then came back yet again as the EZ80. However the EZ80 was not compatible with the Z280, so there you go.
You can still get the Z80 in the form of more advanced chips, and the "death of the Z80" headline was somewhat inaccurate. Yes, they discontinued an obsolete form of the chip, but there are many other instruction compatible versions. I think the most sensible design is to use one of several FPGA versions of the Z80 instruction set. Then you can use a relatively cheap FPGA with a Z80, several peripherals, and a chip with 100's of I/O's to boot.
24 bit? Via bank switching?
@@christopheroliver148 4kb page lookup. So the top 4 bits selected one of 16 pages in the lookup, and each of those pages had a 12 bit page address, so 12 bits page+12 bits page address = 24 bit total address.
@@christopheroliver148 via a 16 page lookup, indexed by the top 4 bits of the address. Each page had a 12 bit page address, so 12 bit page and 12 bit page address = 24 bits. Do a web search and you can still find the manual for it if interested.
I still can remember watching my brother on the family summer vacation writing down z80 assembly by hand and then writing down the binary in hex and finally "linking’ it by counting code length.. at age 12-13! Back home it was tested on an zx81..
That's some hard core stuff. I had a TS1000 at the same age and even had a book on assembly programming but could never get the hang of it.
Assembler instructions on paper, then hand translated to hex machine code, then entered in data statements in Basic. This was the way I wrote lots of Z80 assembler for MSX.
The only think I've done comparable was hand disassembling the ROM from a Kaypro II that along with the video page was a bank swap with the lowest addressed 4K. That little spiral bound reference for Z80 machine code got rather worn.
Z-80 + CP/M will always have a place in my heart. I loved my Kaypro II Z80 at 2.5Mhz 64K of RAM and two 191K Single Sided Double Density full height 5 1/4 inch disk drives, CP/M 2.2. Oh, those were the days. Thanks for making this video.
I used a Sanyo MBC-1000 with a Z80A and CP/M well into the DOS era. Didn’t see a reason to change until lured by a deal for a PC with a 20 MB hard drive.
I still have my Kaypro 4 and 10 as well as an Osbourne 1.
New RetroBytes video? This just made my weekend much more relaxing and I get to learn more about a CPU I knew little about growing up, but had used without knowing.
“A professional Kermit impersonator” is the best jab at someone Ive heard in a long time. Well done Sir. British humor, always prevails.
Love your videos. I hope you keep making them for a long while. I started out with in the C64/Amiga/286 era but didn't really get into things until the 286 so it is nice to see some of this older stuff in detail.
I find it interesting that at that time, all the CPU manufacturers thought creating new products meant new architectures, so the Z800 and Z8000 were completely new designs nothing like the Z80. It turned out that the way forward was evolutionary with extending present architectures, as with the x86 line. Code compatibility and developer familiarity were way more important that radical new architectures (as Intel themselves discovered with iAPX432 and Itanium).
Yeah no. Tell Arm that. The Archimedes was way ahead, but failed due to brainwashing. Yet now ARM is pulling ahead
6502 was incompatible to 6800 . Z80 was compatible with 8080 . Yeah, success made Zilog blind.
@@ArneChristianRosenfeldt We can't blame them really. The technology was developing very fast and nobody thought sticking to a simple architecture was the way forward, the designers all wanted to create new chips that were more advanced architectures, a "mainframe on a chip" etc.
We have the benefits of hindsight that they didn't have.
@@ian_b I read that Zilog did not want to use microcode. As we know (hindsight) this means that we have to use RISC and microcode come back through the back door as cache logic / pipeline stalls/ instruction queue. RISC as a chip is complex.
Zilog could have employed perfectly orthogonal addressing mode and instruction or even load-store. Though what I don’t get: reg-reg 3 stage pipeline is so simple. ARM uses this. Why did Zilog not emphasis reg-reg instructions?
Intel 8008 had a hardware stack. Maybe push and pop so could effectively be reg-reg instructions. Overflow would call a trap to RISC code which swaps out values to a larger stack in RAM. All known at the time.
@@ArneChristianRosenfeldt Not sure I understand all your points. The Z80 stack is very flexible (I write for PIC MCUs these days and the small hardware stacks are a pain). They could indeed have made the Z80 more orthogonal or done many other things but having an accumulator was quite normal at the time and the 16 bit semi-accumulator hl was nice. Also pipelines were a long way in the future.
They were I think focussed on creating an 8080 superset and did a very good job of that. If I had a hindsight criticism I would say they should have pushed for higher speed earlier. I still love the Z80 instruction set and sometimes write 8 bit PIC code as pseudo-Z80 with file registers called a, h, l etc.
😁
Thanks!
Brilliant episode! Probably my favourite CPU. A real marvel that it's still relevant all these years later.
Side-note: like the 4004, the 8008 was also a contracted design. This time for the Datapoint minicomputer company. They wanted a single-chip implementation of their 2200 minicomputer, and asked both Intel and Texas Instruments to have a shot at it, but turned down both designs for being too slow compared to what they'd already built out of discrete chips. TI couldn't find a buyer for their implementation, but Intel managed to sell the 8008 as a general-purpose microprocessor. The 8080 was derived from it based on customer feedback, and the 8086 as a 16-bit design derived from that.
I still, somewhere, have my Captain Zilog t-shirt from when I took their Z-8000 course. Great architecture, shame the Z-8001 had such problems in the first batch.
And the Z-80 was such a pretty chip. As an example, even 45 years later I still remember the LDIR instruction (LoaD Increment Repeat - EDB0) and the registers you needed to load before calling it. Having to do that in a loop on processors that came out years later still seems primitive.
I think current Z80 assembly cowboys no longer use LDIR, as you can move blocks faster with the stack....
@@lovemadeinjapanI don't think so. No matter what kind of trickery you do with the stack, you would still have to run a loop with individual instructions for each byte.
LDIR is the fastest way to move data on the Z80. Only reason some people don't use it is for compatibility with the 8080.
They probably could've made the Z8000 work if they did the Sega Mega Drive/Genesis approach (~10 years later) and integrate the older part as a subcomponent of the new system, for example the Z80 could've done memory management or IO handling for the Z8000 when the Z8000 was running, or it could've run in legacy mode and run Z80 code directly. It could have been as simple as switching from real mode to protected mode on a 286, or even just use interrupts to pass off instructions to the Z80, or it could have been like switching endianness on a PPC in that you'd need a reboot to switch between environments. Either way it would have been a useful stopgap to wean people off the earlier ISA onto the new one. As was the fate of so many architectures, though, it probably still wouldn't have saved it without a killer app for the Z8000. If only Zilog could have ditched Exxon and gotten the IBM PC deal, they could've made MSDOS a Z8000 exclusive.
It's a neat little CPU! The first time I learned about the Z80 microprocessor was my father talking about it in the ColecoVision we had back when I was little.
Finally, a new episode! You made my saturday!
IMHO, their mistake is not making Z80 based microcontroller early, like competing intel 8051 controller. 8051 is not code compatible with massively popular 8085. Considering popularity of Z80 then, such microcontroller would be instant hit.
In 2001, they did just that:
The Zilog eZ80 is an 8-bit microprocessor designed by Zilog as an updated version of the company's first product, the highly-successful Zilog Z80. The eZ80 is binary compatible with the Z80, but operating almost three times faster at the same clock frequency.
I love this background style... lots of slow "beauty pans" over some of our favourite old machines... with that nice wooden table for contrast. Very nice, very gentle.
So much better than the rather hectic computer graphics that I think I've moaned about in the past.
Love all your videos, I often take myself to a cafe at Meadowhall and watch your videos while drinking my coffee and eating my eggs Benedict. Thank you another brilliant video.
29:40 Cue the infamous, possibly apocryphal story about Gary Kildall (the boss of CP/M) bailing on meeting IBM rep's because he wanted to fly his plane that day, & IBM hitting up Microsoft for an OS instead of CP/M, & ending up with MSDOS instead.
This video will definitely help me in my research on the Z80. I, too, hope to create my own homebrew Z80-ish based computer someday. Perhaps the eZ80?
This video is saved in my "watch later" list!
When I was in college, the first thing you did for any project was grab a Z80.
+Retro Bytes this AMSTrad machine you showcase, could we get a closer look at this (resp a link to the video you already do that)?
Every time you say burglar alarm I hear burger alarm. Great video as always I just finished up watching one of your older videos and I was like I wonder if there's anything new and this is the newest one I hadn't watched
I learned to programme in machine code with the Z80, what memories!!!!
Sure was a lot of mentions of the intel 4004 being used in a single chip calculator...
Only problem being it doesn't really work without the 4001, 4002, and 4003 support chips.
If I recall, there were nine chips in the final product.
Early calculators all ran on chips like the Hitachi HD38800 and the NEC µPD7502, 4 bit microcontrollers with built in RAM, ROM and display driving circuits for controlling VFD tubes.
I still like the idea to have ALU, registers and microcode on one chip. Clock generator only needs a few connectors -> different chip . For the system bus a complicated driver is needed anyway.
@@lovemadeinjapan The HD38800 was introduced in 1978, a long time after the Intel 4004. Many other chips were used in calculators before 1978, including many single-chip designs, although the 4004 was obviously not a single chip design.
@@GodmanchesterGoblin The TMS 1000 is one of the earlier all-in-one designs, from 74. The nice thing of the NEC and Hitachi was the ability to drive 30V for the VFD tubes without external circuits.
@@lovemadeinjapan Yes, I am aware of TMS1000 - famously used by Sinclair in the original Sinclair Scientific model (1974).
While the biggest slice of the pie in the retro computing days goes to the z80 and 6502, the 6809 also had a look-in too...
The z80 is also used in loads of early digitally controlled synths too (as was to a lesser degree the 6809, but never the 6502, weirdly)
I haven't watched the episode *yet* - but when I heard a few weeks ago that the Z80 was about to be discontinued the question what the content of your next video would be was a no-brainer 😉
Does anyone else call the Intel 8008 "the boob"? 😅 You know, because calculators. I found myself going "ah yes, the boob" when it came up onscreen...
Guilty.
Ha. I was born before the Z80 and I outlived it. Now I'm definitely old.
A fascinating history indeed! Different architectures, different ways to achieve the same essential result, continually evolving, building on pillars of previous knowledge into the devices of today. ( sorry that was a bit poetic, I know) Thank you for sharing! :)
I think you can safely call 4004 "the First general purpose microprocessor", as the usaf offering was optimised for a specific task perhaps?
Also first publicly available.
Please can you do a video on Hewlett Packard's PA-RISC and a video about the Rise and Fall of IBM's PowerPC and POWER Architecture.
Please. Vote + 1. 🙏
POWER architecture is well and alive .
Alive but a shadow of what it was 20 years ago. Proceed to cope @ArneChristianRosenfeldt
@@legiran9564 I wonder why consoles left Power . IBM sold a lot of fabs. Do they still have critical mass? I thought diversification was a good thing in the past? Higher margins for main frames? Vendor lock-in? That would be evil . Big blue as it was before Microsoft and Intel came along.
Phillips and Siemens concentrate on medical because sick people will pay any price. But now they go abroad for medical services.
We had two RML CPM 380z at school in early 80s. Great way to learn about how computers work.
@@carguyuk7525 the first computer I ever touched was our school's cassette 380Z... hand assembling Z80 code with a pencil and paper. We later got another, running CP/M and that had an actual Z80 assembler. Not as much of a delight as the BBC micro's assembler (the CP/M text editor was quite hard work) but lovely all the same.
My first experience of computers was at school in 1978, computer studies - the classroom had a commodore pet 16k - 6502 processor I think and four challenger IP-8 with 8k and needed a monitor screen - any portable tv would do and they were! We were learning flow charts and BASIC programming - eeh a long time ago now!
1:30 and 12:40 Loved the shot of the Amstrad CPC 6128plus! A machine if it came a few years earlier would have been great. (Working on some hardware stuff for it atm.)
29:55 Wait! So what you're telling me is that we could have been playing CoD: Modern Warfare III on a 24 core 64 bit Z80,000?
64 bit, 32 core Z80M
ARM sort of happened. I would say ARM are the torch bearers.
The ATR8000 for the Atari 8-bit was an add on that happened to include a Z80 and could run CP/M.
These days, the Fujinet has (or had at one point) a virtual CP/M you can connect to.
(Full disclosure: I've never had an ATR8000, nor have I played with CP/M. I did start on a Timex Sinclair 1000 (basically, US version of ZX81), though soon after got my beloved Atari 1200XL.)
One mention of either the Z80 or the 6502 and for some reason I'm reminded of the huge ad for Computer Adventure World above the Wirral Line escalators at Liverpool Central Station. I seem to remember it was still there long after the shop had closed.
Brilliant young lad! Bloody brilliant and damn, I'm old. When the hell did that happen?
31:30 - we did have Unix for home computers in 1985, SCO Xenix. Just no home user was going to pay for it, or set it up - Concurrent had the same issue.
MS had planned in '81 to evolve MSDOS into single user Xenix, which they developed on the VAX, and sell Xenix for multitasking.
When AT&T started selling SystemV (making MS think they couldnt compete with an older Unix of their own) and MS made an with IBM agreement to develop OS/2, all the devs were put on OS/2 and Xenix was sold off via a sort of merger to SCO, who finished the port in ~84.
Also, mass storage back then was limited to floppy disks, and Unix without lots of disk space must be... complicated.
The Commodore 900 was not going to be a home system. It was squarely aimed at re-entering the business machine market that CBM, naturally, had started out in, experiencing success with the PET, but had subsequently abandoned or deprioritised by focusing on home computers.
The workstation model probably wouldn't have done very well against contemporary workstations based on the 68000 and other processors with 32-bit architectures. Faggin admitted that the Z8000, with its 16-bit architecture, was unsuited to dealing with the increasing demands of image processing and other workstation-related tasks. Like all low-cost workstations, by the time they would have been launched, the price wouldn't have been substantially lower than the competition. Commodore experienced that later on with the Amiga 3000UX.
One can argue whether the server or minicomputer model might have done slightly better as a lower-cost replacement for something like a PDP-11, although it would have been competing with systems based on the 68000 and other 32-bit processors, running genuine Unix, these aiming to nibble at the lower parts of the VAX market. It would probably have found itself boxed into a shrinking area of the market, with potential customers paying a decreasing premium to get those 32-bit systems instead.
31:30 -- Wasn't Xenix released by Microsoft for the Zilog 8001 in 1981?
The 40 pin version actually. I think they are still selling +5v devices, just not though holes versions.
You should timestamp the moment of the video that you are replying to or else it just looks like you are having a conversation with the invisible man.
Good job, but I was into Motorola and their 6809 microprocessor.
Just built a 6809 emulator based on an Arduino Due. Brings back a lot of good memories. :)
I'm old enough to remember the guys with 4 MHz Z80 computers telling me how slow my 2 MHz 6502 computer was. Yes, I tried to explain it to them.
@scottlarson1548 The problem with these kinds of discussions is that the comparisons typically involve overly simple non-optimized operations that don't really reflect real-world scenarios. For example in a single shift operation a 1mhz 6502 can potentially complete the task faster than a 4.7mhz 8088 however when running a series of shift operations, as would happen in real software, the 8088 is actually about 2.6 times faster which means it outperforms even a 2mhz 6502. When running code optimized for each cpu a 2 mhz 6502 is roughly equal to a 4mhz Z80, however, for complex operations the Z80 actually pulls slightly ahead. That's part of why vector and 3D games were so much better on the PC and even the spectrum than the C64. Games of the Apple II are a better indication of the 6502s real performance as it lacks the custom chips of the c64.
@@dgmt1 The problem with your discussion is that I was talking about a 6502 running at *2 MHz* not 1 MHz. You're comparing a 4 MHz Z80 with a C64 and an Apple II both running at 1.023 MHz, half the speed I'm talking about. It doesn't sound like you're familiar with the internals of these two machines.
It seams to have been a thing over the years ,of people thinking that comparing clockspeeds across different CPU designs has any great meaning. In 90s I remember a few press release from Apple complaining about people using clock speed to compare PPC and x86
@@RetroBytesUKthat remains an issue to this day. In the 90s, we got a lot of CPUs with a number in the name that wouldn't reflect the actual clock, but the comparable clock a Pentium would run at to get the same performance.
In the 2000s, that continued with the naming of the AMD Athlon series of CPUs.
@@TSteffihello, Marketing has entered the chat. What do you need to sell, we can make up anything to help!
After listening to the bit about traps I remembered Ordinance Survey did the same thing with their maps. They would features on the maps that did not exist in real life to catch out companies copying their maps. Someone told me the AA got caught copying by this method.
I see you are a fan of the 'Map men'.
The American Airlines?
The original Zilog plan was to start out with the Z8 (a microcontroller), then the Z80 general purpose processor and go from there. None were planned to be compatible with the others. They were forced to change the order and release the Z80 first, but the Z8 eventually did come out in 1979. Note that without the MMU the Z8000 could only address 64KB compared to 1MB for the 8086/8088 and 16MB for the 68000. It was the 1986 Z80000 that was Zilog's first 32 bit processor.
Ralph Ungermann went on to found Ungermanm-Bass, a network hardware and software maker.
UB's Ethernet cards and XNS software were in general use at Microsoft when I was hired into the LAN Manager development group in 1990.
They were also the first makers of twisted pair Ethernet interfaces I saw, circa 1992.
Back many years ago I built a Z-80 based system from scratch. It had a Z-80B running at the full 6MHz, 64K of 150nSec static ram, 32K ROM, 24 bits of parallel I/O, and a Z8530 SCC that supported 4 channels of serial I/O. That was a great little machine in its day.
Shockley was heavily involved with the Germanium transistor. The Field Effect Transistor was a bit later.
SPOILER: The 4004 was very much NOT a "single chip" solution for calculators! It was certainly more integrated than previous calculator solutions, but it required a lot of support chips to make a calculator! Eventually, single chip calculators were developed in the 80s, but they were highly integrated solutions that were the 80s equivalent of modern SOCs.
I would love to find an old trainer or better build one like this I think. I don't learn from books but hands on following along with good documentation. I don't know if they teach that way but at the time I was in school computers were in their boring phase at least from a learning perspective. You might get a typing or a basic programming class but that was about it. How to use but nothing explaining how anything really worked.
If you hit the Computerphile YT channel, where they have explanations from CompSci professors, the majority of them had a ZX-80 as their first computer. Multiple of them talked about how they got to know the entire machine, hardware and software, and that contributed to their interest that led to them becoming PhDs and professors.
Those of use who got to cut our teeth on z80 and 6502 based systems I think had it really lucky. It can just become cheap enough that normal people could try it out, and the systems where not so complex that you could not learn the whole system .
The commodore c128 had a z80 on board and in combination with a 1571 disc drive you could run cp/m on it. Not sure anyone did though.
I got a timex sinclair 1000 a while back basicly new in the nice sinclar foam clam box... thing. I guess its time to give it a go. (Amazing what you can find in the Free section on crageslist)
Free? That is a bargain beyond bargains!
@ 31:18 - Coherent was absolutely superb: I used it on PCs from the day it was released for x86 to the bitter end. Never knew it was going to be on Commodore - I'd have bought one instantly if it had appeared. Topic for another video, perhaps?
How informative. I started with a 6502 and the Z80 and its history are a bit alien to me.
My first contact with the Z80 was the NASCOM 1 in 1978. I got to like the Z80 although I did also use 6502 based Apple 2 at work. For some reason I preferred the Z80 over the 6502, but that may be because I encountered it first. The first ever microprocessor I encountered was the Nat Semi INS8060 SC/MP development system with a whole 256 bytes of RAM.
I first used the 6502 on a CBM PET and Acorn Atom. Then used a Z80, but prefer the 6502, probably because I used it first.
Nice piece of history.
The short clips of stock footage on endless repeat are distracting, though. Some shots of logic block diagrams would look better than a fake stock market clip played dozens of times.
You: "... in production the entirety of my life..."
Me: "Wait, are you younger than me? FML I'm old!"
:)
A lot of this is over my head, but I remember my father getting an Amstrad CP/M- based word-processing machine, which was presumably based on the Z80.
In the early 2000s i was working at a computer refurbishing/retailer and would often see z80 driving the OSD on computer monitors and TV PCBs. My first computer was a sinclair clone so I knew the chip label and was surprised of how many I've seen labelled as zilog insteand of the japanese equivalents from NEC and Hitachi.
how come they couldnt see the backward compatibility being important, i believe ic z800/8000 were b.compatible, zilog would be still around and doing well....
So, will the next video be about the 6502, then ? ;)
I loved my zx81, I even had the wobbly memory pack that would crash the computer,
lol. I mainly used it to learn Z80 assembly language and play games.
I got very excited to see a new video on this channel. What a ride! Thanks for sharing
It’s a little crazy how much similarity there is between the z80 and the 6502
Both effectively copies of an existing design, but better and cheaper. And both basically the only chip the companies that made them found success with.
I think that latter part makes a lot of sense. Since they tried to make successor designs by themselves, rather than following the methodology that led to their initial success... For instance, if the Z800 had copied/taken inspiration from the 8086 and the Z8000 a Pentium...
MOS was busy making VIA and CIA and then then Amiga chipset . Yeah, the CPU designers had left. The CPU in the SNES is underwhelming.
Lets see. A zx82, a Spectra video 328 and a 738. Both running cp/m. Did the Epson px-,8 run on z80? Also had dome other msx machines. Yups z80 ead a big part ofy early computing 😊
My first computer was a Nascom 1 with 1K of video RAM and 1K of user RAM, leaving 960 bytes that the monitor ROM didn't use.
This was clearly not enough for BASIC or even an assembler, so I did my coding directly in Z80 hex, looking up each code in the fat Mostek book, gradually getting quicker as my brain cached the frequently used instructions. Then I started learning the instruction group patterns and became a lot quicker and eventually memorized them all! I can still code Z80 now without looking up the instruction codes!
Nice to hear from a Nascom 1 user. Its amazing just how fluent in Z80 opcodes people could become after a while.
Is that a Laphroaig in the background? Good taste in more than one area, I suppose.
Yasss new episode!
IBM did ship with CP/M. They shipped with three OS choices. PC-DOS (Microsoft MS-DOS), CP/M-86 (Digital Research) and UCSD p-System (University of California at San Diego). The problem was that CP/M-86 was late and missed the launch date. It was also vastly more expensive. It also wasn't backward compatible with CP/M for the Z-80 and 8080. While the instructions were similar, it still meant tweaking and recompiling existing software. As Zilog found out with the Z800, almost compatible isn't compatible enough and if customers have to start from scratch anyway...
When the IBM 5150 Personal Computer shipped you had the choice of p-System (really only useful for cross-platform academic projects written in UCSD Pascal) or paying $240 for the promise of CP/M-86 in a few months if it actually did ship (Digital Research was already very late with it), or PC-DOS which was available at ship time for $40.
As a result of these problems, 96% of PC orders included PC-DOS, 3% included CP/M-86 and less than 1% included p-System.
Its that odd case of they shipped cp/m 86 with the PC, but simultaneously basically did'nt for all practical intentions. I get why Kildal was upset, I think he had every right to be, especially when it turned out DOS contained some of his code.
@@RetroBytesUK Actually it didn't. It cloned nonsense stuff just like what you were discussing with clones of the Z80 but no copied actual code.
@@borisgalos6967 That's not what the court found. The code MS purchased that formed the basis of MS DOS, did contain sections of code lifted from CP/M. MS at the time put forward the argument they where not to know and purchased the code in good faith, no one could really prove that statement either way, so the court accepted that argument, but none the less the first version of MS shipped with some CP/M code. So MS paid out.
Even though I know the history of Zilog and their processors almost as well as my own what I want to see is the RetroBytes take on it. Cue 20's music and a PCB way advert (both most welcome by the way- at last a quality sponsor not offering to look after my balls just maybe my reball)
I loved the Z80 and 6502....how about a video on the ESP32? (I have a webcam and home automation that uses this chip)...yeah, it's not a CPU, it's a MC, but I think there's a fine line between the silicon?
27:04 I got a R800 in my Panasonic MSX.
I have the microprofessor 1 plus. and while it's archaic, it's certainly interesting. but I don't use it much 'cos I'm more prolific in the 6502.
Worth noting the Game Boy used a custom Z80 variant with fewer features. Only 7 registers!
Since we are taliking about that "era" (yeah I was there.. I AM 65) and the 8 bit CPUs available there were 2 more popular archetectures. The 8048/51 that was used in most keyboards and (if you were good) embedded systems. The 6812/68xx series used mostly for embedded systems.
Part of the fun of having a retro-style micro is, well, building and maintaining it. At least, in my opinion. But, I can understand why some would see it as a "why bother" time sink.
Knowing that the Z80 survived this long was kinda surprising to me.
Hhehehe. Love the "Shed load of memory" image.
Why do you pronounce "PCBWay" the way that you do?
Its a reference to perifractic retro recipes, his was one of the first channels to be sponsored by PCBWay.
@@RetroBytesUKI knew that the moment you said it :)
Just now discovered your channel and subbed.
the first z-80 i ever saw was in a no name pbx phone system i decommed in like 2012. small business had been using it since the 80s
Whlie the 6800 was popular, the Z80 was my go-to. Used a TI machine with an awful keyboard. Years later, as our Unix minicomputer used z80s as the serial line processor, we taped up a Timex Sinclar 1000 next to the console.
While listening to this, I'm writing 6502 assembler code for an operating system I'm creating 🙂
Intended for a computer with both a 6502 and a z80 (it also runs CP/M), original 1980s era hardware. My new OS also runs on hardware which was available in 1986.
Which machine, Is there a github repo I could nose at, I would be interested to have a look.
@@RetroBytesUK Commodore 128.
There is no public git repository for now, eventually there will be one, but first want to have enough of the core functionality in a somewhat usable state.
@@c128stuff Fair enough, when you're ready for people to have a look let me know.
It’d be kind of funny to put the Video Beast into a Commander X16. I’m guessing it’d be pretty easy with X16’s expansion architecture. It’d be driving a different monitor than the X16’s built in Vera video controller, but it’d be an interesting experiment.
That would be very interesting to see.
3:00 So Fagin went to work for Oliver Twist, I mean, Faggin worked for Olivetti! I'm now going to examine Dickens to see if he wasn't really the Nostradamus of computer tech! ;)
Khm, did anyone mentioned micro beast?
Only like half the video.
While the biggest slice of the pie in the retro computing days goes to the z80 and 6502, the 6809 also had a look-in too...
The z80 is also used in loads of early digitally controlled synths too...
It did power the dragon 32 the most welsh of all the 8bit micros. Quiet of few in the past have said the price point is what stopped it getting more widly used.
@@RetroBytesUK Yep, I've got a dragon 32! The dragon 64 was interesting in that it could run OS-9 on its 6809 which had multitasking etc (needed more than 32k to run)
@@leeselectronicwidgets My friend Tony has that up an running on his, he has quiet the Drangon 32/64 collection. With some interesting hardware, including a replacelment gfx card (well its not a card, but you know what I mean).
@@RetroBytesUK
The 6502 came about because Motorola were not interested in what the customers actually wanted, so the 6502 designers left Motorola and the 6800 to create the 6500 - cut down to what customers wanted and cheaper; a lawsuit later and it evolved into the 6502.
Why haven't you mentioned the Z180 and Z280?
Yes, Z180 (and Hitachi HD64180) took the architecture into many industrial applications due to its memory reach and its rich set of embedded peripherals. Nothing unusual today, but around 1990 when I designed with it, it was a game changer.
I really like this video; thanks for making it. However, I noticed one thing: Federico Faggin's last name is not pronounced correctly. Since the video emphasizes that he isn't as recognized as he should be, I think getting his name right is important. You can hear Federico introducing himself in many videos on UA-cam.
I was firmly more in the 6502 camp (I'm still are) but had many z80 based devices over the years, for example several of my car radios were managed by a z80
You keep saying S-BUS, it's not, it's S100. Another error CP/M was NEVER implemented on the 6502! Most CP/M 80 was implemented in z80 code. The Intel 8085 became very popular as a microcontroller. There were lots of cheaper 2nd sources for the z80 from NEC, sharp, Mitsubishi etc
Well, not NEVER implemented on the 6502. Someone ported it a year or so back, because _of course they did_. davidgiven/cpm65 on GitHub.
It definitely was ported to the 6502, I run a copy of it. Its not spectally useful however due to the complete lack of any software you can run on it.
@@RetroBytesUK It was on ported in the last couple of years as a technical challenge and performed not useful function, and only a small number of developer and demonstration tools were written for it.
Ralph Ungermann and Masatoshi Shima are very underrated
Shouldn't that be called a "quick Segway" at 13:37
Last used it in 1983. Nice processor.
The old Z80 brings back fond memories :)
The IBM 5160 XT hard drive controller made by Xebec had a Z80 on it. I guess they didn't care at that point.