End-Of-Life For Z80 CPU And Peripherals Announced

In a Product Change Notification (PCN) published on April 15, Zilog (now owned by Littelfuse) announced the End of Life for a range of Z80 products, specifically virtually all of the Z84C00 range. This also includes the peripherals, such as the Z84C10 range of MPUs. These are currently already marked as EoL on stores like Mouser, with Littelfuse noting that the last orders with them can be placed until June 14th of 2024. After that you’ll have to try your luck with shady EBay sellers and a lucky box of old-new-stock found in the back of a warehouse.

What this effectively means is that after just under 48 years since its launch in 1976, the Zilog Z80 will no longer be available for sale as discrete components, which is likely to primarily impact hobbyists and people who are trying to keep retro systems going. This does not mean that it’s the end of the road for Z80, however, as the eZ80 will be produced for the foreseeable future.

These new chips will of course not come in easy to drop in DIPs, making the challenge of breadboarding your own Z80-based microcomputer that much tougher. Yet one thing that definitely won’t happen is any of us witnessing the end of the era of the Z80, 6502 and 8051 architectures.

Thanks to [Techokami] for the tip.

57 thoughts on “End-Of-Life For Z80 CPU And Peripherals Announced

  1. Oh no! How will I be able to wire wrap my Z80 machine while listening to my old vinyl LPs under my trusty Coleman kerosene lamp? What is next? Will they come for my vacuum tubes too? Wait! What?!

    1. I can’t tell if you’re as old as I am, or one of those youngins mocking us. But just in case it’s the ladder, your time to be on the other end of this is barreling down on you with all the certainty of taxes even now. Bwa ha ha ha…

    2. Nice try, but Colemen never made a kerosene lamp, at least that I ever knew. But I have a dozen old Coleman lamps and stoves. Before there was “Coleman fuel”, we used what was called “white gas”. Back in the day when gasoline contained lead anti-knock additives, white gas was any gasoline that did not have additives. Coleman fuel has less volatile stuff than white gas and therefore fewer campsites burn to the ground these days,

    1. Does NTE actually make any ICs or other semiconductors, or do they just buy and re-label them?  AFAIK, the only source of 6502’s (actually 65C02, as the NMOS 6502 ended 20+ years ago) is Western Design Center, plus WDC’s licensees that are putting 65C02’s at the heart of custom ICs, ie, ASICs, for automotive, appliance, toy, and even life-support applications).

    2. Having done bare metal assembly on both the Z80 and the 6502 back in the day, can’t say my heart is broken by this. I met the Z80 first, and initially scoffed at the 6502’s seemingly paltry register set – but after delving into some 6502 stuff, soon realized how byzantine the Z80’s timing scheme is. Sure, it’s not impossible to count M-cycles and T-states (watch out for those refresh cycles!), but it was way Way easier with the ’02.

      So if I DO get taken by the urge to light the ol’ kerosene lamp, pull out the milk crate of vinyl LP’s, and break out the wire wrap kit, guess it won’t be the worst thing if there’s no DIP-40 Z80 to be had.

          1. I did 8048 assembly on a C-64 with a Freedom-48 assembler game cartridge and a Jason-Ranheim eprom programmer. Nice, tight little dev system. Freedom later came out with a Freedom-51 assembler on cartridge.

    1. Hopefully, lithography at home would be a reality once.

      Printing a Z80 on ‘glass’ was possible 20+ years ago already.

      https://www.idealine.info/sharpmz/z80glass.htm

      It’s also good that there are so many Z80 clones and derivatives ou there.
      Z80 had been cloned in USSR times already (U880, T34VM1/Т34ВМ1, KR1858VM etc).

      Super sets like the R800 or NSC800 are more interesting, anyway.

      They aren’t timing compatible, though, of course. Software like the firmware of a TNC (TNC2, PK-232 etc) is timing sensitive. Replacing it by a “better” processor would require code modifications, likely.

      The reason why Z80 is reaching EOL now is the 40 years anniversary, I assume.
      The Z80 was designed in ~1974.

      So I can partially understand the decision.
      Zilog wants the Z80 go to rest in dignity.

      Which by the way, is the reverse to making an FPGA-based Z80.
      Z80 was all about elegance and real hardware. You know, assembler programming and such.

      Your FPGA idea is an insult to this principle, I think (no offense, nothing personal). An FPGA is overkill hardware, consisting of an ARM core more powerful than an 80386 processor, paired with millions to billions of function units. It’s the definition of bloat.

      1. FPGAs come in various sizes and only a fraction of available FPGA chips contain an ARM CPU. An FPGA is the best way at the moment to get near identical replica of a Z80. There are several implementations for FPGAs that resemble the original chip in perfection.

  2. Wouldn’t an eZ80 mounted on a small circuit board, like many modern processors, work as a drop-in replacement? Well, boldly assuming it’s electrically compatible, at least.

    1. Z84C00 seems to be a “raw” Z80 you could see in an old computer or game console, while all eZ80s I see on Mouser are MCUs with built-in RAM, Flash, I2C controllers etc.. They still have a traditional memory bus, so maybe you could set the addressing up to not clash with whatever you have in place, or maybe not. Existing programs using illegal opcodes would probably need to be emulated though, at which point a Pico-like board with level shifter seems cheaper.

      1. True. But the whole Z84C00 story is a rather new-fangled thing on its own, I think.

        When Z80 was at its height, the Z80A (and Z80B) was the driving force.

        Those low-power CMOS types were for the sissies, rather.
        Real men™ worked with the rugged NMOS types.

        That was in a time when rock-solid linear PSUs and high-quality switching-PSUs were being made.

        When people had access to plenty of cheap atomic energy and when back then were thinking in amps still, not miliamps.

        When people operated their mobile equipment with d cells, car batteries and lead-gel batteries.

  3. > After that you’ll have to try your luck with shady EBay sellers and a lucky box of old-new-stock found in the back of a warehouse.

    eBay and AliExpress are currently full of desoldered NMOS version of Z80 and the CMOS ones usually look like they’ve been relabeled. Good luck! ;-)

  4. Ironically, I was just beginning to regard the Z80 as one of the most-likely candidates for future projects I want to last and be repairable, in this era of flash-in-the-pans and supply-chains.

  5. My first home computer (DEC Rainbow) I used for college had a Z80 in it, plus an 8088. At that time I never paid much attention to the processor as I just ‘used’ the machine for Word Processing and programming Pascal and C. Wasn’t until I was in the workforce that I again ran across the Z80 in one of our SCADA RTU products and helped maintain/write some assembly code for the Z80 on the CPU board back in the 80s. All ‘ancient’ history now… Time to let the o’ processor go to the dust bin of history.

  6. Is there any platform that I can expect to not be cancelled after some years?
    So that I design my stuff once, and then keep making it for decades without the need to redesign it every 20 years or so just because something is end of life?

    1. FPGA is probably here to stay. Bet you could choose a footprint and keep using that footprint with new iterations of silicon inside

      Also, why the demand for such a slow cycle?

    2. That is an interesting question. After about 21 years the patents have probably all expired, so there is not much preventing a different company or person from implementing a product. On the other hand, it is somewhat costly to produce chips, especially at small quantities. I think open hardware might help, but manufacturing VLSI chips with a home/small business is hard. Sam Zeloof’s efforts are impressive. On the commercial side, Yokogawa’s minimal fab efforts are also interesting. So maybe someday we might just be able to go to a local maker space and get any chip we can create a mask for (probably limited to visible light photo-lithography scale features) created from scratch.

  7. Well someone probably will just make a gate level fpga replica or asic

    Just decap the chip and reverse engineer the schematic

    Then send it to a chip fab

    Get a z80 at 7nm fab scale…

  8. Probably have a couple hundred Z80s laying around. Either on old terminal boards and such or in tubes. They will be available for some time yet on the grey market.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.