Z80 discontinued (after 48 years!)

[Tom Storey]

News making the round is that the Z80 has been discontinued as of a couple of days ago.

The last order date is June 14 this year.

RIP Z80 :-(

https://www.mouser.com/PCN/Littelfuse_PCN_Z84C00.pdf

[Francis Pierot]

Thanks for the fish, Z-80.

--
You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/CAFDgZgU_WkcphNnt_qTOrfpt_cnG%3DefCrWZPEJOUomJiXc%2BtJw%40mail.gmail.com.

[Steve Cousins]

I like the reference

[Nick Tate]

OH NOOO!!! :(

[Andrew Lynch]

Hi

I don't think we'll be running out of Z80s any time soon.  There were literally millions produced and several licensed (and unlicensed) manufacturers.

My guess is someone like Rochester Electronics will pick up the dies and keep producing them.  Maybe at a reduced rate but I don't think Z80s are going extinct within our lifetimes.

Z80s are everywhere; embedded in so many things.  If nothing else, the IP will live on embedded inside FPGAs.  Just like 8051s, they will live forever.

[Jeff Greer]

I hope you are right. I found a kit that contains a new Z80 and its support chips just in case.

Jeff

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/865238690.1086817.1713543633592%40mail.yahoo.com.

[horus falcon]

Hi, Gang,

I did some digging on this.  Here's what I found:

The search led me to:  https://www.reddit.com/r/Z80/comments/1c793o9/z80_discontinued_after_48_years/?rdt=41541

This discussion links, in turn, to:  https://www.mouser.com/PCN/Littelfuse_PCN_Z84C00.pdf

This does look official.  There are a limited number of specific part numbers being discontinued, most of which are 40-pin DIP.  I did go ahead and snag four 20 MHz z80s just to be ahead of the madding crowd.

Later On,
David

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/865238690.1086817.1713543633592%40mail.yahoo.com.

[Bill McMullen]

Actually there's a *LOT* more parts being discontinued than just those in the previous mentioned notice.  The PLCC versions of the Z180 variants were discontinued earlier this year and there's also an end-of--life notice for the Z8S180, including the QFP variants.  The reality is that I can buy an eZ80 from authorized distributors like Digi-Key and Mouser for the same price of the MUCH less capable Z84C0020 or about 2/3 the price of the less capable Z8S180.  For new commercial designs it makes more sense (and cents) to use the newer chips,.

[Tom Storey]

That is my Reddit post and it links to the same document that I posted here. 😉

The parts listed are not just DIP but also the PLCC and TQFP SMD package variants. All speed grades in all package types are discontinued.

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/91b071d4-3850-4b5b-b8d9-88b90ff90faf%40gmail.com.

[horus falcon]

Hi, Tom,

Thanks.  I had missed that, and appreciate the detail in the correction.  I guess my next buy will be some PQFP z180/z280 stuff.

Later On,
David

[Tadeusz Pycio]

In that thread on Redit a user wrote: "The CMOS version of the Z80 , ( Z84C00 ) , which is the only version that is currently made, wasn't even created by Zilog , technically the die design was created by Toshiba."
Which I believe is not true. Toshiba CMOS chips don't pass ZEXALL testing (neither do NEC CMOS, I don't know how it is with Sharp as I don't own an LH5080 chip) unlike Zilog CMOS chips.

[MogensB]

Sad news about the Z80, indeed!!

Some of the designers/creators of several excellent Z80/Z180-based SBC kits are in this group - is it too early to share any thoughts on what's next?
Could/would some of the existing designs be adapted to use the eZ80 in the future?

Also, as far as i can see, Mouser lists Z8S18033FSG as NOT in EOL/obsolete (see https://mou.sr/3WdVjXj)? As an 80-pin QFP for surface mount, it is not the easiest part for hobbyists.

If this is true, this could also be an option in future SBCs.

Regards,

Mogens

[Bill McMullen]

If you look at Mouser's End Of Life notice for the Z8S18020FEG then you'll see that the 33FEG is also included in the notice.  It looks like Mouser just hasn't updated their website.

I've been using the Z8S180 QFP's for the last eight years in several different designs such as the MinZ-C.  It works well but as stated, the hobbyist may need to learn a new soldering technique.  In reality, many of the newer chips that I want to experiment with are only available in SMD packages so it's a good skill to acquire.

I've also been working with the eZ80 for several years now (i.e. eZ-Tiny and Min-eZ).  Besides the obvious SMD and 3.3V issues, the designer must be aware of tighter timing issues.  At 50 MHz there's only about 11 nS for memory access.  Of course you can add wait states but with one cycle processor access, a single wait state effectively halves the overall performance.

Given the long production life of the Z80 and Z180, I'd predict that there isn't going to be any shortage of used chips for a LONG time.  Unfortunately the prices on places like eBay have been rising significantly.  For the novice hobbyist designer who wants a Z80 or derivative, I'd still recommend using the Z80 DIP or Z180 PLCC but they should try to obtain at least a reference chip or two from a reputable supplier with new old stock.

[T Gerbic]

Beyond the typical 40-Pin DIP, which probably there are plenty floating around, there are lots of more 5V modern 44-pin PLCCs, 48-pin J leg and 40-pin SOP packaged Z80s in the 6Mhz to 10Mhz range that are still pretty easy for hobbyists to use. Several vendors made them.  It will be a long time before the Z80 market is exhausted, though there may be a need to migrate DIP designs to PLCC, or use adapters.

[Mark T]

Is there an SOP or SOJ variant? I did find the toshiba part described as SOP but this actually seems to be SSOP with 0.8mm pitch. This is similar pitch to the QFP44 package, but might be slightly easier for pcb layout. PLCC has 1.27mm pitch with J leads.

[T Gerbic]

I should have clarified, the J leg is with the PLCC-48 if I remember right. There are several vendors and different part numbers for the Z80 (and variants of the Z80) so there are other part numbers and packages that have been made. Some that were not quite as popular may have a lot of surplus parts available.  I have at least a tube or two of the PLCC-44 8Mhz versions, along with some other types and some PLCC-44 on little daughter boards. All these variants would suggest more in the market so we should not run out for a long time.

[Bill Shen]

There is already an effort to produce new Z80 silicon clone.

https://github.com/rejunity/z80-open-silicon

On April 15 of 2024 Zilog has announced End-of-Life for Z80, one of the most famous 8-bit CPUs of all time.

It is a time for open-source and hardware preservation community to step in with a Free and Open Source Silicon (FOSS) replacement for Zilog Z80.

The first fabrication of FOSS Z80 is scheduled for June of 2024!

[Peter Wilson]

I'm not sure I see the point of a FOSS Z80 unless they also do the support chips - SIO, PIO, CTC and DMAC. They could all be done on the same ASIC from the looks of it then you could generate as many of each as you wanted from a single set of chips. But I don't really get how the Tiny Tapeout system works.

Pete

--

You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/5efc3e07-07ef-4ff5-9951-0e1e163c83a9n%40googlegroups.com.

[Alan Cox]

On Tue, 30 Apr 2024 at 19:58, Peter Wilson <pwil...@gmail.com> wrote:
>
> I'm not sure I see the point of a FOSS Z80 unless they also do the support chips - SIO, PIO, CTC and DMAC. They could all be done on the same ASIC from the looks of it then you could generate as many of each as you wanted from a single set of chips. But I don't really get how the Tiny Tapeout system works.

I am not sure it's a big issue -

8255 is better than the PIO and still available
16x50 is better than the SIO and still available
Half a dozen different timer chips are available

The only one that doesn't have an off the shelf alternative is the
DMA, and not many things every used that.

[Andrew Lynch]

Yeah, but the Zilog Peripheral chips support IM2 and none of the Intel chips do that.  So, if interrupts are an issue, the Zilog Peripherals definitely have an advantage.  However, if you can live with IM1 then I agree that the Intel chips are generally easier, better, and more available.

If they are going to do a FOSS Z80, then I think they should also do the Zilog Peripherals for completeness.

Thanks, Andrew Lynch

--
You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/CAK9X0%2BuyRc_odqm6DiapSEfxwndWTdmXvTDe7NOD_f%2BoamqdkQ%40mail.gmail.com.

[Tom Storey]

There's an application note from Zilog ("THE Z80 FAMILY PROGRAM INTERRUPT STRUCTURE") which has an example of an IM2 interrupt controller for non-native peripherals using "discrete" logic. It supplies a vector during the IACK cycle, and will decode RETI instructions to clear the interrupt condition.

A simpler implementation without the RETI decode logic would probably fit in a 16V8 GAL.

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/1786972015.4647416.1714512151950%40mail.yahoo.com.

[Tom Storey]

I decided to have a play in WinCUPL to see if I could implement the bare minimum "IM2 interrupt daisy chain controller in a GAL". I think my simulation covers all of the appropriate cases...

It isnt edge triggered like the Zilog example (probably could be with an external DFF or two), and a 16V8 is too simple to permit decoding the IRET instruction, but I think it should work as a "two chip solution" as long as the interrupting device continues to assert its IRQ output until the interrupting condition is cleared.

Name     Z80_IM2 ;

PartNo   00 ;

Date     30/04/2024 ;

Revision 01 ;

Designer Tom Storey ;

Company  Hobbyist ;

Assembly None ;

Location Ux ;

Device   g16v8a ;

/* Control */

PIN 1 = clk;

PIN 3 = !iorq;

PIN 4 = !m1;

PIN 5 = !reset;

/* Interrupt interface */

PIN 6 = !service; /* IRQ input from interrupting device */

PIN 15 = irq; /* Open collector IRQ output to CPU */

PIN 14 = !vect; /* Control output to buffer for driving vector */

/* Daisy chain signals */

PIN 2 = iei; /* From higher priority device in chain */

PIN 19 = ieo; /* To lower priority device in chain */

/* State machine */

PIN 12 = state;

$DEFINE M_IDLE 'b'0

$DEFINE M_IRQ 'b'1

/* IEO mirrors IEI, and is negated when service is requested */

ieo = iei & !service;

/* The IRQ output will drive a logic low */

irq = 'b'0;

/* IRQ output is driven when service is requested */

irq.oe = service;

/* Assert vector output when IEI is asserted, service is requested, and in an

 * interrupt acknowledge cycle */

vect = iei & state:M_IRQ & iorq & m1;

SEQUENCE state {

/* Move to the IRQ state once service is requested, unless we are in an

* interrupt acknowledge cycle, in which case remain in the IDLE state to

* prevent race conditions in supplying a vector */

PRESENT M_IDLE

IF service & !iorq & !m1 & !reset NEXT M_IRQ;

DEFAULT NEXT M_IDLE;

/* Wait in the IRQ state as long as service is requested, or until reset */

PRESENT M_IRQ

IF reset # !service NEXT M_IDLE;

DEFAULT NEXT M_IRQ;

}

/* END */

Pinout diagram:

                               ______________
                              |   Z80_IM2    |
                      clk x---|1           20|---x Vcc                      
                      iei x---|2           19|---x ieo                      
                    !iorq x---|3           18|---x                          
                      !m1 x---|4           17|---x                          
                   !reset x---|5           16|---x                          
                 !service x---|6           15|---x irq                      
                          x---|7           14|---x !vect                    
                          x---|8           13|---x                          
                          x---|9           12|---x state                    
                      GND x---|10          11|---x                          
                              |______________|

An external buffer is required to drive the interrupt vector. Something like a 74x245, or a 74x573 to allow for a software configurable interrupt vector.

Maybe someone would like to give it a try. I am travelling for a few weeks so have no way to test it myself other than in simulation.

[Tom Storey]

Actually, thinking about it some more, I believe the equation for the ieo output should actually be

ieo = iei & state:M_IDLE;

This will prevent the GAL from negating ieo if an interrupt condition occurs during an IACK cycle and potentially causing a downstream device to hiccup in the middle of providing a vector.

[Andrew Lynch]

Hi

Yes, on duodyne, the Z80 processor board has an IM2 conversion circuit for non-Zilog devices and it works well.  It adds 8 vectored interrupts so the whole system can be IM2 even including the non-Zilog parts.  Overall, I think it simplifies things since the ISR immediately knows which device is pulling the interrupt without searching through all the possible candidates.  I think the ZIlog interrupt handling scheme is one of their strengths and I hope it stays.  It is too bad though that there aren't more modernized versions of the rest of Zilog Peripherals especially the DMA and the SIO.

Thanks, Andrew Lynch

[Bill Shen]

I use KIO, Z84c90, in several Z80 designs.  It is integration of SIO, PIO, and CTC in a chip.  It supports IM 2 and several different configurations of daisy chain of SIO, PIO, CTC.  It is rated at 12Mhz, but can be overclocked to about 30 Mhz.  It also works down to 3V.

  Bill

[Tom Storey]

I picked up some KIOs a few years ago, I think there was a bit of a rush after they had been discontinued or some such. 

Haven't used any yet, I found myself sucked in to the 68k world and haven't been able to escape since. 😅

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/0e1ced4a-8bdb-40e0-a845-e56e139b3c45n%40googlegroups.com.

[Tadeusz Pycio]

I don't think the current stock will run out any time soon, and even if it does you can always go back to NMOS chips which are reluctant to sell and there is a good range of them. I am referring to NOS offerings, not official distribution sources.

[Bill Shen]

I received an email from Mouser that the two 20Mhz Z80 I ordered in May will ship Aug 9 2024.  Maybe I’ll have the last date coded Z80.

  Bill

[7alken]

hi Bill, I was long time away ... have you got them? which datecode? :-)

[Bill Shen]

Yes, I received 2 20MHz Z80 from Mouser.  The date code is 2416, so they were made in April of 2024.
  Bill