Zilog MCZ 1/20 Z80 system

This page Last updated Nov 8 2010. COpyright 2010 Herb Johnson This page is a series of Web pages about restoring vintage microcomputers; Check my restoration home page for more. To email me or to order, see see my ordering Web page for my email addresses.

Introduction

A few years ago, I picked up two of these Zilog Z80 MCZ 1/20 development systems at a hamfest (I think). In Aug 2009 I did some review and cleanup of them. These were sold by Zilog in the late 1970's for Z80 development, based on their own bus architecture. Check other Web sites for the history of Zilog and their flagship Z80 processor.

It looks like this system used hard-sectored diskettes - that's consistent with the lack of an FDC chip on the boards, and the date when these boards were first layed out (copyright of circuit board 1975-76, before FDC chips existed). Another system owner, Peter Hill, confirms this, notes from him will be added when available. I have a general description from Ralf-Peter Nerlich. The hardware of my two systems are described below. - Herb Johnson

Zilog system in UK, address PROM issues

"I've got a Zilog MCZ 1/05 development system. A memory decode PROM on the 12/48K memory card (fitted with max 3x16K banks other 16k for 64k is on CPU card) has blown and was running hot. About 10 years ago I made a decoder from TTL but it's 5 chips on a board attached to the socket by a flying lead. I've tried to reconfigure the memory so it bank switches the monitor EPROM for memory on the CPU card thus reducing the decode to a simple A14 and A15 bank select on the 48k card but hit a race [timing] problem.

"I'd like something that can mounted on the socket instead of a card. Any ideas on lower chip count replacement? Even PIC or amtel if it can decode the 5 input lines and select the memory bank in 20ns. [He got supporting responses to create new PROMs in reply.]

"[My] conversion to CP/M required a new memory decode prom for MDC-2 (Memory Disc Controller 2) with I think (and hope) 6Fh in byte 10. And some mods on the MCB CPU/boot card to allow switching of the 4x1K boot proms out of the memory map at 0000h.

The MDC memory decode PROM is where my problem lies, it's burnt out. 20 years ago I made a decoder from TTL that plugged into the PROM socket. 10 years ago I got clever but not clever enough by half. I tried to re-arrange the mapping so first 16K was on MCB and other 3 16K banks on MDC, then I could ignore A12-13 and just decode on A14-15+RFSH. I got it wrong somewhere, so I gave up and tried to reverse all the changes I had made. That failed too, so I don't have a working system. I've checked decode logic by plugging into Tatung Einstein user port and it seems correct, which means I'm looking for a link on a jumper header or track break that's wrong. - Peter Hill

See my notes below about PROMS, for more info.

Peter gave me notes about Zilog systems, including comments on my systems. The notes are below.

Zilog in Germany

Ralf-Peter Nerlich of Germany used the Zilog development systems in the 1970's. Here's his description to me of those systems. He has a Web site at this link about RIO-OS and Zilog brand development systems. - Herb

"There are two families of Zilog's development systems - MCZ and ZDS. MCZ systems are for software development and automation solutions. Several models are exists, for "desktop" use (like your 1/20 model) or rack mount systems. The high level model 1/35 even has a harddisc.

"The ZDS series are systems for soft- and hardware development. They contain additional components to test z80 user hardware (RAM & ROM emulators for user hardware, incircuit Z80-CPU-emulator, bus analyzer).

"Both system families use Zilog's own Z80-RIO Operating System, but with a different ROM resident Kernel. The disc format is Zilog proprietary; 8" hardsector discs, single sided, 32 sectors, 77 tracks, 132 Bytes per sector. (128 data bytes and 4 pointerbytes for linking physical sectors to logical files at OS level)

"You can use any serial rs232 terminal as a console. The interface is situated on the MCB-Board. Optionally alterable to 20mA current loop. 110 to 19200 Baud.

"Some development software is available for RIO: Z80 Assembler & Linker, Z8 & Z8000 cross software packages, PLZ, FORTRAN, COBOL compilers, BASIC. No "office-type" software.

"For MCZ especially, a CP/M 2.2 implementation exits, using a later revision of the MCB-Board or a modified older MCB-Board."

- Ralf-Peter

EPROMs and bipolar PROMs - old tech

Bipolar PROMs

One issue with these systems, is their design's use of fusible link bipolar PROMS to select addresses. In the mid-1970's, a number of these small programmable devices were introduced. In this design, they provide high-speed decoding of the address bits from a microprocessor. Because of size and speed, they can't easily be replaced with TTL logic chips, diodes, etc. But they fell out of favor and use by the 1990's, and even if chips can be found they are not programmable on "EPROM" type programmers.

Peter Hill described these to me in 2010: "It's a Harris 7603 Bipolar tri-state 32x8bit PROM. Here's a data sheet on that device." Peter describes the addressing options for the PROM on the CPU board, in this text document. My CPU board uses an MMI brand 6306-1J chip. Peter's work on these PROMs is described above.

I'm making an effort to obtain more information about these bipolar PROMS and how to program them. I will have a Web page on the subject at a later time. - Herb

2708 UV EPROMS

A similar issue exists with the program UV-erasable PROMS on the Zilog system. They are 2708 1K X 8 PROMS, made by various companies in the 1970's. These (and the earlier 1702) used multiple voltages to operate, namely 5V, -5V, +12V. THey fell out of favor, over later single-voltage UV EPROMS, and later by electrically erasable PROMS. Most of the EPROM programmers developed and sold in the 1980's through today, cannot read or write 2708's and earlier PROMS. But a number of vintage computers - notibly S-100 systems - included boards to program these devices. And a number of early "how to use microprocessors" books included 2708 programmer designs. THere's even a few surviving VERY vintage programming systems for the 2708.

For myself, I'll either build some 2708 reader/programmers, or revive a S-100 programming card. Here's a Web link to my S-100 Web pages where I list S-100 manufacturers. I have some Web pages dedicated to each, which list some PROM programming cards with documentation available. - Herb

Other Resources

Ralf-Peter Nerlich contacted me in July 2010. He's working on a presentation about his MCZ experiences, and here's his Web pages about the Zilog development system. Google Translate provides a reasonable translation from German to English.

Thanks to Peter Hill for comments and corrections to this document, and other assistance and discussion about PROMs.

Documentation for the MCZ may be found on Maben's archives in the Zilog/Brochure section These seem to be 1/25 or 1/40 systems which are different.

Z80CPU archive of Rechner? work on reverse engineering a ZDS 1/25 or 1/30 or 1/40 system. May duplicate Maben's content.

Yunlin University of Science and Technology Museum in Tiawan seems to have "MCZ-1/20,25, MCB board series, hardware user's manual" "QA76.8.Z54 M35 1977".

Apparently there were "clones" or near copies of both the Zilog systems and its DOS, produced in Eastern Europe. These included the EAW P8000, Robotron A5120 and others. They ran the UDOS operating system. These UDOS systems still exist today in museums and private hobby collections. Look at this Robotron Web and UDOS site for an example.

My Zilog systems

System #10XX

The photos in the Introduction are of the system #10XX. The lable on the CABINET identifies it as a 60K system "MCZ 1/20 model 06-0066-01". That does not mean the current configuration or card set matches that model.

Cards

Zilog MCB (Memory Computer Boot) board 09-0013-09 rev F. Stamped June or January 1979 but board copyright is 1976. RAM chips are HM4716A-3 from 1978 (16Kbytes total). ROMs are 2708's, three at 1Kbytes each. Zilog chips include the Z80, CTC, PIO. AMD chip is AM9551. The parts with labels "0053" are bipolar PROMs to address memory.

The circular part in the lower right, is an inductor for circuits to produce +12v and -10v for RS-232 serial, and +12V, -5V for DRAM/ROM. Chips A37 and A41 are AM1488 and AM1489 respectively, RS-232 (EIA) serial interface chips, with traces down to the edge connector. Open pins around the PIO, are likely intended to be wired to the open pins and IC positions on the lower left of the card. the edge connector has pins wired to postions marked Z1 to Z21, which are not connected elsewhere. This is probably where one would wire the PIO or PIO interface chips.

Zilog MDC 2 (Memory Disk Controller) board, 09-0015-07C (c)1977, made in July 1979. Chips have date codes of 1977 and 1978. RAM is Hatichi HM4716A-3 from 1978. the large chip is Zilog but not identifiable. This board also has a circular devices which is likely an inductor or transfomer, to create additional voltage for the DRAMs.

Zilog interface board rev A. has a Z8001A CS "SEQ CPU". Chip date coes are 1979 80 81, the RAMs are Fairchild MB8264-20 from early 1981: 32 of them suggests 256K bytes. Ther are not many connections to the edge connector; there are two IDC (flatcable) connectors at top, 26 pin and 60 pin.

Zilog motherboard or backplane, 09-0018, etch 10-0018 rev H, from 1977. The backplane connectors consist of 122 pins, 61 pins per board side. The backplane is not a "bus", in that each backplane connector is wired for a specific board. On the boards, Pin 1 is rightmost on the component side when holding the board with edge connector down; Pin 62 is leftmost on the non-component side. So pins 1 and 62 are across from each other on the backplane board socket or through the PC board.

The floppy drive is a Shugart 800-2, with a 25102-4 card by part number. Parts date from 1978.

System #12XX

This system #12XX has the same board set as the first, in the same slots, and two Shugart 800 drives. The lable on the CABINET identifies it as a 60K system "MCZ 1/20 model 06-0019-01". That does not mean the current configuration or card set matches that model. Labels and markings in the system give dates of use for 1979 and 1980. The motherboard has a lable: MCZ REV H BACKPLANE MOD???? For Video

The MDC 2 board clearly has a Z80 PIO as the only 40-pin chip. This board is a rev C.

The parts with labels "0053" are bipolar PROMs to address memory. The three program PROMS are 2708's.

The MCB board for this system, has additional chips added on the left side of the board. You can see the wiring on the back of the board as imaged here for those chips. The MCB board on the other system is not wired or chipped as described here. Wiring of this this sort is entirely normal for the era and is called "white-wiring" because most often the wire used is white wire-wrap wire. Peter Hill comments on this wiring in his notes.

Peter's notes about Zilog systems

Copyright © 2010 Herb Johnson