Arthur Zatarain with his Digital Group Z-80 Computer circa 1977

Digital Group Microcomputers

The cocky/nerdy fellow above is me with my first Z-80 "Datarain" microcomputer, circa 1978. The name originated from a piece of interoffice mail I received where my name (Zatarain) had been misspelled, perhaps intentionally because I was the "computer nerd" where I worked at the time. The name Datarain was eventually shortened to "Dataran" when it became the name of my engineering company in 1980. I wish I had kept the pocket protector in case they ever come back in style.

I began working with microcomputers in the mid 1970's, in grad school, using primitive Motorola 6800 breadboards having 256 individual bytes (not K or M) of RAM. In April 1977 my interest moved to a Zilog Z-80 system based on a Digital Group Z-80 kit. After many delayed deliveries, and thousands of solder joints, I had a system that could run simple programs and games in the minimal 2K CPU ram and expansion RAM that eventually reached 24K.

Most of the circuit boards were obtained in kit form and modified to make best use of the quickly changing technology. My DG Z-80 was first installed in an aluminum briefcase/tool case that included the power supply and two (and later three) Phidecks. My unit grew to incude three 8K memory cards and a few self-designed boards for primitive "industrial I/O." I made a "joystick" for the Lunar Lander game using a small empty WD-40 can, four mercury switches, and two buttons.

At night I would "borrow" a portable terminal having a thermal printer from my job at Shell Offshore. Connecting that terminal to the DG Z-80 was where I first learned how to emulate a serial port UART using only one bit of a parallel port. Those were fun times.

Despite the "portable" aspect of the tool case, the computer was not very portable. The homemade linear power supply was hot and heavy. The CRT came from a broken ATM machine, the keyboard was from a discarded bank terminal. A box of Phideck tapes was needed to get anything done. So taking that computer on the road was not going to happen.

CRT and Keyboard taken from salvaged ATM	Original Digital Group configuration in an aluminum tool case c1977

The Zatmon Operating System

The Digital Group operating system was called Phimon because it used automated cassette tape units called Phidecks. The small and very efficient OS was designed to run in the 2K RAM provided on the Digital Group Z-80 CPU board. I devised a simple way to double the CPU memory from 2K to 4K by soldering 16 memory chips piggyback onto the two rows of 8 chips already on the CPU board. All piggyback pins were soldered for mating chips except for the chip select pins which were bent upward and soldered together on each row into a separate bus. That bus was wired to unused outputs of the address decoder. The top row of chips looked like a line of caterpillars doing a line dance high-five. It was an inexpensive way to greatly increase the usable CPU memory for the DG OS. The idea was circulated among the DG user group and was copied on a few other units.

The additional CPU memory enabled me to expand on Phimon by reprogramming the readily available source code (in assembly language). One improvement was paged overlays for the Phimon command line functions such as DIR, COPY, etc. These overlays were originally loaded one at a time from the Phideck, which was very slow. The paged method allowed the modified Phimon to pre-load multiple OS command line functions into  memory blocks. That made the functions instantly available until the space was needed for another function. These improvements, along with low level functions such as primitive multi-tasking, calibrated timer support, and I/O abstraction were rolled into what I called "Zatmon" that was circulated among the user group. The core functions of Zatmon became the root of industrial control system code that I wrote for the next 30 years.

The Cardboard Box

My Digital Group Z-80 running Zatmon eventually outgrew the aluminum briefcase, so I moved it onto a platform with a tilting chassis for the computer and Phidecks. The platform was covered with a cardboard box having cutouts for the Phidecks. It wasn't very pretty, but looked more "finished" to visitors who would always ask "what do you need a computer for." I didn't have a very good answer back then, but I knew that eventually everyone would "get it."

Cardboard box enclosure for Digital Group Z-80 c 1978	Digital Group Z-80 with Phidecks mounted on platform under cardboard box c1978

Digital Group Z-80 with Phidecks on tilting chassis c 1978	Rear view of the cardboard box chassis that included switched AC outlets c 1978

The Rack Mount S-100 CP/M System

After giving up on a "portable" concept, I migrated my Zatmon-based system to a S-100 Bus system mounted in a home-made 19" rack mount configuration. This housed the CRT and keyboard, power supply, S-100 card cage, three Phi Decks, and later two 8" floppies. I recall designing and building an interface card for the floppies, and writing a driver in assembly. However, I don't remember if the 8" floppies were used with Zatmon, or if they were only used after migration to the revolutionary CP/M operating system. My guess is that the floppies were for CP/M only, although one photo does show the Phidecks and floppies mounted in the rack. So maybe Zatmon itself was floppy-capable.

Datarain Z-80 S-100 System in Rack Mount c 1979	Datarain Z-80 S-100 System in Rack Mount c1979

I clearly remember that interfacing the CP/M BIOS with Zatmon's relevant functions was much easier than expected due to the clever concept of the CP/M BIOS itself. I was stunned when the first attempt to boot CP/M resulted in a C:> prompt. Gary Kildall was truly a genius. When it booted I must have said something like "Damn, it worked!" because my wife, who was nearby, sarcastically said "Did you make computer history?" Well, maybe not, but it was an historic moment for me. Access to standard CP/M applications such as Wordstar had turned my homemade pile of parts into a "real computer."

The 19" rack had been fabricated in an 6x10 foot attic space converted into an office/shop. But after my daughters were born my wife converted the space into a playroom. I had my own business by then and didn't need the space anymore. However, the 19" rack was too large to get it out of the tight entrance path to the room. So I had to disassemble the rack and take it out in pieces. I used the scrap pieces of aluminum angle for many years afterward. And the S-100 parts were surely repurposed one way or another.

But I don't remember what happened to my Digital Group system. After leaving Shell in 1980 I became busy building a business that used S-100 systems, and later included manufacture of PC clones. So my use of a Digital Group Z-80 was in my distant rear view mirror. I probably sold (or gave away) my DG gear to someone in the DG user group. I still have a binder full of invoices and letters from those days. I was connected to DG users all over the country via a user group called BRIDGE. I don't recall what that meant, but the "DG" in the name probably referred to Digital Group. It really was an exciting time to be involved with computing.

The names of my DG pals and contacts (that I can still recall) include John McCahon, Chris Callens, Bill Wyzlic, Don Southwick, Joe Gavin, Rick Bourgeois, and Lloyd Kishinsky.

Datarain Z-80 S-100 System with 8" floppies & Pihidecks in Rack Mount c 1979	Datarain Z-80 S-100 System in Rack Mount c1979

Living in the CP/M World, and Beyond

My move to CP/M resulted in building a series of computers for myself, my friends, and eventually my clients when I started Dataran Engineering in 1980. The early micros were constantly evolving because the technology changed rapidly, sometimes on a monthly basis. Most of those systems were accessed with old school CRT terminals and ran the CP/M operating system. Early projects were written in assembly language, BASIC, and PL/M. Later systems leveraged the incredibly useful and efficient realtime TurboDOS multi-user, multiprocessor operating system instead of the single-user CP/M. TurboDos was a marvel at the time, and still impresses me today.

The TurboDos systems prevailed in my applications until the IBM-PC (and its clones) became essential to keep up with times. I abandoned S-100 and CP/M to get on board the bandwagon, eventually producing Dataran AT clones. The first was an 8mhz system called the Dataran 8E, which when spoken sounds like "AT."  The second model was the Dataran AT Plus that ran at 10Mhz (too bad it didn't "go to 11"). The clones were sold through my second company, called Naratad. In case you are curious, Naratad is Dataran spelled backwards.