Development of the 9800 Series

By Robert E. Watson
Engineering Manager, Calculator Products Division

Like most significant programs, the development of the 9800 Series Calculators was truly evolutionary in nature. The outcome is certainly much better than we envisioned at the beginning, and bears the marks of many contributors who added direction and vision at critical points in the definition of the individual products.

One key element was the outcome of early studies conducted by Ed Olander and Fred Wenninger that indicated that a special-purpose machine such as a calculator could be built with a general-purpose processor having a minicomputer type instruction set, with little or no loss in programming efficiency. This promoted the vision of a series of calculators, all using a common set of hardware, being different primarily in the internal firmware stored in ROM. Chuck Near tackled the job of inventing a small inexpensive general-purpose processor which could be microprogrammed to execute an instruction set much like our 2100-Series Computers. Fred Gross, Gerald Reynolds, and Gary Egan assisted with the logic design, while Henry Kohoutek struggled to squeeze into 256 states, all the normal instruction set plus several additional instructions to optimize the processor's performance for decimal arithmetic.

Meanwhile the question of memory technology became all-important. The printed-circuit ROM of the 9100 was still clearly ahead of integrated-circuit technology on a cost-per-bit basis. Thanks to the insight and ability of Ed Shideler, Larry Lopp, and others in our IC area, and the encouragement of Marco Negrete, n-channel MOS was chosen as the technology of the future. Dave Maitland began working on a 4096-bit ROM chip while Larry, Tom Haswell, Tom Ligon, and Virgil Laing were developing the n-channel process which would make it work. For many months, the project hung on the question of whether the design and the process could be brought together successfully in time. It is a tribute to Dave's design and our IC Lab's process that the first circuit worked and required only minor "tweaks" to get to production.

Initially under the direction of Rex James, and later with Geof Chance supplying the leadership, the development of the memory system proceeded in parallel with the ROM development. Calvin Finn, Gene Zeller, Glade Lybbert, and Ron Fuhrman helped devise a system which would allow blocks of memory to be easily added in a modular fashion. This would allow the function and even the language of the calculators to be changed by plugging in additional blocks of ROM.

As the basic hardware elements began to take shape, three different calculators were defined which would all use basically the same hardware. They would be labeled the 9800 Series, models 10, 20 and 30. Lou Dohse had the responsibility for the Model 10 which was to be the first product introduced. As such, Lou had the responsibility tor the major share of the program, Including all of the common hardware. The tact that such a complex project held to a difficult development schedule is a tribute to Lou's direction. Expanding on a concept initiated by Dave Cochrane at HP Labs, Roger Story and John Becker developed a non-contacting, Inductive keyboard which promised to be extremely reliable. Leo Miller was given the job of designing the displays tor both the Model 10 and 20. Jack Walden, Curt Brown, and Samir Gebala were responsible tor the Model 10 firmware. Thanks to their imagination and ability, the machine-- though an extension of the 9100A--contains many features and improvements over its predecessor.

The Model 20 was born with the idea that ROM would now be Inexpensive enough to justify adding higher-level language capability to a calculator. The development of such a calculator was an iterative process and included the ideas of many people. Ed Olander, Fred Wenninger, and Wayne Covington began some early language studies using simulation techniques. Ed maintained the responsibility tor the 9820A firmware throughout the entire project and was also assisted by Frank Yockey, Dennis Peery, and Mike Wingert. Jim Duley of HP Labs also gave valuable assistance with the compiler and basic machine structure.

Throughout its development the Model 20 was a machine which stimulated the imagination of all who were connected with it. The concept and structure of the machine allowed almost an unlimited number of features to be added, each coding "just a few words of ROM." The project leader required a steady hand at the helm to keep the project on course and yet not overlook those things of major significance which should be Incorporated. Rex James inherited the job midway through the program, and brought to the project the necessary direction and management skill to bring it to a successful conclusion. The end result was a calculator which not only has a high-level interactive language, but seems so natural to use that the total concept almost appears to be obvious.

Under Myles Judd's leadership, the Model 30 project team investigated the possibility of a more powerful machine which could effectively marry the interactive nature of a calculator with a common computer language--BASIC. Early in 1970 a computer simulation of the language was done by Myles, Rick Spangler, Fred Wenninger, and Frank Cada. From that effort came many of the concepts which led to the final definition of the Model 30. As the firmware began in earnest, Rick was given responsibility tor directing that effort which included assistance from Wayne Covington, Chris Christopher, Gene Burmeister, Mike Wingert, Frank Cada, and Samir Gebala. Kent Simcoe developed the display and also designed an efficient power supply which was used in all three calculators. Wally Wahlen and Fred Gross contributed the memory design whirs Chuck McAfee, Alan Richards, Gllbert Sandberg, and Perry Pierce developed a tape cassette unit which would take the place off the magnetic card reader used by the other calculators, and could also be packaged separately as a peripheral.

Fundamental to each of the three calculators was the requirement of an alphanumeric printer. Models 10 and 20 required a built-In 18Column strip printer, while Model 30 needed an 80-column, page-width printer. Jim Drehle and Roger Edrinn were given the responsibility of the strip printer. Working with Blair Harrison, Al Antes, and others at Colorado Springs Division they were able to develop a suitable thick-film print head which could be used with existing heat-sensitive papers. An extension of the same techniques was used to design the page-width printer. Russ Sparks headed the project which included Ray Cozzens, Dick Barney, Leo Miller, Fred Wullschleger, and Bob McMillan. Gale Hamelwright led the printer group in the early phases until he transferred from the Calculator division. The magnetic card reader was also developed in Gale's group with Havyn Bradley supplying the electrical design and Fred Wullschleger the mechanical.

From the beginning, the need to hew several key peripherals available at the time of introduction of the first 9800 Calculator was recognized. As the processor neared completion, Chuck Near took on responsibility for the peripheral program. A plotter project was initiated at San Diego Division with Norm Johnson, George Haag and Nilesh Gheewala doing the design. Henry Hetzel, Perry Pierce, and Max Davis contributed the interface for the Facit typewriter. Gary Egan designed the I/O section of the processor, and did the interfacing of the major peripherals. Andy Vogen and Bob Kuseski assisted in the development of the TTL and BCD  interfaces.

There remain several noteworthy contributor that should be recognized. Gary Paulson, Hudson Grotzinger, Ron Fuhrman, and John Bidwell were principally responsible tor the product design, with industrial design help from Don Aupperle and Arnold Joslin. Andy Vogen and Virgil Bennett supplied assistance in scheduling and parts coordination. Ivar Larson, Dave Cole, Homer Russell, and Bill Cummings coordinated activities with the marketing group and participated in the definition of the calculators as well as developing applications tor them. Brian Smith, Ed Miller, and Bill Mueller wrote the technical manuals. Srini Nageshwar and Hartmud Halversheid from HP Gmbh spent several months with Jack Anderson and our production people here helping to develop production test equipment and preparing for a simultaneous production startup in Germany. As the program neared the production phase, It required the combined efforts of many people in the various departments of the Loveland facility. They are too numerous to mention, but many contributed great personal effort to satisfy the logistics necessary to meet the required schedule. Finally, the program received the constant attention and direction of our Division Manager, Tom Kelley. Loveland facility manager, Marco Negrete, gave valuable encouragement and support, particularly in the development of the MOS technology which was needed. Early investigation efforts were directed by Don Schulz before he assumed his present position as manager of the Loveland Instrument Division. I would also like to recognize the many ideas and suggestions of Barney Oliver and Tom Osborne and thank them for their continued Interest.

In total, the 9800 program absorbed almost the entire efforts of the Loveland Calculator Division over an extended period of time. It is impossible to recognize all who contributed to its success. Perhaps it will suffice to say that the end products represent the combined best thinking of many people working together as a team, and as such are finer than anything we envisioned at the outset.

Go back to the HP Journal library
Go back to the main exhibit hall