Blast from the Past

Ain’t Done Yet

The creation of John Nairn’s General and Extended I/O ROMs, Mike Kolesar’s HP 98032A parallel I/O card, and the HP 98034A HPIB card, developed by Dick Barney and Ed Schlotzhauer, extended the HP 9825A desktop calculator’s reach broadly into a wide variety of computing and instrumentation systems. However, these I/O products still didn’t encompass the full gamut of I/O applications envisioned for the HP 9825A, even for these early days of personal desktop computing. More capability was still required.

During the HP 9825A’s development years, many HP instruments introduced in the 1950s and 1960s but still listed in HP’s product catalog in the 1970s had BCD (binary-coded decimal) interfaces. These interfaces came into common use shortly after instrumentation started going digital in the 1950s. Early BCD interfaces drove printers and data loggers to record readings on paper, which reduced the tedium of recording data but didn’t automate data analysis. Later, these same interfaces served as connections between the instruments and computers to create automated data-acquisition systems.

The job of developing a BCD I/O card for the HP 9825A fell to Larry Smith. Though the HP 98033A BCD I/O card (Project Sage) was Smith’s first project after earning his engineering degree, he was no stranger to electronic projects. Before he was 15 years old, Smith had built a Heathkit shortwave radio and designed a burglar alarm from a photoelectric cell and a fire bell. Unfortunately, he forgot that the system was armed at one point and caught himself coming in. It took a few days for his mother to cool off.

Smith spent most of his childhood in Alaska and was filled with curiosity about things electronic. He built his own voltmeter from an ammeter movement and some resistive shunts and he continued to use his homemade meter in college. As a senior in high school, Smith visited an engineering firm and was fascinated by their new desktop calculator (quite possibly an HP 9100A) because it easily computed transcendental functions.

While he was a freshman in college at the University of Alaska, one of Smith’s friends purchased an HP 35 pocket scientific calculator. Smith and his friends stayed up all night comparing the HP 35’s computations to printed tables in engineering handbooks. They could not believe that all that computational power could be crammed into something as small as a pocket calculator. The following summer, Smith saved his money and ordered his own HP 35. When the box finally arrived, Smith was delighted to find that his order had been upgraded to the newer, more powerful HP 45. “It was better,” Smith says, “because it could convert between polar and rectangular coordinates.” That was the year Smith became very aware of the HP brand.

After his freshman year, Smith transferred to Montana State University and then joined HP in the spring of 1975. He was recruited by Geoff Chance, who became Smith’s project manager. Smith took HP’s employment offer over a dozen other offers. Some of the other offers paid better, especially those from oil companies. But Smith selected HP. “It was an innovative, cool company,” he says. Smith had already decided that he’d like to work for HP after his experiences with the company’s pocket calculators. Dick Barney, also a graduate of Montana State, became Smith’s project leader and mentor. “I learned a lot of good practices and work methods from Dick,” says Smith.

Smith says that he knocked out the HP 98033A BCD I/O card in short order—“about two months although I burned a lot of hours on it from start to production release” he recalls. During that time, Smith surveyed the instruments in HP’s current catalog that had BCD interfaces (about 10-15 major instruments Smith says), created a superset definition of all the features needed to connect an HP 9825A desktop computer to these instruments, and then designed the interface card based on those requirements. The requirements included the number of digits in the mantissa and exponent; the location of the decimal point; the handshake mechanism; and the polarities for the data, two sign bits (mantissa and exponent), and overflow indicator lines. In addition, Smith had to fit the HP 98033A’s operation into the envelope already created by Kolesar and Nairn when they defined the HP 98032A parallel I/O card. The HP 98033A BCD I/O card had to work seamlessly with Nairn’s General I/O and Extended I/O ROMs.

Larry Smith designed these two boards for the HP 9825A’s BCD I/O card, the HP 98033A. This card uses 31 integrated circuits (including two DIP switches that are the same size as integrated circuits), nearly the same as the the 30 that Mike Kolesar needed for the HP 98032A parallel I/O card. Smith completed the design of the HP 98033A and released it to production in a quick two months. He was aided by Mike Kolesar’s circuit designs for interfacing to the HP 9825A’s I/O bus and by Tom Osborne’s algorithmic state-machine design methodology, as documented in Christopher Clare’s book, Designing Logic Systems Using State Machines   Information on the I/O pages of this site came from interviews with: Don Morris, Geoff Chance, Chris Christopher, Mike Kolesar, John Nairn, Dick Barney, Larry Smith, Ed Schlotzhauer, and from the memories of Steve Leibson.