The 9100 Project

HP9825.COM

The Story of the Little Computer That Could!

 

Revised 2/21/04

The HP 9100: The Initial Journey

Sometimes it feels like the planets have aligned, the fates are all pointing at you, and “somethin’s just gonna happen.” HP’s VP of Research Barney Oliver must have felt like this in June of 1965 when he was visited by not just one but two inventors who had different but equally radical ideas for building a scientific calculator.

That both of these people should visit HP at all is noteworthy. In 1965 HP is not in the computation business. It’s an instrument company, albeit with a far-flung, international empire of manufacturing divisions and sales offices. Although HP is preparing to enter the minicomputer arena (it has just started a development program that will lead to the introduction of the HP 2116 minicomputer), HP’s involvement in any sort of computation is not yet public knowledge. HP engineers of the day are masters of the dipstick (slide rule) on which they perform daily engineering calculations. For heavier computational work, they buy a little time on a time-shared computer at Bill’s and Dave’s nearby alma mater, Stanford University.

Yet these two visitors, physicist/mathematician Malcolm Macmillan and inventor Tom Osborne, manage to find their way to HP’s R&D lab at 1501 Page Mill Road in Palo Alto within a few weeks of each other. The closely timed visits by Macmillan and Osborne spark Barney Oliver and direct some of HP’s creative engineering energy in an entirely new direction. That energy thrusts HP into the calculator business and creates a new, profitable, and very successful product line for the company.

(To see Tom Osborne’s first-hand account of this story and of the story of the HP 35 pocket calculator, click here.)

Old Line Calculation

By the 1960s, mechanical calculators (high-end adding machines) had been on the market for decades. These mechanical marvels could multiply and divide, making a tremendous racket in the process. However, these machines were not particularly well suited to engineering calculations because they lacked high-powered computational abilities, especially transcendental functions (such as sine and cosine). In fact, many of the more complex calculations performed during the Manhattan project as part of the atom bomb’s development were performed by teams of women, individually referred to as computers, who ran endless simple calculations on mechanical calculators under the direction of a “program” developed by one of the project scientists.

One of the scientists on the Manhattan project was Stanley P. Frankel, who developed intricate mathematical simulations of complex physical systems. Calculations for those simulations were performed by the female computers working for the project. Frankel later wrote computer programs for the ENIAC, the first all-electronic (neither female nor electromechanical) computer. So Frankel was very familiar with computing concepts. During the 1950s, he started to provide computer-design consulting services to companies including Librascope, General Electric, and SCM (Smith-Corona Marchant).

ENIAC, the world’s first electronic computer, was not programed but wired up to solve a problem. It became operational just in time for Stan Frankel and Nick Metropolis to run the calculations for “the Los Alamos problem,” which sought to find out if a hydrogen bomb was feasible. 


In the early 1960s, electronic calculators started to appear from companies such as Friden, Sharp, Wang, Mathatronics, IME, and Olympia. These early machines were designed as electronic analogs of the mechanical calculators that preceded them—performing addition, subtraction, multiplication, and division but not transcendental functions. However, even though they performed the same functions as mechanical calculators and no more, electronic calculators were both faster and quieter than the mechanical versions so the marketplace rapidly adopted the new machines.

The Smith Corona Company, a leading typewriter manufacturer, had acquired Marchant in 1958 as part of a plan to expand from typewriters into a full line of office equipment. It marketed the Marchant mechanical calculators under the brand name “Smith-Corona Marchant” until 1962 when the entire company adopted that name (or SCM for short). SCM realized it needed an electronic calculator too, or it would shortly cede the calculator market to its competitors. The company engaged Stan Frankel for this project in 1963 and either licensed an existing design or had him create a new design for the new calculator. Frankel’s design became the SCM Cogito 240SR, which went into production by early 1966.

Tom Osborne
Photo Courtesy of
Hewlett-Packard

Previously, SCM had hired a very bright electrical engineering graduate from Berkeley named Tom Osborne to help the company enter the electronic age, as its competitors were doing. Osborne learned to program IBM 709 mainframes in the Air Force and he had also programmed an LGP-30, the desk-sized computer that Frankel had designed for Librascope in the mid 1950s. After deciding to license Frankel’s calculator design, SCM assigned Osborne the task of evaluating the design and figuring out how to take it to production.

SCM needed to reduce the design’s manufacturing cost and one of the ways it planned to do so was to buy off-spec diodes for manufacturing the calculator. Frankel’s calculator design used a lot of diodes, which normally cost about 25 cents each. SCM’s bean counters had calculated that they could only afford to spend 5 cents per diode if they were to meet cost targets for the calculator. The off-spec parts were one way SCM planned to reach this cost target.

Osborne studied Frankel’s calculator design and SCM’s production plan for a while. He then decided that the design and the production plan weren’t very good. They would either produce a machine that didn’t work at all, or one that would operate very slowly. Osborne also decided that he didn’t want to work on the Cogito 240SR project at all. In fact, he now wanted to develop a calculator design on his own and he proposed this idea to SCM’s management. SCM weighed the credentials of the freshly minted MSEE from Berkeley against the seasoned veteran from the Manhattan project and unsurprisingly decided to bet on the veteran.

Osborne then proposed a “can’t lose” deal for SCM. He would develop a calculator design on his own, taking no salary and only lab space. He would then allow SCM to buy his design if they liked it, paying him only his back wages. SCM’s management turned down Osborne’s overly generous offer, saying that it didn’t conform to company policy.

At this point, Osborne decided to leave SCM. As is often the case in issues concerning intellectual property, things turned very ugly, very quickly. SCM demanded all of Osborne’s notes and papers, which he turned over to SCM. Then SCM’s attorneys demanded that Osborne give them his calculator design, which did not yet exist. SCM had concluded that Osborne could not possibly be confident enough to propose his “can’t lose” deal for the design of a calculator unless he’d already designed one. When Osborne insisted that he hadn’t designed the calculator yet, SCM’s lawyers become even more adamant. Osborne finally hired an attorney of his own who managed to dispatch SCM’s lawyers with one well-written letter, leaving Osborne free to develop his design.

SCM did finish development of the Cogito 240SR without Osborne and put it on the market in 1966. In the words of HP’s Director of R&D Barney Oliver, the Cogito “was a miserable machine, it took forever to do anything.”

Osborne went off on his own in January, 1964 and developed a calculator that represented a radical departure from any computer or calculator design of the day. He spent most of 1964 developing the design on a card table in his apartment without the aid of engineering test equipment. Fortunately, his wife had a job and could underwrite his personal R&D efforts. Osborne created an electronic calculator that used 10-digit, floating-point numbers when all of the other electronic calculators of the day were fixed-point machines. The use of floating-point data gave Osborne’s calculator the enormous dynamic range required for scientific and engineering calculations—ranging over 200 decades—from very small numbers to very big ones. This characteristic alone made Osborne’s design ideal for a calculator intended to serve the needs of scientists and engineers.

Tom Osborne developed his calculator prototype in his apartment. Here’s he’s assembling the balsa wood case that will house the calauclator’s keyboard and display. He used tools of the modeler’s trade including balsa wood, Elmer’s Glue All, and Pactra spray ‘Namel to create a good-looking case for the prototype. Photo courtesy of Tom Osborne


Osborne used advanced, digital, finite-state-machine design (which he later named Algorithmic State Machine design or ASM) that he had learned and then greatly extended at UC Berkeley. He developed a 64-bit, microprogrammed processor with operating characteristics that would not be duplicated in a microprocessor for another 30 years or so. He then built a prototype calculator, housed it in a hand-built balsa-wood case, and painted it metallic Cadillac green. He finished the prototype on Chrismas Eve, 1964 and then Tom Osborne went out into the world to find a buyer for his brainchild.

Osborne painted his calculator prototype Cadillac green.
Photo courtesy of Tom Osborne.


The CORDIC Algorithm

Long before Tom Osborne graduated from Berkeley with his MSEE, people started devoting considerable effort to figuring out how to perform complex engineering calculations on digital computers. One of these people was Jack E. Volder. While working for Convair’s Aeroelectrics Group in Fort Worth, Texas on navigation calculations in 1956, Volder published an internal report on his work titled “Binary computation algorithms for coordinate rotation and function generation.” This was the first publication of Volder’s development of the CORDIC (COordinate Rotation DIgital Computer) algorithm, which computes trigonometric functions (such as sine and cosine) using iterative repetitions of shifts and adds. Digital circuits can easily perform shifts and adds at very high speeds so Volder’s algorithms really opened the world of trigonometric computation to electronic computers. Volder solidified his work in 1959 by publishing two comprehensive disclosures of his work in IRE (Institute of Radio Engineers) publications.

A physicist named Malcolm Macmillan realized that Volder’s CORDIC algorithm held promise as a foundation for a scientific calculator. Macmillan, who was in Los Angeles at the time, developed a prototype calculator with Volder based on Volder’s CORDIC algorithm. Then Macmillan went out to find a buyer for his design. One of the companies he visited on his quest was HP in Palo Alto. Barney Oliver and Paul Stoft met with Macmillan in June of 1965. Stoft had developed the specifications for HP’s 2116A minicomputer with Kay Magleby at HP’s Dymec division in 1964. In June of 1965, Magleby was off developing that minicomputer at what would become HP’s Cupertino Division. Stoft stayed in HP’s corporate R&D lab with Oliver.

Beehive Hardware Design

Barney Oliver
Photo Courtesy of Hewlett-Packard

Oliver describes the meeting with Macmillan: “He and another guy [Volder] had developed a calculator which could perform transcendental operations, transcendental functions, and he brought this big kluge with him. It was a box about the size of two beehives. They finally got it working and computed a tangent and other trig functions for us. It took over a second to do this.”

Oliver and Stoft realized that Macmillan’s hardware design was not very advanced and was unlikely to produce a marketable product. However, the CORDIC algorithm implemented in Macmillan’s box represented a real computational breakthrough. Without a good hardware design, Macmillan merely had a brilliant idea in search of an equally brilliant implementation technology.

Enter Tom Osborne and his Green Machine made of balsa wood. During the first six months of 1965, Osborne had tried to visit a long list of about 30 companies that made calculators and computing equipment including Friden, Monroe, Honeywell, and IBM. He wrote directly to the presidents of those companies in an attempt to set up meetings. At first, he had quite a bit of difficulty getting the attention of these companies. SCM expressed interest, but Osborne had no interest in working with his former employer after his experience with SCM’s lawyers.

Osborne bought one share of IBM stock, attended a stockholder’s meeting in California, and asked Thomas Watson if IBM planned to get into the calculator business. Watson replied that IBM didn’t plan to get into the calculator business because the sale of just two IBM 709 mainframes represented about the same level of business that IBM expected it would get from the calculator market. (Industry legend also attributes an early prediction that the world would need no more than four or five mainframes, maximum, to Watson.) Osborne then asked if IBM might reconsider and Watson said that it was possible because IBM was always reviewing its business decisions. After the shareholder meeting, Osborne wrote to Watson. His letter recalled the words Watson had said at the shareholder’s meeting and then asked for a meeting to demonstrate his calculator.

No deal, but a good NDA

Osborne got no deal from IBM but he did get a meeting and a prototype non-disclosure agreement (NDA) that worked well for him in his search for a company to build his calculator. Previously, companies had offered to look at Osborne’s calculator design but the onerous NDAs they wanted Osborne to sign would have essentially given these companies the rights to steal his ideas. Osborne was bright enough to pass on these sucker deals. IBM produced a fair NDA and met with Osborne, but still declined to build the calculator. However, Osborne used IBM’s NDA as the model agreement for all future meetings. If it was good enough for IBM, it was good enough for anyone else.

Tom Osborne demonstrated his calculator to several companies in his apartment, where he’d built the machine. Here, the calculator prototype (right) is readied for the showing, along with the  logic module (center box), the power supply (on top of the logic module), and a specialized tester Osborne built (to the left of the logic module) that allowed him to debug the prototype without the expensive test equipment he’d otherwise need for such a project. The white mockup calculator on the left shows how large Osborne expected a finished calculator to be, once it housed the logic and power supply in addition to the keyboard and display. Photo courtesy of Tom Osborne


One company that was very interested in Osborne’s calculator design was Friden, the company whose existing electronic calculators had caused SCM to enter the market. As Osborne puts it, he and Friden “waltzed” for about six weeks without getting anywhere. Friden was very eager during that time to get under the green balsa-wood “skirt” of Osborne’s machine for a better look at the electronics. Osborne firmly kept the cover on his calculator and finally called off discussions with Friden when the deal started to smell. He later found out that Friden’s interest was in acquiring and killing his calculator design to prevent any competing company from eating into Friden’s market share. Osborne had not only developed a calculator, he was getting an MBA in high-tech marketing and some market savvy from his brushes with corporate America.

Osborne had written to the Hewlett-Packard Company as a possible partner and had gotten a polite response saying that said HP wasn’t in the calculator or computing business and declining to meet. However, Malcolm Macmillan’s visit to Barney Oliver and Paul Stoft at HP in June, 1965 seriously altered HP’s attitude. Oliver and Stoft were enthralled with Macmillan’s concept of a scientific calculator based on Volder’s CORDIC algorithm, but they were not at all impressed by Macmillan’s digital-design skills.

Tony Lukes, who was working for HP in 1965 but had previously worked at SCM with Osborne, learned of the HP meeting with Macmillan and told Stoft that he knew someone with terrific digital-design skills who could probably develop a good calculator design based on Macmillan’s and Volder’s ideas. Lukes convinced Stoft to bring Osborne in for a meeting. Stoft called Osborne but had to leave word with Osborne’s newly hired answering service. A fatigued Osborne has gone off for some vacation after trying unsuccessfully to sell his calculator design for six months.

At last, a visit to HP

Tom Osborne finally arrived on HP’s doorstep in June, 1965, some time after Malcolm Macmillan’s visit. Osborne demonstrated his machine to Stoft who then brought his boss Barney Oliver in to see the Green Machine. Oliver and Stoft quickly realized that the combination of Osborne’s wide-word, 64-bit, floating-point calculator hardware combined with the CORDIC algorithm for computing transcendental functions would produce a world-beating scientific calculator. Oliver then asked Osborne if he could redesign his calculator to implement the CORDIC algorithm. Osborne immediately says “Yes.” Osborne believed he could design anything with his ASM design technique.

Osborne shrewdly built the logic and memory for his prototype calculator into a separate box to allow the prototype keyboard and display unit to assume the size he believed would fairly represent the final production product. Otherwise, managers from prospective companies would get the impression that the calculator would be huge and unsellable. Photo courtesy of Tom Osborne.


Oliver then took Osborne to see a read-only memory (ROM) that Chuck Near was developing at HP Labs using printed-circuit technology. It’s only a tiny 64-bit ROM (8 words by 8 bits), but Oliver asks if Osborne’s design can be adapted to use this technology. (The Green Machine’s state-machine design was completely based on diode logic. The first prototype didn’t use a ROM to store its state-machine microprogram.) With a little more hesitation, Osborne again says “Yes.” His hesitation isn’t caused by the thought of converting his design to ROM, it’s caused by his uncertainty that Near’s experimental ROM design can be scaled up by a factor of 500, which is the size Osborne estimates he’ll need for the calculator’s ROM.

Finally, Oliver asks if Osborne can return the next day to talk with Bill and Dave. Osborne replies, “Bill and Dave who?” He rapidly agrees to return the next day to meet the company’s two founders.

The meeting with Bill and Dave went well and Hewlett took the Green Machine home for a couple of days to try it out. When Osborne returned to retrieve the Green Machine, he discovered that Hewlett accidentally poked a finger through the calculator’s balsa wood case. Hewlett apologized profusely and then also apologized for possibly destroying the Green Machine’s electronics as well. It seems that Hewlett accidentally plugged the Green Machine’s external power supply into the calculator backwards and the Green Machine was no longer operational.



Osbornes Green Machine03

Tom Osborne’s Green Machine with separate keyboard/display, logic unit, and power supply.

Photo courtesy of Fred Wenninger





Osborne assured Hewlett that he designed the Green Machine to prevent himself from making the same mistake. A protective power diode in the Green Machine limited the electrical damage to nothing more than a blown fuse. A relieved Hewlett agrees to buy Osborne’s machine, but only if Osborne is part of the deal. Osborne replied that the only way he’s willing to let Hewlett-Packard have the manufacturing rights to the Green Machine is if Osborne comes with it, but as a consultant and not an employee. The two agree to a four- or six-week evaluation period (memories differ on this point) during which Osborne will work with Malcolm Macmillan, who has been paid for his calculator design and hired as a consultant, and HP Labs engineer David Cochran, who had worked with Chuck Near on the HP 3440 digital voltmeter. Cochran will eventually develop most of the algorithms and programs for the new desktop calculator and will then go on to manage the development of the HP 35 pocket calculator and several subsequent pocket calculators.

Osborne will serve as an HP consultant for more than a decade and he will play a central role in launching HP into the desktop and pocket calculator markets, but he will never become an HP employee, by his own choice.

The six-week evaluation is successful and HP buys in. The fuse is lit. Osborne’s initial journey is now over. It was time to develop the HP 9100A.

 

 


 

HP9825.COM is in no way associated with the Hewlett-Packard Company.

Hewlett-Packard, HP, the HP logos, the HP 9825, and all other product model numbers starting
with HP are all trademarks of the Hewlett-Packard Company and don’t let nobody tell you different!

This site exists strictly for educational purposes. Nothing is for sale here. Substantial material from various issues
of the incomparable Hewlett-Packard Journal appears on this site, by permission of the Hewlett-Packard Company.

All text Copyright 2004 to 2017 - Steve Leibson

[Home] [Prehistoric Times] [The 9100 Project] [Prologues] [The 9825 Projects] [Epilogues]