The HP 200CD

HP Journal, Dec, 1952

Everyone knows that Bill Hewlett invented the HP 200A audio oscillator, the instrument that launched the company. But few know that it had an upper frequency limit of only 20 kHz, when first invented. The lab had improved the range up to about 50 kHz, and then ran into circuit parasitic problems, and left it there.

Enter Barney Oliver, returning from years of research at Bell Telephone Labs, in New Jersey. Barney was a classmate of Dave and Bill at Stanford, and was talked into joining HP in the late 1940's. Barney had a way of looking at problems that was unique. His first look at the circuits of the HP 200 convinced him that simply by going to a balanced circuit configuration, he could get well-over a 10 to 1 improvement in the frequency coverage. They tried it, and it worked, to 600 kHz, the first of hundreds of contributions that Barney would make to this company, over a 40 year career.

I recently ran into an interesting retrospective article on the status of HP's main audio oscillator competitor, in 1939. The company was the General Radio Company, of Cambridge, MA. H.H. Scott was later to gain fame and success as a pre-eminent manufacturer of audio systems, and worked at GR before starting his own company.

"In 1937, Scott applied for a pair of patents, covering the application of frequency-selective feedback networks to audio oscillators and amplifiers. One patent covered the resistance-capacitance oscillator, that was destined to completely supercede their older designs, and should have let GR dominate the market for years to come."

"But, for several reasons this did not happen. GR's credo was performance first, price second, sometimes regardless of its customers' wants. In the 1930's, the Beat Frequency Oscillator was a better design, and GR had several models available; Scott, in fact, had helped to design them. In the beat frequency oscillator, two RF signals, one fixed and the other variable, were mixed together to produce the output frequency. This arrangement permitted a wide tuning range, without band-switching. . . . . General Radio now licensed a couple of college kids from California, William Hewlett and David Packard, to use Scott's patent. Hewlett and Packard had been casting about for products that might be manufactured by their newly-formed firm."

"Hewlett was working for his master's degree with a small group of students studying various uses for feedback in amplifiers. Their mentor was the distinguished radio engineer, Frederick Terman. Hewlett's contribution was a feedback oscillator, stabilized by a nonlinear resistance (a tungsten-filament lamp) in one arm of the Wein bridge. Tungsten lamps had been used a year before, in a crystal oscillator by a Bell Labs researcher, and of course, the Wein bridge was not new. But Hewlett was first to bring all elements together in a practical RC circuit. . . ."

"GR belatedly produced an RC oscillator in 1948--their model 1302A. But it was priced at $365. Incidentally, although the unit used a Wein bridge and tungsten lamp, it was not licensed under Hewlett's patent."

From the GR Old Timer's Bulletin, May 1996, Alan Douglas, Pocasset, MA

HP 302A ---------------------------------------------- HP 300A

HP 300A HP Journal, August, 1951
HP 302A HP Journal, Sep, 1959

The HP 300A wave analyzer was perhaps the second instrument designed by Bill Hewlett, in 1941. It was a remarkable instrument, in a huge wooden cabinet, almost 3 feet high. It was essentially a tunable audio receiver, with variable IF bandwidth, achieved using control over a regeneration effect, as well as a degeneration effect-presumably using various feedback methods. But it was VERY sophisticated for 1941, when feedback theory was just coming into its own.

I well-remember my Notre Dame engineering laboratories in the late 40's. Standard instructions for preparing for one lab experiment, called for the experiment leader to come into the lab at dawn and turn on the HP 300A so that it could heat up for at least 4 to 5 hours, before any measurements were made. The alignment procedure for the selectivity adjustments in the IF strip bandwidth, and LO were super-critical. The drift of the vacuum tubes would make all data useless if this alignment procedure were not followed rigorously. It took almost two decades to replace the HP 300A. It is an interesting coincidence, that HP's second transistorized product was the HP 302A Audio Wave Analyzer in 1959. That machine was a wonder.

Having used such a problematic instrument like the HP 300A, the HP 302A was nothing less than miraculous. No drift. It had constant tunable selectivity. It even had automatic frequency control, for locking onto a slowly drifting signal.

Dick Rucker relates that Dave Cochran (BART & HP-35 stories) was, in a small way, legendary before he got famous for his HP-35 algorithm work. He worked part-time, nights Dick thought, in the Test Department on HP 300As -- the old original wave analyzer -- while at Stanford, and he tested them faster and better than anybody. And it was reportedly a BEAR to test.

The HP 9100A

HP Journal, Sep, 1968

The first HP desktop computer was introduced in 1968. Barney Oliver credits three people with the combined idea. Tom Osborne, designed a floating point calculator, while with his Logic Design Co, in Atlanta, GA, and brought the idea to HP. Malcolm McMillan had invented a mathematical algorithm and calculator for transcendental functions. Paul Stoft, and his group in HP Labs, saw the future of such a combined machine for engineering and other uses. In that era, we should remember, that clickety-clack teletype terminals were the accepted human interface, for time-share computers.

The HP 2100 mini-computer had been recently launched, with the purpose of controlling one instrument for each of 14 slide-in interface boards on the bottom cardcage. The new industry applications software language was called BASIC. Engineers were just beginning to anticipate the power of distributed computing, using small computers rather than a central big-daddy IBM mainframe. So the world was waiting for a scientific desktop machine.

Al Bagley and Dave Cochran shared with me some insights on the initial history of how HP got into the desktop calculator (HP 9100) business. Al was approached first because his Frequency & Time Division was the only one with substantial digital work going on. He decided that calculators didn't fit his product charter, so he asked Bob Grimm, manager of the Dymec Division, who also declined due to their business overload. All agreed that HP Labs would be the right place, which is where Dave got involved. "1) Tom Osborne talked to Barney, with his balsa wood, floating-point calculator; and some neat logic, like switching the power to turn on a function, but it was only four-function. 2) About the same time Malcolm McMillan spoke to Hewlett about a spin-off of the navigation computer in the B-58 Hustler bomber that could do transcendental functions, but was only fixed point. It was derived from a paper on pseudo division by Meggitt of IBM in 1962." For an interesting insight into how Osborne built a computer without any test equipment, see the Reference Section at the end for his interview with Steve Leibson.

"Barney and Hewlett compared notes, and Barney called a bunch of us together to see if we could marry the two concepts, with Osborne's help. We decided we could put McMillan's algorithms in Osborne's floating-point architecture. Barney asked, 'Who's going to do the algorithms?' I answered, 'What's an algorithm?' Barney said, 'Well Dave, you're going to find out.' Gee, I should have learned about volunteering from my stint in the military."

Cochran observed that he found out in his first investigations, that algorithms go way back. "The word comes from the village of Kurizimi, where there was a 9th century Arab mathematician named Jabar and "al" means "of the." I found some of the early transcendental function routines while doing some prior art work for legal back-grounding, and what we used was described by Henry Briggs in a treatise in Latin in 1624."

Cochran continued, "During the development of the HP 9100 desktop Hewlett had always talked about how it should fit in the typewriter slot in his desk. Well, when we fished the first prototype, someone remembered Hewlett's wishes so we trooped over to his office to check it out. Darn, it didn't fit, but Hewlett was out of town so we called Johnny Harrison at the carpenter shop. By the time Hewlett came back he had a HP 9100 Calculator in his "modified" desk." Bill said, "I knew you guys could do it."

One must also remember that integrated circuits were still a few years from reality. Thus, the Read-Only-Memory (ROM) for the HP 9100A, was constructed out of a l6-layer printed circuit board, with 31,000 bits (diodes) and transistors. The display was a CRT, so that the four-stack registers could all be displayed. And this was HP's first roll-out of reverse-Polish notation (RPN), a preferred number notation for scientific computations. The sales price was a remarkable wonder too, $4900.

The HP 9100A caught the imagination of the engineering community. I believe one of the first units was given to Arthur Clarke, the acclaimed science fiction writer. The project development was done at HP Labs in Palo Alto, and transferred to Loveland Division, Colorado, for production. This product probably became the first HP product to experience large-scale production, compared to the usual HP electronic measuring instruments. And, without question, it launched HP into personal technical computation.

Cochran designed the algorithms, a monstrous job because of the stringent limits on memory size, and the need to make sure that all calculations were accurate over the enormous range of numerical "arguments." Barney later paid tribute to Dave's genius, "I sometimes wonder if Dave realizes what a remarkable job he did? It took several passes. On the first pass, it appeared hopeless to include all functions."

The HP 9100A led directly to the 1972 family of the HP 9810/20/30A programmable calculators (which would then be called desktop computers), which had full HP-IB automation capability.

The HP-35

At the HPA (components) Division, in early 1970, we had just concluded technical negotiations, for supplying keyboard and digital display light-emitting-diodes (LEDS), with a company called Unidynamics of Phoenix, AZ. And yet, after some months of research and production planning, the company, a division of Universal Match Corp., had decided not to proceed toward building a commercial pocket calculator. The main reason was that their core business was production of super-high-tech electronic sub-assemblies for atomic weapons. Their management had concluded that they should not move into the purely-commercial sector.

But, in retrospect, they had the hand calculator product concept, dead on. They had envisioned a handheld calculator, the size of a king-size pack of cigarettes. It had a 6-digit display, mounted horizontally, and four functions of add, subtract, multiply and divide (what we termed a four-banger). Further, they ingeniously had planned two or four more math functions, to be built into the ROM. These had keys that were to be covered up in the first models, but which could be uncovered, as needed, to confront competition, without having to re-design anything inside, merely put a new faceplate on the front.

The reason Unidynamics had come to HPA, was that we had a first-class ceramic facility for building photo-conductors. Photo-conductors were used for the electronic counters, built by the Santa Clara Division, to decode the BCD logic of counting circuits, into the 10-line code needed to drive NIXIE numeric display indicators. Unidynamics envisioned the need for a cheap keyboard, while at the same time, they needed long-life reliability. A photo-conductor keyboard, with no moving parts, seemed ideal.

Their idea was to use a photo-conductor resistive element at each key location. They had built prototypes, which demonstrated that when a finger was pressed into a shaped depression "key," the photo-resistance would change by about 1 million to 1. They had tested pulse-shaping processing circuitry, to eliminate false key inputs, caused, for example, by shadows of moving fingers across the panel, or variations in ambient lighting.

Unidynamics had done their homework. They had hired an experienced consumer-marketing man from 3M Corporation, who became their V.P. of Marketing. He had lined up international distribution for the product to the tune of 10,000 units per month. The price was to be $200, at the time that Japan's Sanyo Corp. was selling a calculator that was the size of a small cigar box for $400. Japan's Canon was moving to a "pocket" calculator, but it had to be a coat pocket because it was the size of a small book. Their price was projected to be about $250.

In other ways, too, Unidynamics was ahead of their time. They envisioned 4 integrated circuits mounted on the rear of the ceramic "mother-board" which carried the key photo-resistors and display on the front. Production would be so certain, that they planned no re-work, because of the expense, and therefore, would simply throw away bad units.

George Klock was project leader, and visited a number of times, along with Philip Tai, who was project engineer. Naturally, as soon as we in the LED group heard about this large project for keyboards, we immediately began our campaign to change their plan, for displays, to LEDs. We proposed either monolithic numeric chips (later used in the HP 35), or an abbreviated 3 x 5 dot matrix. That deal looked like a huge amount of business. They bought into the LED idea, and there were more visits with Tai, to work out the "LED strobing" technique, which we were just perfecting for use in HP 5300 portable counters.

Suddenly, and without warning, Unidynamics called us with the information that they were suspending their project. Amazing. They had the product concept in their grasp. They were about a year ahead of the market. They would have blown the market away. That year, 1970, Japan built about 150,000 total calculators, desktops, "portables" and "handhelds." The 10,000 per month sales estimate of Klock, for $200 units, was probably low by a factor of 5 or10.

Once we had confirmed that the project was clearly cancelled, we felt we were released from our self-imposed HPA rules, about revealing any sales or technical details to other HP entities. We always held such information strictly confidential during our contract periods. In the components business, such technical and business details were sacrosanct. If another HP entity should discover any such details about a competitor who was our customer, we would lose all our credibility.

The HP 35 Printed-Circuit Boards, showing the click-key assembly and the LED display row.

At the LED department, fortunately, Marketing Manager Milt Liebhaber and Rick Kniss, the LED Product Marketing Engineer, had big plans for outside sales of LED digits, that would leverage on the production of the HP 35. We were sitting around one evening with the thick equipment proposal in front of a group of us, presided over by Division Manager Dave Weindorf. After much discussion, Dave suddenly said, "What the hell, let's make it a proposal for a cool three-quarters million".

So we corrected the equipment lists, to get up to $750K, and went up to see Bill Hewlett the next day. We had our arguments well honed and practiced, for proving the need for supporting the all-important HP 35, with incidental outside sales thrown in. After 5 minutes of preliminaries, presenting our executive summary, Bill stopped us, saying, "I've got another meeting, so is there anything else important, that I should know? If not, let's go with it." We almost wanted to say, "But, Bill, we have all this backup argument material, don't you want to hear it?" It was by far the easiest project I have ever seen sold. But it shows the magic of that calculator project in those days.

In retrospect, it was extremely lucky we sandbagged Bill on the materials processing lab for GaAsP. The 500-1000 per month forecast for the HP 35 was wildly wrong, on the low side. Sales went through 10,000 per month, climbing almost straight up. The reasons, of course, are clear with hindsight. The price was a stroke of genius, because the $395 was easily within the reach of every engineer, not just managers. It was not only a prestigious personal possession, but an amazing drudgery beater. It made better engineers, and it made them faster and more efficient.

The HP 721A

HP Journal, May, 1958

The honor for first transistorized product for HP, was the HP 721A Power Supply. It was a simple concept, a 30 vdc range and 150 ma capacity, with a clever current-limit feature. That safety feature was so critical to a generation of engineers that were just learning that transistors were not as forgiving to electric shorts and other mistakes as tubes. If a test circuit was misapplied, the HP 721A power supply would simply cut off its current to a safe level, that the engineer could dial in. I remember being surprised how many thousands of engineers would buy the product at an introduction price of $145. For only 150 milliamps. But such was the rush to semiconductors of the day.

The circuit design was invented by Malcolm McWhorter and George Barr, both Stanford University professors. They went on to found the Vidar Company, using some of the royalties that were generated from this simple power supply idea. Malcolm McWhorter was my digital circuits professor at Stanford. In 1952, in my undergraduate days at Notre Dame, the transistor was described in mimeographed notes, having just been invented in 1948. So, I really needed to get re-treaded on digital circuit theory, when I hit my graduate school study at Stanford. McWhorter's classroom style was magnificent, and he was perhaps my best technical professor. His explanations and circuit models remain with me today, although I often have to admit that I never really did an honest day of engineering in my life, preferring and enjoying marketing so much more.

The HP 8551/851A

As mentioned earlier, this product put HP into a brand new microwave market area. We had been known for signal generators, power meters and swept-frequency testing before that. But the market for spectrum analyzers was owned by Polarad with about $5 million annual revenue and Panoramic Corp., at about $2 million. Spectrum analyzers had been designed for use with radars during WWII, and Polarad, located in New York City, had been a contractor for the government. Basically, a spectrum analyzer is a tuned super-heterodyne radio, with a swept-frequency or "panoramic" display, signal power plotted vertically against a horizontal scale of frequency.

HP launched into spectrum analyzers, partly driven by pressure from our Field Sales Engineers, who were looking for a new market area. The Polarad analyzers were mostly single-octave-band, hand-tunable klystrons, used as the first local oscillator. The sweeping 2^nd local oscillator (LO) provided the 100 MHz sweep width (dispersion, in the jargon of the day). The Panoramic Company, also on Long Island, was soon building a multi-band unit, capable of 2 to 12 GHz, which down-converted using harmonics of the 1 st LO.

The project engineer assigned to the HP analyzer was Art Fong, who had been recruited from the WWII MIT Radiation Lab. His experience there had involved building signal generators and specialized spectrum analyzers, including waveguide components. By designing a sweeping FIRST local oscillator, using a backward-wave-oscillator (BWO) for the source, it was possible to provide a remarkable 2000 MHz of sweep width. The key technology, needed in the use of the BWO, was the breakthrough use of a tracking, phase-locked, low-noise-VHF sweeping oscillator. The ability to track while phase-locked, quieted the phase-noise of the BWO tube, when in the narrow band sweep mode.

The HP unit was conceived to have a number of significantly higher-performance features, far wider sweeps, an honest and calibrated 60 dB of amplitude range. Yet, it was thus destined to cost much more. In my original market research, I had found that most customers would have been happy to have the same performance as Polarad, but just have a product brand from HP, with HP workmanship, features and reliability. But when we plotted the price-volume curves, the tail of the curve, as usual with most HP products, stretched flat, far to the right. Which simply meant that there would be considerable demand, even with prices that might be twice the available products, or in this case, some real demand even at $15,000.

The HP 8551A Spectrum Analyzer was first demonstrated, privately, to customers during the 1963 IEEE show in New York. Project Engineers Art Fong, Harley Halverson (HP851 display), George Jung (HP 8551 RF section), and others, had finished a prototype, to bring to a hotel room, for showing to VIP customers. The unit was mounted on a draped table, but under the drape was a massive fan, blowing air up through a laundry dryer hose umbilical, to keep the parts cool. It took a lot of air to cool the high-voltage power supply, needed to run the backward-wave-oscillator, requiring 2400 volts.

The HP 524A

A fellowship grant made to Al Bagley, a young graduate student at Stanford University in 1948, led to the development of HP's frequency counter business. Hewlett and Packard personally asked student Bagley to study the measurement needs of the nuclear physics industry. From that study came requirements for a faster nuclear-pulse-counting technology, that could resolve two nuclear events, only 0.1 microsecond apart. Bagley determined that new, low-capacitance semiconductor diodes, just coming on the market might allow faster digital circuitry. He built a prototype as part of his project--and then asked for a job at HP.

Out of that work, came the HP 520A high speed decimal scaler, which was able to condition very short nuclear pulses, occurring at up to 10 MHz. It also divided down the count rate by factors of 10 or 100. Sadly, the 520A had only minimal commercial success. However, Hewlett envisioned a different measurement process, one that gated those scaled-down, high-speed pulses into a slower-speed accumulator (counter). It used a selectable time base, similar to that of the earlier WWII HP 100A time standard. Out of this combination was born the common frequency counter.

Those "combination" frequency counters were a huge commercial success, and in great demand from the 1950's onward. They were used in measuring everything from transmitter frequencies to the accelerometers on which ballistic missile guidance systems were based. HP became the industry leader in electronic counting, in the early 1950's, with the HP 524A frequency counter (ca. 1952), which boasted a 0.01 cps to 10 "mc" measuring range. (Hertz came a little later.)

In 1954, plug-in down-converters were added and introduced as the HP 524B electronic counter, which became an industry standard for some years. Plug-ins eventually measured to 18 GHz, after the introduction of the step-recovery-junction diode. Over the years, this product line generated massive revenues and profits, which funded untold numbers of new products.

This 5-alpha-character cluster used strobing pulses for the display on the HP 9820A, and 9830A Desktop Calculator. Cover of the July 1970 HP Journal, Courtesy of the Hewlett Packard Company

HP Journal, Feb, 1969, July, 1970

In the mid-1960's, HP Labs was working on various display technologies, since virtually every product HP made, needed numeric or alphanumeric readouts. NIXIE® tubes, built by Burroughs Corp. were expensive, and had reliability problems, and needed high-voltage drive circuits, which were not compatible with semiconductor technology. Nor did their high voltage drives lend themselves to portable use. Worse, the technology was available to everyone, and products tended to look the same.

Egon Loebner had joined HP Associates from RCA, and had brought with him, a lot of knowledge of electro-luminescent technology. It wasn't long before HP had work going on with so-called III-V compounds. These roman numerals designate the valence of the elements, in the periodic table. In particular, a certain mixture of Gallium, Arsenic and Phosphorous atoms could join properly in a crystalline structure, and then be configured into an electrical junction diode. If you could do everything right, the diode emitted visible red light at 655 nanometers wavelength, when forward biased.

By the late 1960's, the HPA division was buying crystalline Gallium-Arsenide in cigar-sized boules, slicing, polishing, and growing epitaxial layers of arsenic-phosphorous on the top. It must be remembered as a true black-magic, arcane art. At that time, I was burning out in Microwave Marketing, and asked John Young, to be on the lookout for a somewhat-more technical job, that might appeal to me, and to be a slower pace. It took about two days.

John proposed that I go down to HPA, to head up the LED display group, since it was struggling for direction. Its process shop was tiny, and needed clear product strategies to set materials and assembly direction. I arrived at HPA, to find a department with about 8 people, two equal-level managers, each with a business card which stated they were the official department manager. Apparently HPA Division Manager Don Smith had brought both down from HP Labs, with the promise that each would manage the group, which was the job I just got.

Within two days, the manufacturing man, Gerry Pighini, had resigned. He wanted to move to Fairchild to be with his poker-playing buddy, John Attalla, whom had just left HP Labs. Within about one week, Division Manager (my boss) Don Smith, himself, resigned to join venture capitalist, Jack Melchor. So I started from there, with my new boss, Dave Weindorf. We had one display product, the HP 7000 numeric display, a 5 x 7 matrix character. Using an on-board integrated circuit chip, it decoded the BCD drive data and provided current for each of 21 (abbreviated from 35) diodes on the face. Each die was 10 x 10 mils.

List price of the 7000 was $35 each. That made them interesting for ultra high-reliability, military applications, but a tough sell for commercial. Yet, we soon landed a surprise contract from the Longines Watch Company, in Switzerland. They bought 1000 sets of 4 digits for a handheld sports timer. That single order of $100K, at a negotiated price of, I believe, $25 each, set us on our way. I believe, in the first year, the department had total sales of $250K.

We immediately began working within HP, to find instrument applications, but the price was impossibly high. So we moved out in three directions,

1) a simpler 4 x 7 dot-matrix character in cheaper plastic package, and with a latching on-board driver IC,

2) a full 5 x 7 dot-matrix with 3 to 5 digits per package, and with external strobing circuitry, that would support full alphanumeric characters, and

3) 1/10^th inch monolithic digit clusters of 5 or 7 digits, for miniature digital displays (usually with a simple plastic magnifying lens), ultimately used in the HP 35.

With a lot of missionary work, our product strategy began developing a diversity of applications. The full alphanumeric characters got accepted into the new HP 9820 and 9830 desktop computers from Loveland Division, although the monolithics didn't win in the HP 9810. Instead, Fairchild got that job.

A 4 x 7 dot-matrix modified version of the full-strobed numerics did get chosen for Jim Sorden's HP 5300 family of portable frequency counters. That helped us a lot with engineering refinement of the strobing concept. Joe Diesel and Bob Steward worked the technical applications. Vic Wrooble worked on ceramic materials. Howard Borden ran the engineering, and was a real pro on the esoteric display technology principles.

Howard had worked in the semiconductor industry for some years before coming to HP Labs, and subsequently to our LED group. Interestingly, he fought a lifetime of ultra-sensitivity to certain chemicals used in semiconductor wafer processing. Chemicals like TCE, Tri-Chloro-Ethylene were very common in the industry, and ended up in many toxic plumes underground, which polluted the Bay Area's water supplies.

Our own operation at the intersection of Page Mill Rd and El Camino caused quite a large amount of TCE contamination. The cause was that Palo Alto city code regulations required that the inflammable TCE tanks be put underground, but then they leaked. Our failing was that we didn't have the foresight to install double-walled, leakage detecting tanks. In any event, Howard's super sensitivity remains with him today, even though he has been retired in San Diego for two decades.

One problem with HP applications was the pre-introduction product reviews that each division would hold for Dave and Bill as a final approval meeting. LED's of that day had mono-chromatic light that was SO red that it was at the edge of the visible spectrum. For older people, the tiny red surfaces that emitted light looked blurry, and Hewlett often complained of how uncomfortable it was for him to focus on the numeric displays. Fortunately, he didn't kill the projects, although the same problem surfaced later when the HP 35 calculator was designed.

In the manufacturing area, it was a different story. The process technology necessary to get light out of semiconductor diodes was nasty. Assume it was a process of 32 complicated steps. Slice into wafers, polish, grow epitaxy layers, diffuse zinc, deposit contacts, dice, assemble, etc. And at all steps, clean, clean and re-clean. It may take 4 to 5 weeks for entering material to hit the tester, to find out if you were going to get light.

Grown single-crystal material was purchased from Monsanto and Bell and Howell. In those days, we were happy to get a cigar-shaped boule with a diameter of 1/2 inch. When sawed, you'd get maybe 1/3 square inch of working material. After processing, the outer 20% of each wafer usually wouldn't work, so even though the die were only 0.010 x 0.010 inches, you still wouldn't get many, and for those the brightness was so variable that they had to be measured and selected and matched for brightness so that matrix digits would have even brightness. The human eye is a tremendous brightness error detector.

I can clearly remember the day, probably in late 1970, when we had a coffee-time cake ceremony. On the wall hung a red cardboard sign, which was just 1 square foot. We had calculated that our entire first 1.5 years had been spent manufacturing only that much material area. Some years later, after Ted Larsen's materials people began growing their own Czochralski crystal "boules," or ingots, they turned out a square foot of emitting material every few hours.

Moreover, by the turn of the millennium, there were spectacular technologies which produced red lights, green lights, and yellow lights. They served brake lights in cars, stop lights at street corners, and everywhere a reliable cheap and sun-insensitive light was needed. And now there are "white LED" lights bright enough to serve as flashlight bulbs. Absolutely stunning, when I think of us struggling to figure how to make a somewhat bright far-red LED in 1970, which couldn't be read in sunlight.

A lot of credit for building the massive production capacity for HPA, goes to Bob Zettler, who had moved from HP to Fairchild a few years before, with John Atalla. He returned to HP in 1972 and took over the department after I left. As I mentioned, the reason I had asked John Young for a transfer from Microwave Marketing Manager, was to find a more relaxed pace, and get off the management treadmill. What I found was a terrifically challenging and technologically fun department, but in less than 3 years, we went from 7 people to 105. Another treadmill.

There was also the matter of a totally different personnel culture. HP's instrument culture developed over decades from an almost total dominance over the test and measurement sector of the industry. Thus the environment inside was relaxed and confident, not quite arrogance. There was still stress on project schedules and meeting budgets, but the daily environment was not cut-throat and managers were not sharks.

The semiconductor industry, on the other hand, was total stress. The technology was moving at warp speed, and managers were super-aggressive and difficult personalities. My first introduction to the change I was to face, was a product strategy meeting I attended at HPA, with Milt Liebhaber, the division marketing manager, Dick Soshea, the central R&D manager, Rick Kniss, the product marketing engineer on displays, and several others from my Display Dept. I don't believe that Dave Weindorf, the Division Manager was there.

As I walked in, Milt and Dick were on opposite sides of the table, standing up, and one was pounding his fist on the table, both were yelling at each other. One was smoking a large cigar, and it was what I thought to be a generally disagreeable attitude of the participants. Unfortunately, I was to learn that such behavior was inherited from the semiconductor industry, by way of our hiring people who had previously worked in those aggressive companies. I never got comfortable with those personalities, and that was one reason I moved on after less than 3 years there.

The period around 1970 exhibited a severe economic downturn, and the instrument business was hit like most high-tech companies. The Microwave Division had split, with half moving to Santa Rosa and the remaining group re-named Stanford Park Division. Overlaid onto half the products moving out, the remaining products hit a downturn in sales. In spite of many exciting new products, there were still production employees who had nothing to do. This was the time Bill Hewlett introduced his plan to reduce our production capacity by the simple expedient of taking forced time off every other Friday. Immediately reducing production capacity by 10%.

It wasn't quite enough, and the SPD production managers were looking for places to move some of their people temporarily. Since our LED group was expanding, we agreed to take on a dozen or so instrument assemblers, and make them semi-conductor assemblers. These are people who work all day looking through a microscope, manipulating ultra-tiny parts with tweezers and mechanical levers. It is boring, yet intricate, and tedious work. In the semi-conductor world, hundreds of such workers, sit shoulder to shoulder, for 8 hours.

Needless to say, those transplanted microwave employees got tired pretty fast, and yearned to return to SPD, but couldn't until the business turned around about a year later. Then we would go out and hire assemblers from Fairchild, who thought our production habits were easy, and indeed they were. We were somewhere between the slave-like conditions of Fairchild, and the more moderate conditions of instrument assembly.

The HP 150A

HP Journal, April, 1956

In the 50's, many of the independent HP Reps carried the Tektronix (Tek) product line along with HP. Tek's oscilloscope products were top of the line, and had built an enviable reputation for superb quality and reliability. The inside product design was a dream, with a special ceramic standoff resistor mounting strips and carefully dressed wiring harnesses.

Tek made a decision around that time to take their product sales "direct" meaning no reps. They began on one coast by writing letters of resignation from their business contracts, which were typically a 30-day notice by either party. In the case of HP, and the world-wide sales representatives, we understood that most of those business relationships were on the basis of a handshake and personal friendship, not a written contract.

One after another, as the rep was cancelled, Tek brought in their own people to set up a sales organization. Well, the loss of a general-purpose product line like scopes was serious matter to the HP reps. They prevailed on Dave to agree to design a line of oscilloscopes to compete with Tek. And thereby began decades of frustrating competition. The first two products were a good news-bad news situation. The HP 130A was a slick dc to 300 kHz low-frequency scope that filled a niche that Tek didn't occupy at that time.

But the product that was meant to take Tek head on was the HP 150A. It was a plug-in modular design with dc to 10 MHz, dual channel, which at that time was deemed general-purpose. To avoid a me-too verdict, HP designers chose to use a bottom drawer plug-in which made the plug-ins non-compatible with the Tek design. Some innovations on the CRT made it easier on the eyes. There was a clever "beam-finder" button, which alleviated the nagging problem with the Tek units of losing the trace when overloaded.

But the poor HP 150 suffered the worst reliability problems. First off, even if perfect, customers would have trouble believing that anyone, even HP could measure up to Tek. They were like Gods. The Tek salesmen had all been trained to be scope repairmen first, salesmen second. Tek President Howard Vollum believed that each and every FE must be able to align a scope and most carried small screwdrivers to tweak the amplifier frequency response of a bench unit as they talked with customers.

It took some years for the HP 150 to become more reliable. Meantime, HP was making more innovations as time went on, such as internal graticules on the CRT. This feature was made possible by the decision to build a CRT manufacturing facility in Colorado Springs when the factory moved out there in about 1960.

The HP 185A

HP Journal, Jan, 1960

With a market and technology dominated by Tektronix, HP laboratories was always looking for a new unique oscilloscope technology to plant our product flag in the sand. This came with the step-recovery diode technology, and some signal sampling technology which had been used earlier by some nuclear research at the U.S. atomic operations on Long Island. With these, Barney and his lab began an urgent search for a sampling oscilloscope, which could achieve the equivalent frequency response of 500 MHz.

Although Tek had built some very high priced scopes that did the same job, they were brute force, distributed-amplifier designs which were not welcome by customers. Sampling theory is based on repetitive signals such as a continuing pulse train, which is quite typical of test signals to be analyzed. The sampling point of an extremely narrow sampling window is slowly moved along the repetitive test waveform, and by sequentially plotting those points, a continuous waveform results, exactly reproducing the desired test signal.

When the sampling scope was under active development in the Palo Alto labs, the HP Labs was interviewing a manager prospect who worked at Hughes Aircraft Co. Norm Winningstad was a super-star in their display components labs, which were used in their radar system contracts. It was felt that Winningstad would be valuable in the continuing development of HP scope projects like the sampling scope, which was way ahead in technology. So the lab showed him most of the progress of the project to date. For some reason, I was asked to be one of the interview team. We recommended a hire vote, and an offer was made before he left.

Imagine the chagrin of all of us, when we found out shortly that Winningstad had flown further north, after the HP visit, to Tektronix, and was hired there. The worst was to come, for about 6 months later, as we introduced the HP 185A at Wescon, in Los Angeles, that year, we saw that Tektronix had quickly cobbled together their Type N, a sampling plug-in product for their regular 530 series, which purported to perform to the same sampling specs as our much more refined product system. It was disappointing to see a professional use his HP interview as a spy mission, but such is life. Tek never did catch up with HP on sampling technology, and our products continued to advance.

By the way, the HP promotional gimmick for introducing our first 500 MHz sampling oscilloscope at Wescon that year, were 500 MHz "bills" printed to look like paper money, and handed out on the streets outside the conference. This caused a bit of a problem with the show management because so many of the bills were cluttering the streets as litter.

Wescon also saw HP's first use of commercial billboards, which were rented at sites close to the Convention Center, to announce the new HP 8551. Somewhere, there is a picture record of a screwup by the billboard hanging crew, and I think it might be in the possession of Dean Abramson. The picture shows that several panels were hung in the wrong sequence, such that part of the message was spelled, "Anal Spectrumyzer." My old memory is fading but I think I recall seeing that picture.

The HP ThinkJet Printer

HP Journal, May, 1985

For some years in the 1980s, Barney set up a practice of holding an annual HP Labs Technology Show, in the Bldg 5M auditorium. This was intended to show the employees just what a diverse and dramatic series of R&D projects were underway in his Labs. It was always exciting to attend and browse through the wide variety of innovative and powerful new products-to-be.

About a year before the ink-jet printer was introduced, HP Labs just showed a technology demo, with a print head driving back and forth, and printing a line of alpha characters. Most of us could hardly believe our eyes, and when told that the engineers had succeeded in blasting millions of tiny droplets of ink out through microscopic holes in a process that happened in microseconds. For us engineers, it was hard to conceive that tiny resistors could be heated fast enough to make tiny droplets of ink explode out of tiny holes at thousands per second. But there it was, writing in front of our eyes.

About a year later, the HP ThinkJet printer was introduced, and as they say in the press, the rest is history. HP's dominance of the printing technology is still held today. And, whether we intended it or not, we got ourselves into the "razor blade" business. You recall that selling consumables like razor blades can be so profitable that you can afford to give away the razor for the revenue you get on a constant stream of sales of razor blades. HP's revenues, which come from the ink cartridges, is not ever revealed separately in their financial reports. I do recall being told by someone, before I retired, that HP ships an amount of ink (in cartridges) annually, which would fill 7 railroad tank cars. I feel quite sure it is much higher these years.

The HP 8410A

The HP 8410A Network Analyzer product line of 1968 ushered in a revolution in RF & Microwave component design and test. Multi-band sweeps of Smith Chart characterizations of components and systems arrived at just the right time for exploiting the hybrid microcircuit technology for instruments and communications systems. Paul Ely had coined the product slogan, "Stamp Out Slotted Lines," and the project became an industrial legend.

There was a confluence of technologies in the late 1960's, which resulted in the HP 8410. The step-recovery diode (see below) gave HP a unique sampling technique, and resulted in the powerful HP 185/87 sampling scope. Then the sampling technique was used in a dual-channel RF voltmeter, the HP 8405A, which also measured the phase angle between the two RF signals. This gave RF circuit designers considerable insight into circuits they were extending up to 1000 MHz.

HP then popularized some "homebuilt" reflectometers with dual-directional couplers, which gave customers the ability to measure signal reflections including their phase angle, using the HP 8405A to 1 GHz. But you still had to plot the Smith Charts by hand from the data. When that technique began to catch on, we knew the HP 8410 was on the right track. Meantime, the HP 8708A frequency synchronizer, which disciplined the HP 608F, taught us how to do signal phase-locking to 500 MHz.

Then came the ability to create phase-lock loops, using sampling diodes to lock quiet VHF local oscillators to track sweeping microwave test signals. And the same sampling diodes were employed to be the dual-downconverters of the HP 8410, yielding a two-channel super-heterodyne receiver which could measure complex impedance to 12.4 first and later 18 GHz. Combining amplitude and phase into Smith Chart type displays gave powerful insight into microwave component design and production. And the signal separation boxes with dual couplers put it all into a nice comprehensive package that truly did, "stamp out slotted lines."

HP continued to dominate the vector network analyzer category for decades, with lower-frequency range as well as higher-capability products. The HP 8753A vector analyzer (ca. 1987) covered up to 3 GHz. Sometimes called the HP 8510's little brother, it was the first low-cost VNA operating below 3 GHz. This positioned it perfectly for cellular design and test engineers working in the 800-900 MHz, and later 2400 MHz range. It was also the first RF analyzer to have complete built-in error-correction.

The HP 8540A

HP 1968 Catalog

The great-grandfather of Agilent's 8510C Microwave Network Analyzer was the HP 8540A Automatic Network Analyzer, which was a two-cabinet system introduced in about 1968. It was the HP 2100A computer programming the HP 8410 and several of its plug-ins, the microwave sweeper signals and the transducer signal-separation boxes. It included complete operating software for data acquisition and manipulation. Actually, the first 8540A system only had a teletype for the human interface. The HP 8542A which followed was more refined, with a real CRT display and better software.

Both systems had fortuitous timing, since microwave microcircuit-on-sapphire technology was booming (and blooming) and commercial communication satellite technologies were emerging from the Russian Sputnik surprise of 1957, and the U.S. response of their Apollo moon-shot program.

It can be easily argued that satellite technologies have revolutionized the world's scientific and business accomplishments. Weather satellites provide not just tomorrow's meteorology predictions on the evening news, but give crucial warning of gathering hurricane dangers; mapping satellites sense agricultural and mineral conditions on the earth's surface. The Global Positioning System (GPS) determines navigational position anywhere on earth within 20 feet; and maritime and military satellites provide powerful command and control of our defense establishment. The Hubble astronomical telescope searches for the beginning of our cosmos and the Big Bang.

In the late 1960's, I was flying home from a sales trip to Washington, DC. My flight from Dulles was delayed by one of the jet engines spewing raw fuel onto the tarmac. So, all passengers were removed and sequestered in a metal shed out by the flight line, probably to keep us from finding an alternate flight. I struck up a conversation with a fellow passenger, Harold Rosen, who turned out to be the newly-promoted Division Manager of Hughes Communication Company of Culver City, CA. Hughes had developed an awesome communication satellite payload technology, and was already the recognized leader.

After Rosen found out I was from HP, he revealed that he was carrying home a $69 million dollar contract for the first Comsat communication satellite program. He had just signed the contract with the Comsat Consortium, which consisted of 1/3 U.S. Government participation, and 2/3 from other private communication companies.

In those days, HP employees were permitted to fly first class, if the flight left after normal business hours, so I, as Microwave Division Marketing Manager, had myself ticketed for first class. As we discussed their contract, Rosen noted that he had committed to a penalty clause of $100,000 per day for late launches. It was common knowledge that the inside of their satellites were a rats-nest of dozens of amplifiers, filters, switches, signal processing components, and a maze of cabling, designed for redundancy and reliability. However the bad news was that, if one part failed, at pre-launch test, they would have terrible delays not just to replace the failed part, but re-test everything to assure integrity of all the hundreds of signal path permutations.

The HP 8540A Computer-Network Analyzer System turned out to be a perfect match, even at $200,000 each. It was our first use of the new HP 2116A Instrumentation Computer, which programmed and controlled the signal generators, routing signal switches and network analyzer parts, and collected and corrected all the valuable test data for presentation to the test engineers. And although it was never intended to test satellites, Rosen was highly interested because of his required all-up-around tests on a complete system--especially with his ruinous penalty clauses.

By this time, after a four-hour delay, we got the re-boarding call at midnight. Since I wasn't finished with my sales pitch, I asked Rosen if it would be OK if we talked further on the flight. He said OK. Imagine my surprise when I flopped down in first class, and Rosen walked back into coach class! Carrying a $69 million contract, no less. After takeoff, I asked the stewardess if I could go back to coach, and Rosen and I spent another 2 hours on the jump seat in the rear, where there was an overhead light, to finish with everything I knew.

If I remember correctly, Hughes bought 3 or 5 of those magnificent systems. And, they launched on schedule, a testimony to their engineering prowess. Hughes went on to become a huge merchant supplier for communications systems to countries around the globe, delivering hundreds of flying birds, each with massive capacities of thousands of voice channels. And that HP system never developed any competition. But in its own way, it helped change the world.

The HP 434A

HP Journal, August, 1958

Power sensors using devices like thermistors were limited to about 10 milliwatts on the top end. The need to measure higher powers was clearly necessary since most radar and communications systems used watts and kilowatts. Sometimes this could be handled with directional couplers or accurately-calibrated attenuators, which would sample perhaps l/l00th of the main line signal. So it was deemed necessary to build a unit capable of directly absorbing 10 watts, which is where the HP 434A came in.

The HP 434A was a miniature plumbers' nightmare. To dissipate the incoming 10 watt signal, a flowing silicone oil stream was envisioned. Further, to provide an insensitivity to temperature changes and outside environments, a balanced arrangement was used. The same oil stream flowed past the microwave resistor that handled dc to 12.4 GHz signals and proceeded to a temperature sensor. Then after a heat exchanger it went on to another termination resistor and temperature sensor, that accepted only a balancing dc power. By supplying the dc power in proportion to the unbalance in two temp-sensors in a single oil path, a measurement of the dc power equaled the microwave power.

It worked great and the dc substitution made it quite accurate. It had little drift, which was a considerable problem on the previous HP 430/478 power bridge thermistor technology. The idea of measuring medium power directly without resort to couplers or pads appealed to engineers. The product was a commercial success.

Enter the USAF. These were the years of the fear-the-Russians mode. The U.S. was busy building BMEWS, Ballistic Missile Early Warning System, three massive radars in the Arctic, pointed north and looking for inbound missiles coming over the North Pole. RCA was prime contractor and GE was building the radars. These monsters had antennas the size of a football field tilted up on its side. The radar transmitters themselves, I believe, were 25 Megawatts.

Everything was full military specification design. Including the power monitor system they conceived to monitor the radar system power on a 24 hour basis. For this, they obviously needed a power meter which wouldn't drift, because obviously they couldn't shut down the radar to balance the sensor. The HP 434A was the choice.

The problem; although the commercial HP 434A unit was highly reliable, one reason was that the pump that moved the silicone oil in its path was an ingenious arrangement that didn't depend on a shaft and seal into the oil compartment. Instead it used a magnetically-coupled impeller, operating through a ultra-thin brass shim. The silicone oil path was completely sealed in, no oil leakage.

Well, in the infernal wisdom of the military procurement, GE was told to insist on a "military specification" motor. This in turn required an organic seal on a rotating shaft of a "military specification" pump. The seal and the nasty silicone oil weren't all that compatible. And we looked for future problems. But it met the mil-spec. Confirming that fact, the front panel was specially designed with a golden-colored alodyned aluminum sheet, that was 24-inches wide, and 1/2 inch thick to meet expected shocks to be caused by presumed nuclear blasts outside.

Now, flash to years later. At a trade show, I met an RCA field engineer who had spent some years of his life on a BMEWS station site at Farthingdale Moor in the UK, working on maintenance for the radars. He described the site. The radar transmitters themselves looked like the generator room of Hoover Dam, in a structure as long as a football field. Rows and rows of radar cabinets, high-power ancillary equipment, and rows of instrumentation. And about 15 of the HP 434s sprinkled throughout the floor, some operational, some hot-standby units.

He said he'd be standing on a catwalk, and begin to hear a faint piercing squeal from somewhere on the floor. It would begin a resonance that began to drive the cabinet sheet metal into a resonance. And the offender was the HP pump seal. They would track down the bad unit and with a little oil can full of lubricating oil hit the organic seal and be ready for several hundred hours more of quiet.

The HP 344A

In the early 60's, Bell Telephone (BTL), in their military equipment division at Whippany, NJ, had contracted for an enormous task. They were to build the acquisition radar and the site radar for Nike Zeus, a perimeter defensive missile that was to be placed around hundreds of U.S. cities for defense against the terminal phase of inbound nuclear missiles from Russia. I believe that the ultimate deployment was to be 50,000 missile units and radars. With a project of this scope, operator automation was crucial, and BTL envisioned using an automatic noise figure meter to monitor the radar receiver sensitivity performance, while in full operation. In radar receivers, one important measure of receiver performance is a parameter called noise figure, essentially an indication of whether it was losing its sensitivity to find targets among noise and clutter. Several years before, we had introduced the HP 340A and then B for commercial applications of making NF measurements. It was built with vacuum tubes.

With the possibility for landing such a whopping contract, New Jersey Field Engineer Bob MacVeety prevailed on Microwave Division Manager Bruce Wholey to develop a transistor version of the HP 340B. The HP 340B, being a vacuum tube model, was not suitable for long-term unattended operation that Bell Labs envisioned. And with the bait of 50,000 units in the future, we were beginning to talk of a future need to build the Bob MacVeety "Memorial Noise Figure Building." Actually, that was some time before Building 5 was contemplated on the Stanford hill site, but would indeed have been Bldg 5.

Well, the upshot was that HP decided to design the new transistorized unit on the basis of an order for about 50 units. And as it turned out, that is all the orders that ever came from that deal. Nike Zeus went away as political administrations changed and politics made such defense systems to be deployed in populous areas extremely unpopular.

The good news was that the engineering efforts needed to design this transistorized instrument became a terrific learning experience for a whole team of excellent engineers. Marco Negrete, Nick Kuhn, and Phil Spohn were all engineers who cut their technical teeth on this difficult design project. All went on to greater achievements, and designed significant new reliable transistor designs, although most of that early transistor technology was built on germanium, and some re-learning was needed when silicon took over later.

The HP 5060A

In the post-WWII world of the 1950's, the business and scientific community immediately set to work to commercially exploit many of the scientific breakthroughs which came from wartime developments. And one of the sectors of greatest promise was the area of communications, including everything from wireline to cable multiplex to tactical and mobile technologies, short-wave channels and microwave line-of-sight links. Satellite technologies were waiting just off stage.

The basis of most communication links, which consist of a transmitter and a receiver, is that the transmitted signal containing the modulation and intelligence must be highly stable and in close synchronization with the receiver. (In the receiver, this synchronizer is called the local oscillator.) The most popular way of accomplishing this "signal lock" during that era was to generate the transmitter and receiver signals, and discipline them with quartz crystal oscillators. When properly fabricated, quartz has the property of providing signals of excellent stability related to the physical properties of the tiny slice of quartz.

As accurate and stable as quartz technology can be, it still had one prime failing. The signal still drifts according to temperature changes and physical conditions and simple aging. One method of re-calibrating those millions of drifting oscillators was to periodically compare the operating systems to several standard signals transmitted by the U.S. National Bureau of Standards (now the National Institute of Standards and Technology) from Boulder, Colorado, on Station WWV. The Bureau's "clock" was in turn referenced to the U.S. Naval Observatory in Washington, DC, which by law was designated as the nation's timekeeper using astronomical readings as their standard.

The Bureau's early clocks were a family of unique, pampered, quartz oscillators, cross-compared and monitored continuously. But as scientific progress in atomic resonance advanced, it was early realized that if oscillators could be synchronized to the atomic resonance of certain elements, they would achieve a fundamental standard from nature. Highly specialized systems were developed at NBS, and put into service as the nation's time standards.

Meantime, Hewlett-Packard concluded that there was a large market for a commercial version of the atomic oscillator, and proceeded with development of the HP 5060A Cesium Beam Standard. It was based on synchronizing to the microwave frequency of one of the atomic resonance lines of the element Cesium. It had been designed to be as compact as any bench instrument, and capable of rough handling and rugged environments, even including an internal battery for mobile operation between the times it could be plugged into ac line power.

Len Cutler was the HP father of atomic time standards at HP's Time & Frequency Division, which later became the Santa Clara Division. While he was a world-class intellect and highly productive inventor, Len would also win an award for "The R&D Lab's Nicest Guy." He always made himself available to answer questions from marketing and customers.

The proof was demonstrated in 1964, with a well-publicized experiment, whereby two HP engineers flew from the U.S. Naval Observatory to the official world timekeeping laboratory in Neuchatel, Switzerland. They configured two "flying clocks" and purchased first-class tickets in an international airline, making sure that suitable sustaining power was available from airplane power supplies.

HP soon became the world's timekeeper, with a continuing line of industry-leading atomic standards. Today, Agilent's cesium clocks provide 80 percent of the weighting of Universal Coordinated Time (UTC), and virtually define the world's standard second and the stability of the atomic second. For decades, Agilent has led the industry with superior cesium standards which have advanced communications on a broad front. In fact, it can be immodestly argued that without Agilent cesium products, most of the world communications would stop cold, including our wireless communications of the 21 st century.

Satellite technology is particularly indebted to stable frequency. For example, as a moonshot rocket leaves earth, navigational systems count wavelengths of the disappearing transmitter to measure its distance from earth. All scientific satellites require precision frequency control to achieve their objectives.

On earth, frequency synchronization of the international communications networks and the Internet system are based on lightwave and fiberoptic technology, and have at their heart, Agilent Cesium standards. Broadcast television and cable systems are likewise synchronized with Agilent standards. Even cellular communications and the thousands of base stations at the center of each cell need long-term stability and reference to cesium standards. Military tactical and strategic systems, with their massive information transfers depend on Agilent cesium products. And most of the world's metrology laboratories which keep time and frequency for their companies and national standard laboratories are running on Agilent cesium standards.

If there was ever a single Agilent product with world impact, many would nominate our decades-long love affair with cesium. It's nice to have a unique product and one which customers love. Sadly around 2005, Agilent management decided to sell off the time standards product line.

The HP 2116

HP Journal, Mar, 1967

Some things happen by serendipity. Sometimes faint visions assist in design or business decisions. I believe that the HP 2116A came about that way. Certainly the IBM mainframe business was rolling along in the mid-60's. Instruments were getting digital, typified by the HP 3440A digital voltmeter, and its sales had taken off. Data logging was only as sophisticated as the HP 560 or 561 printer that displayed the readings of our frequency counters.

Comes a visit from a man at a company in Atlanta, who was working on a mini-computer concept. It was probably Dave Packard who realized that there was going to be an intersection of computer control and data processing functions that could leverage the power of instrumentation. In 20/20 hindsight, it seems simple, but my guess is that the alternative at the time would have been to use data logging instruments and relay programmers for sequencing the control of instrument functions. Indeed, military test systems used that very technology, paper tape sequencers for control and paper tape punches for recording the data.

The HP 2116A brought computed measurement and control right down to the engineer's test bench. The so-called "minicomputer's" design strategy was to provide an accessible card cage on the bottom deck of the product, which allowed individual programmable instruments to be connected to the computer. Each interface card controlled one instrument, either stimulus instruments like signal generators or dc power supplies, or measurement instruments such as fast DVMs, counters or spectrum analyzers.

From the year-2006 perspective, these automatic test systems seem primitive indeed, boasting of a human interface which tolerated a clacking teletype with punched paper tape. The original HP 2116A had a memory made using ferrite core technology, with a basic memory size of 4096 16-bit words (expandable to 8K, at $1 dollar per byte.) Storage disks came along a little later, and finally it was modernized with monitor terminals using CRTs.

The real contribution of the HP 2116A was that it was designed so that it didn't need to be coddled, as were most business computers of the time, which existed in temperature-controlled rooms. It was designed specifically for the hostile environmental regime of the factory floor.

The 2116 was extended upward into service as a moderate-priced time-share server, at a time when the General Electric company had created a very user-friendly software application called BASIC. This was quickly accepted by the software-writing professionals, and adopted by the academic community, and ultimately by industry. HP saw immediate applications for BASIC for use in instrument drivers, and appropriate instrument commands were invented to program such instruments.

That system became the HP 2000 Time Share System. But the general explosion of HP's time share revenues was so high it misled our marketing people into thinking that that business was going to go on forever.

Early HP-IB compatible instruments and controling computer

HP Journal, Jan, 1975

The high cost of HP's powerful mini-computer instrument systems were not available to the ordinary engineer, many of whom still needed such automatic measurement capabilities. It was the confluence of programmable desktop calculators (circa 1972) such as the HP 9810/20/30 and a team of creative brains, which changed things for the better for thousands of HP customers.

George Stanley related to me the history of the beginnings of HP-IB. "In the early 70's Bob Brunner attended an internal HP R&D conference at the Broadmoor Hotel in Colorado Springs, CO. At this conference, Bob said it would be very valuable if HP could find a good way to connect instruments to small desktop calculator/ computers like the HP 9820 and the new HP 9830. Some limited BCD interconnections were considered not robust enough."

"Dave Ricci, SCD, took the job of working it out for high speed instruments like counters and Jerry Nelson, LID, did the same for slow instruments like voltmeters (ac voltmeters took only a few readings/sec). This is how the three wire handshake came about, with a technology allowing the speed to self-adjust to the bus speed of the slowest ACTIVE player in the system. Don Loughry concentrated on the industry standards side of things, and I developed the marketing/ training plan. Of course all our work overlapped, but Ricci is the owner of the three-wire handshake."

The team's vision was to develop a system approach to programming and controlling many of HP's individual instuments such as signal generators, signal measurement products and some unique switching matrixes. The concept was a "party-line" communication bus, which interfaced with some added control and data circuitry inside each candidate HP-IB instrument.

By developing the system in conjunction with an IEEE/Industry committee, this "open system" created a powerful design capability for applications ranging from research and development to production and support and even field test systems.

The attractiveness of the system depended in good part on the fact that instruments from many manufacturers with HP-IB functionality could be "daisy-chained" together. This permitted ordinary engineers to assemble stimulus signals and measurement products such as voltmeters, analyzers and others. The bus strategy included bus protocols such as handshake and other signal functions along with the data transfer. The resulting data could be conditioned, corrected, manipulated, analyzed or stored.

As the IEEE-488 standard achieved wide global acceptance, it became known as the General Purpose Interface Bus (GPIB). The functionality was designed into most new medium-complexity instruments. It exists to this day, even as much more powerful desktop computers came into being, and even though many of the newest instruments contain their own powerful computational microprocessors, they are still linkable by the GPIB. This permits customers to automate production testing and accumulate test data centrally.

George attended Dave Ricci's retirement party at HP Labs in 2004, because he had something to give Dave. It was an old, actual Washington State license plate that read HP-IB, originally displayed by the HP-IB field specialist in Washington. It was given to George at his retirement, several years before, because of his involvement with HP-IB. But he always felt that Dave was much more entitled to it. Here is George's citation to Dave:

". . . . .I'm not sure everyone here understands the significance of what Dave did when he came up with the three-wire handshake for HP-IB. Very briefly, we had slow instruments like AC voltmeters from Loveland that took 1-2 readings/sec. Then we had Bagley's high-speed counters from Santa Clara that took 1 thousand or more readings/sec.

"If we had gone with the clocked bus we first considered, we would have been frozen in at the slower rate and could not have changed upward as electronics speeded up with ICs. and better circuitry. Ricci's flexible speed, three-wire handshake allowed the bus to run at the speed of the slowest ACTIVE device and slow instruments could be easily kept off the bus until they were needed by simple programming and then turned on to do their thing. This is why Dave holds the patent on the three-wire handshake." As George described Dave's contribution, MC Tom Saponas's head was doing an up and down 'yes.'

In later years, a simpler communication bus was introduced, called the HP-IL, or HP Interface Link. It utilized a simpler bus structure. Both served different customer needs.

The HP 312B and 313A Tracking Generator

HP also became the world leader in exploiting phase-lock loops for frequency control and phase-disciplining of programmable switched microwave oscillators. Using step-recovery diodes to provide rich harmonics from stable and low-phase-noise LF oscillators, and synthesis techniques such "divide-by-n," sophisticated phase-lock loops were used to discipline MW oscillators and reduce their phase noise. The product features that they provided were exact and programmable output frequencies, of exceptional resolution, and superior phase noise performance.

In an associated application, and certainly an example of Bill Hewlett's "next-bench syndrome," HP innovated a new phase-loop stability measuring technique using the previously-mentioned HP 312A selective voltmeter. With an associated tracking generator that produced a signal equal to and tracking with the tuned frequency of the HP 312A, this combination analyzed the closed loop performance of feedback loops, validating the fast switching response and loop stability under wide environmental performance.

The HP 33312

The next clever design was a programmable step attenuator, which used two of the switches, to switch each attenuator element, in and out of the transmission path. Thus were born the HP 33330-family of step attenuators, 0-120 dB. Later as the more-compact HP 33320-series attenuators, they were used in all signal generators, spectrum analyzers and sweepers to achieve programmable signal control. Then came the HP 33312-family coax switches, which were smaller, 26 GHz, and boasted 80+ dB isolation and 5-port capability.

I was writing a simple application note (AN-332) about that time, which was intended to cover simple applications considerations for using those simple coax switches. I'm not sure when the concept came to me, but I realized that the HP 33312 four-port switch could be sold to microwave automation customers as a "transfer" switch, which could program an external component, say a filter, into and out of a transmission line. So I wrote that application into the application note.

Then I noticed that those same 4-port switches might be used as the "cross-point" switches in a full-access signal matrix. For example, a 4 x 4 matrix (4 input lines and 4 output lines) would require one HP 33312 at each of the 16 line intersections, actually only 12 altogether. So, I included these simple application extensions in an add-on application note, AN-332-1, so customers could build their own full-access matrixes. Soon, customers were requesting that HP quote on building complete matrixes, and we had ourselves a new market.

Some years after retirement I visited Santa Rosa and learned that those matrix switch products have been extended into super sophisticated switching products which have been worth tens of millions of dollars of revenue over those decades.