Nov 30, 2013

On the Hiring of Bob Widlar

This story is so beautifully written that I am basically just quoting it. I’ve deleted a few things but nothing of substance. It was submitted by Joe Malone, the personnel manager at Fairchild in the early days.

I believe the year was 1964. It seems to me it was springtime and either a Thursday or Friday. On this particular morning, I was visited in my office on Whisman Road very early (and unexpectedly) by Bob Graham who was Fairchild's Product Marketing Manager. Bob, along with Don Valentine the National Sales Manager, reported to the Marketing Director, Tom Bay. Bob had in tow a scruffy, curly haired little fellow with impish eyes that simultaneously smiled and bored holes in you. He was clearly self-confident bordering on arrogant, but outgoing and friendly at the same time. He was introduced to me as Bob Widlar a young design engineer currently working at Ball Brothers Research in Boulder, Colorado. I was impressed that Graham had the "stones" to recruit him out of a customer's house. In any event it was clear that Graham "owned" him now and felt responsible for his welfare at least for the day. This was a special guy.
Graham provided me with a courtesy introduction and informed me that he had arranged for Widlar to spend the day with John Hulme, Murray Siegel, Maurice O'Shea and Vic Grinich one of the company's founders who, I believe, was running the Applications Engineering group at that time. He let me know that Hulme, who was the Manager of Integrated Circuits Applications, was the targeted hiring supervisor and I was to aid and abet Hulme in any way possible to see that Widlar joined Fairchild. They then disappeared and I saw no signs of them until late in the afternoon. I simply was on standby to assist Hulme.

To add a little color to the story, it appeared that Widlar was on vacation that day and driving a big, heavy, red Pontiac convertible with the top permanently down from Colorado to Mexico. Graham had somehow convinced him to plan his trip such that he could spend a day at Fairchild in Mt. View. Widlar had agreed to eight hours maximum. Graham knew he had a narrow window with which to work and, master-manipulator that he was, made every minute count. We needed to "close the deal" by sundown.

Sometime late in the afternoon, Hulme and Widlar appeared at my door. We put Widlar in my office while Hulme and I huddled in the hallway outside. Hulme announced that he wanted to hire Widlar and (ever the pristine manager) indemnified that he had an appropriately approved personnel requisition somewhere in my department. We discussed the magnitude of the offer briefly. John stated that Widlar was currently earning (gulp!) $9000/yr and he was comfortable offering him $10,000/yr. but - NOT A PENNY MORE! Hulme was not comfortable extending the offer and asked me to close the deal. Widlar was driving to Southern California immediately upon leaving the premises and both Hulme and Graham wanted him "closed" today!

I then joined Widlar in my office and the dancing began. He knew where we were going and decided to enjoy the ride. I poked around in his early life and discovered that his father ran a radio/TV repair shop in the Cleveland, Ohio area. He learned electronics from the ground up at his father's knee. For some reason that I can't recall, he chose to join the Air Force instead of going to college. Somehow he got stationed in Colorado and was able to attend the University of Colorado in parallel with his military service. I believe he finished college with a BSEE in three years (3.9999 GPA) and met his service obligation at the same time. Ball Bros. was his first and only fulltime professional civilian employment.

At some point, I (naively) asked if he had thought to bring along a resume. In response, he casually tossed a copy of his transcript across my desk. Ever the poker player, I studied it carefully and noticed that it showed ALL "A's" except for one lonely "C" - that being in Colorado History. When I inquired about the obvious aberration, he replied that Colorado History was required to graduate. He said the instructor, on the first day of class, asked the class to take out a blank paper and draw an outline of Colorado. Widlar's response to this request was to write on the paper, (sic) "The map is on the wall behind you, you dumb SOB!" He then smilingly admitted that the Professor never seemed to warm up to him and simply gave him a passing grade. My first clue that this was a different dude?

We then moved into the negotiations. He verified what Hulme had already told me about his current yearly salary. I responded that Hulme very much wanted him to join his group and was prepared to offer him an increase to $10K/yr. His reaction was a blank stare. The silence was palpable. It was one of those "whose going to blink first" moments. Finally, he spoke. "I won't come for less than $12K," he said. Now it was my turn to stare while my mind was whirling. Is this guy playing me? (Of course he was.) Do I dare stonewall him and risk the wrath of Bob Graham if I lose him? What to do?

I broke the silence by launching into a totally irrelevant and wandering discourse on Fairchild; John Hulme, what a trusted guy he was; how if he was as good as he thought he was he'd be making $12K in no time; blah, blah, blah. I was pouring out platitudes non-stop. Interestingly, I don't recall any talk of equity or stock options of any kind. I guess not at that level at that time. Anyway, he was entirely cash motivated then. The equity thirst came much later.

Speaking of thirst, it was getting on to 5 o'clock and he interrupted my babbling by asking where the nearest "watering hole" was. I acquainted him with Walker's Wagon Wheel just down Whisman Road at the intersection of Middlefield Road. He abruptly ended our meeting by asking me how long I intended to be in my office. I said "As long as necessary, why?" He then stood up to leave and committed to me that I would get a phone call with his decision "after six beers." (Somehow I sensed that wouldn't be long. I was right!) We parted ways amicably with much the same affable, smiling eyes I had seen that morning.

Roughly, an hour later, my phone rang. Widlar was on a pay phone at "The Wheel." (No cell phones then) He graciously accepted our offer as originally stated; asked me to put it in writing and mail it to his home in Colorado; committed to give notice when he returned from vacation; said, "Adios, Amigo!" and left for south of the border.

I "processed" him out when he left the company with his counterpart, Dave Talbert to join what was then a struggling Molectro in 1965. Molectro was purchased shortly after to establish a West Coast operation for National Semiconductor. Together they were unbeatable either at work or play. Linear products generated a steady (obscene?) gross profit for NSC throughout its history while the other product lines "eked" out marginal results if any at all.

Widlar was an amazing individual! There are more "Widlar" stories around than can possibly be true. But this one is mine and I've obviously carried it around for over 40 years. It's interesting how a single day can remain this clear among all of the cobwebs of other memorable events.

In closing, in case there remains any confusion as to who the hero is in this tale - it's Bob Graham. John Hulme hosted Widlar all day and made the easy decision to offer him a job. I may have "closed' on him but Graham was the key ingredient. He "sourced" Widlar and convinced him to stop by on his way to Mexico. He set up and orchestrated a smooth process that the rest of us simply implemented. He made a major impact on the industry by introducing Bob Widlar into it. This was not his only contribution to the industry but arguably one of his greatest.
NOTES:
Read the rest at: http://corphist.computerhistory.org/corphist/view.php?s=stories&id=332 "In the beginning… there was Widlar!" by Joe Malone, August 8, 2007 

Nov 28, 2013

Bob Widlar and the Sheep at National

I started out to write the obligatory sheep story because what blog on analog would be complete without it. There are several versions online. Several start out by correcting the New York Times obituary version which claimed it was a goat. But there are other discrepancies: Did he act alone? Was the sheep stolen or purchased? Was it a prank or a protest against the unkempt landscaping? What happened to the sheep? So here is what I believe to be the definitive Sheep Story with a few bonus stories. But first, a few steps backward.

Bob Widlar started working at Fairchild Semiconductor in 1963. He left Fairchild in 1967 to join Molectro which was quickly acquired by National Semiconductor in 1967. The story happened in late 1970 at National. There was no love lost between Widlar and his vice president, Pierre Lamond. Why? A few reasons. Generally speaking, Widlar was difficult to manage (a tremendous understatement, really) and especially so when he felt his boss knew less than he did. Pierre had ascended rapidly – perhaps too rapidly and earned a bit of the disdain from Widlar.

Lamond was born in France and joined Transitron in 1957. The owners, the Bakalars, put him on the production line. In his third week, he was promoted to the head of production to replace his departing boss, and, in a few more months, he was promoted to device development engineer. He returned to France as his one-year working visa expired, and then by 1959 he was back at Transitron as the head of development. Lamond's second stint at Transitron was a short one. In 1961, Lamond joined Fairchild, working for Gordon Moore. After the loss of Lamond's direct supervisor, he was again promoted to manage device development. In 1967, he was one of four key operations people along with Charlie Sporck, to leave Fairchild and restart National Semiconductor.

As a quick aside, Lamond had previously assembled a team of Fairchild managers in preparation to defect to Plessey. When negotiations with Plessey broke down, Lamond and Sporck followed up on Widlar's and Talbert's suggestion to consider National Semiconductor. Ironic, eh?

By 1970, Widlar had been increasingly playing pranks on Lamond. Widlar’s office was near Lamond’s and there was an intercom speaker in the ceiling right between them. The paging bothered Widlar so he had his brother, Jim who worked in facilities, to disconnect it. Lamond had it reconnected. This escalated until Widlar installed a bomb (M-80 or cherry bomb) in the speaker! Lamond saw him on a chair in the hallway. “What are you doing, Bob?”asked Lamond. Widlar replied, “I am going to blow out these damn speakers.” Lamond replied, “Oh,” and turned and walked back out the door. Widlar lit off the fuse and hopped down.

According to Bob Pease, Widlar really was almost always a soft-spoken person. He didn’t have to yell or shout to get his message across. Lamond, not so much. Widlar created a “hassler” circuit for his office. When a person spoke loudly in Widlar’s office, this circuit would detect the audio and convert it to a very high audio frequency. So if you really hollered, it would make sort of a ringing in your ears.

Lamond was a stickler for being on time. Widlar, not so much. So Widlar came up with a circuit that would steal cycles from the 60Hz wall socket for Lamond’s clock, making it lose a few minutes every day. Again, his brother Jim did the deed, cobbling the circuit into the socket. Lamond bought three new clocks before they tipped him off to the prank.

Widlar would also generate prank purchase orders (PO) to see what would get through the system. Widlar put in a PO for a bale of hay, 20ft of rope, a tire & 20lbs of bananas as a prank (and I think there was something about a monkey also). A few days later he came in to work and found the items and was told the bananas (or the monkey) were on backorder. A fitting segue to the story of the sheep.


In 1970 the semiconductor market was in a bit of a slump. Charlie Sporck was the CEO of National Semiconductor and a great operations man, so he decided to curtail landscaping for a while at the Santa Clara facility. Presumably, Lamond was implicated by Widlar for this as well. December in Northern California means the days are cooler and the rains start to return. So the grass grew tall. Widlar saw this as another opportunity for a prank. (I doubt he cared about the unkempt look of the facilities, especially since Bob had supposedly parked on the grass in the past.)

One morning Widlar and Bob Dobkin drove down to Morgan Hill. His friend, John Weiss had a neighbor, Emmett Slaughter, who owned some sheep. Dobkin says they paid $60 for a sheep; Jim Widlar says they borrowed it. Widlar had a Mercedes-Benz 280 SL convertible with a hardtop. There was a little gap behind the seats, so they wrapped the sheep in a blue tarp and tucked her snuggly in the gap so she couldn’t move.
[Image is not Widlar’s car, but you can see a sheep-sized gap behind the seats.]  You can see Widlar's car in the background in one of the photos below.


They tied the sheep up to a tree in front of National’s headquarters and about 20 minutes later a reporter from the San Jose Mercury News showed up to took photos. Dobkin swears that neither he nor Widlar called the newspaper. It must have been someone else. Contrary to Sporck’s version of the story, Pierre Lamond did not coax the sheep back into Widlar’s car. Widlar and Dobkin collected it at the end of the day. They took it to a local restaurant, By-th-Bucket, and left it with the owner (Mike Garcia). Dobkin says that no one from management really had much of a reaction until the story appeared in the Mercury News. Once they saw it in the paper they were furious – and they started having the lawns mowed again.

This was December 12, 1970. Bob Widlar retired from National Semiconductor on December 21, 1970.
















NOTES:

Photo credit for color photos, Fran Hoffart.

Many details from conversations with Jim Widlar.

Many details from conversations with Bob Dobkin.

http://en.wikipedia.org/wiki/Fairchild_Semiconductor

“What’s All This Widlar Stuff, Anyhow?”, Bob Pease – EDN, first published in the July 25, 1991 issue

“They Would Be Gods”, (http://www.scribd.com/anon-52108/d/179382-They-Would-be-Gods), Upside, October 2001


SIDEBAR (why I believe the By-th-Bucket version over the Marchetti’s version): By-Th'-Bucket is still around, but it was called “Buy th’ Bucket then. In those days it was literally a hole-in-the-wall establishment. It was made of cinder block and you collected your food from a glowing window that was, in essence, the only light in the place. You had a choice of wooden picnic tables topped with bowls of salted peanuts or – if you couldn't find a seat – the hood of your car would become your table. Mike Garcia, the original owner, would put large olives in the bottom of your beer, patrons would eat peanuts and throw the shells right onto the floor, which would be there nearly a week later. [source: http://www.btbbarandgrill.com/about-us ] No one would worry about the presence of a sheep – it was probably welcome. Or perhaps it ended up on the menu. No one’s talking.

Per Wikipedia: The San Jose Mercury was founded in 1851 as the San Jose Weekly Visitor, while the San Jose News was founded in 1883. In 1942, the Mercury purchased the News and continued publishing both newspapers, with the Mercury as the morning paper and the News as the evening paper. The story ran in the afternoon paper. Full text, below:

*** San Jose News Sat., Dec. 12, 1970 page 3

Economy Mowing

This is not Mary’s little lamb grown up and following her to work. It is a lawnmower. Bob Widlar, director of advanced circuit development, “borrowed” the sheep for the front lawn of National Semiconductor in Santa Clara to help the firm’s austerity program by cutting mowing expenses. Widlar admitted it is “putting a lot of gardeners out of work,” but notes “at the same time the grass gets cut, it gets fertilized, too.”


Nov 24, 2013

The 10k on Widlar

Bob Widlar arguably started the analog IC industry as we know it. Today, the majority of analog circuits trace their designs to circuits and techniques that he invented. His legendary character may outshine the body of his work. What seems to get little press is how hard he worked. Not only was he a genius, but he probably worked harder than anyone else.

How does one become a genius?

In Malcolm Gladwell’s book, Outliers, he considers how the “elite” got there. Certainly it takes innate talent, but it also takes a tremendous amount of work. K. Ander’s Ericsson conducted a study of elite musicians and couldn’t find any “naturals” who rose to the top without practicing, nor any “grinds” who worked harder than everyone else but lacked the talent. [1]

“The emerging picture from such studies is that ten thousand hours of practice is required to achieve the level of mastery associated with being a world-class expert – in anything,” wrote David Levitin. “In study after study, of composers, basketball players, fiction writers, ice skaters, concert pianists, chess players, master criminals, and what have you, this number comes up again and again. Of course, this doesn’t address why some people get more out of their practice sessions than others do. But no one has yet found a case in which true world-class expertise was accomplished in less time. It seems that it takes the brain this long to assimilate all that it needs to know to achieve true mastery.”
Let’s count the 10,000 hours for Bob Widlar.


Bob Widlar was fortunate to be born in 1937, before World War II but on the tail end of the Great Depression. His father, Walter Widlar, was a self-taught radio engineer of German descent, a ham radio hobbyist and an engineer with Bird Electronics in Cleveland. Bob developed an interest in ham radio and all things electronic. Let’s say he would spend a few hours a week with his dad – who, by the way, designed pioneering ultra high frequency transmitters. [3] A few hours a week even before the age of 10, according to his brother, Jim; maybe that totals a few hundred hours. His dad also repaired appliances and by his teens, Bob had a television repair business after school. At the age of 15, he was billed in a local newspaper article as an "electronics designer and experimenter, who repaired radio and TV sets as a sideline." As a teenager, he also played radio pranks on the Cleveland police. [2] An obsessive consumer of technical information, Bob studied the manuals and gained a great understanding of vacuum tube circuits. It’s probably conservative to say, by age 18, that Bob had close to 1,000 hours of practice with circuits.

Walt died when Bob was just 15. Ray Bird, the owner of Bird Electronics and a family friend, helped out in many ways – including giving Bob a job. [4] I don’t know exactly how long he worked at Bird or what he did but it probably involved circuits. In 1958, Bob joined the Air Force at Fort Lawry in Colorado. He was in the 3424th Instructor Squadron. [5] Simultaneously, he attended the University of Colorado in Boulder (C.U.). While you might think the Air Force would not immerse Bob in electronics, remember that his was an Instructor Squadron. He passed the electronics exam without even taking any classes – a tribute to the self-study as a teenager and time spent with his dad. So they made him the instructor and he flunked everyone in the class. He felt the instruction manual was inadequate and was the reason that none of the students understood the basics. He wrote the Air Force’s six volume training manual in little over a year. Spending many hours in libraries, Bob had full access to the Bell Labs reports on semiconductor technology. He studied incredibly hard, but generally alone. While he rarely attended his classes at C.U., he nonetheless got perfect grades (except for one lone C in Colorado History). Knowing that Bob would devour books and would stay up for several days when engrossed in a topic, it’s easy to think that Bob by now had accumulated about 8,000 hours.

In 1961 he graduated and left the military, joining Ball Brothers Research Corporation (which later became Ball Aerospace). So, by 1962 he had most likely reached the 10,000 hour milestone – age 25.

But accumulating 10,000 hours and having immense talent is not enough, according to Gladwell. You have to achieve it before anyone else, check. Timing also plays a big part. Check. There could not have been a better point in time for the semiconductor industry. In 1957, eight scientists started a little semiconductor company called Fairchild. Also armed with technology originated at Bell Labs, Fairchild was quickly becoming the driving force in the nascent semiconductor industry. First they made transistors and Bob Widlar, as a customer, knew them inside and out. Bob was already a stand-out engineer at Ball Brothers when Jerry Sanders came to call (yes the founder of Advanced Micro Devices, but then a hot-shot salesman for Fairchild). So Sanders lured Widlar over to Fairchild.

Widlar was incredibly smart. He accumulated over 10,000 hours of circuit “practice” by the age of 25. He joined a start-up company full of brilliant engineers and scientists, experimenting freely on new technology. Even within this highly-charged environment, Widlar, with his cohort Dave Talbert, worked secretly on off-hours without management’s knowledge to create the first commercial operational amplifier. Again, rarely sleeping since he was so engrossed in the work. As the story goes, Fairchild introduced the µA702 even though Widlar didn’t think it was ready. It was very popular but it had aspects that made it difficult to work with. So Bob locked himself in a room for 170 hours and reappeared with the legendary 709 op amp as a result. [6] Widlar was known to have extremely meticulous lab books. These were still the days of slide rules, and his books showed the exhaustive math as he worked out the many problems. In addition to design, he became proficient in layout, fabrication, test and the writing of data sheets. Remember, the µA702 was introduced in October of 1964 – one year after he joined Fairchild. [7]  The amount of work done by Widlar and Talbert in that amount of time is unimaginable today.

He was more than a genius.



NOTES

[1] “Outliers, the Story of Success”, Malcolm Gladwell, 2008

[2] Obituary from the University of Colorado, http://www.colorado.edu/engineering/deaa/cgi-bin/display.pl?id=186

[3] http://westparkradiops.org/

[4] Communications with Jim Widlar

[5] http://airforce.togetherweserved.com/usaf/servlet/tws.webapp.WebApp?cmd=ShadowBoxProfile&type=Person&ID=111966

[6] “The man who designed everything”, Paul McGowan, iEyeNews, July 11, 2012

[7] "The History of Semiconductor Engineering", Bo Lojek, 2007



Nov 20, 2013

Computer Labs, Inc.

After taking so many standardized, multiple-choice tests in elementary school, I learned that statements containing “all” or “never” are probably incorrect. So, of course, not “all” analog greats or analog companies trace their footsteps through Boston or Silicon Valley. In the case of Computer Labs from Greensboro, North Carolina, they were home-grown. That is, until Analog Devices acquired them and gave them the appearance of a Boston area analog company.

So who was Computer Labs and why did they appear in North Carolina?

To answer that, we have to go way back to Cleveland in the years before Alexander Graham Bell is credited for inventing the telephone. In 1856 the telegraph was king and George Shawk purchased one of the little shops building equipment for Western Union. Former chief telegraph operator for Western Union, Enos Barton partnered with him and then George sold his share to an Oberlin College physics professor-turned-inventor name Elisha Gray. The Gray and Barton company became Western Electric. Western Electric had very close ties to Western Union in terms of financing, company directors as well as being a key supplier. To prove that timing is everything, on February 14, 1876, Bell filed a patent for his new telephonic device. It arrived at the U.S. patent office only hours before Bell's closest competitor: Elisha Gray (who had sold his interest in Western Electric in 1875 and retired from the business). Western Union sued the much smaller Bell Company on behalf of Gray. And surprisingly, the suit ended in 1879 with Western Union withdrawing from the telephone market and Bell acquiring Western Electric in 1881.

With Bell’s patent eventually expiring in the early 1900s, the pressure was on to continue innovating. So Western Electric's engineering department developed a new “research branch" to cultivate inventors and inventions. Thus was born the organization that would become Bell Laboratories. I will skip over the tremendous achievements by Bell Labs during the two World Wars in the interest of time. In the 1950s, Western Electric had manufacturing facilities across the US, including North Carolina. Known as the “Winston-Salem, North Carolina Works” the facilities in Burlington, Greensboro and several in Winston-Salem were designated for defense electronics. They were heavily involved in manufacturing electronics hardware for Cold War military programs, such as Nike Zeus, Nike Ajax, Nike Hercules. Within the “Works”, the Greensboro “Shops” manufactured missile guidance systems. The designs were initially done mostly at the Bell Labs facility in Whippany, NJ, but Bell Labs established engineering groups at the various Western Electric locations to support the hardware.

And that’s how Bell Labs arrived in North Carolina.

Greensboro native, John Eubanks, received his MSEE from North Carolina State University and went to work at the Greensboro Bell Labs location. Bob Bedingfield, also from Greensboro, was his technician. In the early 1960s they worked on an ADC for the receiver part of the Nike-X radar system. Walt Kester said, “John was one of the smartest engineers I ever met, and I was fortunate enough to work with him for a few years when I joined Bell Labs in 1963.” The converter used a version of Gray code (called “folding” today) patented by F. D. Waldhaur (Bell Labs, Murray Hill, NJ). Walt believes the ADC was supposed to be eventually manufactured by Western Electric, but the calibration procedure and testing was pretty much beyond the capability of the Western Electric support engineers.
Robert Bedingfield
John M. Eubanks

The government acknowledged the importance of Western Electric's defense work in 1956. The culmination of an antitrust case filed by the Department of Justice in 1949, the 1956 consent decree ordered the Bell System to divest all of its non-telephone activities – except those involving national defense. That forced them to agree to confine their manufacturing business to telephone equipment and license any new technologies they developed to others. Of course, that decree ruled out the commercialization of the ADC that Eubanks worked on.

In 1966, John and Bob decided to leave Bell Labs and start Computer Labs. They wanted to do a commercial version of the ADC that not only could be sold to the subcontractors (GE, Raytheon, and MIT Lincoln Labs), but also to other customers. They started out with some financing by a local company that made electronic instruments for the textile industry (Strandberg Engineering) and used their facility as a base of operation. The first employees other than John and Bob were from Bell Labs also, including Buford Goff (engineering/sales), Don Brockman (sales), Gene Boles (draftsman/layout), George Buchanan (manufacturing). Apparently, there were never any hard feelings between Bell Labs and Computer Labs. Computer Labs actually paid royalties to Bell Labs for several years because of the Waldhaur patent. Kester, who also left Bell Labs in 1969 and joined Computer Labs, recalls, “we would get a few $100k NRE to develop a new ADC, and then get production orders, and eventually it would become a standard product. We didn’t have to sell too many systems at $10k each in order to pay salaries to 30 or so employees.”

As the Cold War wound down, most of the Bell Labs employees in NC were transferred to NJ or given early retirement. The Computer Labs products in the early 1970s were mostly rack-mounted instruments, but gradually got into modules, card-level products, and hybrids. Analog Devices purchased Computer Labs in 1978 for the high-speed technology, and John and Bob pretty much retired from the business. Initially it was known as the “Computer Labs Division of Analog Devices,” but that label was dropped. Today, the Greensboro site is a design center focused on integrated circuits.
MOD-815 Video ADC

The footsteps started right there in Greensboro, NC.


NOTES

Photos used with permission of Analog Devices, Inc.

Computers Labs history based on email communications with Walt Kester

"Western Electric and the Bell System, A Survey of Service" edited by Albert B. Iardella, 1964.

"Western Electric", http://en.wikipedia.org/wiki/Western_Electric

"Manufacturing the Future: A History of Western Electric", Stephen B. Adams and Orville R. Butler, Cambridge University Press, 1999; Reviewed for EH.NET by Eric John Abrahamson, The Prologue Group.

Nov 17, 2013

Nike Air Bedingfield

North of San Francisco with a great vantage point on the Pacific Ocean is the only Nike missile site that has been converted to a museum. It is within sight of Russian Hill in the city, not too far from the Russian River and an hour or so away from the southernmost point of Russian occupation long before California was claimed for Spain. For those who didn’t live through the Cold War, it’s hard to imagine that we needed missiles pointed toward the western sky to shoot down incoming Russian bombers.

To me, Nike means running shoes, basketball shoes, celebrity endorsements and Steve Prefontaine at the University of Oregon (and waffle irons – another story). To the U.S. Army, it was a network of installations to protect against the very real threat of 500mph Soviet jets flying at 60,000 feet with nuclear payloads. The inspiration for the names of those two, Nike, refers to the mythological Greek goddess of victory.

I’m well aware of analog computers using vacuum tube to calculate trajectories to shoot down planes in World War II. Twenty years later, that was no longer sufficient. Three radar systems were needed: one to look for enemy targets, one to track the target, and one to track the missile once the others were locked. In 1958, the U.S. Army began the development of the Nike missile system, with Bell Laboratories responsible for much of the early hardware design through Western Electric.

In the Continental United States the sites were numbered from 01 to 99 starting at the north and increasing clockwise. Nike Site SF-88L refers to the launcher area (L) of the battery located in the northwestern part (88) of the San Francisco Defense Area (SF). You and your Nike shoes can wander all over the site, now a museum. Each site had three parts: C for control, L for launcher, and A for administration. Part C, which no longer exists at SF-88 as far as I can tell, had the radar system which was a digitally-controlled phased array radar system for guiding the short and long-range interceptor missiles. I like phase array radar because it uses lots of Analog-to-Digital Converters (ADCs)!

Initially, the logic was resistor-transistor-logic (RTL) in hybrid packages. At the time, ADCs in general were rare and “high-speed” types were virtually non-existent – except to the deep pockets of the U.S. military. Early prototypes for the 8-bit 10MSPS ADC were developed at Bell Labs in North Carolina between 1963 and 1965. In 1966, two pioneers in high speed data conversion, John M. Eubanks and Robert C. Bedingfield, left Bell Labs and founded Computer Labs, Inc. The initial product was a commercial version of this ADC.

I am still researching these two, as well as the history of Bell Labs in North Carolina. Google Patents lists some key contributions in the area of high-speed data conversion:

  • John Eubanks patents a multi-waveform generator, at Bell Labs in Greensboro, NC, US patent # 3,274,396, filed May 16, 1962.
  • John Eubanks patents the peak detector at Bell Labs, US patent # 3328705, July 6, 1964 and a pulse amplifier also at Bell Labs, US patent # 3,439,286, July 29, 1965.
  • Robert Bedingfield patents a sample and hold circuit using an operational amplifier and a high impedance buffer connected by a switched diode capacitor circuit at Bell Labs (Greensboro, NC), US patent # 3,363,113, Aug. 2, 1965.

The first commercial ADC using folding Gray code architecture (Waldhaur's patent from Bell Labs) was the 8-bit, 10MSPS HS-810 from Computer Labs, Inc., in 1966, according to Walt Kester who worked there. Bell Labs researcher Frank Gray introduced the term reflected binary code in his 1947 patent application, now known as Gray Code. (Clever naming: HS-810, for a high-speed 8-bit, 10Msps converter.)
 The HS-810 used all discrete transistor circuits and was offered in a 19" rack-mount box for digital radar receiver applications. It contained its own linear power supply, dissipated nearly 150 watts, and sold for approximately $10,000. The next generation of Computer Labs' designs would take advantage of modular op amps (Philbrick, Analog Devices, et al), Widlar’s μA710/711 comparators (Fairchild) – they must have been so thankful for ICs – and 7400 TTL logic.


In 1978, Analog Devices acquired Computer Labs and continued developing board-level data converter products and hybrid ICs as the Computer Labs Division in Greensboro, NC, and eventually high-speed ADC ICs. And in outright gratuitous name-dropping, I will add that Nike’s biggest celebrity, Michael Jordon and his size 13 footsteps, led the University of North Carolina to a national championship in NCAA basketball in 1984. For which we have a near endless variety of Nike Air shoes. And we aren’t really worried about Russian bombers attacking San Francisco.

NOTES

“ANALOG-DIGITAL CONVERSION”, "MT-024" and "MT-025", Analog Devices, tremendous gratitude for the contributions of Walt Kester.

Wikipedia

Nov 11, 2013

Beginnings of Burr-Brown

The footsteps in “Analog Footsteps” came about because I learned that so many key people spent time in either Silicon Valley and Stanford, or Boston and MIT – and a great many followed each others’ footsteps on both coasts. Literally, up the same steps. So how did a great analog company like Burr-Brown start up in Tucson, Arizona? Well, they happened to add a few footsteps around MIT.

Robert Page Burr was born in New York in 1922 to Robert Page Burr and Lawrence (Hewlett) Burr. Page was his grandmother’s maiden name. (After a quick look, I found no evidence that he was related to Aaron Burr or Raymond Burr, nor that Lawrence Hewlett was related to William Hewlett.) Robert, the son, was active in ham radio and attended Princeton University. In 1943 he left for the V12 Navy ROTC program at Cornell University and received a B.S. in electrical engineering, in 1944. He pursued the midshipman's school at Columbia, and received training in radar at Bowdoin and MIT. It’s not clear when he became friends with Tom Brown but their paths crossed several times. Burr was assigned to the New York Naval Shipyard, where he spent a year supervising fire-control and anti-aircraft radar installations on new vessels before separation in 1946. He worked for 10 years with Hazeltine Corp., (now part of BAE) developing electronics for radio and television.

Thomas Rush Brown Jr. was born in Memphis, Tennessee and raised in Longview, Texas. At the age of 12 he went to boarding school at Woodberry Forest in Virginia. At the age of 16, he was accepted into MIT. But instead, he joined the Navy where he was first introduced to electronics. After the Navy he returned to MIT, this time earning a degree in general engineering in 1949.

His first career was in teaching but when he realized it would never pay, he headed off to Harvard Business School and earned his MBA in 1952.  If anyone knows how he ended up in Tucson, let me know!

Anyway, Burr and Brown became fascinated with the new device – the transistor. It was 1956 and Burr recognized that transistors inherently had the reliability of a "short piece of bare copper wire." They realized that virtually no electronic instruments at that time were using the transistor and a company was born in Tom Brown's Tucson garage. Burr-Brown Research Corporation was incorporated on May 8, 1956 (reminiscent of George A Philbrick Researches and Nexus Research Labs). Burr assumed the presidency and Brown became vice-president. The first Board of Directors meeting was held on May 10, 1956. Corporation law required four directors but BBRC only had three – Page, Tom and Tom's wife Helen. To solve the problem, one of Tom's friends, Joch C. Leonard, agreed to be present at the first board meeting and then resign five days later. Burr once said that Tom Brown was an engineer in the Edisonian tradition – a perfectionist who succeeds through repeated efforts to get it right. Initially, Burr worked on product development in New York and Brown handled sales and manufacturing in Tucson.

Burr-Brown started with “instruments” in wooden boxes. The 1st product was the model 100 AC Decade Amplifier. Other early products, in wooden boxes, included a Differential AC Amp, Square Wave generator, Variable Gain Preamp and AC Millivolt Meter.

As stated on the poster from the University of Arizona conference room named for Brown, he found that customers were removing the circuits and discarding his beloved boxes!

The model 130 was the world’s 1st transistorized op-amp. This was a completely discrete design using just 8 transistors on a PC-board in a 3 ½” long aluminum shell.

Henry Koerner, Rick Gerdes, Helen and Tom Brown, (unknown) and Don McGraw

In 1958, Burr left the company, turned in his stock and Brown assumed the role of president. Burr never expressed any regrets about missing Burr-Brown’s massive success. He started Circuit Research Company, the new product development contractor for Photocircuits Corp. in Glen Cove, New York (acquired by Kollmorgen). There, he developed printed motors, which revolutionized magnetic tape drive systems widely used in medical equipment, welding and automotive equipment. Burr held well over 200 patents worldwide. He was the first recipient of the IEEE Charles J. Hirsch Memorial Award for outstanding contributions to engineering. All in all, he had a successful career and retired in 1987 as the senior scientific officer.

Meanwhile, back in 1962, a Burr-Brown op amp was the first to land on the moon as part of the Ranger program. Of course, the mission was a failure because it actually impacted the moon and never returned any photos – during a period when the program was called "shoot and hope". I should confirm that Ranger 7, which was completely successful, probably also had Burr-Brown op amps aboard.

Burr-Brown relied on its location near the University of Arizona in Tucson for skilled engineers and created a very close and productive relationship.


Much is written about what an inspiring leader Tom Brown was, but I'll leave that for others.  In 1983 the company changed its name from Burr-Brown Research Corp. to Burr-Brown Corporation. In 1984 Burr-Brown went public. In 2000, Burr-Brown was acquired by Texas Instruments for $7.65B. And for a cliff-hanger, did T.I. spontaneously start up in Texas or are there footsteps leading from either Boston or Silicon Valley?



NOTES

1. Burr-Brown company history page, Funding Universe, http://www.evri.com/media/article;jsessionid=tidy7h3co3y0?title=Burr-Brown+Corporation+--+Company+History&page=http://www.fundinguniverse.com/company-histories/BurrBrown-Corporation-Company-History.html&referring_uri=/organization/burr-brown-0x53d97%3Bjsessionid%3Dtidy7h3co3y0&referring_title=Evri

2. Burr-Brown, Wikipedia

3. “R. Page Burr '44”, Princeton Alumni publication, published in July 7, 1999, issue, http://paw.princeton.edu/memorials/42/64/index.xml

4. “Who Is Tom Brown?” By Steve Taranovich, EDN, October 25, 2012

5. “Tom Brown: A Serendipitous Life” December 2002

6. http://www.eller.arizona.edu/eller_times/december2002/Tom_Brown.html

7. “Tom Brown ‘43” [http://www.uakappasigma.org/fraternity_honored/tom-brown/]

8. “Analog: back to the future, part 2”, Steve Taranovich, EDN, July 16, 2012]

9. IEEE Global History Network, http://www.ieeeghn.org/wiki/index.php/Robert_Page_Burr]

10. Robert Page Burr http://www.thepeerage.com/p53428.htm

Nov 3, 2013

Intersil's Origins

Like modern-day Fairchild which is not quite directly descended from the original Fairchild, Intersil has a gap in its lineage. Nonetheless, you can’t trace the footsteps of key players in Silicon Valley and analog circuits in particular, without finding Intersil.

Depending on which way you look, Intersil is connected to big names. When I first heard of Intersil, its data book also had RCA and GE on the cover. RCA (originally Radio Corporation of America) formed RCA Semiconductor in 1950 and was an early player in the transistor industry. In parallel, General Electric moved into the semiconductor business in 1954 with the formation of GE Solid State. In 1980, GE Solid State acquired Intersil. And in 1988, Harris acquired GE Solid State, which included RCA Solid State. [1] Jack Welch was president of GE and felt that the semiconductor business was too competitive for GE to dominate so he sold it to Harris. At Harris, the Intersil name was retired.

But wait, let’s back up even further and find out where Harris came from. In 1895, Harris Automatic Press Co. was formed to make printing presses. In the 1950s, that company, renamed Harris Corp., expanded into electronics. And in 1967, it acquired Radiation Inc., a manufacturer of space and military electronics located in Florida. [1] Fast-forward a few decades to when Harris owns the remnants of the old Intersil product line; Harris spun off the division and the new company was renamed Intersil in 1999. In 1999, then-CEO Gregory Williams cites the meaning of the Intersil name, "The first half of the new name signals our intention of pursuing Internet-related opportunities, while… the second half of the name underscores our legacy in world-class silicon technology." [2]

Too fast? Yes, of course, because I skipped Intersil’s entire first existence! And the original Intersil couldn’t have been an Internet Silicon company.


Credit for founding the original Intersil goes to Swiss-born Jean Hoerni together with H. Gebhatdt, and S. Wauchope (the latter two from Union Carbide) with funding from the Swiss firm SSIH to develop low-power MOS technology for watches and calculators. [3] Hoerni was working at the California Institute of Technology in the early ’50s, when his shy brilliance was noticed by William Shockley. The story of Shockley Labs and the founding of Fairchild is well known. Using what he modestly called “college level” physics, Dr. Hoerni (doctorates in physics from Cambridge University and the University of Geneva) invented the planar process, which enabled Fairchild Semiconductor to produce the first integrated circuit. His inspiration came to him during his shower one morning in 1958, at a time when Fairchild was completely stalled in their research. But plagued by an itinerant’s need to seek new opportunities, Hoerni would leave Fairchild after only a few years, before the company achieved its success with semiconductors. He would go on to found many more companies, the most notable being Intersil. [4]

Because of the international background of the founders and the Swiss funding, Intersil originally meant “International Silicon”. [5] In 1969, Dave Fullagar came over from Fairchild, fresh off the success of the µA741 op amp. He eventually rose to be Vice President of R&D. Fullagar once said, 
“…the Intersil years were the most creative. I developed the first IC logarithmic and antilogarithmic amps and the first monolithic FET-input op amp, the ICL 8007, which dominated the market until the bipolar FETs came along.
I also got to spend a month in Japan designing the first electronic-shutter IC for a single-lens-reflex camera for Canon. This circuit took the logarithm of three inputs—film speed, aperture, and light intensity—summed the result, stored it while the mirror went up, and then took the antilogarithm to generate the shutter speed. It used about 20 transistors in total and was probably my most elegant design. Nowadays, this is done with about half a million transistors in a microcontroller—how prosaic.”

In 1971, Jack Gifford was asked to leave Advanced Micro Devices by Jerry Sanders (that’s another story). Jack wanted to start his own analog company. Hoerni asked Gifford to take over the analog side of his company. In exchange, he would help Gifford raise money to start his own business. [6] Gifford agreed and subsequently built Intersil up to a $130 million company. The analog side of Intersil under Gifford and Fullagar was a pioneer in CMOS analog circuits. The one that immediately comes to mind for me is the classic charge pump, ISL7660.


When GE acquired Intersil, they tried to keep the Silicon Valley spirit and entrepreneurship. By 1982, GE’s new president Jack Welch appeared to be reversing his earlier assurances to let Intersil keep its own stock-option plan. Several conversations occurred between Gifford and Welch until it finally came to a head after a few drinks at a company party. Jack Gifford, in his rapid-fire way, tells the story of how Jack Welch confronted him at the party:


“No sooner does the greetings stop and he says, ‘God damn it. Why the fuck do you guys have to have those stock options? There, you know, jeez.’ And he goes on and on and on in front of fifteen, fifteen. In front of twenty, twenty-five guys, you know. And just, I mean we'd talked about this, five different times, you know. And all of a sudden he's now, you know, coming like it's my fault, like I'm the problem, right? And he's, say he's had a couple of drinks, and finally he starts, ‘Well, what do you think?’ And I said, and I was going to react and I just, I says, I was so mad and I just said, ‘Fuck you.’ And I just turned around and walked out.” [7]

So Jack Gifford was fired from Intersil. Jack went back to farming for a a bit, then started to hear from his former colleages about how bad it was working for GE. So, in April of 1983, he founded Maxim Integrated Products with Fred Beck, Dave Bingham, Steve Combs, Lee Evans, Dave Fullagar, Roger Fuller, Rich Hood, and Dick Wilenken. GE and Intersil immediately filed suit, alleging infringement of trade secrets. The companies eventually settled, when Maxim agreed to let GE/Intersil pick 10 of Maxim's chips during the ensuing five years to manufacture and market on a royalty-free basis. In return, GE granted Maxim rights to certain product know-how, trade secrets and patent rights. [8] I’m told one of the alledged thefts involved the charge pump that went into the MAX232 and part of the final agreement was that Maxim would no longer recruit Intersil employees to join them.

So the original Intersil suffered under the huge corporate structure of GE. Now we can join the story that I started with Harris taking the Intersil group, dropping the name and eventually spinning it out again as a new company.

In 2002, the new Intersil acquired Elantec and Elantec's chief executive Rich Beyer became president and CEO of Intersil. After several more key acquisitions and a couple divestitures, we have the Intersil of today. Crazy as they were, the footsteps led us back to analog circuits.



NOTES:

Fun Fact: Jean Hoerni’s life did not revolve solely around research. A fanatically dedicated mountain climber, he scaled most of the world’s highest peaks, made a partial ascent of Mount Everest, and spent countless weekends in the Sierras. He was extremely strong, hiking at high altitudes with little food and never sleeping in a tent, and he was resourceful as well. To extend the life of an old sleeping bag, he stuffed it with newspapers. On waking up one morning above 12,000 feet, he said he planned to write a letter of appreciation to the editor of The Wall Street Journal saying that it was by far the warmest newspaper. On a trek in his later years, he became devoted to the Karakoram Mountains in northern Pakistan. Together with Greg Mortenson of Bozeman, Montana, (and author of “Three Cups of Tea”) Hoerni donated $12,000 to build the region’s first school and created a foundation called the Central Asia Institute with a $1 million endowment.

[1] “Intersil: Analog contender or pretender?” Mark LaPedus, EETimes.com

[2] http://harris.com/view_pressrelease.asp?pr_id=483

[3] ComputerHistory.org

[4] “Jean Hoerni, 1924 – 1997” by Deborah Claymon, Red Herring, Feb. 1, 1997

[5] “Making Silicon Valley: Innovation and the Growth of High Tech, 1930-1970” by Christophe Lécuyer, 2006

[6] “Jack Gifford: Baseball’s Loss Was The World’s Gain”, Doris Kilbane, Electronic Design, Dec. 7, 2009

[7] Interview with Jack Gifford, July 17, 2002, Sunnyvale, California (http://silicongenesis.stanford.edu/transcripts/gifford.htm)

[8] InsideChips: Maxim Integrated Products, May 01, 2002, Steve Szirom, senior analyst

Jean Hoerni photo credit: Wayne Miller

Jack Gifford photo credit: EDN magazine