David I. Anderson is chief technologist for power management at National Semiconductor. He has spent a long and distinguished career in the electronics business and often provides unique insights on the various panels and forums he participates in as a technology guru. EDN asked Anderson about his past, present, and future.

I notice an accent; where were you born?

My father briefly worked in England around the time I [was born]. We returned to Alloa [Scotland] where I grew up with my four brothers, attending Dollar Academy.

Were you educated in Scotland?

I graduated from Edinburgh University with an honors degree in electrical engineering.

Why did you get that degree?

I have always been fascinated by the workings of the human brain, and my original plan was to study medicine at Edinburgh University and become a brain surgeon. That [plan] did not pan out, and I stumbled on electrical engineering after discovering the intoxicating smell of solder and the thrill of making my own radio. I still remain intrigued by the brain, by the way, and believe that it is the ultimate engineering machine that will one day be replicated in some way.

What was your first job after leaving school? Was it what you expected?

I chose a start-up company, Nuclear Enterprises, turning down an interesting opportunity to join the British Broadcasting Corp research labs because they were in London. My first job at Nuclear was to assist a more senior designer in developing a portable X-ray spectrometer for mineralogy application. Looking back, that [job] was a surprisingly good starter project: The waterproof box included NiCd [nickel-cadmium] batteries and charger, a 2500V voltage multiplier for the photomultiplier, a low-noise capacitive amplifier, a digital counter using brand-new CMOS-logic chips from RCA, which boasted very low power consumption; and a digital ‘nixie’ tube display and driver. [A nixie is a neon tube, typically with 10 stacked cathodes, one for each Arabic digit.] The instrument was featured on a TV show, but fame and fortune eluded me.

After Nuclear, I joined Ferranti Electronics in Edinburgh, designing board-level telemetry products. After discovering that a tiny new semiconductor device called a PLL [phase-locked loop] could realize my board-level design of a 600-baud PSK [phase-shift-keyed] product, I decided the future was in semiconductors, so I joined Siemens in Munich, Germany. Munich was mostly about having a good time and meeting interesting people, but I did get paid for designing consumer-IC applications for cameras at Agfa and audio at Grundig. I shared a laboratory with an eclectic mix of engineers from then East Germany, Czechoslovakia, Northern Germany, Hungary, and Bavaria, so the political discussions were intense.

Why did you join National Semiconductor?

Munich is a wonderful city, but after awhile, you tire of answering the age-old kilt question. So, I accepted a job with National Semiconductor, got married, and returned home to Scotland the very next day. My choice of National was a case of ‘if you can’t beat them, join them.’ At Siemens, it seemed that, whenever I completed a design, I would open a journal and see a similar product advertised from a company called National Semiconductor. The final straw occurred when I was preparing an article on camera ICs and found an ad for a new electronic-timer IC with an integrated photodiode, designed by Dennis Monticelli at National.

At National in Greenock, Scotland, I developed audio ICs, including analog telecom and Dolby ICs. This [era] was the heyday of noisy cassette media, and [company founder] Ray Dolby did a tremendous service to the audio industry, not only by patenting his noise-compression scheme, but by insisting on rigorous standards for any item approved to display the Dolby logo. At one time, National was No. 2 in the field, selling so many ICs to Japanese audio companies that the company received a Queen’s Award to Industry.

About that time, I proposed what I believed was the first analog-cell-based design system, which we called BLOCCS [bipolar library of characterized cells]. This [experience] was my first one of the power of a name. I realized that all people had their own definition for what they believed the technology could do, and it did not really matter if they were correct; the name itself sold the product. Unfortunately for me, National acquired an ailing relative about this time, and it turned out that Fairchild’s Clasic ASIC family pretty much did what BLOCCS could do and had already found a niche in the mushrooming disk-drive market. BLOCCS remained an in-house technology and went on to spawn more than 35 analog ICs—many for the automotive market and a few still in production today.

What was your impression of the Scotland fab? Has that changed over the years?

There is a dirty little secret in semiconductor manufacturing. I remember arriving at National and being told by personnel—human resources to you—that I was indirect. Indirect? Was this an early feedback session? If I were indirect who would they class as direct? In Britain, we had always learned that the United States had abolished the class system. Yet, here we were dealing with ‘directs’ and ‘indirects.’ There were meetings for indirects and separate meetings for directs—and separate reward systems, too. It turns out that the ‘direct work force’ is employed making the actual silicon. These [workers] are the true heroes of the silicon revolution, the last to gain recognition, and the first to get terminated when the business cycle turned down. The ‘indirect work force’ does other stuff and gets most of the perks.

When we interviewed prospective hires at Greenock, we would always pray for good weather. This [prayer] definitely raised the probability of a successful hire. The candidate would be conducted to the fab via the picturesque river route, avoiding the alternate road that passes vestiges of Scotland’s industrial past. However, National’s Greenock fab has been an excellent facility for the corporation for more than 30 years. How many semiconductor plants can boast that record? Now a leaner organization than I remember, National’s Scotland fab is very well-managed and remains a key resource for the company.

So what led to your coming over to the states? What job did you hold? Was it for a promotion?

Promotion? Possibly: There was a reorganization, and Graham Baskerville, who had hired me into National and now headed analog development at National, asked me to move to become ‘worldwide’ design manager for automotive products. (I learned that US citizens like to use the term ‘world’ hyperbolically for many things.)

Were there any trouble signs or things that disturbed or discouraged you?

I was clearly naive when I moved my whole family under protest to the United States, leaving behind the security of Britain’s national insurance and social-security system. I had never taken much note of the stock price as options were few and always ‘underwater’ in those days. This [case] was especially true when I moved to the United States. It was only later that I learned of the precarious state of National’s balance sheet at the time. The golden lining turned out to be the pathetic valuation at the time of the options I received for the ‘promotion.’ Once again, I got lucky.

Eventually you left National. Where did you go?

When I left National in 1996, I joined a little-heard-of company called Semtech. I am the kind of person who needs a new challenge every four to five years, and this one seemed intriguing. [Chief Executive Officer] Jack Poe and [Vice President of Operations] Ray Bregar hired me to start up an IC-design capability in Santa Clara, CA, for a company with more than 20 years’ history as a component supplier to the military. ‘High tech’ was not an appropriate term. Glass diodes were hand-assembled in a less-than-inspiring facility in Newbury Park, CA. I started the same day as [analog-design veteran] Louis Burgyan, now at National, and, together, we bootstrapped an analog-design group in Santa Clara, CA, to address the power market. This [period] was a stimulating one as the processor industry was on the cusp of an insatiable drive for speed and clock rate, necessitating the need for more and more analog-power ICs at lower output voltages and higher powers. Poe was an inspiring chief executive officer and presided over an impressive growth curve, although, admittedly, timing is everything. In four years, we opened four power-design centers, three of which, I am proud to note, are still in place, and developed a solid catalog of power products.

As the market developed, I saw the need to integrate the power switches. We had been working with a well-known FET manufacturer to make the first DrMOS products, and it seemed that this was the way to go. Unfortunately, Poe saw the other company as a competitor and put a halt to the partnership about the same time that I discovered a small start-up, Volterra, which was actually integrating everything onto one chip. So, following the saying ‘if you can’t beat ’em …,’ I joined Volterra to head up its design activities.

What market was Volterra trying to penetrate?

Volterra was my first experience of a real start-up: the ups, the downs, the unreasonable expectations, the amazing revelations of what a small team of smart people can achieve, the emotional outbursts. … To this day, much of the industry is unaware of the capabilities of this company as it still keeps its technical information out of the public eye. Volterra’s founders studied under Professor Seth Sanders at the University of California—Berkeley. Sanders was a well-known evangelist for what has become known as digital power, and the company pioneered the provision of I²C [interintegrated-circuit]-bus access on a switcher. Volterra was one of the first companies to discern the shift from single-brick supplies to POL [point-of-load] IC regulators and the dynamic volume growth that the shift implied. Times have moved on, however: The same strengths that allowed the company to excel in its early stages have, in my opinion, become its weakness. A study of any of the company’s public statements shows a dependence on a very small list of customers and markets.

After eight years, you came back to National Semiconductor in 2004 as a chief technologist for power management. Was it that Volterra did not seem to be going anywhere, or was it that National seemed more promising? Volterra pursued digital-power products. Should National pursue digital power?

Digital power became all the rage just after National had re-established its position as a pure-play analog company. Naturally, a plea for digital expertise would not go down well, so National has not been an early participant. Do I think that digital-power management is real? Yes, in fact it is old hat and is, in fact, present on numerous chips available from National and others using I²C and other bus standards. Digital-control loops present somewhat more of a challenge, but they, too, will become commonplace as topologies move below 0.18 micron. However, there is far more hype in the trade press than in the marketplace, and the ROI [return on investment] is not yet there. At the end of the day, the customer wants a power supply that works efficiently. He does not care how we implement the control—analog or digital. The real benefits of digital are reprogrammability and time to market, particularly for multiple-output switchers. National’s PowerWise technology is a particularly effective form of digital control.

You have worked for all three National chief executive officers: Charlie Spork, Gil Amileo, and Brian Halla. Can you describe what each one did for National? Did they have any shortcomings?

Wow, you really want to test my longevity! The story of National is like an extramarital affair with the so-called digital revolution. First, it was the microprocessor and National’s "first" 32-bit microprocessor. Who today will remember that National beat Intel to 32 bits but misunderstood the value of being the incumbent leader?

I remember Spork as an inspiring leader. People would work to extremes to meet his expectations. But he took his eye off the ball and misread the digital trend. As it happens, I was visiting Santa Clara, CA, from Scotland and was with Sean Hurley, [former analog-device engineer at National], on the day that [Linear Tech co-founder] Bob Swanson, decided the future was analog, and his team resigned from National to start up Linear Technology. I remember the dazed look on Hurley’s face when he took the urgent call. At the time, I did not understand the full implications. Ironically, Hurley later joined Linear Tech.

Gil Amelio put National’s finances back in order after a rocky period; he put water back into the fountains and painted the buildings. But he left for greater things before the job was done, leaving a company in organizational disarray, and shedding talent like a molting bird.

Like Spork and Amileo, Halla, was initially seduced by the lure of digital and the dream of Internet appliances. I was not at National at the time, but from a competitor’s perspective, it seemed as though National, in purchasing Cyrix, had unreasonable expectations of competing with Intel on its home turf—dé-jà vu? But Halla is an enthusiast and a technology visionary. He could achieve better results by practicing a little old-fashioned MBWA [management by walking around], but, under his tenure, National has developed a strong balance sheet.

In the same way that US presidents are allowed to make a few mistakes in the first two years, National’s chief executive officers all come ’round eventually to the realization that the true value at National is good old-fashioned analog.

Can you tell us about what the future holds for the power groups and National?

Power management has come a long way since the advent of the Simple Switcher back in 1995, yet, amazingly, there is still huge potential for further development; many exciting technologies and new applications are appearing all the time. National is blessed with many outstanding engineers, and not only in design groups. For the first time in many years, these folks are being given a sincere invitation and opportunity to innovate. Still, it is an unmistakable fact that much of the innovation in the field still comes from universities and start-ups that ultimately become partners or pieces of larger companies. At National, we do not have a stellar history of successful acquisitions, but that [fact] does not mean we cannot. We have a strategy and the talent and expertise, and we are poised for growth.

What is the biggest issues facing the users of power and power-management ICs?

Energy and its effective management.

Do you think you will be going back to Scotland when you retire?

Unlikely, unless global warming transforms the Clyde [River] into the Riviera.

In the movie Grumpy Old Men, Ariel Truax, the character that Ann-Margret plays said, ‘The only thing you regret in life is the risk you didn’t take.’ You went out to a Silicon Valley start-up. Were there any other risks you wish you had taken?

I suppose I have often taken risks in my career but have been very fortunate in their outcome. But the greatest risks not undertaken have been business ideas I didn’t follow up on. I should really do something about that; I still have a few.

When you left Volterra, your financial status may have allowed you to retire. What drives your passion to contribute to the semiconductor industry rather than go fishing?

I have neither the patience for fishing nor the talent for golf. And the semiconductor world is pretty stimulating most of the time. The real story, however, is that I don’t feel I have made my mark.