40 Years Ago — Linear Technology Corporation

Todd Nelson
6 min readSep 2, 2021

--

On September 10, 1981, Linear Technology Corporation was officially incorporated.

That was forty years ago. Forty years before that was 1941 when wireless meant a radio with five vacuum tubes, when computers were people with slide rules, and when robotics was a new science fiction term coined by Isaac Asimov. A lot can happen in forty years and who among us can predict it? These guys!

Seated Left to Right: Bob Dobkin, Bob Swanson, Brent Welling. Standing Left to Right: Brian Hollins, Michael O’Malley, Clive Davies. (Source, Linear Technology brochure, 1982)

The four founders in the photo (Dobkin, Swanson, Welling, Hollins) had left important roles at National Semiconductor in order to start a new company that was exclusively committed to linear integrated circuits. It was a risk, for sure, but they predicted it would pay off.

On this anniversary, I’d like to look at a few of their predictions to see how they’ve aged. I’ll pull these predictions from a few rare documents that I saved from the shredder: the original Business Plan (October 1981), their first newsletter “Linear Currents” (Vol. 1, № 1, 1982), and text of a presentation “The Other Technology — An Outlook” (Dataquest Inc., Semiconductor Industry Conference, Oct. 1983).

Linear is a business of individual contribution and it’s a business where a few can make an enormous contribution.

It is our plan to attract and hire key design engineers who are responsible for a majority of the present devices which are industry standards. The existing design team even as of today represents an elite corp, and we feel we will have very little difficulty in hiring other key personnel.

Until I stepped away from the industry last year, it was still quite common for a single integrated circuit to be developed by a single engineer from each main discipline (design, layout, test, manufacturing). A single integrated circuit can generate hundreds of millions of dollars over its lifetime.

The engineers in the photo below, with Bob Widlar working remotely, were likely responsible for tens of billions of dollars. While that is less likely today, the prediction stood up well.

Left to Right: Tom Redfern, George Erdi, Carl Nelson, Bob Dobkin, Jim Williams, Nello Sevastopoulos (Source, Linear Technology Prospectus, May 28, 1986)

…in 1983, we’ll have a market size that is in excess of $2B and is predicted to grow at a 20% compounded annual rate for an indefinite number of years.

By 1990 the linear market will be as big as the entire IC market was in 1980. I like to think of the entire linear market as our niche. Linear Technology is our name because serving the linear market is our game — ALL OF IT!

That’s from Bob Swanson reiterating a commitment to lead. I can no longer find the exact figures, but they were basically correct. That 20% growth rate seems a little wild, but sales were doubling about every third year until the end of the 1990s. The linear segment didn’t drive the entire market’s growth on its own but rose along with it.

Linear IC markets in the United States have largely been ignored by foreign suppliers, who are not expected to be a significant factor in linear IC sales…

Still true. Despite recent consolidation, domestic manufacturers still dominate the linear (analog) IC segment.

The power trend is going in two opposite directions. Higher power …more current …more power (and also because of the push to make these digital go faster); Lower power …battery-operated systems or wherever only limited power is available.

Seems obvious now. Whereas they were thinking about 100W as high, and micro-watts as low, now it’s kW and nano-watts.

More systems on a chip [SoC, mixed-signal, higher integration] is an important subject these days. Experience has shown that the addition of standard logic on a linear die is relatively trivial compared to adding linear to the digital chip. However, as the fab process improves for the microprocessor, it probably is being optimized in a direction opposite to where linear wants to take the development.

I can’t leave this subject before commenting on how arrays and standard cells apply to linear. I’m asked that question all the time. My conclusion is that there is not a real good analogy in linear for the standard cell, gate array approach we see today in the digital world. …although high-performance linear LSI will happen, the move will be slow.

Analog arrays never really happened, so they were right. Mixed-signal and SoC integration certainly happened, but slowly. In my opinion, the successful examples of a big digital chip integrating analog function were in application-specific situations like consumer products. The successful examples of a big analog chip integrating digital functions were initially data converters (ADCs, DACs) and later the addition of state machines or microcontrollers onto analog functions for specific reasons.

The bipolar process and all of its variations is the workhorse of linear. The estimate is that some 80–85% of linear sales come from bipolar products. However, anyone who is serious about the linear business must be into CMOS technology, because as the market doubles or triples between now and 1990, a substantial portion of that growth will require CMOS. In order for CMOS to become more useful to the linear designer, linear process people will need to enhance it as they did the original bipolar process to make it: higher voltage, less noisy, provide good capacitors, and in general make it more rugged so it can survive in the hostile real-world of linear.

This also seems obvious now. You could argue that their 1981 view of process technology under-valued CMOS. Given the dominance of bipolar up to then, they can be forgiven. Maybe someone can tell me if Bob Widlar ever designed a circuit in CMOS!

Success in the ’80s will be achieved via a different strategy than the ’70s. During the last decade, …an “economy of scale” advantage could only be achieved by becoming a broad-based supplier. We will see individual market segments, such as linear, microprocessor, and memory become so large in themselves that the companies who specialize will outperform larger companies.

Clearly, Intel and AMD outperformed National and Motorola. One could argue that Linear Technology outperformed National (while both were independent companies). So this prediction held true through the ’90s, as well.

Today, however, is a different story. The industry growth rates are much lower. The pressure for consolidation is intense and therefore acquisitions appear to be the strategy for success — at least for publically traded companies.

… linear — the other technology — represents a large robust business, one that will grow as the digital side grows.

Linear Technology embraced “the other technology” strategically. One cannot overstate how contrarian this idea was in its day. When Robert Noyce first developed the integrated circuit at Fairchild, it was for digital logic circuits. When Noyce and Moore left to found Intel, it was to make digital logic circuits. When Charlie Sporck was CEO of National, he favored digital logic groups at the expense of linear. All the start-ups by 1980 were digital logic.

But Swanson, Dobkin, and the team believed that more digital logic will create demand for more linear. Higher speed and more complexity on the digital side will require higher performance on the linear side. That continues to be true.

The strategic aspect of “the other technology” was that the bigger companies would continue to ignore it. I can’t tell you how many times I heard “fly beneath the radar” or “don’t ride off to the sound of the guns.” The digital side continued to make headlines in the electronics press — much as software does more recently — and the linear side quietly grew along with it.

Forty years on, those predictions look pretty good.

--

--

Todd Nelson

Engineer, sustainability, indigenous history, analog electronics history and anything that supports my belief that bikes can save the world.