Rudy Severns: Lifetime Achievement Award Winner

Sep 1, 2008 12:00 PM
By David Morrison, Editor in Chief

Chances to Innovate

Nearly all of these power supplies for the spacecraft and GPS systems were SMPSs designed solely with discrete devices including bipolar junction transistors (BJTs or bipolars) for the switches. And all of the magnetic components were designed from scratch. These projects required innovation, which would require Severns to develop new circuit topologies and design techniques.

For example, at ATC, Severns introduced a resonant power-supply design for space at a time when most space applications were using flyback converters. In this case, the resonant approach was motivated by the challenge of delivering high voltage at low power (milliwatts), but with very high efficiency. This work became the subject of his second paper.

In this application, Severns says, “You had to supply 20 mW at 400 V from a 28-V spacecraft bus and you had to do it as efficiently as possible. That's extremely difficult to do with hard switching. So I went with the resonant approach, which required some innovation.”

At Hughes, his work with traveling wave tube amplifiers required him to invent circuits that could achieve high efficiency and very light weight. “That's really where I started to invent new circuits and new topologies,” says Severns.

The demand for very light weight together with small size would also be a motivating requirement at Magnavox. “That's where I started doing very-high-frequency work,” says Severns.

In the late '70s, the standard for switchers was 20 kHz to 25 kHz. At Magnavox, Severns pushed the switching frequency beyond 100 kHz, even developing a tiny (for the time) SMPS running at 600 kHz. That work was the basis for two groundbreaking papers that he presented at PowerCon and the Power Electronics Specialists Conference (PESC) in 1978 and eventually led to Severns' appointment as an IEEE fellow.

The PowerCon paper, “Design of High-Efficiency Off-line Converters Above 100 kHz” was one of the earliest to discuss the possibility of higher-frequency operation.

“It was the first real proposal in a large power-supply forum that said, ‘Think about much higher switching frequencies.’” The paper was very popular, even receiving a best paper award, but it was also controversial. Severns recalls that Ron Birdsall, the organizer of the event, suggested that Severns write about something less “blue sky” and more practical for his next conference paper. But within two or three years, high-frequency switching wasn't blue sky anymore.

In addressing the PowerCon audience, Severns took care to explain that high-frequency switching was an option, rather than a necessity in all cases. But this message was not heard.

Severns recalls, “When I presented the paper, I was extremely careful to say, I am not telling you that you should operate at these frequencies. I am telling you that in certain situations you can and you may wish to operate at these frequencies. Of course, within months everyone was saying that I was advocating that we all shift to this frequency.”

Severns notes that in later years, he went so far as to tell designers to decrease switching frequencies when higher-frequency switching wasn't warranted.

At the time Severns presented his paper, the use of SMPS technology was well established within the military and aerospace industries. However, it had not yet been accepted within the commercial industry. It's frequently said that the arrival of the power MOSFET enabled the adoption of switch-mode technology by commercial power-supply merchants in the late '70s and early '80s.

According to Severns, the power MOSFET also enabled the migration to the higher switching frequencies he was describing in his papers. “Almost immediately after I gave my paper, the power MOSFET came on the scene and designers didn't have to go through the effort I did to make BJTs work at 150 or 200 kHz.”

Nevertheless, the transition to higher frequencies did not happen overnight. It would take 10 years before the MOSFETs would become cheap enough that designers would use them en masse and be able to routinely design power supplies to operate at higher frequencies, according to Severns.

Adventures in Semiconductors

When Severns gave his PowerCon paper in San Francisco, the sales manager from Intersil was in the audience. “With some later arm twisting, he dragged me up to go to work for Intersil, because they were simultaneously [along with Siliconix, Hewlett-Packard and International Rectifier] coming out with power MOSFETs.

“Intersil lured me away, but I was easy to lure because it was really different. Developing hardware is one world. The semiconductor industry is another world,” says Severns.

At the time he started working at Intersil as a field applications engineer (FAE), the only design he had completed using a power MOSFET was the 600-kHz switcher described in his PESC paper. “Now I understood the principles, but from principles to writing application notes and giving lectures, it's a long jump.”

However, Severns learned quickly about how MOSFETs work, and how their particular characteristics and flaws could lead to failures in the applications. He then had to instruct customers in how to use the new transistors and how to avoid pitfalls.

“For instance, somebody would put two MOSFETs in parallel and they would start oscillating. That's because these things have gain up in the ultrahigh-frequency region. So my job was to figure out how to prevent that. And there were other strange operating modes having to do with dv/dt,” says Severns.

“The early MOSFETs, for example, they would switch very fast. But when you started doing that with more than one of them in the circuit, a device would be in the off state and have this tremendous dv/dt applied across it. So, the one that was supposed to be off would be turned on due to intrinsic parasitics,” explains Severns.

Severns notes that his counterparts at the other power MOSFET companies were also addressing these problems, though he was among the first to write about dv/dt and other issues such as parasitic oscillation in parallel devices, radiation effects and avalanche-induced turn off. Most of these issues were addressed through semiconductor processing.

“My job was to identify the problem, explain to the world what the problem was and how to avoid it until we fix it right,” says Severns. “And at the same time, I had to tell the device designers in the process department, ‘Change your design so we don't have to tell the world that we have this problem.’”

In 1980, his work as an FAE also gave him his start as a teacher, which included giving presentations to engineers about MOSFETS. “The MOSFET's a new product. We have to sell the engineering community on how to use this product. That's when I started giving lots and lots of seminar talks.”

Though he did this initially for Intersil, he would later present similar seminars for International Rectifier and Siliconix. In time he would expand into teaching power-supply design seminars for Ed Bloom, and for others such as the University of Wisconsin at Madison and Oregon State University.

Acceptable Use Policy blog comments powered by Disqus