Dave Hershberger, Geoff Mendenhall and Cliff Leitch are shown during installation of an MS-15 FM exciter breadboard for on-air testing at WGEM(FM) in Quincy, Ill., in 1977. Mendenhall’s left hand is on the exciter breadboard.
Dave Hershberger’s work with exciters and modulators for transmitters and with low-level signal processing is considered legendary by many in broadcast. In fact, he co-developed the world’s first digital FM exciter as an experimental prototype that was tested on the air in 1986.
The National Association of Broadcasters presented its 2021 Radio Engineering Achievement Award to Hershberger, saluting his work at Harris Broadcast, Grass Valley Group, Axcera and Continental Electronics.
The veteran engineer even worked on X-band uplink transmitters for the JPL/NASA Deep Space Network. He retired in 2017 as senior scientist at Continental.
Hershberger does occasional consulting but is mostly enjoying retirement life. Radio World asked him about his career, which was about evenly split between radio and television design projects, and his thoughts on current events in the world of broadcast engineering.
Radio World: Where did you grow up and what sparked your interest in broadcast engineering?
Dave Hershberger: I grew up in Sycamore, Ill., just west of Chicago but beyond the suburbs. When I was 13 years old I got my ham radio license. After a few years of ham radio, a neighbor suggested that I should look into broadcast radio. That’s what got me going.
[Read: Hershberger Honored With 2021 NAB Engineering Award]
When I was 16, I studied and got my FCC First Phone commercial license. That was my ticket to getting into broadcast. After high school, at age 17 I got my first radio job at what was then WCLR in Crystal Lake, Ill. — AM 850. They had a directional antenna and needed a First Phone licensee on duty.
David at his ham radio station (WA9QCH)
RW: Describe your early days at Harris Broadcast when you started in 1975. What was it like in Quincy?
Hershberger: Harris was a lot of fun. Quincy, Ill., less so. But it was where the work was.
I started out in the TV transmitter group, but after talking with Geoff Mendenhall, who worked in FM, we started brainstorming at lunchtime on what we would like to do in a new FM exciter.
The TE-3 was old, expensive to manufacture and had its shortcomings. Geoff worked with management, got authorization to start a new FM exciter development and I got transferred to FM. We added some more engineers on the project and away we went.
We broke new ground with a new kind of stereo generator, overshoot controlled filters, improved PLL dynamics to eliminate tilt on low-frequency square waves, and many other features.
RW: What do you consider your most important contribution at Harris?
Hershberger: That would probably be the overshoot-controlled low-pass filters. When we were developing the MS-15 FM exciter, we heard that there was a popular new stereo generator with built-in audio processing being made on the West coast. It was called the Optimod. One of its most important features was its ability to control overshoot in the 15 kHz low-pass filters integral to the stereo generator.
Bob Orban had come up with a brilliant solution for dealing with the problem. He took a systems approach, and the filtering and filter overshoot correction were tightly integrated with the audio processing.
We were also developing a stereo generator option for the FM exciter. But if we did not solve the overshoot problem too, then we would not be able to sell very many stereo generators. So my job was to find a solution to the problem.
And it was a difficult problem, controlling both amplitude and spectrum simultaneously. One way to begin to address the problem was to filter, then clip off the overshoots, and then filter again. That would reduce the overshoots somewhat but would not get rid of them completely.
Theoretically the process could be repeated: filter, clip, filter, clip, filter, clip and keep doing that until the overshoots were low enough. Of course such a system would not be practical.
Eventually I figured out that what I needed was something that did more than clipping. A clipper can be analyzed as a gain reduction device, which reduces gain only during overshoot. If I could reduce the gain more than what is necessary to accomplish simple clipping, then I could make the overshoot controller converge in just one pass of filter-clip-filter.
So the trick was to take the overshoots, clip them off, amplify them with a gain of about two and then subtract them from the clipped waveform. Linear phase filtering of that signal resulted in near complete elimination of filter overshoot, while still providing a sharp 15 kHz cutoff.
The advantage was that this was not part of an audio processor. It was a stand-alone function. So you could use any audio processor you wanted, run it into our stereo generator and the 15 kHz low-pass filters would not overshoot and create overmodulation. Any audio processor could gain the overshoot control capability of the Optimod.
We did our first on-air testing of the MS-15 exciter at WGEM(FM) in Quincy in early 1977. The overshoot control, along with greatly improved low-frequency dynamics, allowed a huge increase in average modulation.
Brian Cox was one of the engineers on the MW-1 AM transmitter. He had left Harris and was working for another company in Quincy. He would leave his car radio tuned to WGEM, and one morning he got in his car to drive to work. He turned on the radio and Brian said it was so loud that it blew him into the back seat. That was the first morning we had the prototype on the air.
RW: How about the highlights from Continental?
Hershberger: In general, deployment of digital signal processing in several different product lines: FM exciters, ATSC exciters (including linear and nonlinear adaptive equalization), and VLF and LF transmitters.
RW: Is there a career project that stands out?
Hershberger: Probably the coolest project was the JPL/NASA Deep Space Network uplink transmitters. They generate up to 80 kW in the 7 GHz range
Dave Hershberger
These were nonbroadcast transmitters but it was a most interesting and challenging project. JPL wanted very low phase noise — not for communications, but so they could also use the transmitters for science experiments. Those included searching for gravity waves by detecting phase bumps, and bistatic radar imaging. Bistatic radar has the signal source on earth, with the receiver on the spacecraft.
So everything we did required attention to low noise performance. RF amplifiers, the klystron beam supply, focus magnet supply and filament supply, and even the water cooling system all needed to be very low noise.
The first production transmitter was installed at the Goldstone station and was first used for the Pluto flyby in 2015. In addition to communication, it was used for the bistatic radar mapping of Pluto.
RW: You wrote in Radio World over the years about implementations of HD Radio. Can you discuss that work and assess the state of HD Radio in the United States, and where it may go next?
Hershberger: I’m rather disappointed. There are interference problems. The audio codec cannot be upgraded to more modern technology. There is no “Oh, wow” factor — such as having your radio figure out your preferences, and then find and record shows it thinks you might like — features that are found in some DVRs.
There is little or no ability to provide different commercial announcements and music to different listeners, based on age, interests or location. There is no non-real time transmission and storage capability.
There are many such features that should be part of a new digital sound broadcasting system. As it is, we just have plain old real-time, single-stream radio but transmitted digitally.
RW: You were an advocate for moving AM stations to TV Channel 5 and 6 below the FM band, which didn’t happen. What are your feelings about that now and about the future of AM radio in general?
Hershberger: It was unfortunate yet predictable that it wouldn’t and didn’t happen. But there is a similar opportunity now. ATSC 3.0 includes the ability to carry audio-only programs. And those audio-only programs can be broadcast with optimization for mobile reception, with robust coding, independent of the modulation and coding for the video signals.
Hershberger is shown installing AM stereo at WLS(AM) in Chicago in 1983.
I would like to see AM radio begin simulcasting on ATSC 3.0 signals. A single ATSC 3.0 transmitter could carry all of the AM signals in a market, in addition to TV programs. Car radios could be made to receive at least the audio-only streams in ATSC 3.0. And that would be a solution to the electric car problem, where the drive train makes so much electrical noise that including an AM radio is just way too expensive because of the required EMI suppression. ATSC 3.0 is also a way to make AM programming receivable in homes again.
Meanwhile, I am participating in the AM Improvement Working Group of the National Radio Systems Committee, which is studying ways to keep analog AM viable.
RW: You thanked Geoff Mendenhall and Dan Dickey for their support through the years in your NAB acceptance video. Any other mentors?
Hershberger: Absolutely. There were many. At Harris, there was Hans Bott, Tony Uyttendaele, Terry Hickman, Bob Weirather, Hilmer Swanson, Tim Hulick and there were more. At Continental there was José Sainz, Grant Bingeman, Michael Pugh, Howard Butler and more. And the late Dr. Steve Reyer, an EE professor at Milwaukee School of Engineering, was certainly a mentor.
RW: If you were chairman of the FCC, what one technical change would you want to make?
Hershberger: If I could only make one change, it would be to enforce radiated and conducted emission limits. Don’t let cheap noisy power supplies and chargers into the country. Force recalls of products that are in gross violation of the rules. Make AM radio receivable in homes again!
RW: What do you think is the most important trend or recent development in the management of technical infrastructure for radio? We hear a lot about centralization of engineering departments.
Hershberger: I’m not a manager or an accountant. But I am disappointed to see management make decisions which greatly increase off-air time when there are problems. I am disappointed to hear stations with audio problems that last for years at a time.
Better engineering is sorely needed. And that costs money, but it is well spent.
Letting broadcast infrastructure decay saves money in the short term but not in the long term. You can do the same thing with your car — don’t do maintenance — but it will come back and bite you eventually.
RW: What do you see as the most pressing technical issue facing radio broadcasters today?
Hershberger: Interference. Switching power supplies, chargers, lighting, etc. not only affect AM frequencies but also VHF and even UHF. Allocation and regulatory mistakes create legal interference which only adds to the problems.
RW: As you talk to fellow engineers, what are their most common complaints or career challenges?
Hershberger: Not many broadcasters want to do things the right way. They want to spend as little as possible, even if performance and the on-air product is degraded.
RW: What is your advice to younger folks entering a technical field like this?
Hershberger: Don’t expect to be proficient if your education is digital-only. You still need to understand analog concepts, even if they are all implemented digitally. Learn control and feedback systems, filtering, modulation theory, signal processing and complex math.
Also, beware of computer engineering philosophies invading radio engineering. As more and more digital technologies are applied to broadcast engineering, there have been clashes of culture, which turn into real technical problems. Computer engineers are interested in sending the bits in a proper format without errors.
Dave and his wife Sandy at NAB 1980. She was a software engineer for Harris and worked on remote control systems.
That much is fine. But computer engineers often pay no attention to phase noise in their clocks, or in selection of clock frequencies which are appropriate for broadcast radio, or frequency accuracy of such clocks. And why would they? Data are still transmitted without error, even if the clock is noisy or off frequency.
Bottom line, don’t assume that digital signals have low enough phase noise and frequency accuracy to meet broadcast standards.
RW: You and your wife have certainly been together a long time. Any marital advice to offer?
Hershberger: I have been married to Sandy for 44 years now. She has been a music teacher, a software engineer, for which she won an Emmy award, and a licensed marriage and family therapist.
As for marital advice, I think I would say to change compatibly. We all change as we age, and it’s important to consider our mates as we do so.
Also I think there is some luck involved. I feel pretty lucky. We are blessed to have a gifted 17-year-old daughter who aspires to a career in the sciences.
RW: Retirement life in California seems to be treating you well. Why do you love it?
Hershberger: We live in a log house we built on our rural property adjoining Tahoe National Forest, near Nevada City, Calif. We have deer, bears, foxes, coyotes and other critters for neighbors. My voluntary hobbies include ham radio. My involuntary hobbies include property maintenance — cutting, hauling and splitting firewood from downed trees, and plowing snow with my tractor. We have alternative energy — solar electric and diesel generator backup — for our frequent power outages, but it is beautiful here.
The post Hershberger Honored as a “Renaissance Engineer” appeared first on Radio World.
Randy J. Stine
