Loudspeaker designer John Dunlavy: By the Numbers... Page 4
Atkinson: An analogy would be the similar way your eyes and brain compensate for color temperature. You see white as white, even though under incandescent lights, white is actually more like orange.
Atkinson: For many years, your primary career was in antenna design.
Dunlavy: Yeah. I became interested in antennas because they represented an unusual level of challenge, and I've always enjoyed challenges. Back when I got out of engineering school in the mid to late '40s, antennas didn't have much bandwidth. In fact, an antenna with a 30% bandwidth was considered to be a "wideband" antenna. And that wasn't good enough, particularly in the high-frequency range where, as the ionosphere changes properties at different times of the day, you might have to shift from 2MHz at nighttime up to 20MHz in the daytime. No one antenna could fulfill that requirement.
So the Air Force hired me to design antennas for them. What came out of that was the original Log Periodic antenna. I actually have a picture on my wall of the first Log Periodic, dated 1952. Unfortunately, it was classified "Secret." So I didn't get the credit! Later on, around 1955 or '56, a major university in Illinois came up with a Log Periodic antenna...by that time the Air Force had declassified it down to "Restricted," but I still wasn't able to say much about the original design development. But I do have the patent on the Cavity-Backed Spiral, designed for NASA's ground-base telemetry and communication antennas for the Gemini project.
Atkinson: Did your background in antenna design give you any insights into the behavior of loudspeakers?
Dunlavy: It certainly did, John. Having a fairly good academic background, plus experience in the field of electromagnetic theory, allowed me to see things that other designers probably to this day are not aware of. As an example, some of your readers will be familiar with the term "Thiele-Small parameters." Let's say you have a tall enclosure—tall with respect to its width dimension—and you put, say, a full-range 5" driver dead center along the vertical portion of the column, and you measure the acoustic properties of this system. If you move the location of the driver to one end or the other of the enclosure, then remeasure it, you're going to discover that it will have an entirely different response.
Now to someone versed in waveguide theory and electromagnetic theory, it's very obvious that the behavior will be different. The acoustical impedances that the driver sees behind it are very, very different in these two locations. And yet if you look at Thiele-Small expressions, they don't predict this behavior, because all they say is, Here are the properties of the driver and here is the mass and the volume of the air behind it. But air mass and volume have nothing to do with the impedance transformation that is very obvious to anyone versed in electromagnetic theory. So what antenna design permitted me to do is to gain a real head start, I think, compared to other speaker design approaches.
Also, a cavity-backed spiral is an almost perfect analog of a loudspeaker driver backed by an enclosure of some sort. The theory applies to both, but electromagnetic theory is far, far more advanced than acoustical theory...Not everything is the same. Many, many years ago I thought about working on a distributed line, very similar to a line-array type of antenna for speakers. But now I look back and say, "Gee, how could I ever have pursued that? It was stupid." Because I know a lot more now than I did back then. What a lot of people don't really understand about line-source type antennas, like tall electrostatic sources or something, is the extremely narrow vertical beam-width. If you're seated at a distance equal to or greater than the height, and assuming that the aperture is uniformly illuminated, the vertical beam-width is very narrow.
Atkinson: Many panel speakers also have very narrow lateral dispersion.
Dunlavy: Yeah. Horizontal beam-width is [inversely] dependent upon the width of the radiating area. Which in some rooms might be an advantage, except that the power response of such a speaker changes so dramatically with frequency that unless you're seated very near it, at a distance well less than the [longest dimension] of the radiating area, you're not going to hear natural spectral balance in most rooms. Maybe in an anechoic chamber or something, but [laughs] not too many people have listening rooms that are anechoic chambers.