Roundtable: Inside The Approach
JBL’s development team talks about line array concepts

Pairs of VerTec VT4887 compact line array elements suspended beneath large arrays of the full-size VT4889 elements. (Radio City Music Hall, December 2002, Christmas Music Spectacular, ProMix Inc.)

As we’ve met up at various trade shows, the JBL engineering team and I have had some interesting discussions regarding line array design and concepts.

We decided to present some of these ideas to you, from the folks directly responsible. So let’s check in with JBL Vice President of R & D Doug Button and Director of Tour Sound Engineering Mark Engebretson about line arrays in general as well as the company’s VerTec Series. ­ Keith Clark


Live Sound: At the risk of sounding like Jerry Seinfeld, “What’s the deal with line arrays?” In other words, provide us with an overview of your particular perspective of the topic.

Mark Engebretson: Line array systems for larger-scale PA applications are a relatively recent market trend. While the same acoustical principles govern the performance of all such design attempts, obviously not all line array system designs that end up seeing the light of day as finished products are the same.

There are widespread misconceptions about how they work and perform. Separating fact from fantasy requires taking acoustic measurements under controlled conditions, and an application of scientific principles to interpretations made of the resulting data.

Figure 1: Comparison of relative height of a four-box array (compact, midsize and full-size line array elements), showing distance to near field boundary at 4 kHz

Prior experience in the setup, optimization and use of traditional fan-shaped multi-box arrays does not translate directly to line array systems. Line array deployment and operation is affected by variables in array size, box splay angles, and the input signal processing.

It’s important the choices made in these realms are based upon reliable work that has been scientifically developed. Such an approach will have more validity than either assumptions, or hearsay.


Live Sound: Yes, but aren’t there some basic acoustical truths that govern the behavior of line arrays? Are all line array systems really rather similar?

Doug Button: In some ways they are similar, and in other ways they can differ. In the 1930’s, acoustical pioneer Harry Olson showed us the narrowing of the beam with increasing frequency of a true line source, as defined in the classical sense.

Figure 2: Schematic of a typical line array enclosure integrating MF and HF elements.

At the highest frequencies the beam becomes so narrow as to be pretty useless for audience coverage. The first inclination is to curve the array to get better coverage. But Olson clearly showed the pattern control versus frequency of that is not terribly consistent. This suggests that neither solution (straight, or evenly curved) is ideal.

So when we create a high-powered full-bandwidth vertically oriented system, further array modification through controlled, incremental articulation is required to provide more ideal coverage.


Live Sound: So the current generation of line array systems are basically just large, curved speaker columns? And aren’t the recent smaller line array offerings, like JBL’s VT4888 and VT4887, scaled-down versions of this?

Doug Button: Today’s so-called line arrays are multi-box columns that purport to simulate a line source. They offer wider bandwidth than early vocal-range speaker columns. But, low frequencies behave differently than high frequencies.

If what you want is better performance, the acoustical elements and transducers you design into that multi-box column, and their geometric arrangement, determine what the system is, and is not, capable of ­ just like the type of engine “under the hood” really does matter in an automobile, when you’re driving it on the highway instead of listening to a car salesman.

Figure 3: Horizontal HF performance, shown 0 degrees to 60 degrees, with no bandpass integration devices fitted to the system.

If you want viable performance in more compact arrays, you can’t skimp on the components just because the boxes get physically smaller.

Plus, as we “shrunk VerTec,” we faced a number of mechanical challenges that led to the development of purpose-built transducers to achieve our desired goals.


Live Sound: So when you were looking to scale down VerTec, what factors came into play?

Mark Engebretson:
First we had to evaluate the advisability of scaling down both our acoustical and mechanical technologies. There are clearly applications for smaller arrays that contain the appropriate number of articulated elements. As with full-size line array elements like the VT4889, the challenge for system designer and user alike is to understand how the resulting array will work.

A graphic can illustrate the principle of “near field extension” with three VerTec system enclosures. (Figure 1) At 4 kHz, the on-axis far field begins at 23 meters (75 feet) for an array of four full-size VT4889s. For the same number of midsize VT4888s, which has an enclosure height of about three-fourths that of a VT4889, the far field begins at slightly greater than half the distance, or 12 meters (38 feet).

Similarly, using four compact VT4887s will produce 4 kHz far field characteristics at 7 meters (23 feet). So “extending the near field” isn’t always desirable. Since we always want our audience to be in the far field, yet close to the near field, the choice of what size enclosure to use might be determined by audience area and venue size.

Live Sound: About the near field versus far field... What about the reported sound level loss of only 3 dB for each doubling of distance with line arrays? How do you rationalize this in the design of your compact models?

Doug Button: Much has been written about the near field versus the far field. It’s been purported that line arrays create a cylindrical wave front that at some point in space, transitions to a spherical wave. This description of wavefront behavior in a line source is one many can easily visualize, but it’s largely an over simplification with some underlying falsehoods.

A more accurate way to describe the so-called near-field region is that it’s an interference field. According to classic line array theory, look at the sound pressure level (SPL) versus distance of a line source 4 meters (approximately 12 feet) tall. This shows a transition to the near field at about 100 meters (300 feet) at 10 kHz, 10 meters (30 feet) at 1 kHz and 1 meter (3 feet) at 100 Hz.

Figure 4: Horizontal HF performance, shown 0 degrees to 60 degrees this time with R.B.I.s in place.

If we move toward the array, SPL doesn’t increase as much at higher frequencies. This is because the near field is actually full of destructive interference.

So the notion there is some form of gain that ‘projects’ the near field and ‘increases’ the high frequencies is misleading. The SPL drops off more slowly with line arrays because the interference diminishes as we move away from them.


Live Sound: O.K. ­ regardless of size, how were your basic line array designs derived?

Mark Engebretson: A logical configuration for multi-band line array elements in a horizontally oriented enclosure was suggested by Joseph D’Appolito in 1983. His configuration placed the high-frequency section in the middle of the system, flanked on either side by devices that cover frequencies immediately below.

About 1992, Christian Heil introduced a variant of this D’Appolito configuration, using compression drivers for high frequencies. Subsequently, L’Acoustic began to market systems based on this design. An additional twist with these was the addition of midrange devices to reside inside a waveguide that was energized by a slot-aperture high-frequency section. This allowed the effective lateral dimension of the midrange devices to be smaller, enabling wider horizontal dispersion. (Figure 2)

Figure 5: Cross-section of a NDD loudspeaker motor. Note the long voice coil former to accommodate the dual voice coils fitted to the twin-gap magnet design.

To further this idea and offer a degree of additional control, all of our models include a pair of R.B.I. units (Radiation Boundary Integrators). They’re fitted over the midrange devices and serve several functions, not the least of which is to smooth the transition from the output of one set of bandpass devices to another.

The accompanying figures show how the addition of these units to the system resolves some significant issues. This is especially important in the critical mid- to high-frequency crossover region. (Figure 3 and 4)


Live Sound: JBL continues to develop its own transducers, which is now somewhat a rarity. Can you talk about the advantages of this approach, and perhaps more specifically, the role your dual voice coil design plays in the VerTec design?

Doug Button: The intent of the dual voice coil is to increase power handling, lower power compression, resulting in more dynamic headroom. But it’s not as simple as it might seem, just adding an extra voice coil.

Like many other inventions, one can wonder, if it’s seemingly so simple and offers obvious benefits, why isn’t everybody doing it? As early as the ‘50s, patents appeared describing a loudspeaker with two voice coils wound in opposite directions, spaced apart axially, and connected in series on the same coil former.

Figure 6: JBL 2435 beryllium-diaphragm neodymium-magnet compression driver, performance contrasted with another available compression driver.

Although early designers may not have directly considered it, two coils with twice the surface area can handle twice the input power. This idea appears to have been ahead of its time, as the early designs never really found commercial applications.

It’s also important to mention that once we had this basic motor design worked out, we were able to scale it up in size... or down. For VerTec, we’ve developed cone drivers ranging from 8-inch to 12-inch models up to 18-inch low-frequency elements used in different VerTec systems. (Figure 5)


Live Sound: And what challenges were faced with the high-frequency devices?

Doug Button: Most compression drivers have near-maximum theoretical efficiencies in the midrange, but usually fall short at high frequencies due to excessive mass in the mechanical assembly.

Because all large-format (Ž 3-inch) domes have break-up modes in the usable bandpass region, it’s possible to have efficiencies greater than the theoretical envelope at higher frequencies by designing in resonant behavior. We sought high efficiency and smooth response, with no ringing, and true pistonic motion to at least 15 kHz.

Figure 7: Laser-scanned image of the 2435 driver’s 3-inch diaphragm (left) at 14.5 kHz compared to displacement of a titanium diaphragm at the same frequency and input voltage.

Further, we wanted to balance the various parameters to achieve maximum efficiency in a specified bandwidth. We chose 3 kHz to 20 kHz as the bandwidth we wished for maximize efficiency. Using our controlled testing, it’s interesting to see the performance of this driver, called the 2435 (Figure 6).

It’s also useful in gaining more understanding to see data on the 2435’s diaphragm under load. (Figure 7) You’ll see that it exhibits pistonic characteristics, even at frequencies as high as 14.5 kHz. However, the traditional diaphragm has moved into a serious break-up mode.

So we ended up with a more robust design that has less harmonic distortion. We’ve taken all that we learned on the HF section for the full-size VT4889, and applied that to the newer drivers used on the midsize VT4888, for example.

Live Sound: What’s next from JBL in line array systems?

Mark Engebretson: It’s no secret that we’re working with Crown and dbx on the DrivePack series of electronics modules, which will enable our line array system products to be self-amplified with integrated digital signal processing. Beyond that, let’s just say our R&D program continues to take input from professional users regarding the challenges they face in portable sound reinforcement.


September 2003 Live Sound International

Email this story to a friend.