Designer Notebook: On The Line Array Front
An inside look at the fundamental design concepts that characterize Martin Audio's W8LC compact line array.

By Bill Webb and Jason Baird

Figure 1: Hand-drawn W8L design sketch.

Early in 2002, Martin Audio entered the competitive line array market with the W8L, and a year later its compact sibling, the W8LC is emerging from the engineering department at Martin. Let’s look at the rationale behind the design process and how some of the lessons learned from the W8L have informed the design of the W8LC. Line arrays are perceived to offer benefits over horizontally arrayed systems, such as improved frequency response consistency, increased high frequency throw and reduced set-up time. With only one or two line arrays on the market at the turn of the millennium, the debate within the industry at the time was whether line array was a passing fashion or whether it would be around for a while longer. When a couple of major U.S. manufacturers introduced line array products in 2000/2001, it was clear that the snowball would continue to grow.

With this in mind, Martin Audio needed to decide whether we should enter the line array fray at all or leapfrog to the next “big idea”, whatever that might turn out to be. At a lively product meeting to weigh up the options, we decided to join in, provided that we could enter the market by the end of the year. A tall order time wise, but do-able.

The next step was to figure out a product concept that would be our own “take” on line array and not just a me-too product. Since Martin Audio has always majored on horn design, it was logical that we would want to apply our expertise to the line array format and combine horn-loading technology with line array principles. An early design sketch dated February 13, 2001 shows the basic idea. (Figure 1)

Figure 2: W8L mid horn, measured wavefront curvature at 2.5 kHz.

Because of the fight for real estate within the box, a big problem was how to get a low frequency horn to work down to its target of -3 dB at 50 Hz. Normally to do this it would take a much longer horn than was possible to fold into the space available, so we turned to a technique that we had developed previously, using a short horn with a ported rear chamber. This trademarked, hybrid technique marries the efficiency and attack of a bass horn with the low frequency extension of a reflex design, in our view, the best of both worlds. So even though the horn is theoretically too short, just when it’s output starts to fall off the port output takes over and carries on down.

It was when we turned our attention to the mid and high frequency horns that things really began to get interesting.

TUNNELVISION?

Over the period in the 1990’s when the line array was emerging as the dominant format in touring sound, technical and marketing arguments were being developed by some manufacturers to explain why their line arrays worked better than horizontally arrayed systems.

These arguments focussed on the need for a particular type of waveguide necessary to produce a flat wavefront and the permissible distance between vertically adjacent elements. If your line array didn’t meet these “essential” criteria, then it could not be considered eligible for entry into the line array club.

That may seem like an overstatement of the situation, but it has been borne out by many heated conversations on the floors of trade shows since. If we were going to challenge the prevailing “only this way” wisdom by horn-loading our mid-frequency and high-frequency sections, we needed to be very sure of our ground.

We wanted to use a midrange horn with two vertically stacked 8-inch drivers to cover the range from 220 Hz to 2.5k Hz. This would give us the 6 dB (or so) of horn-loading gain, plus the ability to accurately define a 90-degree horizontal coverage pattern for the mid frequencies, rather than leaving it to chance.

In recent years, we had been working on mid-horn topologies that work higher up the frequency range without narrowing as the frequency rises. So we were confident in our ability to develop a mid-horn that would exhibit 90-degree horizontal pattern control over a full decade.

Figure 3: Progressive curvature array.

The problem was that in terms of line array folklore, we would be challenging one of the “established” line array criteria that call for the vertical spacing between drivers to be less than a wavelength at the highest working frequency. Believing that this was one area where horns and direct radiators differ, we set about designing a mid horn that would have drivers which were greater than one wavelength apart yet still produce a low curvature wavefront.

Figure 4: J-shaped array.

We confirmed this by measuring the phase distribution over the mouth of the horn with the aid of a small microphone and precise positioning jig. Results showed that the mid horn did indeed produce a low curvature wavefront in the vertical plane all the way up to 2.5kHz, its upper limit. (Figure 2) Note the curvature from left to right is due to the 90-degree horizontal dispersion of the horn.

With the high-frequency section, we went for the same philosophy of vertically stacked horns, each with a 90-degree horizontal pattern and low curvature wavefront in the vertical plane. Ultimately, the only difference between the original napkin sketch and the final product was a change from the four high-frequency horns in the sketch to three in the finished product.

Figure 5: A ViewPoint screen offering more than one viewpoint.

PRACTICAL CONSIDERATIONS More difficult than the acoustic design was the design of the flying hardware ­ a learning curve in itself. In previous arrayed clusters, we had used the recognized MAN keyhole system, but in the case of line array we were on our own in uncharted waters and had to develop a proprietary system from scratch. Whenever anyone looks at a bit of fabricated metal, they think it’s an easy design job, but this could not be further from the truth when safety, flexibility and ease of use all have to be reconciled.

After several iterations during the course of the project, we arrived at a system that we were happy with, but it was a difficult task against the clock and the path was littered with a lot of brackets, holes and pins along the way.

Once the W8L was on the road, we really started to add to our knowledge base. In terms of physical and electronic set-up, we soon discovered that line arrays are not nearly as forgiving as previous types of system and we needed to find a way of eliminating guesswork to determine the best array shape, line length and control settings for any particular venue.

MATHEMATICAL TOOLS

Because this question was much too complex to be answered by rule of thumb or simple reasoning alone, we developed a computer model, incorporating the acoustic and electro-mechanical characteristics of each individual low, mid and high element and with each element driven by a virtual crossover.

With this model, we could predict the frequency response at various points in the audience and use the results to optimise the curvature of the array. In nearly all cases, the model yielded a progressive curvature array profile (Figure 3) where the curvature increases gently going down the array. This produced a more consistent frequency response from front to rear than J-shaped array profiles. (Figure 4)

The model also confirmed our suspicions that a “too-flat” wavefront from each high frequency element would produce gaps in HF coverage in a curved line array, and that a low curvature wavefront would give smoother results over a range of splay angles.

Because calculations using the model could take several hours, the progressive curvature rules it established were built into an optimisation program that we called ViewPoint, which automatically optimizes the individual splay angles and overall tilt of the array to give the smoothest coverage. (Figure 5)

Figure 6: Effect of air absorption at 12 kHz as throw increases.

In the venue screen, the venue details are entered along with the nearest and furthest points that the array is intended to cover. Details such as maximum pick height and minimum trim height are entered, together with the number of cabinets in the array, from a minimum of three to a maximum of 16.

Clicking on Design Array will automatically calculate the inter-cabinet splay angles and overall tilt that will give the smoothest coverage over the audience area. The array’s mechanical configuration can then be printed off ready to give to the crew.

Figure 7: Air absorption losses equalised with HF band zoning.

Both the computer model itself and the ViewPoint software took us into the new territory of developing tools to aid our understanding of the complex issues involved in practical use of line arrays.

For instance, we could now use the model to investigate the beneficial effects of hinging at the front versus hinging at the rear (hinging at the front gives a more consistent response, especially for larger splay angles) and look into how best to zone the system to deal with the effect of HF air absorption over distance. (Figures 6 and 7)

LESS OF THE SAME

Large touring systems are overkill for many medium and smaller scale venues. Since it is not possible to scale down a line array by just reducing the number of boxes used, if we were to design a line array for use in theatres and ballrooms, the box itself would have to be scaled down.

Figure 8: W8LC atop its W8L larger sibling.

Again we wanted to produce a system that kicked and again chose the all-horn route. We didn’t need another napkin sketch for the W8LC ­ we just needed to make it smaller and use smaller drivers: a 12 inch instead of a 15-inch for bass and two 6.5-inch inch cone drivers instead of two 8-inch units for the mid-horn. Getting convincing low-end performance out of such a small bass horn was going to be a challenge, but we were confident that with our experience of bass horn design plus computer modelling and Hybrid technology, we would be able to squeeze a quart into a pint. Ultimately, we were able to get down to a -3 dB point of 60 Hz for a single enclosure.

We had already made our take on the line array concept work in the W8L, so all the work on the W8LC would be in the detail design and tooling rather than in the concept itself or the mathematical tools to validate our approach. (Figure 8)

Because it looked like a scaled-down version the full-size W8L, the initial perception was that the W8LC would not take as long to design. However looks can be deceptive. Every piece of wood was different, every driver was different, every horn and moulding was different, and every piece of flying hardware was different. Everything was different except the paint, the wheels and the screws that fixed the handles.

However, this time, we pretty much knew the design parameters from the outset. And the development tools we needed were already in place.

Bill Webb is engineering director for Martin Audio, while Jason Baird serves as the company’s research and development manager.

May 2003 Live Sound International