Live Sound (Part 2)
If loudspeakers and crossovers remain a mystery to you, read on and find enlightenment.
In the second part of our series explaining the elements of the PA system, Paul White explains yer woofers, yer tweeters, yer bag on yer head...
This series is, essentially, about PA systems, but several areas of live sound, including this one, are also common to the recording studio. Both rely on loudspeakers, and though there are some differences between PA and studio speakers, the main distinction is one of scale. The end result of everything we perform or record relies, to a large extent, on the quality of the loudspeaker systems used — which makes it important for us to know something about them.
The purpose of a loudspeaker is to convert electrical energy to acoustic energy with as high a degree of fidelity as is practical — but this isn't as simple as it seems. Humans can hear sound over the frequency range 20Hz to 20kHz, but it is very difficult to make a single loudspeaker capable of covering the entire audio spectrum. Very small diameter loudspeakers can approach this ideal, but small loudspeakers are incapable of generating the high levels of bass and lower midrange required in PA applications, and to a lesser extent, in studio monitoring.
It is an unavoidable law of physics that the reproduction of low-frequency sounds can only be achieved by moving the air in front of the speaker over a large distance, compared to high-frequency sounds, in order to produce the same level of sound. That is why bass speakers have larger diameters than high-frequency or mid-range speakers — and, as we shall discover later, they tend to require larger cabinets. By the same token, large diameter speakers are less effective at reproducing high frequencies, and in most practical systems, the audio range is shared between speakers of different sizes.
A circuit known as a crossover is used to send only the relevant part of the audio spectrum to each speaker. The bass is handled by a so-called woofer and the treble by a tweeter; in larger systems, there may be one or more additional speakers handling the midrange, and these are commonly referred to as mid units. But — not to put the proverbial cart before the metaphorical horse — it would be advantageous to look at how a loudspeaker works.
A loudspeaker without a box or cabinet is referred to as a driver, and nearly all the drivers in common usage are based on the same electromagnetic principles, established decades ago. Designs have been refined and materials have improved, but the underlying principle has not changed. There are other ways of turning electric current into sound, one being the electrostatic loudspeaker but, as these are relatively uncommon, especially in PA circles, they won't be covered here.
The familiar 'cone' loudspeaker comprises a stiff cone of paper or synthetic material suspended in a rigid cage or chassis by means of a flexible surround; a simplified schematic of a typical driver is shown in Figure 1. At the narrow end of the cone is a parallel-sided tube, or former, onto which is wound a coil of thin wire. It is through this coil of wire that an electrical current corresponding to the sound to be reproduced is passed. The coil is positioned in a slot between the poles of a powerful magnet, and whenever a current is passed through the coil, a force is set up between it and the magnet, causing the cone of the driver to move either backwards or forwards from its relaxed position, depending on the polarity of the electric current. In normal use, the driver would be fed from the output of a power amplifier, whose job it is to amplify, or make larger, a small alternating input current. If this input current comes from a microphone or other audio related source, the current fed through the loudspeaker will cause it to move in such a way as to recreate the original sound.
Looking again at Figure 1, it is apparent that the cone can only move so far before it reaches a physical limit — either the suspension will permit no more travel or the voice coil will be driven out of the gap. As the cone approaches these limits, the physical movement is no longer proportional to the electrical input, with the result that the sound is distorted.
The other important factor to appreciate is that the cone will only reproduce the input signal faithfully if it behaves as a rigid piston — if the cone itself starts to bend or vibrate, distortion will again be introduced.
A major physical limitation on the power handling of a loudspeaker is the heat generated in the voice coil. Because the coil has an electrical resistance, the more power that's put in, the more the coil heats up. If more heat is being generated than can be dissipated, the temperature will continue to build up and the coil will burn out.
A driver capable of reproducing a high level of bass sound needs to have a large diameter and must be able to move over as large a distance as is practical. Only then is it possible to directly move the maximum amount of air. The problem is that the larger the cone, the stiffer it needs to be, which in turn makes it heavier. When the cone tries to move quickly, as is the case when high frequencies are being reproduced, the inertia of the cone material opposes it, and the higher the frequency, the harder it is to get the cone to follow the electrical signal. This is one reason why large bass drivers (woofers) cannot faithfully reproduce high frequencies.
The other reason why large-diameter drivers can't reproduce high frequencies very well is purely down to their diameter. If you imagine a 12-inch diameter bass driver that has, somehow, been built with an infinitely light cone, it might seem that it could move fast enough to reproduce all the audio range, assuming that there is enough available power to push it against the resistance of the air. Even if this were true, the high-frequency performance of the driver would still be compromised by its very geometry, because the diameter of the cone is large compared to the wavelengths of the high-frequency sounds being reproduced. If you were to listen from directly in front of the loudspeaker while, say, a 10kHz tone was being played, then you should hear it without any trouble, but what happens if you move a little to one side, or off-axis, as they say in the trade? If you refer to Figure 2, it becomes evident that sound emanating from one side of the cone will now reach you before sound from the other side, with the result that some cancellation occurs. The higher the frequency, the more serious this effect, and to cut a long story short, high frequencies tend to be concentrated into a tight beam along the axis of the cone; the higher the frequency, the narrower the beam. This is clearly undesirable, not only from the listener's viewpoint, but also because tightly beamed, high-frequency sound in a PA situation is inviting feedback problems.
A further problem occurs at high frequencies because the cone of the driver ceases to behave as a perfect piston, and instead starts to vibrate. When people refer to loudspeaker break-up modes, this is what they are referring to — it's nothing to do with the speaker itself breaking up, but describes the way in which the cone is vibrating. These break-up modes cause distortion at high frequencies — some guitar speakers are actually built to encourage these effects, as they contribute to the sound! Some hi-fi speakers use complex damping or bracing to minimise this form of distortion, but this tends to increase the mass of the cone, with a consequent reduction in efficiency.
The solution is to use smaller-diameter drivers to handle the higher frequencies — which is why mid-range drivers are smaller than bass drivers and high-frequency drivers are smaller still. Mid-range drivers may be physically similar to low-frequency cone drivers, the main difference being their size, but high-frequency drivers, or tweeters, generally take a different form. Often, tweeters have small, dome-shaped diaphragms, which may be made from a variety of materials including plastic, treated fabric and even metal. Metal might seem an unlikely choice given the criteria that the diaphragm must be physically light, but metals such as aluminium or titanium are strong enough to enable very thin diaphragms to be built. These often have stiffening ribs formed into them to prevent uncontrolled vibrations. The diaphragm is driven by a voice coil arrangement similar to that employed in the cone loudspeaker.
In PA applications, horn-loaded tweeters are often preferred because of their higher efficiency and greater power handling. These comprise a compression driver (again a magnetically-driven diaphragm device) which drives into a flared horn. This horn has two advantages: firstly, it increases the acoustic efficiency by matching the driver to the air in the room more efficiently, and secondly, the horn shape controls the directivity of the sound, enabling it to be focussed more precisely. Figure 3 shows a conventional, direct radiating tweeter and a compression-driven horn.
Tweeter diaphragms, and to a lesser extent, mid-range driver cones, move over a more limited distance than bass driver cones, and so have to be protected from low-frequency signals which would otherwise drive the diaphragms against the end-stops. Conversely, bass and mid drivers must be prevented from receiving frequencies higher than they are designed to reproduce, otherwise high-frequency beaming and distortion will occur — which is where the crossover comes in.
At its simplest, a crossover is a series of passive electrical filters comprising resistors, capacitors and inductive coils wired between the amplifier output and the drivers. Such systems are invariably located inside the speaker cabinets themselves. In a three-way speaker system comprising bass, mid and tweeter units, the bass speaker would be fed via a so-called low-pass filter — a filter that only allows through frequencies below a certain limit. This ensures that the bass speaker never has to cope with frequencies higher than it can comfortably handle.
The mid-range speaker, on the other hand, has both upper and lower limits of operation and has to be fed by both high and low-pass filters to ensure that it receives only midrange frequencies. At the high end, the tweeter is fed via a high-pass filter, which ensures that it receives frequencies only above a certain limit. Figure 4 shows, in simple graphic form, how the three crossover bands are arranged. Note that one filter slopes away as the next one rises, so that there is a smooth transition from one driver to the next.
Such passive crossovers are reliable, cost-effective and simple to build, but they also have their shortcomings. Because they comprise passive components, some of the amplifier power is absorbed by the crossover circuit, resulting in a loss of efficiency. Furthermore, unless all three drivers are equally efficient at turning electrical energy into acoustic energy (most unlikely), the more efficient drivers have to be fed with attenuated or reduced signals to bring them down to the level of the least efficient driver in the system. Failure to do this would result in a system that reproduces some frequencies more effectively than others — which is not what we want; an accurate system must have a flat frequency response, which simply means that all frequencies within the audio range are reproduced equally. Finally, the filter action of a passive filter can't be made particularly sharp. A filter doesn't simply stop all frequencies beyond its cutoff point but has a sloping response that reduces by so many dBs per octave — the more dBs per octave, the sharper the response of the filter. Simple passive filters usually have slopes of 6 or 12dB per octave, and though higher slopes are possible by cascading two or more filters, the power lost is also higher, making the overall system less efficient.
The world of PA was revolutionised by the invention of the active crossover because, while it may be acceptable to sacrifice efficiency in hi-fi speaker systems, it is less acceptable in a large sound system which may total thousands of Watts. Figure 5a shows a traditional passive crossover system, while in 5b, the same three-way speaker system is fed via an active crossover system. The most obvious difference is that we now have separate amplifiers for the bass, mid and tweeter drivers, and the crossover circuit comes before the amplifiers where the signal level is still small. This means the crossover filters can be built around active electronic circuitry, allowing greater flexibility in design without worries over efficiency. Now the system is no longer restricted by the driver with the least efficiency because the amplifier gains can be adjusted to send more power to the less efficient speakers. This gives the designers far greater flexibility when selecting drive units for use in a system, while the ability to design steeper filters helps ensure that each driver only receives the frequencies that it is best able to handle.
Active crossovers also make it less likely that high-frequency drivers will be damaged in the event of the system being overloaded. In normal music, the energy at the bass end of the spectrum far exceeds that at the high end, and in a passive system, this can pose a danger to the tweeters. If this seems unlikely, consider what happens when something like a dance record is played too loud so that the amplifier is driven into clipping distortion. Every time a loud bass note or drum beat is played, the amplifier clips, producing square waves rich in high-frequency harmonics. These pass through the crossover to the tweeter and, if they are high enough in level, they cause excessive heating of the voice coil and may cause it to burn out. In an active system, on the other hand, the crossover comes before the amplifiers, so an overload at the bass end won't affect the mid-range drivers and tweeters at all — they'll still receive clean signals from their own amplifiers.
So far I've completely neglected the subject of loudspeaker cabinets, but as will be discovered, these play a major part in the overall performance of the system, so next month's instalment will be devoted entirely to this important topic.
Feature by Paul White
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!