Loudspeakers: construction and characteristics

It is clear that the overall quality of a loudspeaker will be determined by the nature and complexity of the design itself, the standard of construction and finish of the enclosure, the quality and number of drivers used, and that of the critical components of the crossover. Even if we are seduced by the sonic qualities of a high-end, and therefore expensive, model, we should bear in mind that a high-performance model will sound worse, especially in the bass, than a smaller design in a room whose size or acoustic conditions are not suited to the larger model. But whatever its level of quality and price, every loudspeaker will have a series of characteristics of its own, some acoustic and others electrical, which taken together will determine its sonic personality and how difficult it will be for the associated driving electronics. This time, we will look at these characteristics.

The arrival of home theater brought about a real revolution in the design and development of loudspeakers, with the veteran electrodynamic driver as the main protagonist despite its limitations. Nevertheless, there have been numerous technological advances in the field of electroacoustics. These have made it possible to develop new materials for manufacturing driver cones and more efficient crossovers built with fewer but better-quality components, as the most effective way of combating harmful phase rotations. As for enclosures, we find the most innovative designs aimed at reducing both spurious diffraction and internal resonances. In short, it is the ceaseless pursuit of the most linear frequency response with the lowest distortion, while increasing efficiency and achieving good sound pressure levels, even with small loudspeakers.

Robust, rigid construction is essential in any loudspeaker, which is why all kinds of internal bracing are used to prevent vibrations that would translate into spurious sound waves. The most widely used material is wood in its different forms: solid wood, particleboard, MDF or solid wood panels joined using various techniques to give them the most varied shapes in the best high-end designs. In issue 25 of CEC we had the opportunity to review a set of loudspeakers built with glass panels. For small loudspeakers, metal and even plastic materials are often used, in some cases of dubious quality, which should be rejected because of their very poor acoustic behavior.

The joints between the panels must be completely airtight, as must those of the drivers and the connection panel with the cabinet, since it is imperative to avoid uncontrolled air leaks. The inside of the cabinets is lined with sound-absorbing material, and in some bass-reflex designs internal structures based on labyrinth or transmission-line techniques are built, while sealed designs are usually filled completely with absorbent material to eliminate the energy of the rear wave of the cone. The midrange and treble drivers are enclosed in a separate chamber to prevent the rear wave of the bass driver from interacting with their diaphragms.

As we mentioned at the beginning of this article, every loudspeaker has a series of characteristics, which we will now describe.

FREQUENCY RESPONSE. This is how faithfully the loudspeaker follows the electrical signal it receives. This figure is always provided by the manufacturer, specifying the range of frequencies it can reproduce and the limits within which it maintains linearity -minimum deviation- in its response. By way of illustration, the graph in figure 1 shows a typical curve from which a frequency response from 30 Hz to 20,000 Hz could be deduced, although with a significant attenuation of 10 dB at 30 Hz relative to the sound pressure level, or SPL, of 90 dB. It would be more honest to say that this loudspeaker offers a fairly linear response curve between 60 Hz and 20,000 Hz, with an estimated deviation of +/- 1 dB. Catalog specifications should be read with caution, although manufacturers usually express the frequency range with a deviation of +/- 3 dB.



IMPEDANCE. This is the resistance that a circuit, in this case the loudspeaker's voice coil, offers to the electric current it receives from the amplifier. It is measured in ohms, with nominal values at a frequency of 1,000 Hz of between 4 and 8 ohms being typical for hi-fi and home theater loudspeakers. But this nominal value says very little about real-world behavior, since the impedance of a driver is determined by a constant value, the ohmic resistance of the voice coil, and by its inductive reactance, whose magnitude depends on frequency. As a result, the resulting impedance is not a fixed value but varies with frequency, producing peaks and dips along the whole curve, and it could drop to values even lower than half the nominal value, and not only in the most unfavorable cases.

Every time the impedance is halved, the associated amplifier will have to double its energy output, so in certain cases it may be advisable to choose a loudspeaker with as "comfortable" an impedance as possible or, conversely, an amplifier that guarantees sufficient energy supply capacity. Bear in mind that 8 ohms is the most comfortable load impedance for a solid-state (transistor) amplifier.

The graph in figure 2 shows a typical impedance curve for a loudspeaker with a nominal impedance of 8 ohms. Manufacturers do not usually provide this curve, limiting themselves to stating the nominal value and, at best, the lowest value it can reach. There may be the paradoxical case of a loudspeaker with a nominal impedance of 8 ohms that is harder to drive than one rated at 4 ohms, a typical example being electrostatic loudspeakers, in which the minimum value can fall below 1 ohm in the high-frequency range. For readers less familiar with these magnitudes, I should point out that a value of 0 ohms would be equivalent to applying a direct short circuit to the speaker outputs of any amplifier.



SENSITIVITY. The low efficiency of a driver is well known, and this can be extrapolated to the loudspeaker that houses it; to quantify this efficiency we measure its sensitivity. It is expressed in decibels (dB), and is equivalent to the sound pressure level obtained by applying 1 W, measured at a distance of 1 m on the loudspeaker's axis, with values of between 86 and 93 dB being typical for most loudspeakers on the market. Every 6 dB increase means doubling the sensitivity, so a loudspeaker rated at

86 dB will require twice as much amplifier power as one rated at 92 dB to achieve the same sound pressure level. Whatever a loudspeaker's impedance and sensitivity values, they will say nothing about its sonic qualities, although it is likely that a loudspeaker with excellent performance will have low sensitivity and/or impedance. We can already deduce that a difficult impedance curve, combined with low sensitivity, will mean choosing the driving amplification with the utmost care.

POWER HANDLING. The current delivered by the amplifier goes to the drivers' voice coils, so the power a loudspeaker can handle will be limited by the heat-dissipation capacity of those coils. If the current is excessive, the heat generated may melt the insulation of the coil windings and consequently destroy it. Therefore, depending on its capacity, it will be able to handle all or only part of the power the amplifier is capable of delivering, with which the loudspeaker will generate a given sound pressure level, or SPL. Catalogs usually express this figure with two values, for example 30-200. The first indicates the minimum recommended amplifier power, while the second refers to the maximum power handling.



DISTORTION. It is easy to tell the sound of a violin from that of a trumpet, even when both are playing the same musical note. This is because the fundamental frequency of that note is joined by other secondary frequencies of different intensity, called harmonics, and together, fundamental plus harmonics, they define the timbre of each instrument. Harmonic distortion occurs when a signal of a given frequency is amplified by a non-linear device -such as our driver- and therefore creates spurious signals whose frequencies are integer multiples of the first. Distortion generated by odd-order harmonics is much more unpleasant than that produced by even-order harmonics, and it is the cause of the aggressive, shrill sound of some loudspeakers of dubious quality. Intermodulation distortion is produced by the interaction of the aforementioned non-linear frequencies when they are close to one another. Catalogs usually state total distortion as the sum of all of them, although in some cases they specify that of particular harmonics.

The human ear is unable to detect distortion below 1%, a figure that can be exceeded in the very low frequencies, even in very good loudspeaker designs. This level of distortion decreases as frequency increases, reaching levels in the order of 0.5% or even lower at the highest frequencies, the range in which distortion is particularly unpleasant and can cause listening fatigue. This is very noticeable when, on switching off the system, we feel a genuine sense of relief.



DISPERSION. The listening position is the place in the room where the sound is perceived with the greatest sense of realism and "on-stage" presence of the performers. This is because sound waves have different propagation properties depending on their frequency. Low frequencies propagate in all directions, while high frequencies travel practically in a straight line, so it is worth analyzing to what degree each is attenuated in the horizontal and vertical planes of a loudspeaker. For the former, different measurements are taken through 360º, while for the vertical plane they are taken over an angle from +30º to -30º relative to the loudspeaker's axis, since dispersion beyond that range is out of reach for any loudspeaker.

The amplitude of the measurements obtained can be shown in what is known as a polar diagram, from which the loudspeaker's degrees of dispersion at different frequencies can be deduced with complete accuracy. In practice, no manufacturer publishes the polar diagram of its loudspeakers, offering us instead -and not all of them do- the range of degrees within which we will hear all frequencies with minimal attenuation. In general terms, it is not advisable to move the listening position more than +/-15º from the loudspeaker's vertical plane, or to allow for a listening area more than 1 m wide at a distance of 2 m. From this it follows that it is advisable to angle the loudspeakers toward the listening position.

Some of the above characteristics are closely related to the quality of the loudspeaker, so a very high-end model will -with exceptions- offer a wide, highly linear response curve, low distortion, good power handling and a good dispersion angle. However, some high-end designs based on electrostatic or ribbon transducers are very critical in terms of placement because of their directivity.