What Makes a Good Crossover?

A crossover is often described as the part of a loudspeaker that sends low frequencies to the woofer and high frequencies to the tweeter. That is true, but it is only the beginning.

A good crossover must do much more. It affects frequency response, phase, directivity, power handling, impedance, lobing, and the apparent position of sound. Most importantly, it determines whether several physically separate drivers behave like several sound sources or combine into one believable loudspeaker.
The real objective of a crossover is therefore not simply to divide frequencies. It is to make several drivers speak with one voice.

Why Do We Need a Crossover?

No practical loudspeaker driver reproduces the entire audible range equally well. A woofer is designed to move enough air to reproduce bass, but its relatively large diaphragm eventually becomes directional and enters breakup at higher frequencies. A tweeter is small enough to reproduce high frequencies well, but usually cannot move enough air safely to reproduce deep bass. The crossover allows each driver to operate primarily within the range where it behaves best. Low frequencies are gradually reduced before they reach the tweeter, while high frequencies are gradually reduced before reaching the woofer. The word gradually is important. A crossover is not a brick wall. There is usually a region where both drivers are operating at the same time. That overlap is where things become interesting.

Two Drivers Must Become One

Suppose a woofer and tweeter are both producing sound near 2,000 Hz. The listener does not want to hear two separate 2,000 Hz sources. Their outputs must combine acoustically into one convincing result. Whether they do depends upon more than their individual frequency responses. Their relative phase, physical spacing, acoustic centers, directivity, and crossover slopes all influence the sound that finally reaches the listener. If the two drivers combine properly, their outputs reinforce one another smoothly. If they do not, parts of the spectrum may become too loud, too soft, or uneven with listening position. This is why crossover design cannot be reduced to choosing a frequency and calculating a few components.

The Crossover Frequency Is Not Arbitrary

A designer should not begin by deciding, "I want to cross these drivers at 2,000 Hz," and then force the drivers to cooperate. The drivers should determine the sensible crossover region. A woofer may appear capable of operating to 4,000 Hz according to its published frequency response, but its diaphragm may begin breaking up or its radiation may become too narrow well below that point. A tweeter may reproduce 1,500 Hz on a laboratory graph yet produce excessive distortion or require too much excursion if asked to operate there at realistic levels. A good crossover works within the comfortable operating ranges of both drivers.
The crossover should protect the drivers, not challenge them.

Phase Matters

When two sound waves meet, their timing relationship affects how they combine. If they arrive appropriately aligned, they reinforce one another. If their relationship is wrong, they can partially cancel. This timing relationship is called phase. A crossover changes phase as well as amplitude, and the physical positions of the drivers introduce additional differences in arrival time. For that reason, electrical crossover behavior and physical loudspeaker geometry cannot be considered separately. The objective is not necessarily to make every phase measurement mathematically perfect. It is to create predictable acoustic summation throughout the crossover region and useful listening area.

Crossovers and Lobing

Because the woofer and tweeter occupy different physical positions, sound from them travels slightly different distances to different points in the room. Where both drivers are reproducing the same frequencies, their outputs may reinforce in some directions and partially cancel in others.

This produces areas of stronger and weaker sound called lobes. Some lobing is inevitable in most conventional multiway loudspeakers. The problem arises when it becomes severe enough that the speaker changes character significantly as the listener moves vertically or horizontally. Good crossover design, sensible driver spacing, and appropriate crossover frequencies can reduce these effects considerably. This is one reason placing drivers close together is often desirable. The more nearly several radiators behave as though they occupy the same physical position, the easier it becomes to create a stable acoustic image.

Directivity Must Match Too

Imagine a woofer whose radiation has become very narrow by the time it reaches the crossover frequency. Now imagine handing the sound abruptly to a tweeter that radiates very broadly. Directly in front of the loudspeaker, the frequency response might still look excellent. Elsewhere in the room, however, there may be a sudden change in radiated energy at the crossover. The room hears that discontinuity even if the microphone directly in front of the speaker does not. This is why crossover design and directivity are inseparable. A good transition requires not only similar sound levels but reasonably compatible radiation patterns.

More Components Are Not Automatically Better

A complex crossover can perform useful functions. Additional components may correct response irregularities, control impedance, suppress resonances, or produce a desired acoustic slope.

But complexity is not a virtue by itself. Every component introduces cost, tolerance, resistance, and another opportunity for unintended interaction. Conversely, an extremely simple crossover is not automatically superior either. A single capacitor may be elegant, but elegance does not excuse poor integration. The correct crossover contains as much complexity as necessary and no more than the design requires. Neither the fewest components nor the greatest number wins.

Passive or Active?

A traditional passive crossover operates between the power amplifier and the loudspeaker drivers. It uses components such as capacitors, inductors, and resistors to divide and shape the signal. An active crossover performs this job before the power amplifiers. Each frequency range is then amplified separately and sent directly to its appropriate driver.

Both approaches can work very well. Active systems offer greater flexibility and remove some difficult interactions between passive crossover components and driver impedance. Passive systems can provide simplicity and allow the loudspeaker to operate from a single conventional amplifier. As always, the appropriate method depends on the objectives of the design.

A Crossover Cannot Rescue Bad Decisions

This may be the most important point. A crossover can shape the electrical signal reaching a driver, but it cannot transform an unsuitable driver into a suitable one. It cannot eliminate severe diaphragm breakup, restore wide dispersion to a driver that has become directional, or make widely separated acoustic sources occupy the same physical location.
By the time crossover design begins, many of its possibilities have already been determined by driver selection, cabinet geometry, and physical layout. Good crossover design therefore begins long before the first capacitor or inductor is chosen.

One Loudspeaker, Speaking With One Voice

When a crossover works properly, the listener should not be conscious of it. Nor should the listener hear a woofer operating below a certain frequency and a tweeter taking over above it.

A voice should remain a voice as its harmonics pass from one driver to another. A piano should not appear to change size with pitch. Instruments should occupy stable positions in the stereo panorama instead of shifting as different drivers become dominant. That is the real test.

A good crossover does not merely divide the music into pieces and distribute them among the drivers. It puts the pieces back together.