Crossover 101

By Marty McCann
© 2001

What is a Crossover?
A crossover is a filtering device that separates the audio frequency spectrum and directs portions of that spectrum to a specific audio device. As an example, all sound reinforcement quality two-way loudspeaker enclosures have a passive crossover as part of the design that separates the audio spectrum into two bands of frequencies. This passive crossover is a bit like a traffic cop that directs the lows to the woofer and the highs to the horn. All of the frequencies below the designed crossover point (Low Pass) go to the woofer, and all of the frequencies above this crossover point or frequency (High Pass), are directed to the high frequency compression driver and the horn. This is accomplished via a network of inductors, capacitors, and resistors that direct or pass the proper information to the proper transducer.

How does the crossover work?
Inductors pass low frequencies and capacitors block low frequencies. Capacitors pass high frequencies and block low frequencies. You can have both a very high quality woofer and horn/driver combination, but it takes an equally high quality crossover circuit to make the units perform well as a system. The more complex the crossover network, the greater number of poles or orders you will find in the design. The more poles or the higher the order of crossover the more components you will have in the low and high pass filter stages of the design. The more filter stages, poles, or orders incorporated in the design, the more phase shift you will introduce to the audio signal. The phase of the output signal of the crossover can either lead or lag that of the original incoming signal. When signal passes through an inductor, the current flow is said to lag or be behind the voltage in regards to time. When an electrical signal passes through a capacitor, the current flow is said to lead or be ahead of the voltage in time. So in a simple single pole or first order crossover network, the voltage output of the high pass port will be shifted -90 degrees in phase, while the low pass will exhibit +90 degrees of phase shift, with the maximum of 90 degrees being at the extremities or farthest away from the crossover point.

Each crossover filter network stage (order or pole) will also introduce a specific rate of roll-off or attenuation of the signal, above and below the crossover frequency point. The rate of attenuation or roll-off is measured on the decibel scale in the form of so many dB per octave of attenuation (signal reduction) above and below the crossover frequency. For instance a 1st order filter network will attenuate at a rate of -6 dB per octave. Let's use the example of a 1600 Hz crossover design. The crossover frequency is designated by the -3 dB down point from flat, so in a two-way 1600 Hz crossover design, the Low pass output would be down -3 dB at 1600 Hz, as would the High pass. Above and below this crossover frequency the system would be flat (with the exception of CD horn attenuation and EQ, which will be addressed later).

Below is a chart to represent the filter order or number of poles, the rate of attenuation and the maximum amount of phase shift in degrees (+/-):

Note: It is possible to build 5th through 7th order but they are not generally used in audio. Most passive crossovers do not go beyond 3rd or 4th order filters, and some passive designs may chose different orders for the low pass than the high pass, (example 2nd order low pass & 3rd order high pass).

The following graphics will depict the roll-off and phase shift apparent in typical filter orders. I will use 800 Hz as the crossover frequency to keep the graph more symmetrical:

What is Constant Directivity Horn EQ?
Constant Directivity high frequency horns require a special form of equalization that permits them to exhibit proper high frequency response. In this article we will explain the need for the high frequency compensation known as CD horn EQ. The current Peavey models RX-22 and 44XT compression drivers require this form of high frequency equalization when used on our constant directivity high frequency horns, or for that matter, any manufacturers constant directivity high frequency horn.

All high frequency compression drivers perform more efficiently or play louder than their paper cone loudspeaker counterparts. The efficiency of a loudspeaker is measured by driving the loudspeaker with one watt of input power while measuring how loud in sound pressure level (SPL) it will be at a distance of one meter from the loudspeaker enclosure. This is called the One Watt, One Meter Sensitivity rating of the loudspeaker.

A typical compression driver may have a one watt at one meter sensitivity rating of 112 dB of SPL, while a typical paper cone loudspeaker used for sound reinforcement may exhibit a one watt at one meter sensitivity of 100 dB of SPL. In order for the two transducers to produce the same acoustic level from a loudspeaker enclosure, the crossover must provide for -12 dB of Attenuation or reduction (Pad) in the signal level of the high frequencies going to the compression driver.

Every two-way loudspeaker system provides this high frequency pad or attenuation in the system's passive crossover, and is standard throughout the industry. However, in addition to this standard pad or attenuation that is designed into the loudspeaker's passive crossover, it is also necessary to provide a special high frequency equalization when the compression driver is used on a constant directivity horn.

In today's professional audio, the constant directivity high frequency horn allows us to obtain uniform high frequency response with dispersion or angle of coverage. Before the introduction of the constant directivity horn in the mid seventies, all high frequency horns exhibited the same common problem, i.e., the horn may have measured very flat on axis or directly in front of the horn, but as you moved off-axis of the horn itself, the higher frequencies would not be equal in level to the mid range of frequencies that the horn produced. This narrowing of the beamwidth at high frequencies was due to the very rapid flare rates associated with these earlier exponential horns. An exponential high frequency horn is one whose flare rates or taper increases proportionally to the square of the distance away from the throat entry to the horn.

The very small wavelengths of the higher frequencies could not cling to the rapidly expanding side walls of the exponential horn to be directed off axis; therefore the high frequency energy radiated directly down the center of the horn and exited in a pattern about equal to the diameter of the entry to the horn throat. Constant directivity (CD) high frequency horns were first introduced in the late nineteen seventies. Using computer assisted design (CAD) the internal parameters of the side walls of the horn were manipulated resulting in flare rates that were more gradual allowing the smaller high frequencies wavelengths to be directed off axis. Today many manufacturers make these constant directivity (CD) horns that offer uniform frequency response with dispersion.

Since the CD horns are now able to direct the high frequencies off axis, the amount of high frequency energy formally available directly on axis is less. Therefore the CD horn no longer measures flat directly on axis without its needed signal processing in the form of a special high frequency equalization that is designed to be the reciprocal or mirror image of the horn/driver high frequency roll off response. This is what is meant by constant directivity horn equalization (CD EQ). All CD horns roll off the higher frequencies at about -6 dB per octave, and the CD horn EQ is usually in the form of a +6 dB per octave boost beginning at about 3 or 4 kHz.

The high pass section of a passive crossover in a loudspeaker system that employs a constant directivity high frequency horn provides two additional functions beyond that of the high pass filtering circuitry. The first additional function is the necessary pad or attenuation to match the sensitivity of the paper cone loudspeaker. Function two is the equalization necessary to allow the driver to have a flat response in the last two octaves. How is the equalization or boost of high frequencies accomplished in a "passive" crossover? Essentially there is a primary crossover circuit that crosses over the audio signal at the designated crossover frequency and provides the proper amount of attenuation. Then a secondary crossover circuit uses the voltage window of the un-attenuated high pass signal to provide an additional signal path for the high frequencies with less and less attenuation as the frequency rises.