| Audio Crossover |
Article Index for Audio |
Website Links For Audio |
Information AboutAudio Crossover |
| CATEGORIES ABOUT AUDIO CROSSOVER | |
| linear filters | |
| analog circuits | |
| audio electronics | |
| audio engineering | |
| filter theory | |
| speakers | |
|
OVERVIEW Audio crossovers (filters) are characterised by a linear '' Phase Response '' and constant Group Delay (typical of Bessel Filters ) as opposed to the maximally flat Frequency Response and variable Group Delay (typical of Butterworth Filters ) of most industrial and data acquisition filters within their Pass-band . A crossover is constructed as a combination of Low-pass , Band-pass and High-pass filters (LPF, BPF and HPF respectively). The number of Pass-bands determines the "way" of the crossover. Thus, a 2-way crossover means it has one LPF section and one HPF section. A 3-way crossover has one LPF section, one BPF section and one HPF section (the BPF section is in turn a combination of HPF and LPF sections). The filters are so designed that each section has a −6 DB amplitude at the crossover point(s). Thus,
Although the 2nd order Butterworth Filter can be used to construct an audio crossover, it generally is not considered the ideal choice, and is fast being replaced with the Bessel or '' Linkwitz-Riley '' (L-R) alignments. Butterworth filters when used in a crossover, are 3 dB down at the crossover frequency(ies), which means that the summed frequency response of these filters exhibits large peak(s) at the crossover point(s) - this is not an ideal (i.e., flat) frequency response. As far as audio is concerned, a 2nd order Bessel or 4th order L-R filter characteristic is preferred. The reason is that the Bessel and L-R filters, when used in a crossover are '''6 dB down''' at the crossover frequency(ies). For 2nd or 4th order crossovers, the phase difference between the signals in the bands related by the crossover point is always 0° (for 4th order) or 180° (for 2nd order, which is corrected by inverting). This becomes important at the crossover frequencies, which in the case of Bessel or L-R is when the amplitude is −6 dB relative to the pass-band gain — this causes the summed frequency response at the crossover point(s) (and their vicinities) to be flat. Thus the overall the summed response is also flat and 6 dB lower than (at −6 dB relative to) the pass-band maxima over the entire audio spectrum. With a 2nd order Butterworth, since the crossover point is the same as the −3 dB point, there is 3 dB peak in amplitude response at the crossover point — note that a Butterworth filter has a maximally flat ''amplitude'' not phase response. Note: The crossover points (frequencies) must not be confused with the −3 dB Cutoff Frequency of filters. While the −3 dB frequency means the same thing (as in Bandwidth , for example) for all filters irrespective of their order or type, the crossover point is the point (frequency) where one filter stops (or fades out) and the other filter takes over. Specifically, the 2nd order Butterworth filter must have the crossover point to be the same as its −3 dB frequency in order to work, otherwise the summed response will exhibit a large or wide dip in the frequency response — this means we cannot simply design two 2nd order Butterworth filters and some how cross them over at −6 dB. 2nd order Bessel filters will have a −3 dB frequency away from their −6 dB points and crossing them over at their −6 dB points gives a flat summing — a result of the Bessel characteristic. CROSSOVER CLASSIFICATION Classification based on filter order Just as filters have different orders, so do crossovers — depending on the filters they use. Most audio crossovers use first to fourth order filters. Higher orders are generally not used for audio as there are less gains corresponding to the resulting high cost and complexity. First order crossovers 1st order filters have a 20 dB/ Decade (or 6 dB/ Octave ) slope. All 1st order filters have the same characteristic and are therefore not classified into Bessel, Butterworth etc. 1st order filters are considered by many Audiophiles to be ideal for crossovers. This is partially true because this crossover uses minimum parts, has relatively less loss (if passive) and is more predictable than higher order passive crossovers. The predictability factor stems from the fact that a typical Loudspeaker is not a resistive load but reactive (''See Electrical Characteristics Of A Dynamic Loudspeaker ''). This means that the impedance that the loudspeaker presents to the passive crossover varies with frequency — affecting the crossover point of the crossover itself. A 1st order crossover is simplest in this respect since the factors that impair the crossover's performance are minimum. However, 1st order crossovers are generally not found in hi-fi professional systems. They are more common in off-the-shelf component systems or micro-systems and that too in 2-way speakers. A 1st order crossover may contain nothing more than a simple capacitor connected in series with the Tweeter , the Woofer being directly connected to the amplifier. The crossover point is determined simply by the woofer's natural response and the combination of the tweeter's impedance and the series capacitor. A 1st order crossover tends to allow relatively more unwanted frequencies to get through in the LPF and HPF sections. While woofers can easily take this, the tweeters may be damaged since they are not designed to handle large powers and lower frequencies. Second order crossovers 2nd order filters have a 40 dB/decade (or 12 dB/octave) slope. 2nd order filters can have a Bessel or Butterworth characteristic depending on design and choice of components used. This order is most commonly used in passive crossovers as it offers a good balance between complexity, response and tweeter protection. Where a 2-way 2nd order crossover appears, the tweeter is wired in opposite phase to the woofer to correct for crossover phase. In active form this crossover is very simple to design and build and usually has the HPF output inverted thereby not requiring the tweeter to be wired out of phase. Third order crossovers 3rd order filters have a 60 dB/decade (or 18 dB/octave) slope. These crossovers usually do not have a Bessel or Butterworth characteristic and phase response is neither flat nor abrupt. These are not commonly used in passive form, although some active crossovers employ them. Fourth order crossovers 4th order filters have an 80 dB/decade (or 24 dB/octave) slope. These filters offer the best performance but are more complex to design in passive form. They are more commonly found in active crossover form and are considered superior in all respects for audio. A 4th order crossover with −6 dB crossover point and flat summing is also known as a ''Linkwitz-Riley'' crossover (named after the inventors of this crossover). It can be constructed by cascading the individual filter sections of a 2nd order Butterworth crossover. Classification based on components Crossovers can also be classified based on the components they use. Passive crossovers A passive crossover is made entirely of passive filters. The passive filters employ passive components like Resistors , Capacitors and Inductors for their operation. These are relatively cheap when the cost of the entire equipment is considered. On the negative side,
Active crossovers An active crossover contains active components in its filters. The most commonly used active device is an Op-amp and active crossovers are operated at line-level powers meant to drive amplifiers in contrast to passive ones which operate at the high-level powers meant for speakers. Active crossovers always require the use of power amplifiers for each band. Thus a 2-way active crossover needs two amplifiers — one each for the woofer and tweeter. This means that an active crossover based system might end up costing more than a passive crossover based system. It also requires the use of a tweeter protection capacitor (≥ 22 µF) since the tweeter is now directly connected to the amplifier and may be damaged due to DC or the short thump an amplifier produces as the amplifier is powered on. The cost disadvantage is offset by the following gains:
Mechanical crossovers This crossover is entirely mechanical in design and uses the properties of the materials used in a speaker to achieve filtering. It is commonly found in Full-range speakers which are designed to cover as much of the audio band as possible. A mechanical crossover is constructed by coupling the diaphragm of the speaker to the voice coil through a compliant material and directly coupling a small light-weight cone called ''whizzer'' to the voice coil. The compliant material ensures that the diaphragm responds only to lower frequencies while the whizzer which is directly coupled to the coil can respond to the rapid movements of the coil at high frequencies. This combination results in the diaphragm having an upper cut-off frequency while the characteristics of the whizzer and voice coil set the lower limit to the whizzer's response, thereby approximating a crossover action. The choice/weight of materials used for the compliance, diaphram, whizzer and the speaker's suspension determine the crossover frequency and the accuracy of the crossover. This crossover does not suffer from the disadvantages of its passive electrical counterpart(s) but is more complex to design, especially if the highest degree of performance is desired. This type of crossover (rather the speaker that uses it) can considerably simplify the overall setup since it eliminates the need for an electrical crossover, at least as far as the range of frequencies this speaker is designed to handle is concerned. Also, due to the way this crossover is fabricated, speakers using it are almost always 2-way. ''See Full-range Speaker for construction details''. Classification based on topology Parallel crossovers These are by far the most common. Electrically the filters are in parallel - with the exception that the outputs of the filters do not connect to each other, but to the next stage (power amplifier, if active or speaker driver, if passive). Series crossovers Crossovers using this topology are almost always passive. This is because it is easiest construct in passive form. In active form this topology has no particular benefit and hence active crossovers using series topology are not found (or are very rare). In this topology the individual filters are connected in series, with one driver connected in parallel to each filter. As can be seen in the image, a low-pass filter in shunt with the tweeter results in a high-pass response for the tweeter, since the lower frequencies are shunted by the LPF via the woofer. Similarly, the HPF in parallel with the woofer shunts away the higher frequencies via the tweeter - a low-pass response for the woofer. A particular advantage of this crossover, when used in 1st order passive form is that the crossover compensates for variation in Driver Impedance with frequency, thereby maintaining the crossover point. The disadvantage is that the drivers are not well-isolated from each other. {Link without Title} REFERENCES
SEE ALSO
|
|
|