Subtractive Crossovers
Douglas Self · 2012
What you might call the standard crossover architecture has a lowpass filter to generate the LF signal for the bass drive unit, and a corresponding highpass filter to create the HF signal for the tweeter/midrange driver. (Obviously that only describes a two-way crossover and a three-way crossover has more filters.) That is not, however, the only way to do it, and one of the alternatives is the subtractive crossover. In this you just have one filter, say the lowpass which gives the LF signal, and you make the HF signal by subtracting the LF signal from the original input, so that HF = 1 − LF. This process is illustrated in Figure 6.1, which shows a first-order subtractive crossover. Higher-order subtractive crossovers can be implemented by replacing the first-order filter with a higher-order version; Figure 6.3 shows a second-order subtractive crossover. There is no necessity for the one filter used to be the lowpass type. It is equally possible to use a highpass filter, and generate the LF signal by subtracting the highpass filter output from the original input. Subtractive crossovers are also called derived crossovers as one output is derived from the other instead of being filtered independently, and constant-voltage crossovers because of the way that their outputs sum to reconstruct the original waveform. The constant-voltage crossover, first properly described by Dick Small [1], one of the great pioneers of scientific loudspeaker design, is a subtractive crossover.