THE LOAD REACTIVITY
Now we should speak to you of reactance, phase angles, argument and compound numbers. But if you already throw up in front of real numbers, what a frozen heart we could have to speak of imaginary ones? We shall seek to explain it with an oversimplified example hoping that such a simplification won't cause Albert — may he rest in peace — to turn in his grave.
Imagine pushing a wheelbarrow: it represents your load. While on the plain there will be full correspondence between the force applied to it and the achieved speed, going uphill you'll see the wheelbarrow inexorably slow down. On the contrary, going downhill the speed will progressively increase and you'll have to do your best not to strike your snout to the ground. Almost likewise, in order that whole power is dissipated by the load this has to be purely resistive — the barrow on the plain. If the load presents also reactive components — uphills and downhills in the example, inductive and capacitive components in real life — then only a part of the power will be transferred to the load, the other one will be stored by these components. Where it will end up, you do not need to care a damn! What you do need to know is that the presence of these reactive components brings a shift between the tension applied to load and the current flowing in it. This phase-shift — so it is called in full, take a note in case they'll invite you again to that cocktail-party, thing that we doubt — in many cases can cause serious perturbations to listening.
If the speaker alone is reactive, think what it has to be with tons of capacitors and inductors to its connecting terminals! For this reason first-order filters are also called minimal-phase filters. They are preferred by many designers to cross multiway systems in which the overlap margin of unfiltered responses is adequately wide.
Filters of even order generally output a signal shifted by 180° — btw, phase shift is measured in sexagesimal degrees, from -180° to +180°, depending if the reactive component is capacitive or inductive respectively. To restore the correct phase it's advisable to alternatively reverse filter/speaker's connection polarities between contiguous ways.
Filters above second-order are definitely more difficult and they require skill and experience. What's more, their use isn't so frequent after all.
Now a brief return to attenuation slopes. Is it possible that in some cases this attenuation, that we have described as soft or steep depending on the filter order but constant with the frequency, does assume characteristics varying with the frequency? The answer is yes, and it calls into question a particular configuration, the cascading filters