There is no "summing" of anyting in a speaker/crossover/amplifier/cable interface. Quite simply, there is a low pass crossover network usually consisting of a series inductor shunted on it's output (and across the woofer's input terminals) by a capacitor. This yields a 12db per octave low pass filter for the woofer that eliminates the high frequencies going to that driver.

Then there is another (completely seperate) network consisting of a series capacitor that is shunted on it's output side by an inductor (across the terminals of the HF driver). This yields a 12db per octave high pass filter that eliminates the low frequencies going to the tweeter driver.

There are also sometimes other components to change the frequency and phase contour of the signals being fed to the individual drivers. These fall into the category of components used in order to "voice" the speaker.

The two "ends", the inputs to these two independent crossover sections are tied together - this is really like a "Y" adapter. This constitues splitting the signal, not "summing" it. Huge difference.

That tying together point can be at the back of the speaker via that bar that connects two sets of input terminals to the two crossover network sections inside the speaker. Those two sets of terminals can be fed either by removing the shorting bar and using two seperate amplifiers, or by two seperate wires. Or you can leave the shorting bar in place and feed the whole schbang with one wire and one amplifer.

The woofer is fed a signal that goes basically from DC (let's hope not though, as that means you have a fried amplifier channel!) up to the crossover point - anywhere from around 250Hz to around 3kHz. It is the crossover network that limits the high end extension - the upper limit of frequencies being fed to the woofer. The tweeter crossover network works in a similar manner except that it filters out all frequencies below the crossover point and feeds the remaining signal to the tweeter.

An active crossover network just does this frequency division with active electronics and usually also contains components like equalization and phase compensation (at least in some crossovers) to perform the "voicing" task. In these systems, the passive crossover components inside the speaker are eliminated or bypassed, and the output terminals of the high and low frequency power amplifiers are connected directly to the input terminals of the actual speaker drivers, without having to pass through the original passive crossover network.

[This message has been edited by soundhound (edited May 16, 2004).]