On 09/06/04 22:46 ET, Theaterworks said...
|
Really? How cheap are we talking about? All
the speaker switchers I've peered inside of (Sonace,
Niles & the like) had a large wattage 4 ohm resistor
and some switches, nothing more. Nothing there
(as far as I know, correct me if I'm wrong) that
could alter the frequency response of the amp/switcher/speaker
circuit.
The problems are more subtle than those listed.
Yes, every switch box I have seen has a resistor in it. I think I once saw one with 2 ohms, just enough to keep the load above 2 ohms if eight or so speakers in parallel were in series with this resistor.
Let's look at the frequency response.
First, one speaker. Speakers have different impedance values at different frequencies, so putting a resistor, which has the same resistance at all frequencies, in series with a speaker will affect the amount of power getting to the speaker at each frequency, and thus the frequency response.
In a typical speaker system, there is a bass resonance (or two) that usually peak well above twenty ohms. I a, being conservative; I have seen 60 ohm peaks. If you feed your amp output through a resistor, the response at those peaks will be changed very little.
But an octave above those peaks the impedance is usually around 8 ohms. That means that 1/3 of the power going to the speaker at that frequency will be absorbed by the resistor. Result? Lower output in the midrange.
The crossover should keep energy out of midrange and tweeter drivers at their resonance, so the effect will be similar to that on the woofer in its eight ohm-ish range.
But I have also seen well-designed speaker systems vary in midrange and tweeter impedance from six to ten ohms. Thus the resistor will affect frequency response.
Add to this that old bugaboo, damping factor. A speaker is a motor, and it should move, then stop, to put out the sound that is fed to it. To do this, the amp and all associated wiring has to have zero ohms impedance. This is impossible, but the lower the impedance of the source feeding the speaker, the better controlled the woofer motion will be. (Midrange and tweeter motion are better controlled for various reasons.) When you put a resistor in series with a speaker, you lessen the woofer's ability to stop moving when the voltage drops to zero, so the bass can become less clear.
On all of the above, there is one last issue: can anybody hear any of this? It is real, and it is measurable, but most of the distributed audio systems I have heard sounded pretty good!
Now, take two speakers, put them in parallel, and use the same resistor; this will be like putting an eight ohm resistor in series with a single speaker. Make it four speakers in parallel, and this will be like putting a 16 ohm resistor in series with each speaker. Thus the effect of that single series resistor increases as the number of connected speakers increases.
Turn the volume up and down, and the volume control will offer slightly different impedances to the switch box. Now the frequency response varies as the volume varies, in addition to not being accurate.
Again, can we hear it? Don't forget that the frequency response of our ears varies with volume. At lower volume, we generally hear low bass and high highs less well. How can you tell if a speaker's frequency response is varying with volume when your ears vary with volume?
To sum up, I personally would go with the cleanest method, especially if there is any chance that different rooms will EVER want to play different audio: use the multiple channel amp with volume controls in each room, or line level volume controls in a preamp before each pair of amps, as in a multi-room situation. Sometimes such an amp is not in the budget, or is not in the budget for this year: I have put in many single-amp, big transformer, impedance-matching volume control systems, which are not the best. But they sure sounded good!