I'm designing crossovers for my rear channels; a 2500 Hz 3rd order butterworth between an 8 ohm ribbon tweeter and two 4 ohm woofers in series (8 ohm load).

While my original plan was to use two DPDT switches for L-Pad attenuation circuits; in one position there would be no attenuation, in another there would be a ~2dB attenuation. However, I found On-Off-On DPDT switches that would enable me to 'remove' the tweeter, woofers, or both from the circuit and this intrigues me.

However, there will be an inductor and a capacitor in series remaining on each channel when in the 'off' configuration.

I'm wondering:

  1. When the drivers are removed from the circuit, will the remaining LC/CL circuit be forced to dissipate the power (up to 175W)?
  2. When both drivers are removed from the circuit, what 'load' would be 'seen' by the amp?
  3. If only one of the drivers are removed, how will the LC/CL circuit affect the remaining driver?

Here's the circuit:

Crossover Circuit


If you want to remove drivers from your system and not have any horrible impedence changes then the safe way is to switch a resistor in when the driver in question is switched out .The resistors should be the same as the speaker impedences ie 4ohm and 8 ohm .The power rating for the woofer should be for the average program power of the amplifier .This could be half of the amps output power .The tweeter resistor can be about a tenth of the amplifier power because most of the audio power is in the bass.In other words the resistors wont be too expensive .Remember that reactive power can make amplifiers produce more distortion than specified ,in fact most people test the amp into a resistive load to save thier ears.Sure your passive crossover wont have a perfectly flat impedence Vs frequency curve but the amp will drive it .When you substitute the resistors the curce will not get any worse .Finaly the resistors mean that you will not blow up the amp!.

  • \$\begingroup\$ That would require two additional switches though, would it not? \$\endgroup\$ – Ehryk Sep 29 '15 at 19:35

A few notes:

  1. It's important to remember that Butterworth passive filters, etc, still have frequency response dependent upon load. So your current design will likely shift a bit in frequency when the load goes from 8 to ~28 Ohms. I don't have the time to run the numbers on it right now, but something to be aware of.

  2. The power won't be "forced" to be dissipated - the output voltage (usually) or current (occasionally) will be dictated by the amplifier, but it would be highly unusual for it to dictate both. Answering question 2 will help you.

  3. (C1 in series with L1), in parallel with (L2 in series with C3). The load impedance will be a function of frequency, but almost entirely reactive (there will of course be series resistances from the non-ideal components). If the amplifier drives more outputs than just this one, reactive power could be an issue as Autistic mentioned.

  4. Likely not very much, although depending on amplifier design you might get some distortion. Assuming an ideal amplifier, this would have no effect (your mileage may vary). Switching in a resistive load is a good alternative solution.

  • \$\begingroup\$ Thank you. I cannot find any 2P3T rocker switches, and I'd prefer not to have four switches instead of two. I think I'll just try this and see how it works out (and likely never run them with drivers bypassed once I satisfy my curiosity). \$\endgroup\$ – Ehryk Sep 29 '15 at 20:43
  • \$\begingroup\$ A suggestion: if that's the case, buy normal 2P2T/2PDT switches without the "off" position, and don't hook up the attenuator until you've satisfied your curiosity? \$\endgroup\$ – Kikanaide Sep 30 '15 at 9:04
  • \$\begingroup\$ Do you know of any online or downloadable calculators where I could attach the components and do a frequency sweep from 20-20k Hz? I have seen screen shots of this type of thing, but I'm not sure where to get them from. \$\endgroup\$ – Ehryk Oct 1 '15 at 18:01
  • \$\begingroup\$ You are looking for a "SPICE" simulator. There are a few free ones (see here electronics.stackexchange.com/questions/1206/…), notably TINA-TI and LTSpice. Either should do what you want, though I've never used either one. \$\endgroup\$ – Kikanaide Oct 2 '15 at 18:54
  • \$\begingroup\$ Okay, technically there are some non-SPICE circuit simulators as well (one of which is referenced in that thread). This being just RLC circuit stuff, you might be better off with that. \$\endgroup\$ – Kikanaide Oct 2 '15 at 20:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.