All speakers have a little EQ built in

As one who occasionally dabbles in crossover design, often with one eye on the frequency response curve and the other on the impedance curve, I’d like to toss out an observation:

Equalizing a frequency response anomaly without introducing an impedance curve anomaly often calls for twice as many crossover parts as equalizing the frequency response curve alone. The impedance curve is arguably of greater consequence when designing with tube amps in mind than when optimizing for solid state amps.

It would never occur to me to hand over my crossover parts count to my marketing department lest they attempt to ascribe virtue to the sheer number, but I got a chuckle out of something that happened at an audio show:

This guy came into the room and sighed as if he was finally hearing something relaxing, and said to me, "I can tell you’re using a very simple first-order crossover." He was close! There were actually twenty-seven elements in that crossover, but the acoustic rolloff WAS first-order for about 2/3 octave on either side of the crossover frequency.

Duke

dealer/manufacturer

@akgwhiz wrote:

"@audiokinesis can you expand on how low, or how much of the phase curve that’s low inductance, at low impedance is acceptable for most tube gear?"

The interactions between the speakers’ impedance curves and the tube amps’ output impedances have too much variation for me to make a general recommendation of a specific minimum impedance value.

In general, a tube amp will put out LESS power (less wattage) into an impedance dip than a solid state amp; and MORE power (more wattage) into an impedance peak than a solid state amp will. So it’s not just the dips you have to be careful of; if there is an impedance peak in the crossover region (for example), you can end up with a frequency response peak in the same region when switching to a tube amp.

In fact when you read about someone trying a tube amp and it made their speakers sound worse, it’s very often because the interaction between the speaker’s roller-coaster impedance curve and the tube amp’s output impedance resulted in unwanted frequency response peaks and dips.

So I can’t really answer your question with specific numbers. The higher the tube amp’s output impedance, the more benefit from minimizing swings in the impedance curve and from keeping the average impedance fairly high.

Note in particular that the amp’s output impedance modifies the woofer’s effective electrical Q, raising it as if the speaker’s DC resistance was increased by the same amount. This has a greateer net effect on vented boxes than on sealed boxes. But it is also a potential "free lunch" that can extend the bass lower than it would have gone with a solid state amp, if the speaker was designed taking that higher output impedance into account.

Imo the thing to do is, use a modelling program that allows you to specify the amplifier’s output impedance, so you can get an idea of what the net result will be up and down the spectrum. It is possible to design speakers that work well with a wide range of amplifier output impedances, but woofer choices will be reduced, and you may need to make provisions for changing the port tuning frequency or for sealing up the port(s) completely.

Duke