Dear TWL and Teres:
With respect to the answer given by Mr. Crabbe, he's right when he says that the voice-coil impedance can be considered to be in series with the amplifier's output impedance. However, that's not the point. The ideal amplifier's mission is to work as a perfect voltage source, regardless of the load impedance presented at its output terminals. This ideal can be reasonably approximated using many technologies -vacuum tubes, bipolar transistors, mosfet transistors, etc.- However, using vacuum tubes this task is almost unsurmountable, simply because tube 99.9% of current designs depend heavily on the use of an output transformer that further introduces anomalies, distortions, colorations, saturations, phase shiftings, etc. etc. By the way, perhaps at this moment the word OTL will come to your mind, but I could mention many aggravations they have too, but that is another matter that would be subject of a different thread. Anyway, assuming that all these problems could be solved, tubes generate big amounts of harmonic content not present in the original signals, and I see this as the biggest obstacle in our road to true fidelity. Tubes will always add sweetness to the music, at the expense of accuracy, that to many music lovers as myself means fidelity. Today's solid state devices, on the other hand, have current linearity far more linear and and noise characteristics far more desirable, all in favor of a better musical experience. This is the ultimate goal of listening to music.
And regarding Teres arguments, I have several points I would like him to consider: Tubes are voltage devices, whereas transistors are current devices. Therefore they operate in completely different fashions, not similar. Any electronics designer can confirm that. Although tubes and transistors both present big amounts of distortion if not corrected, transistors have the advantage of having complementary devices, that when properly matched, can virtually eliminate alinearities, harmonic distortion and IM distortion without needing any global feedback. Unfortunately, there are not and will never exist such thing as a complementary tube; they simply can't have this fundamental advantage. So when you say that transistors depend on big amounts of feedback, I can see that you have been simply ignoring the state-of the-art in modern solid-state design, where non-feedback is almost a gospel.
Ultimately, there will never be such thing as a zero-feedback amplifying device, and this includes single-ended triodes (where the electron flow and the space charge are the feedback parameters working in conjunction with the grid-cathode voltage) and transistor emitter-follower configuration (where the flow of minority carriers is the feedback parameter). I agree that a zero-feedback amplifier would be great, but it's not achievable yet.
Regards and enjoy the music,
Raul
