Moth Audio Cicada loudspeaker
A fair question, and one of the few that attracts wisdom and selflessness as readily as their opposites. My own answers aren't set in concrete, but readers should know that the following survey falls into the latter camp. That doesn't mean I set out to write a Valentine to each of these companies; I did, however, set out to invite them all to the same Valentine party, and to give them all a chance to meet the same girls—meaning you (no offense).
Assume the position
During the years I edited Listener magazine, hardly a week went by without a phone call or e-mail from some reader who wanted to know: What affordable speakers can I use with my low-power, single-ended-triode amps? (We SET devotees are a mostly cheap lot, sometimes out of necessity, but just as often by inclination.)
From that you can assume that my own position on SETs is a sympathetic one. True: I've committed a chunk of my own time and money to this quaint little subset of an already quaint little hobby, with consistently fun results. A low-power tube amp in combination with the right loudspeaker is a fine way, though not the only way, to enjoy recorded music.
What makes one speaker more or less appropriate to the pairing? In looking for the answer, we'll bow to empiricism, common sense—and John Atkinson's expertise: Notwithstanding his lack of enthusiasm for SETs, an ailment for which listening to my harangues is surely the cure, JA's background in audio engineering and teaching will help us interpret loudspeaker specifications correctly (pronounce that last word as did Grady, the butler who tidied up Jack Nicholson in The Shining). Besides, JA hates to see the language abused.
What we're interested in are two things: loudspeaker sensitivity and loudspeaker efficiency.
Think you know how to find them? Here's a test: Take a look at the brochure or Web page describing your favorite speaker, and find the number that expresses efficiency. If you're pointing to the number of decibels the speaker can produce when fed 2.83V and measured from 1m away, you've given the wrong answer, and Grady must correct you. That number expresses the loudspeaker's voltage sensitivity. Sensitivity and efficiency are as different from one another as gain and power—an imperfect analogy, I know, but a helpful one, I hope.
Briefly, sensitivity describes how loudly a speaker will play given a certain voltage; efficiency describes how well the loudspeaker converts electrical power into acoustical power. In other words, efficiency is how easy it is or isn't for an amp to develop and maintain that voltage across the speaker, as a load. Sensitivity is measured in decibels relative to a given voltage and measuring distance (typically 2.83V and 1m, respectively), and efficiency can be assessed by looking, not only at sensitivity but also at a speaker's nominal impedance, in ohms. Efficiency can therefore be gleaned from a graph that plots a speaker's impedance, in ohms, vs frequency, in hertz.
The Wilson Audio Specialties WATT is a notable example of a loudspeaker that's reasonably sensitive (at 90dB/2.83V/m, it's 2-3dB more sensitive than what most people reckon is average), yet, because its nominal impedance is less than 6 ohms, and because its impedance curve contains drastic dips at audio frequencies (2.4 ohms at 78Hz, for example; see JA's measurements in the September 2003 Stereophile, Vol.26 No.9), the WATT is a load across which most amplifiers find it difficult to develop power consistently, with respect to frequency. And if the WATT challenges an average amp, it fairly torments a SET amp with pointed heels, a dog collar, and a riding crop: Because it lacks feedback and speaks to the world through an output transformer, the average SET has a high output impedance—often notably higher than the WATT's own impedance, and that of other, similarly wiggy-load loudspeakers. Thus the undoubtedly sensitive Wilson WATT is unabashedly inefficient.
Inversely, there are also speakers that exhibit very low sensitivity, yet whose predominantly high impedances make it easy for an amplifier to drive them—and, thus, we should think of such speakers as very efficient. The classic BBC LS3/5a falls into that category: an 82dB speaker with a nominal impedance of 16 ohms. I once had unexpectedly good results driving a borrowed pair of Rogers LS3/5a's with a 300B SET (roughly 7Wpc), leaving me to conclude, anecdotally and unscientifically, that efficiency may be more crucial to the SET experience than sensitivity.
For the most part, however, to be considered a good mate for a low-power amp, a loudspeaker requires both high sensitivity and high efficiency. The latter is easy to predict: 8 ohms minimum is the magic number, and where the curve dips, as all curves will, make sure it doesn't go too much lower than 8 ohms at any frequencies corresponding with musical notes or overtones that hold some interest for you. The former number, I dare say, is more open to interpretation: Depending on your installation, 89 or 90dB may prove sufficient. (With a 300B amp in a smallish room, I've had fine results using the 90dB-sensitive, 8-ohms-nominal Spendor SP100—surely one of the greatest "all-around" speakers you can buy.) Other users may find that their amplifier choice, room, and listening habits require something in the high 90s or beyond.
It's interesting to note that some hobbyists dismiss the SET craze solely on technical grounds, suggesting that any amplifier this loudspeaker-dependent is incompetently designed. Allow me to suggest that it's equally fair—quite possibly more so—to observe that any modern loudspeaker whose impedance dips below 3 ohms within the audio passband is no engineering marvel, either, an opinion that gains strength as the price of the thing goes up. It's horses for courses, as good ol' Grady would say, and, as always, the only real cure for any of our ailments is to listen, preferably to great musical performances and not railway sound effects.
And listen I did—at the end of my usual chain (see Sidebar), with Fi 2A3 Stereo (3Wpc) and Audio Note Kit One (7Wpc) power amplifiers.