The Sound of Transistors

Editor's Note: Those of us who cut our engineering teeth on tubes still remember the advent of the solid-state amplifier with mixed feelings. Yes, they were lighter and cheaper per watt than the thermionic hulks we loved so much, but they broke all the time (thanks to the germanium transistor) and sounded like—well, let J. Gordon Holt tell us what they sounded like in an "As We See It" article from Vol.1 No.10, first published in May 1965. We also develop the theme with a JGH review of an early transistorized amp, as well as a selection of readers' letters from the early days of Stereophile. Enjoy.John Atkinson

Ever since the first solid-state high-fidelity component (a Fisher phono preamp) made its appearance ca 1957, we've been hearing more and more about some wondrous phenomenon called "the transistor sound"—a quality of sharpness and crispness that, allegedly, makes tube components sound muddy by comparison. We've been auditioning a number of transistorized components recently, and while we are now fully prepared to assure skeptics that there is indeed such a thing as "transistor sound," we are equally prepared to state that it ain't as hi-fi as it's cracked up to be.

A typical transistor amplifier will make any loudspeaker sound crisper, as though its transient response has been improved, but it is not the same kind of "crispness" we can get from a tube amplifier by just cranking up the treble. In a tube amplifier, treble boost will increase the "harshness" of disc distortion. The typical transistor amp, on the other hand, does not seem to exaggerate "hash" as much as average tube equipment does, so the sound is often as clean or cleaner, despite the increased crispness.

Prolonged listening, however, brings a gradual realization that all program material has this razor-sharp clarity, including recordings that we know don't go much beyond 9 kc [kHz] at the high end. And when the same transistor amp feeds a speaker that has inherently excellent transient response, the crispness is just too much to endure. Highs come through with subtle "whiskers" of high-frequency impulses riding on them—impulses that are not heard either in live music or in the best tube or transistor equipment.

This is why we've been stressing "average" transistor equipment, for the better the transistor amplifier—that is, the lower its measured distortion—the more it comes to sound like a top-grade tube amplifier, and the better it sounds on a top-grade speaker system.

Unlike tube amplifiers, transistorized power amps generally yield their lowest distortion at output levels just below the overload point. This is because, in their output stages, the signal is "switched" back and forth between the two halves of the output circuit with every half-cycle of the signal, and this "switching" is not always perfectly smooth. As a result, a "kink" is introduced into the waveform at the switchover point, and the size of this kink remains constant regardless of how much power the amplifier is delivering. At high power levels, where the signal is much stronger than the "kink" content, the distortion added by the "kink" is but a small percentage of the total output. At reduced output levels, the signal amplitude goes down while the "kink" energy stays put, so the percentage of total distortion rises.

The presence of this "kink" probably accounts for most of the "transistor sound" in current-model amplifiers, because the less of it there is in a given design, the less "transistor sound" it will have. It isn't the whole story, though, for we've heard "transistor sound" from equipment that didn't employ any switching circuits—transistor preamps, for instance—but again, whenever the unit's measured distortion at low output levels was reduced to the vanishing point, it started to sound very much like a topnotch tube component. So that, in a nutshell, is the much-touted "transistor sound." It is distortion.

Comparing the very best tube and transistor components, we did not find either type to have any clear cut advantage. The overall sound from the tube components was felt to be a shade more pleasant to the ear, but this advantage may be lost when designers learn how to reduce transistor distortion even more. Tube and transistor preamps had almost identical bass sound, but the lack of the output transformer (and the much higher damping factor) gave the transistorized power amps slightly tighter, cleaner bass. This difference was particularly noticeable on speaker systems which are highly sensitive to amplifier damping (the AR-3, for example), and in their case it could truthfully be said that the transistor amps had it all over the tube types.

Again, though, this was evidently due to the transistor's inherently high damping factor, rather than to any essential difference between the characteristic sound of transistors and tubes. Undoubtedly, transistor design techniques will be improved, and most transistor components are already capable of giving better sound than most tube components, particularly in the very-low-power (10 watts and less) class, where transistors simply run rings around their typically miserable-quality tube competition. But there are some moderate-priced tube components that are better than the best of their transistorized competitors (some Dynakits, for instance), so just because something is "solid-state," this doesn't necessarily mean it's better than any tube component. It may not be as good.

The topnotch transistor units—the Acoustechs and Citations for example—are another matter altogether. They are almost totally free from the "transistor sound," which means their distortion, and their overall sound, can be comparable to that of the best tube types. In time, we hope their prices will be comparable, too.—J. Gordon Holt