On Vacuum Tube Power Amps…

By Bryan Geyer

As commonly credited to Mark Twain*…

“It’s easier to fool people than to convince them that they have been fooled.”

—and—

“The trouble with the world is not that people know too little; it's that they know so many things that aren't so.”

Were this not the case, the production, sale, and use of receiving-type vacuum tubes would have permanently perished throughout the world. Instead, some tubes still cling to life in the modest audiophile and amplified guitar markets**.

The continued use of vacuum tubes in guitar amps is easy to comprehend. That’s because guitar buffs treasure the sound of 2nd harmonic distortion, and tube amplifiers present high levels of that noxious stuff. Plus it’s very easy to create more whenever you overdrive (clip) a vacuum tube amplifier’s output stage.

The on-going acceptance of vacuum tube design in high-end audio power amplifiers is more puzzling, especially given the premiums paid for power amplifiers that always…

            …operate at high temperatures, at all output levels,

            …utilize hazardous (potentially deadly) high voltages,

            …perform poorly when compared to solid-state equivalent***,

            …require endless costly maintenance†,

            …and violate every precept of the “go green” environmental initiative††.

Yes, it’s a puzzle. Science seemingly gets smothered by expectation bias and confirmation bias; refer Audioholics: “Do Our Expectations Determine Our Experience of Sound More Than We Realize?” at https://www.audioholics.com/room-acoustics/mind-over-music.

The audiophiles of today that use tube-type power amplifiers tend to be < age 65. This tells me that they’re not likely to have owned a power amplifier during the era when all electronic equipment was vacuum tube dependent. The genial glow of those orange filaments is not always reminiscent of happy hassles for those of us that serviced such stuff. We recall the hours devoted to trouble tracing and repair; maybe charred salvage.

A significant number of audiophiles say that they prefer the “warm” sound of a tube-type power amplifier. What’s generally overlooked is that the source of this warm coloration traces to a design limitation that’s implicit with all traditional tube-type power amplifiers. Because the output stage of a vacuum tube power amplifier operates at high source impedance (several kΩ), it’s not optimum to directly drive a low impedance load (e.g.: a loudspeaker). A load that presents low impedance is best driven by a source that exhibits still lower impedance. Indeed, a value of zero source impedance would be ideal. The classic means of curing this circuit-to-load impedance disparity is to introduce an output transformer between the tube circuit’s final stage and the external load. This (big, heavy, and expensive) magnetic device will then, by virtue of its differing internal winding ratios, transform the high impedance state into a low impedance source, so that the signal can better mate with its intended low impedance load. Audio engineers conclude that it’s probably this processing path through the windings of the output transformer that create the perception of warmer sound, so “tube sound” likely doesn’t trace directly to the presence of vacuum tubes—it’s the consequence of coupling the circuit’s output stage to the load by means of a transformer. What you’re hearing is “transformer sound.”

Tung-Sol Ad, 1955

As always, this beneficial design fix (add transformer) is not without restrictive limitation. An audio output transformer is a non-linear device by nature, with low end frequency response that’s largely dependent on the mass of its magnetic core, and high frequency response that’s subject to the vagaries of leakage inductance and stray capacitance. It’s also prone to waveform saturation at high signal amplitudes, and that tends to create undesired harmonic distortion due to soft clipping. Further, design constraints generally limit the output source impedance to some 3 or 4 Ohms. Lower Zout is impractical, and higher Zout options must sometimes be applied (when increased load impedance permits) in order to linearize a tube amplifier’s voltage gain, as the latter will often vary (by several dB) with changes in the loading impedance (refer “Special Footnote,” at end).

Modern solid-state power amp circuits don’t have to contend with any of these messy generic issues that plague tube-type power amplifiers. A solid-state power amplifier’s natural internal source impedance will be very close to zero; normally somewhat < 0.1 Ohm, and it will be stable. That near zero value is more than an order of magnitude below the source impedance of any transformer aided tube-type power amplifier, so there’s negligible undesired source/load interactive variance. In addition, the solid-state power amplifier can provide direct coupling to the load, so there’s no transformer interface imposed. This assures better transient damping, with wider, flatter frequency coverage; also less distortion, firmly fixed voltage gain, lighter weight, and (potentially) lower cost. As a result, the signal that gets delivered to the load will be a highly accurate representation of the input, and the sound that’s perceived will be determined purely by the input source and by the load, not by the compound interaction of a higher driving impedance in tandem with the load impedance. This is why a well designed solid-state power amplifier has no intrinsic sonic signature, it’s functionally transparent.

Vacuum tube users sometimes recommend a particular power amplifier + speaker with a preferred connecting cable, with choice of cable based on listening tests. Beneficial cable effect is potentially possible when a tandem source/load termination happens to form a euphonious (aurally pleasing) filter. However, do realize that any termination so marginally sensitive that the niggling impedance variance conveyed by a few feet of cable can cause audible change must be highly unstable to start. Merely moving such cable might provoke further change. This shaky state is precisely why audio engineers extoll the load invariant advantage assured by driving the speaker from a near-zero (≤ 0.1Ω) source impedance, something that’s naturally inherent with solid-state power amplifier design.

Despite all of the noted technical and environmental deficiencies, there’s no ethical deceit implicit in the promotion and sale of high-end vacuum tube power amplifiers. Natural acceptance and approval of such product reflects the innocence of human trust, just as with faith in a deity or a belief in astrology. Blind trust defies rational explanation, but many regard trust as noble—no reasoning required. Others are more realistic—they want to see the data. How you side in this issue is your choice, but don’t let the inexorable tide of audiophile groupthink††† swamp straight thinking and good science.

BG (February 25, 2020)

*A popular attribution; probably apocryphal.

**All of the U.S., British, Dutch, and German producers of vacuum tubes are now either defunct or ceased making tubes some four decades ago. The entire world market for new tubes is presently served solely by obscure sources in China, Russia, and Slovakia. New tubes that get labeled with the names of long deceased and respected sources (like Tung-Sol, Mullard, et al) exist simply because a Russian entrepreneur bought the right to reuse dormant copyrights. (Tung-Sol died in the early 1960s.)

***Ruler-flat power response, near-zero (<  0.1%) total harmonic distortion (THD) at full rated power, and ultra-low output impedance (less than 1/10th the Zout of a typical tube amp) is now routine in the case of solid-state power amplifiers. Identical measurements made on the very best vacuum tube models show that they can’t approach those values. For example, typical tube-type power amplifier THD limits are ~ 16X to 50X worse (1% to 3% THD instead of 0.06% max.) than as specified for a popular “mid-market” solid-state power amplifier (Parasound Halo A23+).

†A matched pair of vintage “NOS” Tung-Sol 6550 output pentodes = $220. Refer…http://vintagetubeservices.com/pentodes/. Back in the era (1963-1976) when I used a Marantz model 8B stereo power amplifier (four EL34 output tubes, rated 35 Watts/channel), I had to replace the output stage pairs on the order of every 30 months to maintain optimum “in spec” operation. I also readjusted the bias settings 2-3 times/year to minimize IM distortion (SMPTE) at full output, using a Heathkit AA-1 analyzer and a ’scope.

††A typical 150 Watt/channel stereo tube-type power amplifier consumes more power (about 240 Watts) when in benign standby than a 55 inch Sony LED/LCD TV set does when in actual use. Consider: 240 Watts of standby (just idle, no signal output) power consumption is equivalent to continuously burning four 60 Watt incandescent light bulbs without providing any light; just heat. That’s conspicuous waste.

†††Refer https://en.wikipedia.org/wiki/Groupthink

Special Footnote: Unlike solid-state power amplifiers, vacuum tube power amplifiers do not exhibit stable fixed voltage gains. Their intrinsic gain will typically vary by several dB, dependent on nominal load impedance. To minimize this undesired variance, the output transformer often exhibits multiple output taps, e.g. 4Ω, 8Ω, 16Ω. When the nominal rated load impedance permits (i.e., rises), the use of these higher Zout taps will assist in stabilizing the voltage gain. (Refer AudioXpress, issue dated Feb. 2020, p. 64, fig. 5)