The Joule Electra's warmth comes from more than just its use of vacuum tubes, and more than just its class-A design. It runs warm because it's an output-transformerless (OTL) tube amplifier.
That's a really horrid oversimplification. Let me back up a bit...
As a power source, an audio tube has a naturally high impedance: It isn't good at supplying high current for a given voltage, so it has a hard time developing power across a load, at least on its own. A loudspeaker, on the other hand, is a low-impedance load, and it requires a fair amount of current in order to do any real work.
The classic means for bridging those two otherwise unbridgeable qualities is an output transformer, the primary coil of which is also used to conduct DC to one or more of the tube's electrodes. The drawbacks of such a thing are obvious, and while there exists a wide range of quality from the good ones to the bad ones and back, it simply can't be denied that output transformers tend to compress amplitude, limit bandwidth, shift phase, and ring like cowbells (mu!) to one extent or another. Good ones also cost a lot of money.
For as long as output transformers have existed, there have been designers who've tried to dispense with them altogether. None is better known than New Yorker Julius Futterman, whose pioneering work in the 1950s gained sufficient fame that, among some hobbyists, the terms OTL amp and Futterman amp are used interchangeably—to the great discomfiture of still others.
Futterman's predecessors in the OTL genre strove to create an output section whose impedance is low enough to drive a loudspeaker directly—and, of course, the most natural way to reduce a tube's output impedance is to configure it as a cathode follower. That's precisely the trick they used, in all of the very first OTL amps. But because a cathode follower requires a much larger input signal than other configurations—thus opening the door to the same level of distortion that one hoped to avoid by ditching the output trannie in the first place—none of those early amps can honestly be considered successful. Besides, when a designer goes from a single-ended cathode-follower output to a push-pull cathode-follower output, in a reasonable effort toward generating reasonable power, he or she doubles the output impedance by comparison. Back to the starting line.
Futterman's idea was to use two separate output tubes—or groups thereof—in such a way that one was a cathode follower and the other was a regular "common cathode" output device. That arrangement, known as a single-ended push-pull (SEPP) output section, had been used to create OTL amplifiers in the past, but Futterman added a clever twist: He tied the cathode resistor of his full-wave, auto-bias input tube to the top of the loudspeaker load, to create a signal-imbalanced drive for the inherently signal-imbalanced SEPP output section, thus producing a signal-balanced waveform of reasonably high current. Notwithstanding certain drawbacks, such as a lack of immunity from DC offset and the need for at least some negative feedback, Futterman's solution was a good one, and his amplifiers are generally considered the first truly successful commercial OTLs.
A Different Architecture
Throughout this time, there was yet another, very different output architecture floating (footnote 1) around in Tubeland: the bridge amplifier—or, as the Electro-Voice company called it in 1954, the Circlotron. In this remarkably elegant design, two output tubes are arranged in a push-pull cathode-follower scheme (ie, the loudspeaker load appears in the cathode circuits of both tubes), with their anodes tied to two separate power supplies that float above ground.
The Circlotron wasn't conceived as an OTL, although its inherently low output impedance (think: cathode drive) was touted with the suggestion that the amp could be built with a significantly smaller and thus significantly cheaper output transformer than usual. And although the ends of the loudspeaker load were tied to the bottoms of two power supplies, the opposing polarities meant that DC voltage was canceled out across the load. Very tidy, all in all.
The Circlotron's praises did not go unsung, and while it never gained the popularity of other output architectures, its potential was appreciated by a few inquisitive designers. Chief among those was Minnesota's Ralph Karsten, who created the first commercially viable Circlotron OTL amplifier in 1989. Karsten's amp, the Atma-Sphere MA-1, didn't have quite as low an output impedance as other OTLs, and in fact required an auxiliary autoformer for use with loads of very low impedance. But the single-stage MA-1 was the first truly reliable OTL amplifier on the market, owing in large part to its sheer design elegance: Going from a SEPP circuit to that of an MA-1 is like hearing an especially bothersome chord find its resolution at the end of a phrase. As a consequence, Karsten's place in OTL history was secured. (Sadly, he has also become known for denigrating the achievements of fellow designers past and present, including the late Julius Futterman himself, whose circuit "should never have seen the light of day," according to Karsten. See www.atma-sphere.com/papers/otl.html.)
That brings us to Joule Electra, founded in the early 1990s by retired chemical engineer Jud Barber and named in honor of his lovely wife, Marianne Electra Barber. (It was originally going to be called Muse Electra, but someone else got to the Muse part first.) Barber became interested in electronics as a kid, and built a number of amps from scratch in the 1950s. But it wasn't until the late 1980s, when a friend loaned him some contemporary tube electronics, that Barber considered making amplifiers for a living. ("Why screw up a perfectly good hobby?" is how he explains his earlier point of view.) As luck would have it, the borrowed amps included a modern OTL design—and the spark was, well, sparked.
Barber's first commercial product was the Joule Electra LA-100 preamplifier, which has sold well since its release in 1991. Key to the LA-100's performance is Barber's unique implementation of a classic mu-follower circuit, in which a tube configured as a common cathode drives a tube configured as a cathode follower, the latter functioning as a constant current source for the former. The mu follower produces a lot of voltage gain, and Barber says it gives him the operating range he prefers for his tubes—a conclusion he reached after countless hours of good old-fashioned listening. Imagine.
Footnote 1: Believe it or not, that's a joke—albeit not a very funny one.