This review page is supported in part by the sponsor whose ad is displayed above
As paralleled push-pullers, the CA-339s rely on accurate biasing to perform as intended. Grab your volt meter once a week until the tubes settle down (tighten the ground probe inside a binding post). This guarantees that the phase/anti-phase circuits are perfectly matched to sum for an undistorted wave form. As per Canary's John Laurel, because the 300Bs derive their power from the secondary transformer primary, the recommended 0.55V point remains identical for 120 or 240V units. While higher bias current settings may introduce sonic advances, they will also shorten tube life. He feels that WE 300Bs do seem stouter and are comfortable anywhere
between 0.60 and 0.65V if you want to push them. For this review, he interfaced with Western Electric's Charles Whitener who promptly dispatched the requisite sets of matched output glass. All sonic commentary is thus based on what most of us would have to assume was the very best valve scenario possible. I'm not in the habit of stocking octets of Russian, Chinese or Czech 300Bs. Nor could I dream up a feasible scenario whereby to solicit them now. After all, who in their right mind would volunteer goods for comparison against the reputed class leader in 300Bs while the price for such a questionable privilege was to be out eight loaner bottles for a few months? Genuine Western Electrics they were then. Four matched premium pairs. I felt like going to an AA meet brandishing a full bottle of Scottish Schweppes. "Cheers chaps. Let me indulge my addiction with a loaded grin rather than apologize whilst all of you choke." Gracias, Charles!

Even if you're not in the habit of worrying about bias, you'll have to with the 399s. They're delivered without 300Bs, trim pots set to minimum. With my WE loaners, minimum meant seriously underbiased at 0.30V. Counter clockwise adjustments were necessary to dial in the recommended 0.55 start settings (I'd later elevate those to 0.65 to report on hotter bias). How precisely you'll be able to lock in all four tubes will also depend on AC stability. You could notice your multimeter's display fluctuate a bit doing nothing. Also, getting just the desired figure without skipping over the right value will take some repeat twirling on your flat blade.

One amp kept blowing its mains fuse right after one of the rectifiers flickered briefly. Replacing the rectifier netted blown fuse #2. An e-mail to John Laurel suggested that though the WE300Bs were brand-new, one could have suffered in transit. The same had happened to Edward Barker who kept suspecting his amp. John's suggestion sorted me out in a flash -- pun intended -- and I was up and running in glorious stereo.
Did I mention that the 339s are positively giantific? It takes 20 screws spaced three inches apart around the edges to affix their bottom plates. Twenty each. That translates to more than 2 feet of sprawling mass behind the stout 1" face plates. Now add the amp's foot-wide stance. Side by side, the monos will exceed the usual shelf width. Never mind that their combined 140 pounds will crush lesser support contraptions. No, the CA-339s are a case for dedicated amp stands. Or oversized equipment racks with manly load ratings. Or good old floor placement. Which is what I resorted to, permanent carpet marks included. Titanic stuff, this.

But no mechanical noise. Despite their size, the American Transcendar-issue iron and twin chokes beneath the integral cages don't hum or buzz. That's of course in line with a $14,000/pr amplifier of bespoke parts but sadly not always the case. Canary seems to rather go the extra mile. As we already saw for their premium Hovland caps which are fastidiously mounted to scalloped Teflon blocks, the bigger transformers that occupy the full height of the amps' casings are likewise decoupled from the chassis with thick Teflon braces. Canary doesn't enjoy the prestige of more established US brands. Yet - there's nothing about the 339s' mechanical execution that suggests anything but upmarket top drawer. Make no mistake, Audio Research and Conrad-Johnson would be proud of the 339s' build quality.

Perhaps it struck me right away because I hadn't heard 50 watts of 300Bs before. The slightly fuzzy edges and associated languorousness of my CA-308 conditioning were replaced by increased focus and overall mass. Still, the musical action remained distinctly embedded in relaxation. It wasn't as microdynamically fast off the line as my EML solid-plate 45s. I was hoping for a bit more - um, hair when things get hairy; a bit more bite and speed and kick for the really fine dynamic gradations and transients. The 399s on my high-efficiency speakers accelerate to loud plenty fast but seemed to skip over the really minuscule stuff. Vocal inflections, bow and reed pressures were less articulated than I was ultimately hoping for. Time to walk away and let break-in of circuitry and valves do its thing.

That's the trouble with first impressions. They tend to be very astute right at the cusp of leaving the familiar for the new. It's the fresh contrast. Then break-in muddies the waters. Our bio computer keeps resetting. The keen knowing from before must now be revitalized by returning to the beginning, the prior reference. Without that, true north drifts and all subsequent headings are off by definition.

I must confess to being rather smitten with the luminous and veined halos that gathered at the top of each Western Electric output tube in those early days, noticeable really only with very low ambient light or -- best -- none at all. I was told it would eventually wear off. So I took a few long exposures to document the effect for posterity. Lightning in a bottle.

Electrically too, the CA-339s are very quiet, making them eminently suitable even for high-efficiency speakers like mine which will not be widespread mates (though Zu Audio is fond of using the 225wpc Rogue Zeus on my speaker model).

I clocked break-in progress by how initial desire to return to my customary Yamamoto amp faded. The 339s firmed up nicely. In the process, they also further honed the good type of edge which plucked strings -- spinet and banjo are excellent examples -- possess. How to portray transients with realistic rather than annoying bite is one telltale sign of superior amplifiers. The Canarys walk that line with a very fine degree of caution. Still, nobody could accuse them of cotton taffy behavior. I fancy the timbre and personality of the Arabian oud. It's capable of startling transient vigor combined with a darkish timbre on which can ride a sharp metallic tinge depending on who plays it. From Anouar Brahem to Adel Salameh, John Bilezikijan to Taiseer Elias, Simon Shaheen to Mehmet Bitmez and Cagrihan Erkan, these amps covered the variety of oud playing styles and how much fire and steel their players evoke with aplomb. Leading edges weren't turbo-charged but neither unduly dull and superior in fact to the CA-308s.

The Yamamoto does rather exceed the 339s in zip factor, partially due to more top-end energy, partially because it's faster and more detail oriented. Yet testing the detail and incision aspects with Klezmorim/Roma clarinet captured live -- on the fabulous Consensus by Beyond the Pale [Borealis 165] -- you'd not miss the Yamamoto's extras. Unless you compared. The deliberately squeaky aspect of the clarinet which Klezmer music calls for on occasion had good wasabi action on the 399s. Alas, it's clearly not the leading edge why audiophiles would flock to the Canary birds. It's the kind of fulsome tonal body such as Martin van de Ven on the same recording celebrates with his swaggering bass clarinet: spittle on top, elk-in-heat roar below, clacking keys in-between. Ditto for Bret Higgins' sawing bass accents. They turn a Gypsy waltz number from swirling to limping affair. So it's what happens between rise and fall of tones where the 399s perform their particular magic.

Naturally, that seems a given with 300Bs. Still, if you haven't heard a really good push/pull implementation, you might be surprised by just how much tone can coexist with fine articulation across the board and not just in the vocal range. Then add dynamic reach and mass to the recipe. Developed tone doesn't seem unduly thick because dynamics keep up jump and spunk and pepper the stew. Yes, bass quality is slightly rounder and fatter than center neutral -- a figure of speech; nobody knows what true neutral is -- but bass never drags or ponders to somehow seem behind the beat or mushy.

It could be 2nd-order cancellation which push/pull is said to perform automatically; it could be particularly linear behavior of my choice Western Electric glass; truly wide-band transformers; monstrous headroom given my speakers; or likely all of the above. Be that as it may, these 339s are not romantic or genteel by nature. Post conditioning, I didn't think too much about needing blood thinning or more whip cracking. I certainly never felt like entering fog or an echo-y cave. In fact, peaks of soloists stepping up to the lead mic, with the audience egging them on to ever more reckless risk taking, routinely took me aback. You just don't expect this level of harmonic density together with expert scaling of amplitudes over half a bar. Long-winded crescendos are one thing; truly violent sudden outbursts quite another. Such switch-blade displays belonged to a very different (under) world than where 300Bs are commonly believed to meander, sashay and strut their stuff. Color me duly impressed and recalibrated.

At this point, I had progressed to 0.65V bias. It seems trite to confirm that higher bias equalled more intensity. But so it was, predictability be damned. I'm leery of openly endorsing stressing such costly tubes. I have no idea what the punishment at the tail end might be. How many precious weeks or months will higher bias current steal from the usual life expectancy? However, even Edward who doesn't benefit from review loaner tubes confesses to running his WEs at this setting. Kevin Scott of Definitive Audio -- Edward's bane and devil's advocate with the Kondo gear -- fancies them at 0.62V. Three listeners agree. Make of that what you will.

Reverting to the single-ended Canarys, I had a definitive preference for the pushier pullers. And that's not just cute literary license. The 339s were pushier, feistier and less laid-back. They were more articulate and unquestionably more dynamic. Yet regardless of how raunchy the material I threw at them, the 339s never lost their innately relaxed composure. They never got nasty or out of sorts. While they can do edge, they're the antithesis of edgy. They don't fatigue even at high levels. That's probably evidence that they're sufficiently free of high-order harmonic artefacts or mask them successfully.

Returning to the Yamamoto for reference was confirmation. The 399s are tonally thicker, fleshier and macro-dynamically more endowed. The 45s shine more light on struck metallic objects, violin flageolet and the fierce uppermost registers of pan flute and operatic sopranos. They stage less densely though with more intense holography and telegraph additional ambient detail that eludes the canaried 300Bs. While I thus felt that this -- 45-based -- single-ended amp held the clear trump card in the microscopic world of truly tiny audio stuff -- and conversely, the push-pull amp owned a clear advantage in the macro world of voltage swings if not necessarily speed -- the SE-vs-PP 'religious rules' of our introduction were far from clear-cut when considering the same tube type. In fact, it made me downright curious to hear a superior 45-based phase-splittin' example to learn whether the changes from 300B SET to P/P would translate accordingly. My preference for the 399s over the 308s was wholesale, without qualifications or exceptions. That surprised me. But I quickly adapted.

For one session, I deliberately imbalanced the push/pull circuits. With the amps side by side, I set the four outer tubes to 0.55V, the four inner ones to 0.65V. This caused blurring of edges and focus, akin to but still different from partial phase cancellation with speakers. This served as reminder that maintaining equal bias is a prerequisite to hearing these amps at their best. A total of 16 tubes is also a considerable source of heat. Prospective owners need to factor this -- and their fully exposed nature -- into the list of things to consider. There's also the significant cost to first acquire these colossi, then maintain them with quality output tubes. This is a prospect not for the faint of heart or wallet.

What makes it appealing are the sonics. The CA-339s are seriously ballsy machines that sing. Generally held preconceptions about 300Bs -- too enhanced, romantic and diaphanous -- seem mostly suspended by push/pull operation. I dare you to hear these from behind a black curtain and identify the alleged evils of phase splitting. Those arguments read compelling in theory. In practical execution however, they don't seem to hold up too well (at least not in this iteration). Another reasonable myth -- that high-efficiency speakers won't benefit from unnecessary power -- also proved outdated. My Zus showed no signs of minding the extra headroom. Dynamic range clearly expanded.

In short, a mass showing of 300Bs is a beautiful thing indeed; extravagant and not exactly common but beautiful. Unfortunately the b-word among audiophiles often suggests a kind of dishonesty, a divergence from the noble ideals of high fidelity. When overdone to the extent of homogenizing various music with a recognizable patina or hue, I'd concur. But when it's simply about better tone colors and an utter absence of brittleness and whitishness, it makes you listen more and longer. It gets you closer to the heart of the hobby. And that's higher fidelity to why we listen to music in the first place.

Violin and piano are two notoriously tough acts to follow when you get really picky. The 339s were string tone revelers of the first order. Whether South-Indian violin virtuoso Kunnakudi Vaidyanathan and landsmen L. Shankar and L. Subramaniam; whether Turkish maestro Nedim Nalbantoglu or fiddler Mark O'Connor; whether concert hall idol Anne-Sophie Mutter or the Kronos Quartet with assorted guests; whether Russian wizard Oleksandr Klimas, Lebanese ace Claude Chalhoub (whom Barenboim made concertmaster of the Weimar-based Western-Eastern-Divan Orchestra) or the Hungarian avantgarde fiddler Zoltan Lantos who plays on 5+16-string customs;
violin was clearly an instrument these amps adored in all its many different guises. They embraced con arco octave doubling, pizzicato, spiccato, glissandi, the deliberately dirty fundamental-plus-flageolet edge walking popular in Arabia and the bird-chirping and other sound effects beloved by Russian gypsies. Listeners who limit their violin consumption to the Western concert hall miss out half the fun. The Canarys demonstrated why and how very effectively.

Unlike the Silvertone Audio Model 3.2 300B SET with the permalloy iron from Singapore, the 339s are not deliberately weighted towards the bottom to compensate for the lack of extension and weight the kind of single-driver speakers suffer which are traditionally mated to micro-power amps. This becomes obvious when you consider O'Connor's overall leaner timbre and how it plays off Jon Burr's walking bass lines on the seriously inspired Hot Swing! tribute to Stephane Grappelli and Django Reinhardt [Omac Records 4]. Whatever kind of guitar Frank Vignola plays, it always exhibits a goodly dose of banjo-ish nasality and thinness. This the 339s rendered intact rather than glossed over. The same restraint showed for fiddle and bass. They sounded the way I expected them to from years of listening to this album. Conversely, the rare Stradivarius played by Chalhoub on extended loan from the Stradivari Society of Chicago sounded more elegant and expensive. As probably conservatively rated 50-watt monos, the Canarys clearly aren't voiced to compensate for bandwidth-restricted speciality speakers. They are relative muscle amps for full-range transducers.

Though clearly gifted with tone, they're not inherently thick and opaque. They do tone without cloying punishment. Where they do fall short of exceptional outside of speed and incision power is in the uppermost treble. On the brilliantly recorded Jacques Loussier Trio outings which were mastered by the pianist himself, brush and triangle work lacked the uppermost harmonic sheen. This relates back to the previously cited minor caution around transients. Certain attacks that emerge like viper strikes from the thicket of the background jungle with the Yamamoto remain more shaded. They're less bared and out in the open. As with any good amp, these are matters of degree only, minor personality traits. Reducing bias current shifts the amps into slightly cooler terrain. That also affects how this particular balance plays out subjectively. Mechanical noises like nails on ivory, finger slides on strings and key clacks gain a skoch more prominence. Ultimate treble extension remains unaffected, naturally. That might in fact be a trait of this output tube type. I doubt even diehard fanciers of the 300B would name stratospheric treble extension as one of its key virtues.

Wrapping up
If you asked me to name another 50-watt 300B amp, I don't think I could. Ayon Audio's Reference produces 55 monaural watts from a pair of single-ended paralleled 52B-S valves yet those aren't traditional 300Bs. Substituting WEs would significantly lower that max power rating. The Canarys seem to suffer little competition then. Outside the obvious impracticality of size, weight and stocking needs of designer glass -- and given that the Signature 30 by Red Wine Audio achieves most of it on 1/10th the tab without a single valve -- Canary Audio's CA-399s are formidable brutes with high class. Powering up and down without fuss, pops or shrieks, they appear to be true no-nonsense designs that are done up by the book with superior parts which all seem conservatively spec'd. The chassis don't start to boil nor does the heavy iron on deck hum. User-adjustable bias means you can compensate for uneven tube wear/ageing and experiment with various bias points to tailor performance. 50 watts per side are truly righteous artillery for most real-world speakers (due to benign overload behavior, tube watts are often counted double vs. transistor watts).

It means that the special virtues of 300B sound are finally open to far more listeners than just those few who play the high-efficiency fringes. Don't let the push/pull soul of these machines lead you astray. Not only is this kind of output power impossible otherwise, it appears that common 300Bs complaints are handsomely minimized with this type of 2nd-order distortion-cancelling circuit. I wouldn't have thought so myself. Yet given the evidence of simple listening, I'd take the push/pull CA-339s -- and, I'd have to believe, its half-powered CA-330s siblings -- over the single-ended CA-308s any day of the week. And twice on Sundays. There go my initials. Drat. On second thought though, good riddance. For today.
Manufacturer's website