Allnic Audio A-5000 DHT monoblock power amplifier
Even at its humblest, a 300B is a fine thing. And at its best, this classic triode output tube can deliver some of the most intoxicating music playback imaginable. If tubes are liquor, the 300B is clearly absinthe. (The 2A3 is Cognac, the 45 is Armagnac, the F2a is Tequila, and the EL34 is vodkawhich is to say, you can make almost anything out of an EL34, from the repulsive to the sublime.)
In the 20 years since the 300B reappeared on the home audio stage, the tube has appeared in a few thoroughly original designs; of those few, the ones I've heard have sounded magnificent. The 300B has also shown up in a great many other, generally less expensive amps, the designs of which are decidedly me-too. Even they can sound good, but in 2012, I think there's little point in creating another one of the latter, notwithstanding the expense of the former.
In the 15 years since the People's Republic of China became a major player on the audio-manufacturing stage, I've tended to lump most new 300B amps in the latter camp. And, I admit, when I was offered a chance to write about a new 300B amplifier designed and built in South Korea, I agreed with mild optimismbut not the hope of hearing anything new beyond what has become the staid and stolid norm. That turned out to be a mistake.
Allnic Audio is run by a Korean engineer named Kang Su Park, whose experience with the 300B and other tubes appears considerable. Beyond that, Park has something that may be of even greater worth: a point of view. He has strong ideas about how to drive an output triode and, just as notably, how to make an output transformer. In the seven years they've been in business, Park and his coworkers have worked to refine their solutions to those and other challenges.
The new Allnic A-5000 DHT ($19,900/pair) is said to be a distillation of that work. This 10Wpc single-ended mono amplifier has three gain stages, the first two intended to swing far more volts than usual in driving the output stage. (Kang Su Park's disdain for the "distorted" sound of an underdriven 300B is a matter of record.)
From the input jackthe user has a choice of RCA or XLR, the latter tethered to a switch that can invert signal phase between pins 2 and 3the signal travels to a Marconi CV1673 direct-heated triode. This vintage tube, which was completely new to me, has a distinctive-looking metallized envelope, and its heater requires only 2V at 0.1 amp of current. From there, the amplified signal travels to another direct-heated triode I'd never seen before: the British/Australian 3A/110B, which appears to be a rather pricey thing at the moment. In Park's amplifier circuit, each of these triodes has its own separate heater-regulation circuit, as does the output triode itself.
At first glance, on seeing a smallish transformer cover in the Allnic amplifier's "front row," pointedly placed between the driver and output tubes, I assumed that the 300B is transformer-coupled. That turned out to be another mistake on my part, as the cover conceals a Permalloy-core driver choke: a high-efficiency, high-impedance load that seems to be another key element of Kang Su Park's unique driver-stage design. A 0.47µF Audio Cap capacitor is also in line with the grid of the output tube.
The 300B direct-heated power triode is used in a fixed-bias circuit, with a user-friendly potentiometer and meter for monitoring and adjusting bias current. I observed a plate voltage of 420V, which is supplied through the primary winding of a proprietary output transformer, also with a Permalloy core. Switch-selectable secondaries of 8 and 16 ohms are tied to a single pair of five-way binding posts on the A-5000 DHT's rear panel. A global loop is also tied to the hot post, providing what Allnic describes as a modest amount of negative feedback, intended to raise the damping factor and lower distortion without sacrificing the output triode's character. The output triode supplied with the Allnics is the Electro-Harmonix 300B, which my limited experience tells me is both reliably good sounding and reliably consistent in quality.
The rail supply, based on a 5U4G full-wave rectifier tube, is very solidly made but otherwise unremarkable. The mains transformer appears to be a fairly standard frame type. (The output transformer and driver choke are inaccessible.)
Construction quality is first rate: I have never seen a more well-built amplifier of any type. And I have never seen neater, cleaner, and altogether sparer solder joins in a hand-wired amp. Of special note are the apparently proprietary sockets for the small-signal tubes, which use gel dampers to isolate the tubes both vertically and laterally: very nice. The only thing over which I raised an eyebrow was the direct soldering of the ground wires to the threads of two connector posts, rather than using separate lug washers: a very minor complaint, and one I had to work hard to find.
The case also deserves special mention: It comprises individual top, bottom, front, back, and side pieces, all made of aluminum alloy and held together with machine screws, but so well machined and fitted together so tightly that the chassis appears to be made of far fewer pieces. As both a protective measure and, I assume, a bit of aesthetic whimsy, each of the four tubes is contained within its own cylinder of clear plastic, supported by two slender brass tubes and topped with a light cover of alloy, the vents in which echo those in the chassis below. Finally, the front-mounted meter looks wonderful: retro, but not preciously so. (A side-mounted toggle switch allows those who feel otherwise to darken the meter while the amp is in use.)
Installation and setup
Despite their abundance of user controlsrelative to other single-ended triode amplifiers, that isthe Allnic A-5000 DHTs required little attention. Because I used only their RCA input jacks, I was spared the drudgery of flipping their XLR polarity-inversion switches. And because the listening impressions that follow are all based on the 300B tubes suppliedaccording to the Allnics' bias meters, they didn't drift at all during their time hereI didn't need to adjust bias current. (But I did anyway, to see if it's easily done. It is.)
I also limited my use of the Allnics to their 8 ohm output mode, driving Audio Note AN-E/SPe HE loudspeakers. One of my two Quad ESL loudspeakers remains unusable at this time; notwithstanding the theme of this issue's "Listening" column, I just couldn't see the point, or the relevance to most of our readers, of trying that combination.
The Allnics performed reliably during their stay. Heat dissipation, though noticeable, was never excessive. With the A-5000s switched on and warmed up, I heard hum close to the loudspeakerssurprising for an amplifier with such apparently stringent heater-voltage regulationbut not at the listening seat. I was mildly dismayed by the audible pops the amps made through the loudspeakers when shut off, but these were never so bad that I feared for the health of my tweeters. The A-5000s did not appear to invert absolute signal polarity.