These trends have passed long beyond the cult stage. While tube amplifiers still cannot provide the high damping factors or woofer Control that is possible With transistor amplifiers, no one still claims that they soften highs or somehow romanticize recorded sound (footnote 1). There is a consensus among high-end magazines that the best tube units are exceptional in the areas of transient and harmonic detail, high frequency information, and imaging and soundstage data. Further, there is a similar consensus that the days when tube units provided superb midrange at the expense of the frequency extremes are long gone.

This brings me to the units under review: the Conrad-Johnson Premier Three preamp and H-la head amp combination. Unlike many tube units, they can be used With most moving-coil cartridges (more on this in a moment), and the preamp has sufficient output to drive most amplifiers.

The very best transistor units still win out in area of noise. The Premier Three preamp alone cannot be used with even moderately low-output moving-coil cartridges like the Accuphase, Argent Diamond, and Dynavectors without the noise and gain mismatch seriously coloring the sound. Even a Koetsu Black is marginal and tends to have its highs rounded and masked, becoming slightly dry and lacking in detail in the upper midrange. The Alpha-l is also marginal, although the problem tends to be a drying up of the upper octaves and a lack of proper warmth and midrange balance.

Like the Audio Research SP-l0, or Premier Three used with the HV-1a (which are remarkably close in price) has just enough gain so that you can use any moving-coil cartridge with a minimum of 2mV output. "Just enough," however, means detectable hiss and noise. I have recorded such noise and superimposed it on the sound of much quieter transistor units (this is a rough-cut approach since it is impossible to do this without adding some colorations to the sound), and it is clear that it matters. Ironically, one has an expanded impression of depth or air; the unpleasant aspects are a tendency to mask the softest musical detail and harmonics, while adding a dry coloration in the spectrum where the noise is dominant. Worse, the imaging tends to alter. There is a feeling of expanded soundstage size and detail, but it is not natural and, although it is initially impressive, it eventually has the same irritating effect as exaggerated highs.

The revised Conrad-Johnson HV-1a head amp offers more gain and less of this coloration than the SP-10. It is by far the best commercial step-up device I have heard, although Murray Zeligman has a new tube head amp in prototype that may well be major competition, and ARC's William Z. Johnson is working on a new transistor head amp design for use with the SP-8 and SP-10. Unlike most ARC and C-J gear, you can mix the C-J head amp with the ARC preamp and amp. Try using C-J preamps with ARC amps, or vice versa, and you get the worst of both sets of gear, not the best.

The HV-1a does not, however, permit input impedances much below 100 ohms without changing sound character and diminishing performance. This presents problems, because I am increasingly unhappy about high-impedance loading as a general solution to the moving-coil cartridge problem. Many of the best cartridges must be loaded down to well below 100 ohms for the best imaging as well as linearity of frequency and dynamic response. Furthermore, high-impedance loading often combines with tube noise to produce even more exaggeration of soundstage size, so that imaging detail is unnatural and irritating, with a tendency for the image to wander or appear in an unnatural location.

I should also note that both the SP-10 and the C-J head amp are sensitive to shock.

Sound Quality
I tested both the Conrad-Johnson and Audio Research combinations using top quality turntables like the Goldmund and SOTA-Sumiko arm combination, top quality cartridges, and speakers like the modified Quad ESL-63s, Thiel CS3s, Spica TC-50s, and Fuselier 3.3s. I have used them with Compact Discs and a wide range of FM Tuners. I have mixed and matched cables, including the Straightwire, Petersen, Discrete Technology, Audiosource, and Monster Cable Interlink Reference.

In discussing how these terms apply to high-performance audio, I suggested that if audio designers are to succeed in this highly competitive field, they must be sharpeners: they must pay attention to differences that may be very small but that, in the aggregate, could produce a significant improvement in sound quality. I've encountered a good many audio designers whom I would describe as sharpeners, but perhaps none more than Ken Stevens, of Convergent Audio Technology (CAT).

A case in point: A few weeks after I'd received the review sample of CAT's SL1 Renaissance preamplifier, Stevens called to tell me that, because of an error by their parts supplier, some capacitors in the phono section of a few Renaissance units weren't exactly as specified. (It has since been determined that only three samples of the Renaissance were affected, my review sample being No.3.) The difference wasn't something obvious, such as capacitor value, but the metal used in the capacitor lead wire. That wire was supposed to be made of copper, but for some of the capacitors copper-clad steel was used. Stevens admitted that the sonic difference produced by the lead wire was small, and would not show up in any conventional measurement of the Renaissance's performance, but it was one that he could hear in his own high-resolution system.

He said that only a practiced eye could tell by looking at the capacitor which kind of lead wire it had, and he wanted to ensure that the review sample represented current production. Stevens offered to visit me in Toronto (the CAT factory is in Rush, New York, near Rochester, about a three-hour drive from my house) and, if necessary, replace the affected capacitors. That was fine with me—the visit would also allow him to confirm that the Renaissance was working as expected in my system. (By that time I'd already listened briefly to the phono section, and couldn't tell that there was anything amiss.)

And so it came to pass that, a few days later, Ken Stevens showed up. After a bit of listening to CDs, which satisfied him that the Renaissance was working well in my system, he removed the preamp's top panel and pronounced that while one channel of the phono section had the proper capacitor, the other didn't. He proceeded to unsolder the capacitor with the wrong lead wire and install the correct one. However . . .

One of the important design features of all CAT products is vibration control. To help damp any vibration of the circuit board, the larger capacitors are glued to the board using a special adhesive. Stevens had not brought a tube of this adhesive with him. He could have left the replacement capacitor unglued—again, none of the preamp's measurable attributes would be affected—but that would not do. So we made a trip to a nearby Home Depot, where Stevens spent considerable time determining which of the adhesives they had in stock had the right characteristics: not too hard when dried, but not too soft, either. Finally, he found an adhesive that, while not the same brand used by the CAT factory, had properties sufficiently similar. We drove home, and he completed the installation of the new capacitor.

Now, that's what I call a sharpener.

Renaissance SL1
The original CAT SL1 was introduced in 1985, and quickly established a reputation of being one of the best—if not the best—preamplifiers on the market. I bought a later edition, the SL1 Signature, in 1993, and although I've since heard some other excellent preamps, the SL1 has remained, in various incarnations, my reference preamp.

The basic design of the SL1 Renaissance ($9995) is pretty much the same as that of the original SL1: It's still a two-channel tube amplifier with volume and balance control, phono stage (a version without phono stage costs $7950), no remote control, single-ended rather than balanced inputs and outputs, and a separate power supply connected by a thick, hardwired cable. I've always found this cable a pain to deal with, and during Stevens' visit I asked if he might consider using a detachable cable.

He told me that although hardwiring the cable has some audible benefits, his original reason for this design was a practical one: reliability. When designing the original SL1, he'd wanted to ensure that it not only sounded great but was reliable as well. He asked a major dealer for Audio Research if there was any one problem with ARC's SP-10—widely considered the top preamp at that time—that regularly required warranty service from the factory, and the dealer had an answer: the connections between the detachable cable connecting the preamp to its power supply. Stevens decided that this was one problem his preamp wasn't going to have; the sonic benefit was something of a bonus.

Over the years, there have been numerous changes in the SL1, some of which led to changes in its name: Reference, Reference Mk.II, Signature, Signature Mk.II, etc. The last revision resulted in the SL1 Ultimate, which I reviewed in August 1999. At that time, I'd kidded Stevens about the Ultimate designation: If you say that a product is the ultimate, what do you call its successor? Is it possible to have something better than the ultimate? Does, say, "Ultimate Mk.II" even make sense? It would be like a football coach saying "The players gave 110%."

The single-ended-triode movement of the 1990s may not have changed everyone's point of view, but its aftershocks endure. Low-power amps sell well. The market for high-efficiency loudspeakers is healthier than ever. Even vinyl's most recent renaissance can be linked, in part, to a rekindled interest in old tubes and technologies. Best of all: Some of the thinkers and designers who came for the Revolution decided to stay for the Great Leap Forward.

Though hardly a household name, Arthur Loesch is revered in SET circles for being among the first to rediscover the glories of hand-wound transformers and directly heated tubes, chief among the latter being the well-loved Western Electric 300B. But Dr. Loesch did something that may prove to be a greater legacy: He designed a perfectionist-quality phono preamplifier, references to which often share sentence space with the word legendary, then actually built a few of the things—sometimes for commerce, usually just for fun. (Another aftershock: designing for the love of music more than for the love of money.)

Then as now, Arthur Loesch had little enthusiasm for manufacturing, and, despite being aided now and again by able partners, examples of his preamp are thus thin on the ground. That shortage is now undergoing correction: Dr. Loesch has allowed a real company, Tempo Electric, to build and sell his preamplifier—now a full-featured phono stage with line stage—along with other of his audio designs.

Tempo Electric has existed for 10 or so years, but their ramp-up to full production was bumpy. The designer came up with various refinements, as designers often do, and set about testing them in the field. But early in the last decade, Dr. Loesch was slowed in his work by a bout with cancer. Health problems came to the fore again when Joe Levy, Tempo's founder and chief builder, needed cardiac-bypass surgery and a long period of recovery. (Musical Coincidence No.1: The world-renowned bluegrass mandolin player Frank Wakefield—who, like Joe Levy, is a resident of Saratoga Springs, New York—had cardiac bypass surgery at more or less the same time; during the therapy phase of their recoveries, Wakefield and Levy found themselves exercising every morning on adjacent treadmills.)

And as so often happens when a small company and a magazine writer try to schedule a review, Joe Levy and I exchanged phone calls and e-mails for years, to little avail. Finally, at the end of 2009, the product now known as the Tempo Electric Arthur Loesch 1.1 Control Preamplifier ($7100, base model) arrived at my doorstep (footnote 1).

Loesch is more
When talk turns to handmade, limited-edition, legendary products, I envision small, messy-looking things of decidedly rugged quality, whose manuals are nonexistent and whose shipping cartons bear such phrases as U-Haul or Little Friskies. Imagine my surprise when the Arthur Loesch 1.1 arrived in a total of four nice, new cartons: one for the tubes and cables, one for the preamp itself, and one each for its two outboard power supplies (although the preamp can also be had with just one).

When I beheld the 1.1's build quality, surprise turned to wonder. Whereas some domestic audio products impress with sheer bulk or silly opulence, this one was notable for the intelligence of its sturdy but light chassis, comprising various nicely finished aluminum panels held together with nuts, bolts, and lock washers. Metal beams span the inner width of the preamp proper, and dual-mono circuit boards—another surprise for this hobbyist, who expected hand-wiring—are suspended from them with nylon fasteners and strips of damping material. Phono sockets are low-mass WBTs (in silver), while the sturdy, six-conductor plugs and sockets for the power-supply umbilical are made by Amphenol Aerospace. (Musical Coincidence No.2: The banjo player in my band is a longtime employee of the Amphenol plant in Sidney, New York.)

Although the Arthur Loesch 1.1 preamp can be had with an onboard step-up transformer in front of its phono section, my review sample lacked that option, relying exclusively on tubes. The first phono gain stage is built around the legendary—there's that word again—Western Electric 417A miniature triode (which, you may remember, was also the tube of choice for the excellent Lamm LP2 phono preamp). Original 417As are rare and expensive, and their potential for microphony and noise are such that re-sellers often wind up having to discard at least a portion of their stock. That said: If it's true, as some say, that the first stage of an MC phono preamp must respond to signal variations the size of a single electron—a colorfully useful if overly sci-fi way of putting it—then this sensitive tube, above all others, is probably the one to have.

But while the kWP's fit and finish were jewel-like, its appearance was cartoonish. The front panel was dominated by two large knobs about as subtle in their susceptibility to alternate interpretation as the butter nipples topping the pancakes in IHOP's TV ads. The gargantuan remote control, apparently milled from a solid block of army tank, made every visitor to my listening room laugh. Most of us, when we buy pricey audio gear, aren't going for laughs. In fact, Musical Fidelity's Tri-Vista kWP preamp and kW monoblock amplifiers was the homeliest bunch of expensive audio components I've ever owned.

I loved how they sounded, especially together. That was good enough for me. But I never understood how someone with Musical Fidelity chief Antony Michaelson's finely honed aesthetic sensibilities could produce such ungraceful-looking kit, or why, with so many years in the business, his company had yet to develop a unified look for its premier products.

That all changed with the introduction of the Titan power amplifier, which I reviewed in June 2009. But while its attractive art-deco looks are unmistakable, they've so far had little effect on the design of other Musical Fidelity products. Rather, the company's sleek new AMS Primo preamplifier ($10,999), launched at about the same time as the Titan, introduced a new, understatedly elegant look that's now found throughout MF's AMS and M6 series of components.

Built in the UK
The AMS Primo is a fully balanced, class-A preamplifier with 14 dual-triode tubes (ECC81/12AT7). It employs zero global feedback and weighs 37 lbs. An internal enclosure of mu-metal shields the signal-carrying circuits from the two onboard power supplies, one each for the signal path and the control functions. The power supply rectification is dual-mono and solid-state. Each channel has its own fully regulated high-voltage and heater circuits.

Why so many tubes? According to Antony Michaelson, because the Primo is fully balanced in operation from input to output, it includes four mono amplifiers attenuated via a four-gang ALPS volume control. Low-current, small-signal tubes don't drive loads effectively, becoming variably nonlinear when presented with an ever-changing load such as a volume pot. When you change the volume, you change the tubes' operating conditions, hence their performance. Paralleling the tubes gives you twice the current capability in each stage, as well as lower the coupling impedance between them, both of which produce greater linearity, "average out" tube performance, and better deal with tube aging. Michaelson warns against tube rolling with the Primo: the linear performance of its zero-global-feedback design depends on carefully matched sets of tubes.

Peer through one of the two mesh-covered vents on the top plate and all you'll see will be the crowns of the tubes—seven per channel—isolated by a secondary enclosure that runs from just behind the large, central volume control all the way to the rear plate, on which are mounted the various high-quality input and output jacks and other connectors. All circuit boards and associated wiring are hidden beneath the internal subchassis, making for an extremely neat and orderly look that complements the Primo's exterior.

The tidy and understated front panel is dominated by that big, motorized volume knob at the center. There are also two rows of small pushbuttons, each surmounted by a small blue LED. The five to the right of the volume knob select among the five inputs: CD, Tuner, Aux 1, Aux 2, and Tape. The three to the left are a single Power/Standby/Mute button, as well as ones for Gain and Tape Monitor. Gain is used to program the individual inputs to compensate for the 6dB difference in gain between the balanced and single-ended inputs. The panel's simplicity and ease of use are welcome, but the lack of an LED on the volume knob makes it difficult to ascertain the setting from a distance or in the dark.

Musical Fidelity's inclusion of a tape loop is an interesting choice in 2010. Few audiophiles today use tape recorders, and fewer still have decks that include separate record and playback heads, but those who do will welcome this loop. The Tape input, of course, can be used for any source, as can the other four inputs.

The rear panel is equally orderly. Each of the five inputs has both high-quality single-ended RCA and balanced XLR inputs, selectable via a slider switch. (The Tape Out jacks are single-ended only.) Trigger in and out jacks allow the Primo to be turned on via remote trigger, and to turn on other remote-triggered products, including MF's Titan and AMS power amps. The Primo is built in the UK.

Included is an equally sleek-looking, ergonomically pleasing remote control machined from a solid billet of aluminum—large, but not tank-like. While not backlit, the remote's 10 buttons are logically placed on its spacious surface, and I found their functions easy to memorize. In addition to repeating the front panel's Input, Gain, and Power controls, there are also Volume Up and Down and Mute buttons.

Setup and Use
With the exception of the Gain button, setting up the AMS Primo was straightforward. Just plug in your sources, select balanced or single-ended input for each, run single-ended or balanced cables out, plug in the power cord, turn on, and play. To set a single-ended input for additional gain, just select that input and push Gain. The Primo will then remember to raise the gain each time you select that particular single-ended input, even if you lose power. I drove the Titan in balanced mode and ran a combination of balanced and single-ended sources to feed the Primo, which performed without a glitch for the many months it was in and out of the system.

The eternal question
Anyone who's vacillated between tubes and transistors in his or her long and sometimes confusing audiophile journey can't blame a manufacturer for doing likewise. Antony Michaelson began with tubes—1977's highly acclaimed Michaelson & Austin TVA-1 power amp is now a classic. Later he switched to solid-state exclusively for Musical Fidelity's premium products, and then to hybrid designs using military-grade, low-noise, high-tech, long-lived mini-tubes, hardwiring into his circuit such tubes as the metal-cased nuvistor triode and the glass-cased 5703WB. Tubes were also used in MF's X series. Musical Fidelity claims to have made and sold more electronic components employing small-signal tubes than anyone else in the past 30 years.

The Primo's published specs boast a very low THD+noise of 0.005% from 20Hz to 20kHz, a signal/noise ratio greater than 105dB, and a relatively wideband frequency response of 5Hz–50kHz, –0.5dB. The claimed maximum output voltage of 32V (!) and the 30dB overload margin are impressive. In fact, many of the Primo's specs read like those of a good solid-state design.

Between the Primo's high output and the volume control's taper, I never had to turn the knob above 11 o'clock—even when using a single-ended input without additional gain. Beyond that, it was too loud. But with no signal, turning the pot almost all the way up produced through my speakers nothing but the faintest suggestion of hiss. "Tube rush" was nonexistent.

Wow!
After breaking in and warming up the Primo with the Sooloos music server set to Swim (ie, Shuffle) mode, the first LP up was a test pressing of a vinyl reissue of Donald Johanos and the Dallas Symphony's justifiably famous recording of Rachmaninoff's Symphonic Dances, engineered by David B. Hancock in 1967. The entire series of Johanos/DSO recordings are audiophile spectaculars recorded with four custom ribbon microphones at 30ips in Southern Methodist University's McFarlin Auditorium. The original LP edition, on Vox Turnabout, was made from the master tapes, and David B. Hancock's initials are inscribed in the dead wax. Unfortunately, Vox, a budget label, pressed on poor-quality vinyl; quiet copies were hard to find. Still, if you see any in the $1 bins or at your local Goodwill, don't pass them up!

The new reissue of Symphonic Dances (2 45rpm LPs, Vox/Turnabout/Analogue Productions) beats the original in every way. It's as warm and harmonically complex, but sounds more open and less boxy, and dynamically it's in a different league entirely. Through the AMS Primo the imaging and soundstaging were ultra-expansive, the harmonics richly and vividly drawn, the dynamics thunderous. Most significant, the bottom-end extension and control, particularly in the big timpani wallops, gave up nothing obvious to solid-state sound, nor did the Primo's top-end extension sound truncated. Mostly, the midrange plumped up, but by a subtle amount that never caused bloat, or produced ballooning images like those in a Macy's Thanksgiving Day Parade. At no time did I think, Yes, this is really pleasing, but I wish I had more weight and/or definition on bottom, or better top-end extension or transient attack. The Primo was fast and extended on top, and well defined and pleasingly punchy on bottom, with enough weight to anchor the low-frequency end of every kind of music.

Going from the twice-as-expensive darTZeel NHB-18NS solid-state preamplifier's precise, some might say clinical, control and finely drawn images to the Primo's harmonically vivid, exuberantly drawn sound pictures was jarring and invigorating. The Primo elicited from me an immediate and enormous "Wow!" that I still felt even after weeks of listening. No solid-state device in my experience can produce the sense of effortless musical flow that all-tube designs—even poorly designed, tonally colored ones—deliver with ease.

Yet, clearly, the Primo's overall sound was altogether different from that of the darTZeel NHB-18NS: wetter, stickier, slappier (not sloppier), and altogether more generous. But unlike with some other tube preamps, with the Primo I never had a sense of darkness creating a desire for more light, or a lessening of bottom-end resolve creating a desire for more intense wallop.

Sanctimonious though it sounds, I prefer things that, by whatever method, allow me the luxury of the former, sunnier view of myself. The newest product from Fi is one of those.

It is also, as far as I can tell, the first Fi to come with an instruction sheet. "You do have to know which tube goes in which socket," says designer and builder Don Garber, "and you've got to know what the switches are for." He apologizes for the scant labeling on the metalwork: the result of a change in suppliers, itself occasioned by the steady erosion of manufacturing jobs in Garber's beloved Brooklyn. "The company that made silk screens for me went belly up a little over a year ago," he tells me. "I thought I'd found a new one, and I had them do the screen for the back panel, which came out well. Then they folded."

The real subject of our conversation is the Fi 2b, Garber's long-awaited successor to the simply-named Fi Preamplifier, a wildly loved and widely sought product. (I've turned down more than one generous offer for my own original.) In the 14 years since the Fi Preamplifier hit the bespoke audio scene of the 1990s, Garber experimented with various changes, some of which—onboard moving-coil transformers and an all-tube power supply among them—made it into various prototypes. But the 2b is the Fi Preamplifier's first commercially available replacement (footnote 1).

The earlier pre was based on an original circuit by the late Nobu Shishido, in which all voltage gain was provided by four 6DJ8 dual-triode tubes. In each channel's moving-magnet phono stage, the two halves of a single 6DJ8 were arranged plate-to-grid, followed by a passive RIAA network. The line-stage tubes were configured in series-regulated, push-pull (SRPP) fashion, to keep the output impedance as low as possible. The Preamplifier's dual-mono outboard power supply used tubes to regulate the rail voltage—after rectification by a pair of solid-state diodes—and provided a trim pot for maintaining a steady, highish rail. (Three houses and 14 years later, I've virtually never had to adjust mine at all.)

Yet the soul of the Fi Preamplifier was in its construction—or, more specifically, in its circuit layout, which some colleagues describe as one of Garber's greatest talents. (Among the nicest-sounding amplifiers to come from Wavelength Audio's Gordon Rankin was the Baby Ongaku 2A3 monoblock he created for Issue 9 of Sound Practices magazine in 1995; when it came time to lay out the circuit, SP contributor Frank Reps turned to Don Garber.)

Just as important, the Fi Preamplifier defied convention by looking unlike anything else on the market. Both the preamp and its outboard power supply measured a little less than 9" square, viewed from above, and instead of stuffing the parts into a metal box of the usual sort, Garber mounted everything on a series of rigid aluminum tiers, for optimal wire paths (the Fi was a star-ground design) and good looks.

As Garber describes it, the brand-new Fi 2b—priced at $8200 for the full line-plus-MC-phono version that I have in hand—starts where the original left off, then goes in a slightly different direction.

The 2b's gain stages are similar to those of the original Fi Preamplifier, although the new model comes with Electro-Harmonix 6922 dual triodes: functional twins of the 6DJ8. "Exotic NOS tubes? I've just never gone there," Garber says. But the new preamp marks the first time he's used certain other premium parts, in places where he's convinced they make a difference in the sound. Among the more expensive bits are Vishay nude resistors—"They come as close to disappearing as anything," Garber says—and the Teflon and tin-foil V-Caps, from VH Audio. I was surprised to hear of the latter, given Garber's long-standing skepticism about the audible differences among designer capacitors, not to mention his famous dislike of tedious phone conversations about same. But his enthusiasm for V-Caps is real: "They're made by a guy called Chris VenHaus," he says, "and I think it's all straightforward and honest, the way Chris describes his cap-making process."

Don Garber also remained keen to have a pair of onboard step-up transformers for MC phono gain, so he spent a considerable amount of time over the past few years just listening to all the commercial units available. He chose the split-primary HM-3 from Hashimoto Electric of Japan, implemented in the Fi 2b with switch-selectable inputs for low impedance (40x gain) and high impedance (20x gain). As a bonus, the Hashimoto cans are finished in a gorgeous shade of hammertone green that looks more vintage than vintage.

Footnote 1: Fi, 30 Veranda Place, Brooklyn, NY 11201. Tel: (718) 625-7353. Fax: (718) 875-3972. E-mail: dgfi@earthlink.net.

Designing
Dejan Nikic has designed tube electronics for NAT Audio since 1993. Their current product line consists of three tubed preamplifiers, a tubed phono stage, five tubed monoblock power amplifiers, a hybrid integrated amp, and three power cords. The visually striking Magma, a single-ended, class-A monoblock, uses a single TH450 tube and costs $45,000/pair. NAT claims that this amplifier, at 160W, is currently the most powerful single-ended amplifier ever made to use a single tube in its output stage. At $8000, the Symmetrical line stage is not the most expensive preamp in NAT's product line; that honor belongs to the dual-chassis Utopia line stage ($9000).

The Symmetrical's dual-mono, fully balanced circuit includes 12 tubes (6 per channel) operating in pure class-A triode configuration with zero loop feedback. The signal path for each channel has a 6N30P-DR and a 6N1P-VI tube, uses a 6X4WA tube as a power-supply rectifier, and an OA2 as a power-supply plasma-effect stabilizer. All tubes are military-spec new old stock (NOS) and are available from NAT and other tube retailers. The Symmetrical also includes custom-made toroidal transformers, as well as high-capacitance, audiophile-grade capacitors in the signal path and polypropylene capacitors in the high-voltage power supply. Nikic explained to me the Symmetrical has a sufficiently low output impedance to allow it to drive even a 600 ohm load, which he feels is unusual for a pure tube design. He also believes that the preamp's performance will not vary with either volume-control setting or under a wide range of power-supply voltages.

Describing
On the rear panel are five inputs, a tape monitor loop, and two outputs, with single-ended and balanced options for each input and output. All inputs and outputs have gold-plated, Teflon-insulated RCA jacks and gold-plated XLR (XLR) connectors. There's also a neat switch that gives you the option of grounding the line stage to its chassis, in case that results in less hum. In terms of tube noise and hum, I found the Symmetrical dead quiet throughout my listening sessions.

The Symmetrical's rugged-looking yet elegant faceplate of solid aluminum has two knobs: One controls the volume, and is also a pushbutton for the Mute function; the other is an input selector. Each knob is encircled by two rubber O-rings; handling the knobs, all I felt was the rubber—a cute feature. The knobs are flanked by two toggle switches, one for power, and one to switch between single-ended and balanced operation. The small remote control, of solid aluminum, sports only three large buttons: Volume Up and Down, and Mute (footnote 1). Two large heatsinks also serve as the sidepanels.

I used all single-ended input sources into the NAT, and a balanced output into my Audio Research Reference 110 amplifier. I found two aspects of the Symmetrical's operation to be somewhat unusual. Both the volume-control knob on the line stage and the buttons on the remote trigger a series of computer-controlled relays that set the volume level, which is displayed via a series of blue lights along the bottom of the faceplate. As changing the volume causes a series of relays to switch on and off, with no signal being passed during the relay switching process, there will be a few milliseconds of silence between each two steps of volume level. Thus there is a bit of stuttering chatter as the preamp cuts in and out while the volume is being changed, which I got used to rather quickly.

A more annoying feature was the Symmetrical's silent turn-on mode. When the preamp is turned on, it remains Muted while its tubes warm up. During this period the volume setting gradually decreases to zero from whatever the previous setting was; then the preamp comes out of Mute. Thus, every time you turn on the preamp, you need to chatter back up to find your original volume level. A much more convenient and elegant solution would have been to implement something like Audio Research uses in its VSi60 tube integrated amplifier (review to come): during warmup, the amplifier remains in Mute with the volume level constant; the user then needs to manually unMute it.

Listening
Listening to the NAT Symmetrical stimulated me to think about how much the state of the art of tubed preamps has advanced since I began listening to and comparing them nearly three decades ago. In the early 1980s, three manufacturers dominated the tube-preamp landscape: Audio Research, Conrad-Johnson, and the now-defunct Counterpoint. Each company's flagship products had a strong personality, a sonic signature that represented the designer's interpretation of tonal balance, soundstage reproduction, and rendering of transients. And each of these manufacturers had a cult following—only rarely would an AR guy consider buying a C-J product, and vice versa. I never joined any of those cults, however; I've owned multiple preamplifiers and amplifiers made by AR and C-J over the years, and at one point came close to buying a Counterpoint MC step-up preamp.

First, designer and builder Ken Shindo isn't willing to tailor his amps, preamps, loudspeakers, and other playback gear to produce only those sounds that might be heard from a single, mythically perfect concert-hall seat, or to highlight spatial effects for the entertainment of those who know much about sound but little or nothing about music. (Some day, listeners who respond to the sound of Shindo gear may help reclaim the art of critical listening from the ninnies who think it has something to do with "locating images in space.")

Second, and more subversive, there isn't much of a cheaper-equals-lesser dictum in Shindo's product line. Ken Shindo, who's also a noted collector of new-old-stock vacuum tubes and other vintage parts, works to exploit the musical strengths of the many different tubes of which he has chapter-and-verse knowledge; he doesn't appear to regard any one amp in his line as "better" than another, preferring instead to think of each design as being suited to a different set of parts, a different sort of playback system, and a different mood. (In Shindo's current amplifier line there are, for example, two 10Wpc models, two 20Wpc models, two models that use EL34 output tubes, and two that use 300B output tubes; of the ones I've heard, each has different voice, yet all of them sing.) Far more than anyone else with whom I'm familiar, Shindo approaches audio design as an art; if for no other reason, there could never be a best Shindo amplifier, any more than there could be a best recording of Purcell's Dido and Aeneas, a best Bordeaux, a best guitar solo, or a best woman (footnote 2).

Yet for all that, Shindo's preamplifier line is less in keeping with the company's blessed defiance of that Second Commandment: Whereas Shindo's line of power amplifiers flatters the proletarian image I like to keep of myself—their least expensive power amp, the Montille ($3995 when reviewed in 2007), is still among my favorites—I think that, as their prices rise, the different Shindo preamps sound more and more beautiful.

Recent experience with some of the dearer Shindo preamp models has provided a more focused explanation: It may be that I'm just more responsive to the sonic flavors conferred by certain vintage parts, the use of which is a Shindo hallmark. Think of it this way: The amplifier shopper who enjoys the distinctively warm yet muscular sound of a push-pull amp built around the EL84 tube—a power pentode that's relatively plentiful and cheap, and that doesn't require an exotic power supply or output transformer—is in luck. The preamp shopper who falls in love with the organic textures and deep, black internote silences that come from designing a product around a rare, hand-wound, 70-year-old output transformer, is not.

Pretty and sweet
The Shindo Vosne-Romanee preamplifier speaks to the world through a pair of such transformers (footnote 3) which surely take some of the blame for its motorbike-like price: $17,900. Other such factors are the Vosne-Romanee's custom-wound Lundahl moving-coil step-up transformers; its Telefunken EF800, Siemens C3m, General Electric 6072, and Philips 6189 and 6X4 vacuum tubes; its many vintage capacitors (Sprague Black Cats, Sprague Vitamin Qs, a pair of extremely large Sprague oil capacitors); and, of course, its gorgeous metalwork and fully hard-wired circuitry.

At 15.6" (400mm) wide by 6.4" (165mm) high by 10.5" (270mm) deep and weighing 23.1 lbs (10.5kg), the V-R is wider, taller, and significantly heavier than my own Shindo Masseto preamp ($12,000), if similarly styled. Its steel chassis—which, according to Shindo, sounds better than an aluminum chassis, with or without the industry-standard stupid-thick faceplate—is no simple box, but rather has a number of integral partitions for shielding and structural integrity. The steel top is fastened to the outer box and to each of those partitions, requiring a total of 26 machine bolts for one chassis surface—and all of the bolt openings are tapped with the appropriate threads. Unlike the products from some of perfectionist audio's best-known makers of tube amps, Shindo's enclosures can actually be put back together the same way they came apart.

There are also different levels of steel platforms, for various different parts. Most of those surfaces are ventilated, and all of them—every partition, every platform, even the straps that hold the Lundahl transformers in place—are beautifully finished on all sides in Shindo's trademark metallic-green paint. The rear panel contains several pairs of Switchcraft low-mass RCA input jacks—Shindo's favorite—and a pair of XLR output jacks. A 0.2"-thick acrylic panel, silkscreened from the back, adorns the front. (The front panels of Shindo's preamplifiers used to be made of glass, but frequent breakage during shipping has discouraged that practice.)

Inside, the Vosne-Romanee's power supply is build around a larger-than-average Denki mains transformer, custom-made for Shindo. As in the Masseto preamp, the power supply's main rail is rectified by a series pair of 6X4 full-wave rectifiers, but the V-R has much greater storage capacitance than that model, and double the number of silencing chokes (a Raymond Chandlerism if ever there was one).

Footnote 1: Shindo Laboratory. Web: www.shindo-laboratory.co.jp. US distributor: Tone Imports. Tel: (646) 425-7800. Web: www.toneimports.com.

Footnote 2: Note to my wife: I do not mean that.

Footnote 3: These transformers' metal enclosures, significantly larger than the ones that contain MC step-up transformers but not quite as large as those of amplifier output transformers, are painted light gray, presumably by Shindo Laboratory, and presumably to disguise their make and model from the prying eyes of circuit plagiarizers, who plague Shindo still.

In fact, Bohlmeier has done far better than Albert Einstein's father. In the late 19th century, when Hermann Einstein opened an electronics company in Munich, he bet on direct current. Fourteen years later, by which time alternating current became the dominant way to distribute electricity, he was forced to close his business.

From their beginnings until today, Einstein Audio Components' products have been built to an heirloom standard set many years ago by McIntosh, and impeccably finished in a distinctive visual motif of chrome and black lacquer. The packaging (wooden crates) and overall presentation are equally impressive, and foster in the buyer a pride of ownership befitting an expensive, limited-edition product. It may not be essential, but it's a nice touch.

Einstein Audio's initial offering in 1990, a solid-state amplifier called simply The Amp, was an immediate critical and commercial hit that remained in the company's product line for six years. They followed with a series of equally well-reviewed and commercially successful tubed and solid-state products with similarly generic names, including, in 1998, the original The Tube line-stage preamplifier.

The Tube Mk.II
The Tube Mk.II is handsome, looks essentially identical to The Tube, and sells for $18,400. The fully balanced, true dual-differential, pure class-A design features point-to-point wiring and a shunt-to-ground volume control that's outside the short signal path. Its true dual-mono topology is indicated by the two large, chrome-plated power-transformer enclosures protruding from the chromed chassis.

The volume level and the five choices of source can be selected via a plastic remote control unbefitting so costly a product. Another $930 gets you a curvaceous, beautifully finished model called—what else?—The Remote, which also controls Einstein's CD player, The Source. If you can afford The Tube, you'll want The Remote.

The Tube Mk.II incorporates eighteen E88CC/6922 dual-triode tubes and one ECC82/12AU7 dual-triode. Ten of the E88CC/6922 tubes function as input pairs and are lined up along the chassis rear, inserted in sockets affixed to a spring-suspended subchassis that's one of the modifications made for the Mk.II. The remaining nine tubes handle primary line-stage gain and buffer functionality.

Einstein claims a high-current output stage, an ultralow output impedance of 50 ohms, greater than 98dB channel separation, a signal/noise ratio in excess of 95dB, and an unusually high (but unspecified) bandwidth. The total harmonic distortion at 1.5V RMS is claimed to be less than 0.05%. Although less than complete, this is an impressive set of specs for a tube preamplifier. The Tube runs hot. You might be able to fry two eggs sunny-side up atop those chromed transformer enclosures.

Uniquely configured
The Tube is relatively compact, with a minimalist feature set and only two knobs, with which you can switch inputs (sort of, as you'll read), regulate the volume, and that's it. A rectangular cutout on the front panel does not contain a fluorescent display, and there's neither a mono switch nor a balance control.

Because The Tube's narrow rear panel already contains five sets of inputs (three balanced on XLRs, two single-ended on RCAs), a single set of balanced outputs (The Tube can be ordered with single-ended outs), and two pairs of singled-ended tape outputs, the IEC power socket and On/Off switch are on the chassis's underside. Although the resonance-absorbing feet are unusually tall, the AC cord's female connector must be short and right-angled if it's to fit. You can forget about trying most audiophile power cords with The Tube.

If not, not. In my experience, the electronics offered by Conrad-Johnson, Mark Levinson, Quicksilver, and Audio Research have achieved this most perfect synthesis of technology in the service of music. Each of these companies offers products which force a reviewer to search for new words to help communicate to the reader a quality of sound reproduction which transcends excellence. The SP14 preamp is such a product.

The last Audio Research product to grace the pages of Stereophile was the original SP9 preamp, which got a rather lukewarm review at the hands of both J. Gordon Holt and John Atkinson (Vol.10 No.8). That was 2½ years ago! A follow-up to that review is in progress; only the flagship SP15 preamp now remains a stranger to Santa Fe.

The SP14, being a single-chassis unit, shares an outward appearance with the SP9 Mk.II, albeit with half again as many knobs and twice the number of toggle switches. Retailing for $1200 more than the SP9 Mk.II and just half of the SP15's $5995, it occupies an increasingly volatile market niche, with strong competition from both this country and abroad. The SP14 is a hybrid design, and follows in the tradition of the SP9 Mk.II in using one 6DJ8 dual-triode vacuum tube in an intermediate gain stage of the phono section. The line section is all-FET, using a circuit similar to that in the SP15.

The onboard power supply features extensive electronic regulation for both low and high voltages and an excellent, shielded toroidal power transformer (located at the rear of the left-hand sidewall so as to be as far away from the sensitive phono circuitry as possible). Construction quality is first-rate—just what we have come to expect from ARC. Ergonomics are also excellent, with highly legible markings identifying intelligently placed controls. I appreciate the fact that no squinting is required to read the black lettering on the silver front panel. (I wish more companies would adopt this style. I'm getting tired of the ubiquitous black with gold lettering. For those who want black, however, Audio Research components are available, at extra cost, with anodized black front panels and knobs with silver-gray lettering.)

One characteristic which distinguishes the SP14 from the SP9 Mk.II is the degree of control flexibility offered. For starters, it has separate gain and attenuation knobs for optimization of the signal gain setting. Separate input and record output switching offers the listener the opportunity to listen to one source while recording from another. Perhaps most important, the SP14 incorporates a bypass switch which directly connects the attenuation knob and gain control to the source selected by the input selector. All other controls and switches are removed from the active signal path when this switch is thrown. Note that, unlike other competing designs, this configuration applies to the line inputs as well as to the phono stage. There are two separate sets of outputs (for ease in bi-wiring) and three convenience AC outlets.

Six knobs grace the front panel. From the left is Gain, a stepped control which normally should be used around the 12 o'clock position, the main volume level being set by the Attenuation knob next to it. This provides 6dB of attenuation in each of four increments (–6, –12, –18, and –24dB). Using these controls makes easy the optimization of input levels relative to output. The next knob is the Balance control, also stepped. Beside it is a control which I feel essential to any high-end or other preamp—Mode. In addition to stereo, this control provides the all-important mono position, which will eliminate the pseudo-stereo effect in early-'60s reissues of recordings originally recorded monaurally. This control will also reverse the stereo channels and send either left- or right-channel information to either channel as selected. Next is the Record Out selector, which shares a menu with the Input selector. Input source selection is provided for Phono, CD, Tuner, Video (!?), and Spare. With separate inputs and outputs for two tape recorders, I feel the SP14 will offer all but the most obsessed enthusiast more than enough source selections. Independent control of these functions is a handy feature, though purists may look at it scornfully.

Below the control knobs are two sets of sturdy toggle switches. To the left of the green LED are Power, Outlets, Bypass, and Mute. The Outlets switch controls power to the three grounded receptacles on the rear panel. Two of these are switched, the third unswitched. I did not use these convenience features during the course of my audition, though, preferring a separate, computer-grade protection device such as Data Shield's Model S85. The Mute switch is not only handy when listening must be interrupted, but helps safeguard the power amp and speakers from unexpected transient signal surges. I activated it each time I changed signal sources. In Mute position the green LED glows faintly, returning to full brightness when Mute is disengaged. The Bypass switch, as mentioned earlier, provides essentially a "straight wire with gain" conduit for either line-level sources or phono. All of my serious listening was done with the Bypass switch on.

The other four toggles deal with taping. Bidirectional dubbing can be accomplished with ease, even while listening to another program source. Source/tape monitoring is done with the flick of a switch.

The only features missing on the SP14 (found on the SP15) are absolute phase, phono high-pass filter, and front-panel impedance selection for cartridges. This last can be accomplished on the SP14, though, with a little soldering. I asked Audio Research about this somewhat awkward arrangement for optimizing cartridge loading; they feel that in a product at this price point, a hard-wired connection is the best compromise, in terms of mechanical stability and signal quality, between various pin-socket arrangements and the expense of a precision, gold-contact switch (as is found on the SP15). Unless you're a compulsive cartridge swapper, I feel this inconvenience to be minor. My sample was loaded for 100 ohms for my AudioQuest 404i-L cartridge.

In my review of the SP8 in the February 1983 edition of Hi-Fi News & Record Review, I described a topnotch performer that was equally adept at reproducing both vinyl and early CDs. It moderated some of the harshness and hardness heard on some recordings and I recommended it highly.

With hindsight, I feel that the SP8's moderating influence resulted from the summation of certain subjectively complementary low-level errors that resulted from the SP8's single-ended triode (SET) design. These helped make the "push-pull," wholly symmetric digital audio signal paths sound a little more natural. That same effect can be easily observed in many Cary designs, and not least in Conrad-Johnson's ART zero-feedback triode line preamplifier.

Now, some 17 years after the SP8 made its appearance, Audio Research has released a close parallel: the triode LS8. Priced at $1495, the LS8 physically resembles the SP8, but with two of the latter's four rotary controls omitted, as well as two of the four front-panel levers. The LS8 is substantially less expensive (in real dollars) than the older model, but the ubiquity of digital sources has led to a dramatically simpler design.

Gone are the stereo/mono mode switch, the balance control, the tape-monitor loop and its toggle switch, and the phono section. What you get instead is a line-control unit with an accurate chip-controlled volume control—"64 steps of 0.5dB,"—five single-ended RCA inputs, and two outputs: one for Tape Record, the other for the power amplifier.

While the SP8 was built on a steel chassis, ARC's current series of electronics have been made significantly lighter through the use of aluminum alloy casework. This can result in two benefits: aluminum is typically less resonant than steel, meaning that you can expect less reaction to vibrations and the local sound field. In addition, all metal surfaces react to some degree to electromagnetic audio fields; ie, those signals passing through the electronic structure of an amplifier. As a nonferromagnetic material, aluminum interacts with such fields to a degree an order of magnitude lower than that of sheet steel.

From another viewpoint, the LS8 can be seen as a successor to the Audio Research LS7, one of my favorites in this market sector—or, alternatively, a stripped-down, single-ended-only, nonremote version of the ARC LS15, another favorite of mine.

A moderate fourfold signal gain of 12.2dB is promised by the LS8, appropriate for line-level sources. Input impedance is quite high at 50k ohms, a fairly gentle loading for signal sources. Signal inputs of up to 3.5V are suggested; some of the few very-high-output digital audio sources are thus contraindicated, and might need a line attenuator to avoid overloading the LS8's input section. To get the best performance, Audio Research suggests output load-matching criteria of 60k ohms and 100 picofarads. Many power amplifiers will meet the 60k ohm part of this spec, but almost none will manage the latter figure for capacitance if you include a typical interconnect cable. Hedging their bets, ARC gives minimum suggested values of 20k ohms and 100pF; perhaps the LS8 isn't quite so load-sensitive after all.

Rated output is specified at 2V for an IHF 0.5V input, 0.5Hz–100kHz, ±0.5dB. Other specifications quoted are a very wide overall frequency response of 0.1Hz–250kHz, –3dB ("solid-state" performance), and a rated distortion of –80dB, or 0.01% in the midband.

Technology
To achieve that exceptional low-frequency bandwidth, a capacitor-coupled tube amplifier requires some special circuitry. ARC uses one of their established techniques, which places that output capacitor in the feedback loop, effectively making a modest-valued component that has been selected for good sound act as one of much larger value, offering greatly extended low-frequency response.

Each channel has a pair of wide-bandwidth double-triode 6922s from Sovtek, an Audio Research favorite. These tubes are readily available, sound very good in suitable circuits, have a long life (typically 5000 hours), are cheap to replace, and, in the case of the LS8's circuitry, don't require special selection or matching. Vibration-absorbing rings are fitted to the tubes.

That may sound like fan hyperbole, but it's what I immediately heard a few years ago, when I first encountered Ypsilon gear at a hi-fi show. Though the company was then new to me, nothing I've heard since, at shows or at home, has deviated from that very first impression.

Ypsilon models look beautiful, even dramatically so, in their cases of thick, milled aluminum, and perhaps that's what first drew to them the reviewers and civilians who attend audio shows. What kept them there was the sound, or the lack thereof.

Many listeners tempered their initial enthusiasm with caution: A sound that good must be based on sonic tricks that only time will reveal. I found myself almost wishing that to be true, given what Ypsilon products cost. But having spent a great deal of time with Ypsilon's VPS-100 phono preamplifier ($26,000; I reviewed it in "Analog Corner" in August 2009 and March 2011), I'm convinced there are no tricks.

As is often the case in high-performance audio, less artifice comes at a high price. Ypsilon's products are very expensive, and deceptively simple in design. The PST-100 Mk.II will set you back $37,000. If you don't need the active stage, the completely passive PST-100 TA can be had for $26,000. (The active tubed stage can be retrofitted at the factory.) But either way, and considering that a preamplifier's basic job is to switch and route low-level audio signals without adding to or subtracting from any signal fed to it, these are high prices to pay for what is, essentially, nothing.

Of course, there's more to a preamp's job: It must also provide signal attenuation and, usually, gain, as well as an output impedance low enough to drive cables and interface with a power amplifier of high input impedance.

A Purist Approach
Over the past year or so, a few impressively neutral, dynamic, quiet, wide-bandwidth tube preamplifiers have passed through my listening room that rival the quiet and tonal neutrality of my reference, the solid-state darTZeel NHB-18NS. The best solid-state and tubed preamps these days are more sonically alike than different, though of course the subtle differences are the basis on which listeners who can afford such products choose.

Ypsilon's co-owner and chief designer, electrical engineer Demetris Backlavas, believes that the key to a preamplifier's sound is the means by which it attenuates the signal it's fed. Instead of the more commonly used resistor attenuation, Backlavas uses what he says is a very linear, 31-tap transformer that Ypsilon winds in-house. By comparison, he says, attenuators that use even the finest-quality resistors tend to sound grainy and discontinuous because the in-series resistor converts voltage into current, while the parallel resistor turns current back into voltage.

Not that Backlavas and his partner, Andy Hassapis, didn't try to build a better resistor-type attenuator, using a variety of materials. The problem, according to Backlavas, is that, in order to resist, a resistor must be made from a bad conductor of electricity. Copper and silver are good conductors and small-value resistors can be made from these metals that, not surprisingly, can sound very good. Unfortunately, it's impossible to use copper and silver to make high-value, wideband resistors because of the parasitic inductance that goes along with the need to use coils of very many turns. In addition, resistor-based attenuators waste signal energy by turning the attenuated energy into heat.

Nonetheless, Backlavas admits that attenuators of reasonably high quality can be built using carefully chosen resistors. You're probably listening to such a device as you read this. And, as anyone who has spent time listening to transformers (and Backlavas has spent more time listening to them than most) knows, even those with identical specs can sound remarkably different from each other, and some can ring unpleasantly or sound bad for a variety of reasons.

In fact, transformer-attenuated preamplifiers—or, more precisely in this case, autoformer-attenuated preamps, in which the primary and multi-tap secondary overlap—aren't new. Hobbyists have advocated and built them over the years, but few are commercially available. The advantage of such an attenuator over one that uses resistors is that energy is transformed and not lost as heat. Backlavas gave an example: starting with a source impedance of 1200 ohms, attenuating the signal 10dB (or 3.16 times), produces an output with lower voltage and higher current and an impedance of 120 ohms (1200/3.16²), which has an easier time driving loads, unlike the less amplifier-friendly results produced by a passive-resistor attenuator.

That said, of course, transformer attenuators have their own problems that must be solved before they can produce good sound. The core material must have low hysteresis (hysteresis being like unwanted "magnetic memory") at both low and high frequencies, and linear magnetic permeability with flux and frequency.

We all can agree that a Bose Wave CD player sits at one end of that continuum, a Koetsu Jade Platinum phono cartridge at the other—but what of all the products in between? Scarcity, mode of manufacture, appearance, even sentimentality ("This is just like the one my father used to have!")—each plays a role, but there's no doubt that price tops the list: The more we pay, the more we love (footnote 1).

If we remove price from the equation, what then? Is sheer quality enough to tip the balance? Given the availability of audio goods designed in America or Europe and manufactured in the Far East, the answer may already be at hand: Although hand-wired vacuum-tube electronics have long been among the most cherished audio heirlooms, those sorts of thing are now available at popular prices, as the saying goes.

Consider PrimaLuna, who design preamplifiers, power amplifiers, and integrated amplifiers in the Netherlands, and manufacture them in their own factory in the People's Republic of China. Far from being just a me-too maker of undistinguished gear or—worse still—a forger of other people's designs, PrimaLuna has impressed me and other reviewers with their innovative circuits and very high build quality, not to mention a level of performance that, nuances aside, can be summed up in three beautiful words: classic tube sound. Thus, when John Atkinson asked me to audition their new top-of-the-line DiaLogue Three preamplifier ($2599), I was, as they say, all over it.

Description
The first thing I noticed about the DiaLogue Three—before the nice paintwork, before the point-to-point wiring, before the sound—was its considerable weight: 53 pounds of iron, steel, and copper, not to mention all those vacuums. This preamplifier weighs half again as much as some of my favorite tube power amplifiers, and three times as much as my dog.

That's because the DiaLogue Three is designed and constructed as a dual-mono preamp—all the way through, including the power supply and, unusually, the mains transformers. Said trannies appear to have been made specifically for the Three, and are protected with a circuit that detects excess heat and temporarily interrupts the primary feed, should things require cooling down. Shoppers in hell, take note.

The DiaLogue Three is housed in a steel chassis whose shape also calls to mind that of a classic tube amp. A vented transformer cover towers over the rearmost portion of the low-slung surface, the latter interrupted by six tubes: two rectifiers and four dual-triodes. The tubes are protected from the user, and vice versa, by a removable cage shaped like a roll-top desk's most distinguishing feature, and the front of the chassis is fitted with a modestly sized faceplate. The latter is aluminum alloy with a satin finish, while the remainder of the DiaLogue Three is painted with five coats of hand-rubbed blue-gray lacquer. (Black is also an option.)

The preamplifier circuit is straightforward: In each channel, the line-level input signal is conducted, by means of a shielded cable, to the signal grid of a 12AU7 dual-triode tube, the two halves of which are operated in parallel. The ganged plates of that tube are tied, via a Solen capacitor, to the input of a motorized Alps Blue Velvet volume pot, the output of which is sent to the signal grid of a second 12AU7, also operated with its two halves in parallel. The plates of that tube are capacitively tied to the output jacks . . . and there you have it. Except for that very last connection (the printed circuit board for the power supply is also home to the chunky Solen signal-output caps), the preamp section is hand-wired, point to point, with liberal use made of nice-looking ceramic terminal strips, which are cemented to the inside of the chassis.

The DiaLogue Three's power supply, which is constructed using both point-to-point wiring and the remainder of that PCB, uses a single 5AR4 tube for the rail voltages of each channel. Filament voltages for the rectifier tube are supplied by dedicated secondaries of the mains transformers, while full-wave rectifiers constructed with discrete diodes are used to heat the signal tubes.

A modestly sized solid-state preamp board ($199) can be used to change one of the DiaLogue Three's inputs to a moving-magnet phono input, but that option wasn't provided with my review sample. As typically supplied, the preamp's rear panel has five line-level inputs, two system outputs, one tape-monitor output, and pass-through jacks for driving a surround-sound device in a home-theater system. The front panel sports only a volume knob and a source-selector knob, which of course prompted my usual teary whine: I wish the PrimaLuna had a balance control and, especially, a mono switch. I would happily trade that surround-sound thing for the latter.

Included in the price is a remote-control handset that incorporates all the user controls for this and, apparently, all other PrimaLuna products—including the Triode/Ultralinear mode switch for the DiaLogue Seven power amplifier. The handset is built into a chunky aluminum-alloy housing of reassuring heft, although a seam between the two halves of the housing had an edgy feel that, in time, discouraged me from using the remote. (I needed the exercise anyway.)

The DiaLogue Three's external fit and finish were very good, and its internal build quality was nothing short of magnificent.

Like most audiophiles, I eventually discovered just how difficult it is to find a preamp that sounds good and still leaves some of one's audio budget for other parts of the system. Designing and building a first-rate preamp must not be an easy matter, or there'd be more of them around. For the past several years I've been using a Conrad-Johnson PV5 and have felt no great urge to replace it, reading with a certain smugness in JA's review of the PV9 (Vol.12 No.5) about the fellow who swapped his PV5 for a Class A phono stage/passive controller combination that was ostensibly more accurate, but somehow he wasn't enjoying listening to music as much as he had with the PV5!

As the PV11 is a lineal descendant of the PV5, I was particularly excited to get it as one of my first pieces of equipment to review. Does it retain the musicality of the PV5 while offering lower levels of coloration? Then there's the Threshold FET nine/e that I also review this month, junior sibling of the FET ten/e, a solid-state preamp that has earned a rave review from noted tubeophile Dick Olsher (Vol.14 No.3). And what about the Audio Research SP9 Mk.II, sent along by Thomas J. Norton to provide a reference? Does this hybrid design combine the best of vacuum tube and solid-state technologies? Can a tissue be soft and strong? Will Spenser ever meet Kinsey Millhone?

The Conrad-Johnson PV11: $1895
Conrad-Johnson advertises the PV11 and the MV52, its companion power amp, as "classics revisited." Indeed, there are obvious similarities to previous C-J models: the brushed-gold front panel, reassuringly large knobs, two sets of main outputs, the much-criticized tape-monitor arrangements (I've never had a problem with this, but it is possible to induce potentially speaker-damaging positive feedback with the wrong combination of switch settings), and phono inputs optimized for moving-magnet rather than moving-coil. The tape outputs are not buffered, so, for optimal performance, any tape decks connected to these outputs should be turned on even when you're listening to another source; alternatively, the plugs should not be connected to the tape-out jacks except when making a recording. In design, C-J seems to be moving toward a less-is-more philosophy: there is no global negative feedback (the PV5 had about 30dB feedback), and the line stage now consists of one rather than two gain stages. Since the line stage inverts absolute polarity, speaker connections should be reversed (assuming that the amp is non-inverting).

In comparing the PV11 with the PV5 and the Audio Research SP9 Mk.II, all of which are non-inverting, I took care to make the speaker connections so that the system as a whole would be non-inverting. (The original Quads are polarity-inverting themselves; ie, what's marked positive on the Quads would be marked negative on other speakers. In this case, two wrongs do make a right.)

The balance control of the PV11 is a multiple-position switch rather than a potentiometer; resistors are switched in to provide a maximum of 12dB lateral shift, and they are out of the circuit in the neutral position, thus removing the temptation to do a balance-and-mode bypass modification. (Still, I do wonder if bypassing the mode and balance switches could provide at least half a smidgen of improvement.) Conrad-Johnson has eliminated the use of electrolytic capacitors in their latest preamp designs; all capacitors, including the ones in the power supply, are the same high-quality polystyrenes and polypropylenes made to their specifications.

Quality of construction is to a higher standard than in earlier models (switches and the volume-control pot seem smoother than on the PV5), and the overall look is one of restrained elegance.

Listening tests
At the 1990 Stereophile Writers' Conference, someone (PWM or LA, I think) made the point that audiophiles and audio writers have a tendency to use hyperbole in describing sonic differences between components, so that any noticeable difference is described as "huge." Meridian's Bob Stuart, in his Vol.14 No.9 interview with Bob Harley, has suggested that it is almost a law of human nature to emphasize the importance of small differences. Taking into account both points, I can say that the sonic differences among the auditioned preamps were "huge"; alternatively, that the differences were small but important, perhaps because they were small.

The PV11 quickly impressed me as being distinctly superior to my somewhat-modified PV5: less colored, tidier at the frequency extremes (sweeter highs, but instruments with a lot of overtones retain their characters; less bloated bass), a more layered definition of depth, more precise focus, greater sense of ambience, and, when the music featured these qualities, an exceptional sense of rhythm and drive. In fact, it was this "rhythmic" quality, allowing one to more easily follow small variations in tempo, that I found to be the PV11's outstanding positive attribute, present in both LP and CD reproduction, an attribute not quite matched by any of the other preamps (footnote 1).

Listening to something like "Under the Sea" (from The Little Mermaid, Disney CD-018), Paul Simon's "Kodachrome" (There Goes Rhymin' Simon, Columbia XC 32280, LP), or Dick Hyman playing Fats Waller's "Viper's Rag" (Reference Recordings RR-33DCD), my toe (which does a lot of listening) could not keep from tapping. The rubato and subtle dynamic variations in the Bolet/Chailly Grieg Piano Concerto (CD, London 417 112-2) were more apparent than I've heard before, and, again, better than with any of the other preamps in the system. Voices had a rounded, "present" quality; listening to Pavarotti's Carnegie Hall recital (CD, London 421 526-2) in the dark made me wonder if Luciano had somehow teleported himself into my listening room. (Come to think of it, the floor did creak in a mysterious manner during "Pietè Signore.")

Rumor has it that in the mid-1990s, when William Z. Johnson set out to design the original Reference 1 preamplifier, his jumping-off point was the SP10. What's not a rumor is that the Reference 1 carried on the tradition and lineage of those classic preamps. It was fully tube driven, it was the best preamp that ARC could build with the technology available at the time, and, like its predecessors, the Ref 1 redefined what was possible in a tube preamp—and, to some ears, what was possible in a preamp of any type. It thus became an instant classic. Time and technology marched on, and the Ref 1 was replaced by the References 2, 3, and 5, each one in turn carrying on the lineage and setting a new standard of performance (footnote 2). The latest in this series is the Reference 5 SE ($12,995), introduced in late 2011 to replace the Reference 5.

The Reference 5 SE
The Reference 5 SE benefits from several advancements that Audio Research developed for the cost-no-object, limited-edition Anniversary Reference preamplifier ($24,995), including two readily visible changes from the Ref 5. First, the 5 SE's front panel mirrors that of the Anniversary: the buttons are now tidy, round, metal, and set directly into the faceplate, rather than the Ref 5's square, recessed plastic buttons. Added on the inside, to the left, is a vertically mounted circuit board, stuffed to the gills with capacitors to nearly double the Ref 5's already massive energy storage. Other changes include new Teflon and hybrid coupling and bypass capacitors, and changes in the internal wiring.

Otherwise, the Reference 5 SE's basic circuitry resembles that of the Ref 5: it's mostly contained on a single-sided board that stretches from wall to wall. Additional, smaller boards handle I/O, switching, and display functions. As in the Ref 5, the Ref 5 SE's transformers—an R-core for the audio circuits, a toroidal for the display, relays, and microprocessor controls—are mounted on the sides of the chassis, above the board, to reduce vibration in the audio circuits. The audio circuitry's power supply is a hybrid-regulated design employing both FETs and a 6H30 dual-triode tube controlling a 6550 pentode tube. The audio circuit itself is "as simple and good as we could make it," according to ARC's David Gordon: a single gain stage with a cathode follower tube in a zero-feedback, fully balanced, pure class-A triode layout. There are six tubes: four 6H30s in the analog stage, and the 6550 and 6H30 in the power supply.

The Ref 5 SE's cosmetics and layout are handsome in Audio Research's traditional fashion, and it's the same goodly size and weight as the Ref 5: 19" wide by 7" high by 15.5" deep, and weighing 30 lbs. The chassis is of heavy aluminum, with a thick, elegantly machined aluminum front plate and two rack-mount handles. The review sample's top plate was of tinted polycarbonate, with slot vents. A cover of perforated aluminum is also available, but ARC recommends the polycarbonate, perhaps because of lower chassis resonances. The Ref 5 SE is available in natural aluminum or black.

Like the Reference 5, the 5 SE has at the center of its faceplate a large window that prominently displays the volume setting with both a bar graph and large numeric display. Also displayed, in far smaller characters, are the selected input, whether it's single-ended or balanced, and the status of several functions: mono/stereo, mute on/off, and polarity non-inverted/inverted. The display can also be used, via the remote control, to show the hours each tube has been used, handy for confirming what your ears will tell you: that you've reached the 600 hours of break-in time that ARC recommends. Via the remote, the display's level of illumination can be set to one of eight levels, from off through bright. But even when the display is fully dark, any change in status lights it up for 10 seconds to display the change.

Flanking the display are two return-to-center knobs, attached to a microprocessor, that turn through an arc of about 60°. The one on the left changes the volume across 103 steps; the one on the right selects among the six inputs. Below the display are six small buttons: Power, Mute, Mono, Balanced/SE (input selection), Processor, and Phase Invert. All front-panel controls are duplicated on the utilitarian remote control, which also includes buttons for adjusting balance and display brightness, and to display tube hours.

On the rear panel are six line-level inputs, a processor input that bypasses the volume control to allow the Ref 5 SE to be integrated into a home-theater system, a set of line-level output jacks for recording, and two sets of outputs. All inputs and outputs have both single-ended and balanced connections. For each input, the type of connection is selected with the Balanced/SE button on the front panel or remote.

Use, Handling, and . . .
The Reference 5 SE was a delight to use. All of its controls were intuitive and positive, and the single-ended and balanced connectors were all nicely solid. An interesting feature of the Ref 5 SE is that it's supplied with a 20A IEC power cord, not one of the standard 15A cords we're all used to. I would have liked to experiment with different power cords, but no dice: Although ARC's extremely heavy cord is removable, it takes a different type of IEC connector. I used the Ref 5 SE with both single-ended and balanced sources, and though I experimented with both types of outputs, I found that I marginally preferred the sound with the Ref 5 SE's balanced outputs driving my VTL amps. In any configuration and throughout my listening, the ARC was completely quiet.

Performance
The Audio Research Reference 5 SE's lineage and $12,995 price led me to expect a lot from it, and it didn't disappoint. I can sum up my impressions of it in one word: "Wow!"

The Ref 5 SE was no shrinking violet of a preamp. It was big, bold, brash, and brassy, in exactly the way live music is big, bold, brash, and brassy. While listening to Howard Hanson and the Eastman-Rochester Orchestra's Fiesta in Hi-Fi (CD, Mercury Living Presence 434 324-2), I was repeatedly taken aback at how the Ref 5 SE reproduced this recording's full dynamic spectrum and harmonic richness. Bass drums exploded and boomed with a wonderfully round, taut tone. Castanets cracked with perfect woodiness. Cymbals rang and shimmered, maracas hissed and clattered, violins sang, cellos swooned, basses groaned ominously. You name it, the Ref 5 SE nailed it.

Nor did the Ref 5 SE shine only with sonic spectaculars. It was equally impressive with small, intimate recordings, such as Doug MacLeod's Brand New Eyes (CD, Reference Fresh FR-703CD). Its incredible resolution and recovery of fine detail made it seem as if several layers of the recording chain had been wiped away and I was in the room with MacLeod.

Regardless of the genre or scale of the music, the Ref 5 SE brought about a fundamental, not-subtle jump in my system's performance. It was like hearing a good direct-to-disc record for the first time, or maybe an analog master tape. It wasn't just a question of being a little bit better, or better in one or two ways—it was a broad-brush, wholesale improvement, like going from two dimensions to three, or from black-and-white to color. It was almost the sort of fundamental change in the fabric of the music that (dare I say it) separates live music from recorded. This isn't to say that my system magically became the real thing with the Ref 5 SE installed, but it was definitely a step closer.

The ARC's reproduction of timbres was uncanny; it seemed to be painting the music with a tonal palette both bolder and more nuanced than I'd heard before. Tonal colors were richer and denser, harmonic structures more complex. With the Ref 5, instruments sounded more like themselves, more alive. Violas were more distinct from cellos and violins, and each woodwind retained more of its inherent characteristics, whether it was the woody, airy tone of a clarinet or the reedier buzz of an oboe. The rich way the Ref 5 SE portrayed instruments also made it easier to discern individual musicians in a chorus or orchestra, the differences in timbre and texture enhancing the image specificity to create a nearly holographic effect. With recording after recording, I noticed more complexity in background voices and passages—and not just across the midrange, but from the very bottom to the very top of the audioband, and across the full range of volume, from the faintest pppp shadings to thundering ffff climaxes.

Footnote 2: The Reference 1 was reviewed by Wes Phillips in the November 1998Stereophile; the Reference 2 by Michael Fremer, in September 2000; and the Reference 3 by Paul Bolin, in December 2006.

After a yearlong scrutiny of the state of the art of line-stage design, two underlying principles suggest themselves. First, the Gods, in having decreed that man shall labor long and hard in search of the perfect preamp, must surely be crazy. In the trek toward sonic perfection, mistakes are frequently made. The attempt to coax the signal from the program source and nurture it to its full musical potential is fraught with labor pains. Like the political process, the audio signal is subject to corruption. Small sins early in the chain may become capital offenses by the time they reach the loudspeakers.

That there is only a handful of great-sounding preamps out there (line-level or otherwise) is evidence prima facie of the difficulty in caring for the audio signal at its formative stage. Aided by these few centurions of sonic truth, the music can bloom, filling the soundstage with the fire, drama, and power that only live music can communicate. Too often, however, the preamp sinks the ship, and the "illusion of live" descends from the realm of the plausible into the realm of yellow cling peaches. Most preamps can hope to simulate the flavor of a fresh peach only to the extent afforded by the canned variety.

Second, I feel it essential for a preamp to incorporate the magic of the vacuum tube—especially where digital source material is concerned. In my experience, the ultimate sound of any CD player or digital processor is dependent upon the type of associated line-level stage. If you doubt this for even a moment, I invite you to audition the Theta DS Pre Generation III processor/preamp in my listening room. Through the DS Pre's own solid-state line-level stage, the sound quality deteriorates to the point that the bloom and dynamic breadth of the music are largely squashed. The resultant harmonic textures are convincingly solid-statish, the overall effect being to subdue the vital link between perception and belief. Route the DS Pre's analog output from Tape Out to a good all-tube preamp or even a good hybrid design, and the sound quality changes dramatically for the better. The illusion of a real musical event ebbing and flowing before my ears becomes enormously heightened. For whatever reason, it's clear (to me at least) that digital sources require a tube buffer prior to the power amp. This is yet another manifestation of "Futterman's First Law of Audio": Thou shalt use a vacuum tube as early in the amplification chain as possible.

"Wait a minute," I hear some of you complaining. "If you're so hot about tubed preamps, why has the Threshold FET-10/e line stage lasted so long in your reference system?" Good question. Let me remind you that nothing is sonically right if it's harmonically wrong. And my master tapes have long told me that the FET-10/e line-level preamp really got the upper-midrange/lower-treble tonality right. This is the frequency range that makes or breaks soprano voice, and the Threshold didn't let me down. But deep in my heart I knew that its mastery over soundstaging and dynamics was less than perfect. Memories of the Conrad-Johnson Premier Three periodically flooded my consciousness. The way the C-J sculpted image outlines was a sound to behear. Certainly, no solid-state preamp—including the Threshold—came close in this respect.

The final straw was my exposure to the Convergent Audio Technology SL-1 preamp (reviewed in December 1992 by Jack English). During the several weeks the CAT resided in the system, I connected with the music like never before. My level of sonic expectation would never again be the same. The CAT showed me that harmonic integrity, palpable imaging, and dynamic range can be bundled together in one package.

It was into this environment that the Jadis JPL made its grand entrance.

Technical details
The Jadis JPL is a thing of beauty, a beguiling French damsel. The gold inner fascia set off against the chromed chassis finish looks positively luscious. Sitting as it did amid some pretty drab-looking neighboring gear in Bright Star Audio's "Rack of Gibraltar," Lesley had no trouble at all picking it out.

"Oh, what's that?" she asked.

"That there, my love, is the El Dorado of preamps."

Having looked over the JPL's schematic, it's difficult for me to objectively identify the source of its sonic magic. The design, by Jadis's Andr;ae Calmettes, is pretty conventional. The four line-level inputs and one tape loop are routed through three 12AX7 dual triodes (footnote 1). Voltage gain on the order of 35dB is provided by the first two 12AX7s, which are cascaded together. The final tube in the chain, used as a cathode-follower buffer stage, allows the use of long cable runs to the power amp without the danger of treble rolloff. There's also a dedicated CD input that uses a single 12AU7 as a buffer stage (unity gain). This input is DC-coupled to the 12AU7's grid, while the regular line inputs are AC-coupled via a 1;uF capacitor. The quite beefy power supply deploys solid-state bridge rectifiers followed by a capacitive filter network. As a final touch, active regulation is provided for the tube plate voltages.

A large circuit board accommodates the entire active signal path, and construction quality and part selection appeared to be nothing short of excellent. Stereo volume and balance pots adorn the front face. The mute switch—an especially useful feature for someone like me—allows record and interconnect cable changes without adjustment of volume. The unit mutes automatically for a couple of minutes when powered up. For best sonic results, Jadis recommends that the unit be left on continuously.

So where does the JPL's magic live? In my opinion, it's in the details: the power supply, the selection of passive parts, and the execution of the circuit. You meter-readers out there—you know who you are, you whose modus operandi can be summed up in the motto "parts is parts"—please take note: As H. A. Hartley put it many years ago, the sonic difference between a Stradivarius or an Amati and a mass-produced fiddle is literally in the stuff of which the instruments are made. Ditto for the difference between a Steinway and a Yamaha. It's not easy to measure sonic differences between violins or pianos, yet the musical ear has no problem at all in instantly resolving such differences. Build the same circuit with Radio Shack parts and with premium parts selected on the basis of active listening tests. Those who feel that the Radio Shack version would sound as good as or even better than the audiophile alternative are directed to read Ben Duncan's "Harmonic Convergence" article in the October 1992 Stereophile (p.78), where he discusses the measured results of just such an experiment.

Sonic impressions
The JPL spent its time exclusively in my reference room, where it was complemented by the Sound-Lab A-1 ESLs and a variety of power amplifiers, most notably the Air Tight ATM-3 (review forthcoming) and the Fourier Components Sans Pareil OTL monoblocks (reviewed in June '92). I tried CD program material with both the CD and line inputs. At least with the Theta DS Pre Generation III, I found the line input to give me a fuller palette of dynamic shadings, so I stuck with the line input for the duration of the evaluation. I also used the JPL in conjunction with the Threshold FET-10/e phono preamp for all of the analog listening sessions. The JPL did benefit from being left on continuously, particularly in terms of detailing and textural purity. Still, I'm a bit nervous about leaving tubes to cook indefinitely. Tubes are thermionic devices, depending for their operation on electron emission from a very hot cathode surface. The mere act of electron emission means slow but sure disintegration of the emissive surface. It pains me to think of all those premium tubes suffering so.

Memorable first impressions happen occasionally in this business, but nothing like this. It would not be an exaggeration to say that I had to pick my jaw up off the floor. I was immediately and overwhelmingly won over by the Jadis.

I've heard the occasional Nagra product at audio shows, but infrequently. When I have, I've been in awe of the gorgeous but understated quality of their handmade construction. And the rooms containing Nagra gear always produced superb sound.

My fondest memory of Nagra gear dates back almost a decade. I'd booked a gig for the John Atkinson Trio to perform jazz standards and originals at my local country club in New York City. Our usual drummer, Allen Perkins, of tonearm/turntable maker Immedia, was on the West Coast, so I had to find a drummer to replace him. I called on Mark Flynn, who's now the drummer of my jazz quartet Attention Screen, to sit in. Although Mark and I had long been friends, I'd never heard him play, and took the chance based on his reputation.

It was a warm day, so we set up under an awning next to the outdoor pool. JA asked if it was okay to record the gig; I said, "Sure, why not?" These were the days before the proliferation of Zoom portable digital recorders, and JA had shown up with his Nagra-D open-reel deck and a pair of $2000 DPA microphones. The gig went well and the recording sounded great—but we'd set up midway between the pool and the bar's takeout window. As we played, we were besieged by tweens running from pool to window and back as they loaded up on liquid refreshment, their bodies splashing chlorinated water within inches of the Nagra. I was a nervous wreck for the entire gig.

So when the opportunity arose to hear in my reference system Nagra's Jazz tubed line stage ($12,250), I was, well, jazzed. I promised Nagra I'd keep their baby dry at all times.

Design
As soon as I opened the box, I knew I was in for an interesting experience. The packaging was immaculate, with inner and outer boxes separated by triangular Styrofoam supports. Inside the inner box were the preamp and its ACPS II outboard power supply, each in its own silky sack tied shut with a fabric sash. This packaging made Tiffany seem like Walmart. Although the Jazz and the ACPS II are unbelievably small (respectively, the size of a telephone answering machine and a laptop battery), the preamp's construction is rugged, with a faceplate machined from a solid billet of aluminum. The look, size, and shape are cosmetically identical to those of Nagra's CD players, which are all derived from the company's legendary IV-S portable open-reel deck. I was tempted to pop off the top to see what was inside, but refrained—I didn't want any of my tools to go near the Jazz's gorgeous case.

If I had wandered inside, I would have discovered a 12AX7 dual-triode tube, one per channel configured in differential topology, followed by a gain stage using a 12AT7 tube. The gain can be set to 0dB or 12dB. I tried both and heard no difference. I used the 0dB setting 80% of the time, the +12dB setting the remaining 20%. The Jazz has three groups of four-layered circuit boards of military-grade epoxy glass, gold-plated and festooned with polypropylene capacitors. There circuit boards also a have ground plane, intended to screen out disturbances and radiation, and to stop static loops that might give rise to hum. (Nagra claims an A-weighted signal/noise ratio, ref. 1V, of 105dB.) The ribbon cables linking the boards are as short as possible. The main board is mounted on elastomer blocks to filter out vibrations. The Jazz is also available with the optional VSP, a shock-absorbing double-plate isolation platform ($1950) made of solid aluminum with Alpha-gel feet.

The Jazz's suite of inputs consists of one pair balanced (if the optional input transformers are fitted, otherwise this input is single-ended), four pairs single-ended, and a bypass that routes the inputs directly to the output; the outputs comprise two pairs single-ended and one pair balanced if the optimal output transformers are fitted. On the front panel is an incredibly sexy-looking Modulometer—basically, an output-level meter derived from Nagra's original 1952 tape recorder, the Nagra II, with a red needle for the right channel and a green needle for the left. The modulometer has six levels of illumination (it can also be turned off); I heard no sonic differences among the various levels. The motorized volume, balance, mute, and input selector controls are also accessible via the compact, minimalist, and ergonomically robust remote control. The only incongruity among the otherwise sexy and Swiss-precise front-panel controls is the recessed volume knob, which resembles the heater control on the dashboard of a 1985 Honda Accord.

Listening
As much of my listening for this review occurred during December, I spun some of my favorite Christmas CDs. A Jolly Christmas from Frank Sinatra (CD, Capitol CDP 7 48329 2) is from Sinatra's best period, in my view: the albums he recorded in the 1950s for Capitol. The Nagra's ability to develop an organic and voluptuous uncolored midrange made it a natural match for well-recorded voices. Sinatra's voice in "The First Noel" was warm, open, rich, and vibrant. Although I thought his voice sounded quite natural, the Nagra demonstrated its ability to clearly delineate differences between recordings when I moved to The Doris Day Christmas Album (CD, Columbia 513539 2), a sonic masterpiece. Day's voice in "Silver Bells" was larger than life and startlingly realistic—much more so than Sinatra's. The entire recording was enveloped in an airy, silky sheen of ambience. It was, however, easy to tell that the reverb had been poured on a bit excessively for the Day recording.

The big splash about the VTL gear demonstrated in Chicago almost 30 years ago was that while I'd expected to hear the silky, luscious liquidity I normally associate with tubed components, these had amazingly low coloration and deep, solid, tight, forceful bass. These characteristics aren't uncommon in today's tubed products, but were very unusual in 1986, even in the most expensive models. Since then I've attended many more audio shows, and it seems to me that, of those manufacturers that don't make tubed gear themselves, more demo with VTL amps than with any other brand of tubed electronics. VTL must be doing something right.

The TL-5.5 Series II Signature preamplifier is VTL's most expensive all-tube preamplifier ($7000, or $9500 with integral tubed phono stage), and it's the latest version of the original TL-5.5, which Chip Stern reviewed for Stereophile in November 2002.

Design
The TL-5.5 Series II Signature retains the basic all-tube circuit topology of the original TL-5.5 with a lower-gain, high-current 12AU7 tube circuit, a 12AT7 tube buffer, minimal negative feedback, and a low-impedance output stage. The TL-5.5 II also features a new, precision-regulated power supply trickled down from VTL's flagship TL-7.5 Reference hybrid preamplifier. Moreover, as the original TL-5.5 used a standard resistive potentiometer as its volume control, the Signature II has a 117-step, chip-based differential volume control, with a 3-digit display.

The new version of the TL-5.5 operates in fully balanced mode from input to output. The TL-5.5 II has two pairs of inputs that can be operated in either balanced mode (XLR) or single-ended mode (RCA), as well as six additional single-ended RCA inputs. There are three pairs of outputs, one each of balanced (XLR), single-ended (RCA), and single-ended Record (RCA). The preamp can be switched between Normal and Low gain (I listened at the Low setting), and has two programmable trigger outputs. Also included is an RS-232 connector, to enable the preamp to be remotely controlled.

The TL-5.5 II's own remote has more functions than any non–home-theater remote I've seen: Power, Source Select, Volume, Mute, Balance, and Invert Phase. All of these are also accessible on the attractive front panel, which displays the volume level in increments of 1dB. The optional phono stage is a hybrid design (parallel JFETs and 12AU7 tubes) originally developed for the more expensive TL-6.5 preamplifier. The sample I received didn't include the phono stage; my auditioning comments here refer to only the line stage. For LP playback, I used the Vendetta phono stage.

Listening
The VTL's midrange—voluptuously rich, detailed, and liquid—made it a natural match for well-recorded voices. My favorite Beatles recording of solo voice, musically and sonically, is John Lennon's soulful rendition of Arthur Alexander's "Anna (Go to Him)," from Please Please Me (CD, Parlophone). The TL-5.5 II brought out every low-level dynamic nuance in Lennon's vocal style; even his raspy upper register was bathed in a golden glow.

The VTL's midrange, however, was revealing and uncolored enough to differentiate among the sound qualities of various recordings. Jennifer Kimball's forceful yet delicate alto on Oh Hear Us (CD, Epoisse 786851 1092 29) was much more realistic and involving than the otherwise excellent Carole King on The Living Room Tour (CD, Rockingdale rcd2-6200-2), which sounded relatively more drab. But both recordings were more enjoyable than the slightly edgy-sounding Sarah Vaughan on The Divine Sarah (CD, Bluenite Records BN050).

But the best vocal recording wasn't a vocal recording at all. Olivier Messiaen was a fan of the Ondes Martenot, an electronic synthesizer from the early 20th century that has a cameo role in his Turangalîla Symphony. In his Fête des Belles Eaux, Messiaen puts the instrument in a solo role within a chamber sextet. On the LP (Erato LDS 3202), this unique instrument sounds like a cross between a woman's voice and the purest woodwind you've ever heard. Through the VTL, the phrasing, delicacy, and dynamics of the Ondes Martenot were mesmerizing.

The integration of the VTL's midrange with its delicate and uncolored highs made it a good match for well-recorded guitars. Bill Frisell plays a wide range of acoustic and electric instruments on his solo album Ghost Town (CD, Nonesuch 79583-2), and through the VTL each had a perfect integration of mid- and high-frequency timbres, for a rich, airy sound.

His style of hitchhiking is simple: Step into the path of a moving car and see what happens. Until recently, my own response was one of avoidance and acceleration—to flee, as it were, a shadow (which, as already mentioned, I once suspected him of being). But on a morning last month, as I returned from driving my daughter to her part-time job, I decided it was no longer sensible or even acceptable for me to avoid him. I pulled over, rolled down the passenger-side window, and asked where he was headed. The Old Soldier, who that day looked about 70, replied by opening the door, sitting down, and fastening his seatbelt without prompting—the last an unexpected concession to life in the 21st century.

We exchanged first names and shook hands, and I carried on driving. My passenger asked what I do for a living, and I told him that I write magazine articles. I asked what he does for a living, and he replied, "I'm a collector. I collect old military uniforms and 45rpm records." So much for the mystery of the Old Soldier's anachronistic appearance. And as for the man's interest in records, he spoke for a few minutes, at my urging, about the music he enjoys ("Country and western. The good stuff. Not that new shit. Jesus.") and the reasons for our shared love of vinyl ("I don't have a CD player. Don't want one, either.").

I admit, with shame, that my materialistic streak often surfaces when I meet someone—especially an older someone—who has in abundance those things I covet: vintage guitars, vintage audio gear, vintage books, vintage fly rods, and, of course, vinyl. I've resolved, in recent years, to try harder to banish from my mind all thoughts of material goods when meeting such folks, and to simply enjoy their company; unfortunately, like my record collection itself, my personality is a work in progress.

But on this late-summer morning I scarcely had to try at all: I have no interest whatsoever in 45s. In 59 years of life, 49 of which have been spent as a vinyl buyer, I have accumulated no more than two dozen of the things, and while I appreciate a good one when I hear it, the 7" record is forever in my growing blind spot.

Or so I thought before Tuesday.

Another veteran of vinyl pursuits—an older man whom I've known, through other friends, for about a dozen years—got in touch to say he planned to vacate his Manhattan apartment, and wondered if I could help him dispose of his collection of approximately 1000 classical LPs. The man was convinced he'd have to pay someone to haul away the 20 cartons of vinyl now scattered throughout his tiny home—an impression that, I'm sorry to say, at least one less-than-scrupulous NYC-area record dealer did his best to encourage. I offered to drive down and buy them for as much as I could afford to pay. A deal was struck, and, to the best of my recollection, as I was running out the door, I whispered to Janet that there might soon be 20 more cartons of LPs in our living room. Honest.

The number proved to be fewer than 20: There wasn't enough room in my Volkswagen Tiguan for the last four cartons, which now await my return before winter starts in earnest. And the media proved to be more than just LPs: I spied in one of the cartons a small pile of 78s and, surprisingly, about a dozen 45s. Perhaps, like many record collectors who were active in the 1960s, this gentleman bought the occasional single by the Beatles or the Beach Boys, just for the fun of it? That was the first carton I emptied on my return, and I was surprised to see, at the top of the pile of 45s, the rounded-triangle logo used by Decca during their golden era of Full Frequency Stereophonic Sound (ffss) records. Good grief: I was looking at a 7", 45rpm record of classical music!

The record, which bears the catalog number SEC 5045, is a perfect miniature of an English Decca LP from the early 1960s, right down to the fold-over sleeve and the back-cover ad for the Decca ffss Stereophonic Pick-Up and Arm ("a triumph of British engineering and craftsmanship"!). The music engraved on its two short sides is J.S. Bach's Brandenburg Concerto 2, with Karl Münchinger and the Stuttgart Chamber Orchestra—a recording that also appeared on a Decca LP (SXL 2126), paired with the rather longer Concerto 5. Unable to contain my curiosity a moment longer, I fired up the hi-fi and was treated to some of the most sublimely clear, colorful, and tactile sound in the Decca catalog. And that's saying something; in addition to showcasing an impressive roster of world-class performing artists, Decca's stereo classical recordings from 1957 through 1967 are simply the best such things ever made, surpassing by a slight margin even the superb products of EMI, and leaving everyone else in the dust (a fitting metaphor in the case of other highly touted records that, to my ears, sound somewhat grungy by comparison).

Having read that Decca stereo LPs of that era were, as a matter of course, mastered at half speed, I wondered if the same might be true of this 45rpm microgroove single. I also wondered if there were any more at home like her. I searched for other numbers in the series, and learned at once that Decca SEC 5044 is 45rpm record of Ravel's Boléro—which, like my Bach single, has as its cover art a photograph of a small bit of pottery, under which appears the series title Classical Miniatures. Other numbers brought up other titles, also containing excerpts from various LPs from Decca's hallowed SXL series—somewhat more expensive things, by comparison. It may be time to expand my collection in a new direction.

Incidentally, my new pile of 45s includes classical titles from other labels, including Columbia, RCA (Jascha Heifetz on the same shade of red vinyl as the 1971 Nazz album Nazz Nazz), Royale (an American label that seems to have specialized in European recordings, also on red vinyl), Disques de l'Oiselet (a French label whose logo suggests a connection with the EMI offshoot l'Oiseau-Lyre), and Cantate, a German label with an apparent passion for publishing the religious works of Heinrich Schütz on 7", 45rpm records. (I can just imagine the parties . . . )

Coffman Labs G-1A preamplifier
Back in the 1980s, when I landed my first job in the audio-industry press, I was surprised to learn how many well-known high-end companies farm out their assembly work to manufacturers of medical equipment. Surprised, but not horrified: After all, if a factory can produce a functioning CT scanner (average price: $200,000), or even a no-frills liposuction aspirator (average price: $2000), how hard could it be to turn out a halfway-decent CD player, or a turntable with a vacuum platter?

I'm reminded of that time-honored relationship by Coffman Labs, a new company (footnote 1) whose founder, Damon Coffman, has served as both CEO and CTO for a prominent West Coast manufacturer of medical electronics. More interesting is the means by which Coffman earned a living while doing his graduate work in physics: as a concert violinist. It's as if an aspiring maker of frankfurters had experience in the interrelated disciplines of the culinary arts, animal husbandry, and highway maintenance: If there's a more perfect background for an audio manufacturer, I can't imagine what it might be.

Coffman Labs' first product is the G-1A preamplifier ($5495), a line stage plus phono stage that uses seven vacuum tubes: a 5AR4 full-wave rectifier plus two each of the 12AX7, 12AU7, and 5687 dual triodes. Because much of his debut product's circuitry is built around vintage parts—including old-style, metal-encased oil capacitors and a new-old-stock (NOS) rotary switch from eastern Europe—Coffman is following Shindo Laboratory's lead in offering the G-1A as a limited-edition product (footnote 2).

Although it lacks a balance control and a mono switch, the Coffman G-1A tempts the hobbyist with other amenities—and it is nothing if not headphone friendly. A rotary output switch allows the user to choose between double pairs of rear-mounted output jacks and the G-1A's front-mounted, ¼" headphone jack, the latter with settings for headphone loads of greater than 100 and less than 75 ohms, and loads that are best driven directly by the preamp's line stage. The output switch also has a position labeled Out, which disconnects all outputs and, at the same time, shorts whichever input is currently in use. (When desired, all inputs can be shunted to ground with a separate Mute control.)

Footnote 2: Coffman Labs and their sole distributor, Echo Audio, say that production of the G-1A will be limited to 500 units—which would, in fact, be a very large run for some companies that don't limit their production. My review sample was No.34.

The Leto's four tubes are mounted on an exposed black platform backed by an S-curved rear wall of mirror-finish stainless steel that reflects the tubes' golden glow. Below the platform is an anodized aluminum silver front panel with a curved top on which are three knobs, for volume, balance, and Input selection. At the left end of the front panel floats a small, rounded triangle of silver that displays the Zesto logo and the model name. Around the corner, discreetly tucked away on the left side panel, is the On/Off switch. The Leto stands on Isonode anti-vibration feet.

The Leto's sexy, curved, understated, almost feminine looks remind me of architect Eero Saarinen's 1962 design for the TWA Terminal at New York's JFK airport. Without exception, everyone who entered my listening room said, "That's gorgeous! What is it?"

But, of course, it's the sound that matters. When the Leto arrived, I itched to put it through its paces.

The Building of a Design
The Leto's beautiful countenance is the brainchild of Carolyn Counnas—her husband, co-founder George Counnas, takes care of Zesto's electronics side. I asked George to sum up his design philosophy for the Leto: "A preamp should be as neutral as possible and not add any components that could compromise the signal. For example, there are no transistors anywhere in the signal path. All the switching is done through reed relays."

The word that best describes the Leto's design is simplicity. The signal for each channel travels through one 12AX7 input tube, one 12AU7 driver tube, 1% metal-film resistors, and polypropylene capacitors. The Leto has three single-ended inputs and two true. balanced transformer-coupled inputs, as well as two single-ended outputs and two true transformer-coupled balanced outputs. Linked to one of the single-ended inputs is a pair of cinema-bypass outputs, to enable the Leto to be connected to a surround-sound processor. (The bypass function works without the Leto having to be turned on.) The remote control is the most minimalist I've ever encountered: it has four buttons, for Volume Up/Down, Mute, and Mono. As the Leto defaults to Mute when turned on, there's a remote bypass switch on the rear panel that enables the Mute function to be bypassed in the event the remote's batteries need replacing or the remote is lost. Finally, there are left and right ground-lift switches, which adjust the grounding of the single-ended outputs.

An internal switch enables the user to switch between 12 and 3dB of gain. The factory default setting is 12dB; the 3dB setting can be used if the normal operating volume falls below the volume dial's 10–12 o'clock range. I used the Leto with its gain set to 12dB. I used its single-ended inputs for my phono stage, CD player, and Blu-ray player, and its balanced outputs for my Audio Research Reference 75 power amp.

Sound Architecture
Some aspects of the Zesto Leto's sound reminded me very much of what I liked about the considerably more expensive Nagra Jazz line stage ($12,250), which I reviewed in April 2013. I would have liked to have directly compared the Leto and Jazz, but the Nagra had vacated my listening room nine months before; I had to rely on aural memory. One word that appeared many times in the notes I took while listening to the Leto also applies to my recollections of the Nagra's sound: delicacy. In the Bill Evans Trio's Sunday at the Village Vanguard (CD, JVC JVCXR-0051-2), drummer Paul Motian's subtle cymbal work was detailed but involving, and didn't distract from the others' playing. And throughout his piano's registers, Evans's piano sounded open, coherent, relaxed, and natural. As I mined my collection of jazz piano recordings, another word that popped up in my notes was liquidity. In the Tord Gustavsen Trio's Changing Places (CD, ECM 1834), the piano's airy, holographic image retained the warm flowing quality of a live performance, whether Gustavsen was playing subtly or during his more intense solos.

With well-recorded acoustic music, the Leto was also able to reproduce layers of room ambience. In the title track of David Chesky's Jazz in the New Harmonic (CD, Chesky JD358), the sound of Brooklyn's Hirsch Center, which engineer Nicholas Prout has perfectly captured, was reproduced in refined detail by the Leto—I felt transported to the hall. The retrieval of ambience went hand in hand with the Leto's ability to resolve detail, even in recordings of difficult, densely orchestrated works. During Knee 1, from Philip Glass's Einstein on the Beach, in the second recording of the work by Michael Riesman and the Philip Glass Ensemble (CD, Elektra Nonesuch 79323-2), the opening passage features churning woodwinds, electronic keyboards, and singers executing a series of rapid, dizzyingly difficult contrapuntal passages. Through the Zesto, not only was every passage reproduced cleanly and without blurring, but I could hear the air between the notes.

The Leto's flawless lower midrange made me want to listen to rich, breathy tenor saxophones, from George Coleman's work on Miles Davis's My Funny Valentine (CD, CK 93593 Columbia/Legacy), to John Coltrane's readings of standards on Stardust (CD, Prestige PRCD-30168). Regarding Coltrane's playing on this disc, my notes read "wide, deep, and round." Further up in the midrange, the Zesto's reproduction of women's voices was holographic and tactile; it unraveled individual phrasings, from Janis Ian's husky breathiness on Breaking Silence (CD, Morgan Creek/Analogue Productions CAPP-027) to Alison Krauss's ethereal, angelic vocalizing from her and Union Station's Live (CD, Rounder 11661-0515-2).

The Leto's pristine, uncolored, extended reproduction of the high frequencies also reminded me of the Nagra Jazz's similar qualities. The massed strings in Prokofiev's Romeo and Juliet, with Stanislaw Skrowaczewski conducting the Minneapolis Symphony (CD, Mercury Living Presence 432 004-2), can easily sound strident or blurred through lesser preamps. Not so with the Zesto—the violins didn't lose a jot of definition during the more heavily modulated passages, yet remained silky and airy. And in the more sparsely orchestrated sections of Stravinsky's Pétrouchka, with Antál Dorati conducting the same orchestra (CD, Mercury Living Presence 431 331-2), the biting, holographically imaged solo trumpets seemed to pop out of thin air toward the rear wall of a wide, deep stage.

Percussion and plucked strings were startlingly lifelike through the Leto. Listening to Tchaikovsky's The Nutcracker, with Dorati this time conducting the London Symphony (CD, Mercury Living Presence 432 751-2), I found myself focusing on the sparkling highs of the harp and triangle in the ballet's less densely orchestrated sections. And Daniel Druckmann's battery of percussion in George Crumb's Spanish Songbook 1: The Ghosts of Alhambra (CD, Bridge 9335) shimmered with perfect definition and extended upper harmonics in the recording space.

Woodwinds in jazz recordings also impressed me—I focused on the upper register of Daniel Carter's airy but metallic flute in "X-ray," from Matthew Shipp's Nu-Bop (CD, Thirsty Ear 57114.2). The preamp's superb definition of plucked strings extended to well-recorded solo jazz guitar, from Bill Frisell's warm, involving solo take on John McLaughlin's "Follow Your Heart," from Ghost Town (CD, Nonesuch 79583-2), to the sparkling, ringing upper harmonics of Derek Bailey's unorthodox guitar techniques on his Improvisation (CD, Ampersand Ampere 2).