Thursday, May 1, 2008

Krell FBI integrated amplifier

Krell FBI integrated amplifier Krell FBI integrated amplifier Krell FBI integrated amplifier Krell FBI integrated amplifier Krell FBI integrated amplifier Krell FBI integrated amplifier Krell FBI integrated amplifier Krell FBI integrated amplifier Krell FBI integrated amplifier Krell FBI integrated amplifier HI-FI, Sterio, Home Theater, Audiophile, Amplifier, Speaker

There's something a bit oddball about the notion of a $16,500 integrated amplifier—until you stop to consider that the market is fairly drenched with preamps and power amps that, together, cost that much and more. And putting both pre- and power amp in a single chassis cuts down on storage (one less shelf), accessories (one less pair of cable), and electrical outlets (one socket freed up).

Still, it must be a tough sell; it seems a novelty, along the lines of the $100 Kobe beef burger or the $1000 caviar omelet offered at certain Manhattan restaurants. The kind of people who tend to stick with burgers and omelets think such prices are comically insane. The kind of people who don't chafe at these prices figure they might as well go all the way and order the dry-aged porterhouse or the blini and Beluga.

Maybe it's the categories—burger, omelet, integrated amplifier—that are off-putting—or, more to the point, the usual associations with these categories. I've never eaten Kobe beef, but some who have say that it tastes fabulous, as sumptuous as a good steak in its own way. Just because most burgers are diner food doesn't mean they have to be. Just because most integrated amps are designed as compromises doesn't mean they have to be, either.

So here we have Krell Industries, maker of no-holds-barred, power-pumping behemoth amps, brazenly treading this uncertain terrain with the FBI—a Fully Balanced Integrated (hence the initials) amplifier that, judging from its appearance and design (to say nothing of that $16,500 price tag), aims to give no ground to its separate-components brethren.

Description and Design
The FBI isn't Krell's first stab at an integrated amp. In the mid-1990s they put out the KAV-300i, which evolved a decade later into the KAV-400xi. Priced at $2350 and $2500, respectively, the KAVs were pitched to the entry level of the high-end market, a realm that until then Krell had largely bypassed. (The less expensive KAV-400xi was scaled down but not dumbed down; Wes Phillips, in the February 2005 Stereophile, called it his "favorite Krell"—by which he didn't mean the "best Krell," but still...)

The FBI is something else entirely. It weighs 104 lbs, and puts out 300Wpc (600Wpc into 4 ohms, 1200Wpc into 2 ohms) with vanishingly low distortion at frequencies ranging from subway rumble to dog whistle. The preamp and amp sections have separate circuits and separate toroidal transformers; the preamp's is rated at 25VA, the amp's at 3000VA with a capacitance of 40,800µF per channel. Each channel has 20 TO-3-cased output transistors and 10 TO-3-cased regulators.

The chassis—compact but beefy, an anodized aluminum faceplate covering massive heatsinks—looks like that of Krell's FPB-300cx power amplifier, with a huge, analog volume knob tacked on the front. This is no coincidence—the FBI is an FPB-300cx with a line-stage preamp built in.

Krell's cx series of amps, introduced in 2004, was a modification of the c series, which came to market in 2000 and featured two innovations: Krell Current Mode and Current Audio Signal Transmission (or CAST). The first involved circuit topologies that manipulated the audio signal in the current domain as opposed to the standard voltage domain. Dan D'Agostino, Krell's proprietor, hit on this idea while designing A/V processors. The high bandwidth of video signals forced him to work in the current domain (voltage-based circuits aren't optimal for video's high bandwidth), which led him to wonder if the current domain might expand the bandwidth of audio signals too—and it did. According to Jim Ludoviconi, Krell's technical manager, the audio bandwidth in the current domain exceeds that of the voltage domain by "an order of magnitude."

This discovery led to CAST, which used Krell Current Mode not only in the circuits within a component but also in the connection between components. This approach would have two advantages, in theory. First, there would be no need to convert the signal from current to voltage and back to current; it would flow as one continuous stream of current. Second, whereas signals in the voltage domain go from low to high impedance, signals in the current domain go from high to low impedance. As a result, factors that inevitably (but unpredictably) corrupt an audio signal in the voltage domain would be sharply reduced, if not eliminated. These factors include stray capacitance and inductance, which build up on circuit boards, and strange interactive effects caused by the impedance of cables connecting a preamp and power amp. (D'Agostino has reportedly sent a signal through 5000 feet of Krell's proprietary CAST cable with minimal loss or distortion.)

In the c and cx power amps, the CAST input was moot at the time unless you happened to have a Krell KCT line-stage amp, the only preamp with a CAST output. In the FBI, the line stage and power amp are in the same enclosure; the signal between them goes down a multipinned header from one circuit board to the other. The CAST circuitry is activated as a matter of course. If you go further, and hook the FBI up to a CD player with a CAST output (more about that later), the signal passes in a continuous, unaltered stream from source to speakers.

No surprise, the FBI's line stage is based on the KCT, though it also employs some of the circuitry in Krell's later Evolution 2 preamp. Especially notable here, according to Ludoviconi, is the Evolution's Current Mirror, which makes a "copy" of a signal in the input stage, in order to isolate it from the inevitably intruding noise from a high-bandwidth signal. The line stage, like the Evolution, also uses LEDs, which have a much higher tolerance than traditional diodes; as a result, Ludoviconi claims, circuits perform more predictably at the critical low-level stages.

The catch is that running in the current domain requires a lot of transistors—up to twice as many, watt for watt, as the voltage domain—which means that an amp like this must be heavier and more expensive, and run hotter, than it otherwise would.

A few convenience features: First, power is activated by a toggle switch on the back panel. Pushing the power button on the front panel turns the amp from stand-by to full-tilt; the warm-up time, before you hear its full sonic bloom, is negligible. Second, the wireless remote has a 12-volt trigger, which lets you control other components in the system, even non-Krell ones. (I was able to select tracks from a Simaudio CD player, for instance.) Third, a "Theater Throughput" circuit (which I never used) allows the signal from a surround-sound processor to pass through the FBI at unity gain.

I did all of my listening through Verity Audio's Parsifal Ovation loudspeakers and Nirvana speaker cable. I listened to LPs, CDs, and SACDs. For a digital source, I began with the Simaudio Moon CD5.3, which has only RCA outputs. After I got used to the FBI's sound (it didn't take long; I've heard many Krell amps over the years, including, not long ago, the FPB-400cx), I switched to Krell's Evolution 505 SACD/CD player (borrowed from Wes Phillips), which has RCA, balanced, and CAST outputs. I began with the RCAs, to introduce the new elements one at a time. (I spent a few months last year with Krell's SACD Standard player, which is sonically similar.) Then I moved briefly to the balanced outputs. With my Nirvana interconnects, I've never noticed much, if any, difference between balanced and unbalanced hookups (after having adjusted the volume knob to compensate for the 6dB gain); the same was true here.

Finally, I switched to the CAST outputs, using Krell's proprietary CAST cable. That's when everything locked into focus.

As noted, I spent a few months last year listening to Krell's FPB-400cx, the beefier version of the 300cx that serves as the FBI's foundation. Certain things about the FBI sounded familiar: the tight deep bass, thunderous dynamics, and snappy transients. But, even at the outset, when I was using the single-ended inputs, there was something new: greater detail, even delicacy, in the timbres and textures of instruments.

Long gone are the days when Krell electronics were marred by what I felt was a cold glare in the upper midrange. In the last five or six years, Krell amps have warmed up in the middle octaves while preserving the pizzazz. But the FBI boosts Krell's traditional strengths while diminishing the weaknesses, both to new levels, in my experience.

For instance, even with the warmer midrange, silky violins weren't quite a Krell trademark. Yet when I put on Michael Tilson Thomas and the SFSO's recording of Mahler's Symphony 9 (SACD/CD, San Francisco Symphony 821936-0007-2), or Andrew Manze and Rachel Podger and the Academy of Ancient Music performing the solo and double violin concertos of J.S. Bach (SACD/CD, Harmonia Mundi HMU 807155), I heard not just silky, but shiveringly silky violins.

Some amps achieve tonal beauty at the cost of transient detail, but the FBI delivered both. With the Bach recording, it revealed the subtlest quivers of vibrato, the slightest ebb and flow of rhythm and dynamics. With Keith Jarrett's Carnegie Hall Concert (CD, ECM 1989/90)—a concert I attended—the FBI captured not only the creamy percussiveness of the Steinway grand, but also the slightest hesitations of Jarrett's rubato and the lightest shadings of his pedal work.

I've always put a high premium on how an audio component handles microdynamic contrasts—the slight variations in loudness or softness when a singer stresses a note, a violinist bows a bit harder or softer, or a drummer hits a cymbal in some different way that's barely perceptible. The FBI handled these as naturally as any amp I've heard. This isn't merely the sort of detail that audiophiles like to show off; it's the sort of detail that reveals the rhythm and the soul of music—and the presence of a human being blowing, bowing, singing, pounding, or otherwise playing it.

On Miles Davis' Cookin' (SACD, Analogue Productions LAPJ 7094 SA), when the band breaks into a faster tempo on "My Funny Valentine," listen to the way drummer Philly Joe Jones lets up on the hi-hat cymbal after he taps it with his stick. The effect adds an extra layer of rhythm and cool that I hadn't heard through other amplifiers the previous hundred or so times I'd played this album. Meanwhile, Miles' trumpet sounded golden, and Paul Chambers' bass was tightly strung and woody.

Bass, of course, is Krell's longstanding strong point, but I hadn't before heard such musical detail within the bass. The London Symphony's recording of Górecki's Third Symphony, conducted by David Zinman (CD, Elektra/Nonesuch 79282-2), begins with a very low, rumbly bass line—the double basses state the melody—which on many stereos you can barely hear. Listening to this passage through various systems over the years, I'd thought I'd discovered everything there was to hear in it, but the FBI unveiled more: dynamic variations in the bowing, two contrapuntal bass lines about three minutes into the first movement that I'd never before noticed, and simply more clarity and focus in the bass section—and all other sections of the orchestra—throughout.

The FBI also unraveled subtle harmonic variations with ease. On jazz trumpeter Dave Douglas' Charms of the Night Sky (CD, Winter & Winter 910 015-2), when he plays in unison with Mark Feldman's violin and Guy Klucevsek's accordion, I could hear their distinctive harmonic overtones much more clearly than I had before; in other words, I could better distinguish the three instruments, tonally, spatially, and harmonically. Yet they didn't appear as some overly analytical Etch-a-Sketch; the overtones blended in the ambience, as I've heard them do in concert.

Voices were also uncannily clear. If you've ever found it difficult to decipher lyrics sung by Bob Dylan, Donald Fagen, Randy Newman, or Rickie Lee Jones, take a listen to their recordings through the Krell FBI; it should be a clarifying experience. Dylan's Blood on the Tracks (SACD, Columbia CH 90323) sounded so clean—his voice so naked and articulate, the guitar strums so propulsive, the bass line so distinct—I would have guessed the recording had been dramatically remastered, had I not heard this disc so many times before.

Lest I give the impression that Krell has crafted an instrument of mere litheness and delicacy—a chamber ensemble among amplifiers—I should emphasize that the FBI could also let loose and roar. At the crescendo in the first movement of Mahler's Ninth, the FBI passed along the full force of the San Francisco Symphony without breaking up, backing down, or clutching in any way. Everything remained clear and distinct at high and low decibels.

I see that I haven't written anything about how the FBI handled soundstaging, so I'll say this: The soundstage was as wide, deep, precisely imaged, and densely layered as the recording and the rest of my system allowed.

Was there anything wrong with the FBI? Well, the highest frequencies sounded a bit truncated—or, if not cut off, a little less transparent than all the other octaves. This shortfall wasn't immediately obvious; to the extent it was noticeable, it wasn't at all annoying—and I'm one who is annoyed by chopped-off highs. I'm scrounging here for flaws.

I do have one caveat: As suggested near the beginning of this review, the FBI sounded much better with Krell's CAST circuit fully activated—that is, when the music source was plugged into the amp's CAST inputs. Using the balanced or single-ended inputs, the FBI still sounded very good; I wouldn't disavow anything I've written about it so far, though I might tone down some of the accolades a bit. In an A/B comparison adjusted for volume differences, the balanced and RCA inputs (with Nirvana interconnects) were hard to tell apart. But in A/B comparisons with the CAST input, both the balanced and RCA inputs fell short. They didn't reveal quite the rhythmic agility, dynamic detail, or bass clarity of the CAST circuit: I didn't hear all of Andrew Manze's and Rachel Podger's subtle fingerwork in the Bach violin concertos, or the full force of Dylan's strumming, or the obvious distinction between violins and violas in the Górecki and Mahler recordings.

Then again, CAST doesn't work outright miracles. I compared the CD and 180gm vinyl pressings of Miles Davis' Live Around the World (Warner Bros. 46032-2 and 46032-1) and Donald Fagen's Morph the Cat (Reprise 49975-2 and 49975-1). My turntable (which, of course, has only single-ended outputs) was plugged into the FBI's RCA inputs. Yet in the case of both recordings, the LP-into-RCA was superior to the CD-into-CAST. The two formats were closer than I've heard in similar analog-vs-digital face-offs, but the LP's ambience was a bit airier, Miles' trumpet and Fagen's voice a bit breathier, the background instruments a bit more 3D.

This comparison pointed to two conclusions. First, very good analog still has it over very good digital, in certain respects. Second, CAST or no CAST, the Krell FBI is an excellent amplifier; the difference between its inputs is one of degrees of excellence.

I can't presume to know what an economist would call your "utility function." That is, I can't know whether or not a certain product at a certain price—say, $16,500 for the Krell FBI integrated amplifier—is worth your money, given your tastes and income. But I can say that you won't hear everything the FBI has to offer unless you also buy a CD player that has CAST outputs, and the only such player that currently exists, as far as I know, is the Krell Evolution 505, which retails for $10,000.

In for 16 grand, in for 26? If that's where you're sitting (and if it is, mazel tov), go for it. Alas, it's not where I sit. So I petition Dan D'Agostino, on behalf of my fellow well-heeled but not that well-heeled music-lovers and audiophiles: Please, sir, can you make a digital player with CAST outputs (or maybe some kind of CAST converter) for a bit less? That will make the full package of the FBI's magic tricks accessible to many more of us.

Related Links...