Krell KPS-20i CD player Page 2

The DSP56002 is clocked at a blazing 66MHz, and has more computing horsepower than the DSP56001 used in previous Krell processors. Consequently, a single DSP56002 does the work of two DSP56001s. The digital filter section is located on the main printed circuit board, with the DAC and analog output stages mounted on a board stacked above the main pcb.

The digital-filter output is converted to an analog audio signal by four Burr-Brown PCM63 DAC chips. Four DACs are used for true differential operation: the digital signal is split into positive and negative phases in the digital domain, then converted separately by the four DACs (+L, –L, +R, –R). This method of using four DACs for balanced operation is much better (and more expensive) than simply putting a phase splitter in the analog output stage to create the balanced signal.

Unusually, Krell hand-trims the PCM63 for the best low-level linearity. Although the PCM63's "co-linear" architecture confers inherently good linearity without trimming, Krell says they get better sound from the DACs by trimming them. Current-to-voltage (I/V) conversion is handled by PMI 2131 op-amps. The gain of each 2131 is hand-set at the factory by adjusting trim pots, ensuring exact gain-matching between the two halves of the balanced signal. I haven't seen this technique used in any other processor.

The analog output stage is a refinement of Krell's previous analog stages—pure class-A, fully discrete, and all direct-coupled. Note that four analog output stages are required for the KPS-20i's fully balanced topology. In addition, the single-ended output signal is created from the balanced signal with a discrete differential amplifier, which converts a balanced signal into an unbalanced signal. Consequently, users of the unbalanced outputs still get the benefit of the balanced DACs. The other technique of creating a single-ended analog signal from a balanced signal simply connects the positive phase of the balanced signal to the center pin of the RCA output jack, obviating any advantage of balanced DACs. Krell did it right in the KPS-20i, both in making the circuit fully balanced, and by giving single-ended users the benefit of the balanced digital topology.

The transport section of the KPS-20i is based on a Philips CDM-9 Pro mechanism. The laser-pickup assembly is mounted in a brass casting weighing 7 lbs. This casting is machined to accept a mechanical isolation mounting that holds the laser pickup, spindle, disc, and disc clamp. This assembly sits on additional isolation mounts attached to the chassis bottom. The entire chassis is further isolated by the unit's external supporting feet, which are designed neither to store energy nor transmit it to the transport mechanism. The KPS-20i has the most extensive vibration-isolation system of any Krell transport.

Krell throws out the spindle (the small turntable on which the CD rests) supplied with the CDM-9 Pro and replaces it with a custom-machined type. Krell has found that they can reduce the tracking servo activity by reducing the effective disc eccentricity (out of round) with their custom spindle. Each spindle is installed with the aid of a dial indicator to assure accurate positioning. Krell claims a twofold reduction in disc eccentricity with their spindle compared to the stock Philips unit.

A unique star-shaped clamp, which extends to the CD's outer edge yet has very low mass, holds the CD to the spindle magnetically. The large, lightweight clamp is a departure from the small, heavy clamp used in previous Krell transports.

The CD and transport mechanism are bathed in the light of green LEDs, which you can see when you open the motorized door which covers the disc-playing chamber. Krell claims the presence of green light helps the laser recover the data from the CD, and improves the sound quality (footnote 1). Further, the disc-playing chamber is sealed from outside light by the motorized sliding door.

The transport, display, and other functions are controlled by a Motorola microprocessor running Krell's own code. The circuit boards are all four-layer and densely packed. Passive parts quality is very high, with 1%-tolerance metal-film resistors and high-quality film bypass caps used in the power supply.

The KPS-20i's build quality is a work of art in itself. The chassis is made from very thick extruded aluminum that gives the KPS-20i its very solid look and feel. From the mounting feet to the motorized top cover, the KPS-20i exudes solidity.

On the down side, I found the transport control buttons small and difficult to use, particularly in the dark. Putting the clamp over the disc correctly also took a bit of getting used to. The owner's manual could have been better written, and would benefit from larger and clearer call-out diagrams. I also found that, when entering a track number directly, I had to make sure I released the button immediately. Otherwise, the unit would mistakenly search track 22 rather than 2, for example. Finally, the first review sample had no audio output—a relay problem that was not present in the second sample.

I went into the auditioning expecting the KPS-20i to sound like a lower-quality version of Krell's $14,000 Reference 64 processor and $7900 DT-10 transport (see my reviews of both units in January 1994, Vol.17 No.1). Although I liked the Reference 64 and DT-10, I thought they were eclipsed in sound quality by the similarly priced Mark Levinson Nos.30 and 31. To be frank, while I've found previous Krell Digital products to be excellent, they were not better than the very best digital-playback components.

Imagine my surprise, therefore, at discovering during the first evening's auditioning that the KPS-20i not only beat the performance I remember from the Reference 64 and DT-10, but also that this $9000 transport/processor combination held its own in many respects with the processors I hold in the highest regard.

One area in which the KPS-20i clearly beat the Mark Levinson No.30.5—and especially the Sonic Frontiers SFD-2 Mk.II—was bass reproduction. The Krell had unbelievably deep, tight, powerful, and authoritative bass. In fact, I would characterize the improvement in bass performance rendered by putting the KPS-20i in my system a transformation. The entire bottom end was big and powerful, making listening to music more of a physical, visceral experience. Bass drum had an extension and impact I hadn't heard before from the Genesis II.5 speakers. The bottom end seemed to extend another octave lower, and the power in the lower registers was awesome.

Even though the bass was powerful and forceful, I heard an amazing amount of articulation and finesse. More bass isn't necessarily better, but when combined with excellent agility, pitch resolution, and detail, the result is a greater involvement in the music. I found myself thoroughly enjoying the music through the KPS-20i in a new way—particularly music with strong rhythmic drive. One disc that serves as a reference for a component's ability to express pace and rhythm is Robben Ford's Robben Ford and the Blue Line (Stretch STD-1102, footnote 2). Some products tend to make the rhythm drag, as though the bass and drums are slightly out of time. But through the Krell, the rhythm locked-in with a snap and speed that made this music so much more involving. The KPS-20i's ability to convey musical energy, pace, and drive was extraordinary by any measure. I greatly enjoyed this aspect of the KPS-20i.

This remarkable bass performance was heard when playing CDs on the KPS-20i's transport. Driving the KPS-20i with an external source (the Mark Levinson No.31 transport) resulted in the bass becoming less tight and defined, and softer and a little bloated. This isn't knocking the No.31; it's a result of the jitter introduced by the necessary S/PDIF interface.

It wasn't just the bass performance that made the KPS-20i so rhythmically involving. The Krell had a transient crispness and snap that imparted a tight, rhythmically incisive quality to the music. Snare drum had a wonderful sense of pop and weight, with just the right balance between the drum and chains (footnote 3). The music had an overall rhythmic coherence and upbeat quality I found captivating.

The KPS-20i's overall tonal balance was warm and full, but slightly brighter than that of the Sonic Frontiers SFD-2 Mk.II and the Mark Levinson No.30.5. The KPS-20i had some midrange and lower-treble grain that was revealed in comparison with these other processors. The mids and treble were rather less liquid, smooth, and refined through the Krell. Consequently, instrumental textures were a little less palpable and realistic. I also heard a bit of an edge and hardness in the upper mids that was manifested as glare on horns, hardness on the transient attack of high piano notes, and a slightly synthetic quality to woodwinds. This character was heard more easily on naturally miked classical music than on jazz and blues. The treble also lacked the delicate, finely filigreed quality I've come to associate with the UltraAnalog DACs used in the No.30.5 and SFD-2 Mk.II.

I must stress that these characteristics were very slight, and only noticeable when compared to the smoothness of the SFD-2 Mk.II and HDCD-fitted No.30.5. I spent many enjoyable hours with the KPS-20i; it was only after analytical listening and direct comparisons to characterize the KPS-20i's sound that these traits became apparent.

The KPS-20i's soundstage was spacious, deep, and had excellent transparency. The overall spatial perspective was more up-front and incisive than that of the No.30.5, but a little less immediate than the SFD-2 Mk.II. Although the soundstage didn't have quite the degree of depth and width as the No.30.5 or SFD-2 Mk.II, I heard a remarkable clarity, and was able to hear deep into the soundstage. The KPS-20i was the antithesis of veiled, murky, or clouded. The Ry Cooder and V.N. Bhatt A Meeting by the River CD (Water Lily Acoustics WLA-CS-29-CD) was reproduced by the KPS-20i with a startling immediacy, clarity, and impression that the performers were in the listening room.

A comparison of the three processors when playing HDCD-encoded discs was no contest: the HDCD decoding provided by the No.30.5 and SFD-2 Mk.II elevated these two processors far above the KPS-20i. How significant a factor this should be when making a purchasing decision will be determined by how many record companies adopt HDCD encoding. It's just too soon to tell.

The KPS-20i's build quality is outstanding. The thick, beautiful metalwork, motorized disc-chamber door, transport-isolation mechanisms, and custom-machined disc clamp suggest a high degree of mechanical refinement. The KPS-20i has the look and feel of a product that will last forever.

The $9000 KPS-20i is the best-sounding digital product yet from Krell, in many respects offering world-class performance. I was particularly taken by the KPS-20i's powerful bass, stunning power and extension, wide dynamic contrast, and its physically involving sense of pace and rhythm. The KPS-20i's soundstaging was also superb, with a great feeling of transparency and clarity.

In relation to the much-more-expensive Mark Levinson No.30.5/31 combination, or the No.31 driving the Sonic Frontiers SFD-2 Mk.II, the KPS-20i had some shortcomings. Specifically, I heard some grain and hardness in the upper mids that affected instrumental timbre, and the treble somewhat lacked the smooth liquidity and refinement heard from these two other reference processors. I should add that the KPS-20i's bass, rhythmic coherence, and dynamic slam were a notch above the performance offered by the No.30.5 and especially the SFD-2 Mk.II. Note that the KPS-20i sounded considerably better when used as a CD player, and less good when fed from an external digital source.

Is it possible to call a $9000 CD player a bargain? Yes, when you consider how much it would cost to achieve the KPS-20i's sound quality with a separate processor and transport.

Footnote 1: I am not sure why this should be so. The infamous green ink treatment has been shown to absorb the infrared laser light, thus reducing common-mode light that may or may not affect the operation of a CD pickup's photodetector, which is usually a differentially operating device. I don't understood how bathing the disc in green light, which represents a strong common-mode signal, albeit one with a very different wavelength, will have an effect on the recovered FM data signal. Perhaps Krell would like to go into this subject in their Manufacturer's Comment. But please note that, as light rays do not interact with one another—a tenet of modern physics—the green LED light has no effect on the pickup laser's emission. It sure looks pretty, however.—John Atkinson

Footnote 2: This disc features the great Roscoe Beck on bass and Tom Brechtlein on drums. Their frequent live shows in Los Angeles are a real treat—sometimes with Vinnie Collaiuta on drums and Russell Ferrante on keyboards.—Robert Harley

Footnote 3: Snare drum is a challenge to record. Getting enough of the drum itself without ringing or fatness is difficult, as is making the snares or chains on the bottom skin cut through the sound without sounding like trashcan lids. A well-recorded snare drum should have an explosive pop, combined with some sense of the drum's pitch. Too many snare drums today sound like nothing more than bursts of white noise, with no indication that the sound was made by a stick hitting a drum head. My three favorite snare sounds all happen to be on the Sheffield label: James Newton Howard and Friends (Sheffield Lab 23—the direct-to-disc LP, in particular); Michael Ruff's Speaking in Melodies (Sheffield CD-35); and The Sheffield Drum Record (Sheffield Lab 14).—Robert Harley

Krell Industries
45 Connair Rd.
P.O. Box 0533
Orange, CT 06477-0533
(203) 799-9954