Conrad-Johnson CAV50 integrated amplifier

The search for signal transparency has led to much experiment and debate concerning losses in fidelity that can be traced to the preamplifier or—as it's more often and awkwardly called these days when the phono stage is omitted—the "line controller."

Few preamps these days are burdened by analog disc inputs, and in consequence should logically be either cheaper or better, or both. While in general this is true, some losses in signal quality are inevitable when using a preamp. These losses result from the necessary cable runs, the plugs and sockets, the input selector, the volume control (however well executed), and the active amplifying device.

Some hair-shirt enthusiasts have gone all the way, attempting to eliminate the line controller altogether. I say "hair-shirt" because, on the face of it, audiophiles require the facilities a line controller offers: control of system volume and input selection. Some enthusiasts even resort to changing signal source by manually switching out interconnect cables: sources are chosen with sufficient power and voltage to drive the power amplifier directly. This part of the chain is completed, albeit inflexibly, by a hand-soldered pair of high-quality resistors per channel, these serving as a system's ultimate, fixed-level volume control.

There are two practical alternatives. One sidesteps the problem of input selection and lets the aspiring audiophile concentrate on his main source of audio replay, usually digital: an input-versatile CD player or DAC. When these include their own volume controls, either in the digital (eg, Wadia 850 and 860, dCS Elgar, Meridian 518) or the analog (Krell KPS-20i/l) domain, their audio outputs can be directly linked to the power amplifier, eliminating the controller and its cabling.

There are unmistakable gains in fidelity with such systems, and it takes a truly exceptional combination of separate line controller and interconnect cable to compete. On the other hand, sensible choices have to be made when analog disc is a key component of a system; this source, in addition to amplification and RIAA equalization, generally requires a form of line-level control.

An intermediate strategy is to use a passive controller. This may well look and act like a preamplifier, but it lacks an active gain stage or unity-gain impedance-matching buffer. Though such devices are highly demanding of source and load matching on grounds of both impedance and signal voltage, it is nevertheless possible to achieve very good sound in compatible audio chains, to the serious embarrassment of many costly active-control units.

The second alternative has become popular over the past decade and is frequently seen in moderately priced audiophile-grade integrated amplifiers (footnote 1). The original practice of integrating a preamplifier and small power amplifier in one casing, generally for reasons of economy, is modified by removing the active part of the onboard line preamplifier, so shortening and simplifying the signal path. Simply, the power amplifier gain is raised to increase the input sensitivity from (typically) a volt or so to a quarter of a volt, accommodating a standard range of audio sources. All you need is a volume control and input selector and the amplifier is complete. The important difference between this approach and the external passive control concept is merging the two sections into one unit solves the impedance-matching problems. It also avoids the expense and quality of the additional cables required.

High-end audio companies are beginning to embrace this concept, as in the case of Conrad-Johnson's new CAV50 integrated tube amplifier. Levelheaded, conservative, and generally evolutionary rather than revolutionary in its engineering, Conrad-Johnson sets enduring standards for build quality and performance—the more successful C-J models have remained in production for years on end. Take, for example, the long run they've enjoyed with a generic 50Wpc push-pull ultralinear power amplifier design. It commenced in the early '80s with the MV45, which developed first into the MV50, then the '52. Then came the Premier Eleven and the Eleven A (which were actually based on an entirely new circuit). C-J's experience with the Premier led to the addition of the improved, moderately priced MV55 (a Sam Tellig favorite), which is frequently partnered by the PV10A/L line controller at a combined price of $3290.

Enter the CAV50
Now enters the long-awaited CAV50—to all intents and purposes, the PV10/MV55 combination in a single chassis and competitively priced at $2495. C-J calls it, somewhat portentously, a "vacuum-tube control amplifier." The usual output-stage operating mode is ultralinear push-pull, the tetrode-pentode tubes providing the best technical compromise of output power efficiency, operating life, and distortion.

This design may also be configured for triode operation if desired, at a loss of 50% of the (3dB) maximum output power. The ultralinear mode is rated at 45W into an 8 ohm load for 1% distortion over a tolerably good 30Hz-15kHz full-power frequency range; the triode setting gives about half that, 22Wpc, with a slight increase in distortion to 1.5%. While these distortion figures are higher than for many solid-state amplifiers, if the harmonic content is low-order (typical of this technology), then these levels are actually at or below the threshold of audibility.

Input impedance is high, rated at 50k ohms (an easy load for signal sources), while the sensitivity is 500mV for full power. This is not as high as some designs; I favor 250-300mV for full compatibility with some lower-output sources, though I doubt that this will lead to any compatibility problems.

As with the MV55, the output tubes are EL34s driven by an octal-based 6SN7 dual-triode driver. A close equivalent is a miniature all-glass 6FQ7, as used in the Premier Eleven A. But as these are now in short supply, the Sovtek 6SN7 is proving very serviceable in this application, and is both well priced and available. The input and splitter section uses the 12AX7Q or ECC83. A 12AU7 is used for the preamp output, one half for each channel.

A polarity-inverting amplified output is available with a gain of 10x, or 20dB, if the CAV50 is required to drive other components; eg, a second, lower-gain amplifier or subwoofer, or a tape recorder input.

A slant-fronted protective casing is provided, though I greatly prefer the appearance without, assuming it's safe (no small children or pets with wet noses). Fortunately, quick-release fixings are used for the tube cage.

Features include a tape-monitor button obliquely coded "epl" (signifying external processor loop), a true power on/off button, a simple rotary volume control, and an input-selector switch. You have the choice of Phono doubling as Aux 1 (for an external disc preamp equalizer), Tuner, CD, Video, and Aux 2. Relay-muted for surge protection, the CAV50 waits for 90 seconds before turning itself on. As in the other C-J power amps, output-tube bias worries are smoothed away by an LED indicator and bias adjuster for each tube, these easily set using an insulated driver (supplied by C-J). Matched tubes are not essential with this locally fixed-bias, tube-specific system.

Speaker outlets are gold-plated binding posts, insulated and color-coded. My UK sample suffered from some mild transformer hum on 50Hz supplies, while the US version is presumably silent on 60Hz. All low-level inputs are normal-quality, gold-plated phonos, while the power cord is a captive three-core cable.

Footnote 1: I believe this concept was first commercially realized in the Acurus DIA-100, favorably reviewed by Corey Greenberg in the November 1993 Stereophile (Vol.16 No.11).—John Atkinson
2733 Merilee Drive
Fairfax, VA 22031-4407
(703) 698-8581