Krell KSL line preamplifier A Matter of Balance

Sidebar 1: A Matter of Balance

Much has been written about balanced operation, but few real facts have been established except for its frequent use in professional studio applications, where it was primarily developed for low-level microphone signals and long cable runs, sometimes as long as 100m. By using an intimately twisted pair of conductors (which offers good rejection of external electromagnetic interferences when driven in a push-pull balanced mode), then shielding the assembly, often with two successive dense-wire braids, external hum and noise fields could be suppressed. The zero voltage point or ground reference is implied and allowed to float separate from the ground line proper. This isolates the signal path from the ground line, which often carries stray currents as well as any induced electrical noise. Any electrical noise which does get into the signal path is present on both; this is the common mode.

At the receiving end of the cable, the +V and –V signals are differentially summed to produce a single voltage twice as big as either, while the noise signals present on both are supposedly identical and are therefore canceled out in the differential summing. This is termed common-mode rejection. Professional equipment often defines a figure for common-mode rejection at its input, a figure somewhat dependent on frequency according to the differential balance achieved over the frequency range. 60–80dB figures are possible with fine professional tolerancing for microphone circuits, while 40dB is generally considered sufficient for line-level interfaces.

Consumer audio-cable runs are generally short, electromagnetic interference levels are fairly low, while signal levels are high. The vast majority of interconnections are successfully made with a simple unbalanced single-ended connection. However, at the highest-quality level it is argued that balanced working can improve quality—at a moving-coil phono input, for example—by reducing induced hum to the vanishing point. Likewise, any reduction in hum and noise is generally no bad thing; balanced working cannot harm line-level links and may well be beneficial in difficult situations. One of these would involve poor house power-supply quality with a noisy harmonic content, and with the potential for chassis current to flow between interconnected pre- and power amps. Balanced working allows for a high rejection of these chassis currents which could well carry signal-modulated components—in plain words, potential distortion.

"Tacked-on" balanced circuitry may well reduce the performance of a given amplifier system. However, if properly designed from the outset, with true bi-phase differential inputs and symmetrical phase-matched bi-phase outputs, balanced system design and use has the ability to help achieve the maximum performance obtained from audio components. It is, however, no magic recipe for improved sound quality. This has to be right from the start.—Martin Colloms

Krell Industries, LLC
45 Connair Road
Orange, CT 06477-3650
(203) 799-9954

Axiom05's picture

Love these old reviews, brings back a lot of memories of reading Stereophile and trying to figure out how I could buy some of this stuff. It doesn't seem that long ago. Seems there were a lot more reviews per issue back then too.

jmsent's picture

would not be acceptable in a $500 preamp, much less one that sold for 2 grand. Odd that Krell didn't have a manufacturer's comment addressing this. High quality controls with near perfect tracking were certainly available in the early '90s. Most Japanese receivers did far better than this.

tonykaz's picture

You make a valid observation but it probably wouldn't bother me if I owned the piece, which, I did just look up on eBay to find one for under $1,000.

I was selling Audible Illusions PreAmps that had separate volume knobs for each channel, I felt it was a good design.

Krell is good gear, I could live with it.

Tony in Michigan