Accessory Reviews

By Robert Deutsch

As far as I've been able to determine, the first Stereophile review of a power-line conditioner (PLC) appeared in the April 1988 issue: Lewis Lipnick's review of the Adcom ACE-115 power-line filter/spike suppressor. Since then, devices designed to improve AC power have become a major audiophile product category, and by 1994 they'd grown popular enough to be allotted a separate section in Stereophile's "Recommended Components."

In 1999, the PLC world was stood on its ear by the introduction of the PS Audio Power Plant (currently available in 300W and 600W models, with 1200W and 2000W models in the works). The Power Plant is not really a power-line conditioner but a power synthesizer—or, as PS Audio refers to it, an "AC regenerator." It's actually a class-AB power amplifier with a highly regulated, low-distortion AC sinewave output.

I put the P-300 through its paces in the December 1999 Stereophile, and was most impressed with the results. My enthusiasm was matched by that of John Atkinson (Whew!), who found that the P-300 allowed his $6495 Mark Levinson No.380S preamp to more closely approach the sound of the $14,950 Mark Levinson No.32.

A unique feature of the Power Plant is that the frequency of the sinewave output can be varied from 50Hz to 120Hz. The idea is that increasing the frequency from 60Hz (or 50Hz, where that is the standard) enhances the effectiveness of the power supplies of equipment connected to the Power Plant, recharging capacitors at a higher rate and allowing transformers to operate at a more optimal frequency.

Indeed, I found that increasing the AC frequency produced better dynamics and an improvement in overall clarity—but only up to a certain frequency. Above 90Hz, the sound seemed to thin out in the midbass, and I eventually settled on 80Hz as the optimal setting. This observation has been confirmed by other Power Plant users, most of whom run the Power Plant at 80 or 90Hz. McGowan's explanation for this effect is that frequencies above 90Hz are not a problem for transformers—in fact, their performance continues to improve. It's the inexpensive diodes commonly used in power-supply diode bridges that have trouble keeping up with these higher frequencies.

Fixing What Ain't Broke
The Power Plant was such a success in the marketplace that most designers would have been content to leave the design unaltered and devote their energies to developing new products. Paul McGowan has been working on new products, including amplifiers based on the Power Plant, but he also thought he could improve the performance of the Power Plant itself. The harmonic distortion of the AC sinewave produced by the Power Plant was about 0.05%—much better than the 3-10% distortion of the AC line—but McGowan wanted it to be even lower. He determined that most of the distortion was caused not by the amplifier but by the DSP-based oscillator generating the AC frequency. A new D/A converter on the oscillator board, using a higher sampling rate, resulted in cutting the distortion in half: down to 0.025%.

Then, he wanted to give consumers the option of tweaking the AC voltage. This would be useful to owners of electrostatic loudspeakers, which tend to perform better with the voltage increased to 120V. Varying the AC voltage is also helpful in situations where a transformer in a piece of equipment connected to the Power Plant exhibits mechanical hum. In many cases, the hum can be reduced by lowering the output voltage by a few volts from the normal 117V. Previously, there was an internal adjustment for this; on the new MultiWave-equipped P-300/P-600, voltage can be selected from the front panel.

Finally, McGowan wanted to solve the problem of thinning midbass with higher AC frequencies. Working with Doug Goldberg, he set out to find a way of getting the benefits of higher AC frequencies without any sonic or technical downside. Combining waves was considered, but it became apparent that combining, say, 60Hz and 120Hz would result in the production of a beat frequency, which would impair transformer performance. They came closer to a solution by having a sequence of waves of different frequencies; eg, 60Hz followed by 120Hz. Listening to the results confirmed that they were on the right track.