Switching amps and RF noise.

I read the special report in TAS 166 with close attention. Reducing radio frequency (RF) noise artifacts in my system has been my main concern for several years. I recently installed a pair of Gilmore Audio Raptor switching amps. The result is a level of performance beyond my expectations, which has allowed me to make substantial progress in identifying and controlling RF noise.

Up front, I should make it clear that I've known Mark Gilmore for 12 years and have a great deal of respect for him. The chance to audition the Raptor amps made me worry that I would not like them because of RF noise, and that I would have to tell this to Mark. When they arrived on short notice, I had to scramble to rewire my speakers (modified Magnepan MG-20) for single amp per channel operation, as they were previously bi-amped with OTL tube monoblocks and a stereo solid-state amp.

The first impression was positive. The Raptors were as clean out of the boxes as my home-made OTL monoblocks, but had much better control of the MG-20 driver resonances. The Raptor's rated output impedance is 50 milliohms (0.05 ohms), and I understand the actual impedance is even lower. Applying this level of control to the Magnepans improved the sound in ways I could not imagine. The bass got much better, of course, but the subtle nonuniformities in tonal response that I had lived with for a decade without noticing became obvious through their absence.

The TAS special report contains several descriptions of sonic problems with switching amps that mirror my experiences with RF noise. We have a digital video recorder/satellite receiver in another room that has always been the major source of RF noise in our house. Turning it off and unplugging it would cause a drastic improvement in the audio system performance, but that is impractical. My improvement efforts have been along the lines of treating the audio system and the AC power to the DVR to reduce the level of noise. Every effective step of treatment produces a reduction of treble glare, midrange chalkiness, sound-stage compression, etc., that is easy to detect with careful comparative listening.

Not only did the Raptor amps not make the RF noise situation worse, they provided a higher level of resolution that allowed me to focus on AC and interconnect cable problems with my source, a Wadia 861. The Raptors are much less sensitive to power cord issues than my previous Class A and A/B amps. I can't rule out some switching noise effects on the interconnect cables betweeen the Wadia and the Raptors, but it has been easy to bring these cables to a new level of quietness (I make my own cables).

It is interesting that most of the switching amps sampled by TAS use switching power supplies. This is reflected in the size and weight: switching power supplies are small and light, but are also very strong sources of RF noise. Keeping this noise out of the audio path is a significant technical challenge, and the simple way to avoid it is to use a linear supply. The Raptor amps use linear supplies for this reason. However, the reports on the Cary A 306 make it clear that this choice alone is not all that is required to get good performance.

In summary, the TAS special report and my experience suggest that switching amps can provide better performance than conventional linear amps as long as the rest of the audio system is selected or treated to be immune to external RF noise. Such systems will also be immune to residual switching noise from the amps.