Solid State Power Amp Reviews

The kindly delivery guy hefted the original review sample of the Adcom GFA-7805 (serial no. FCS0044) out of its crate and placed it gently on the floor in front of my equipment rack—it was too tall and deep for the slot available. He left with my gracious thanks for having saved my back. I then transferred the RS Audio Solution multichannel interconnects and the five Goertz speaker cables from my Bel Canto eVo6 amp to the GFA-7805, plugged in the power cord, and fired it up. After a few tense seconds as the front-panel LEDs flickered (as is normal for the power-up sequence), they dimmed and the music poured forth.

That didn't last. The internal fan soon switched on and, one after another, the distortion LEDs lit up. Shortly thereafter, the thermal-protection LEDs glowed and the indicated channels shut down. Curiously, this happened in a predictable sequence: RF channel first, then center, then LS.

With three channels down, I shut off the amp and checked my connections. Everything seemed right. But from then on, each time I powered up, after five to ten minutes of listening the problem would repeat itself. Trying to avoid the conclusion that the amp was at fault, I switched around which amp channel was connected to which speaker cable and speaker. Whatever I did, the result was the same: the same three channels shut down in the same sequence. I concluded that the amp was faulty.

Adcom's Matt Lyons had to drive all the way to Connecticut to deliver the replacement GFA-7805 (serial no. FCS0000), which, he averred, had been thoroughly checked out. We hooked it up, turned it on, and...same problem. Lyons squinted at the AlphaCore cables and suggested that they might be at fault. He said that the GFA-7805 is a wideband amp and is definitely not rolled off anywhere in the vicinity of the audible range. Using "certain" cables—those that permit RF to get into the output stage—might cause the amp to oscillate and overheat, even in the absence of audible distortion. I replaced the stereo channel cables with Straight Wire Teflon-12 (kept from my GFA-555 II days), and all was well in those channels.

AlphaCore, on the other hand, acknowledges the possibility of such problems in their literature and on their website. They state that the resistive load of loudspeaker voice-coils can rise to 10 times the DC resistance at 20kHz, along with an even higher, almost purely inductive impedance. Unfortunately, the feedback loops in certain solid-state amplifiers does not satisfy the Nyquist or Bode criterion for stability. The amplifier might at times exhibit gain bandwidths in the internal feedback loop in the MHz order, meaning that instability can arise above the audible range. The problem arises when this type of amplifier is connected to speakers that exhibit high impedance at high frequencies via an otherwise ideal low-impedance matched cable. Under certain circumstances, instead of stabilizing the gain via the internal feedback loop, as is intended for a low-impedance output, the amplifier is turned into a high-frequency power oscillator operating at a frequency well above the audible range. This may load the amplifier excessively, "resulting in a higher-than-normal operating temperature or overload shutdown." Exactly!

Goertz's remedy is a Zobel network (R/C Match Links), now provided free of charge with Goertz cables, but not back when mine were delivered. A 0.1;uF capacitor in series with a 10 ohm resistor across the speaker terminals becomes an effective noise shunt at frequencies well above the audible range. The Zobel network did the trick. Matt Lyons, your trip to Connecticut was unnecessary. With the cable-compatibility problem cured, the Adcom fairly loafed along with the Alphacore cables, no matter what the levels. The fan and the LEDs never came on again.

I don't know where to point my accusing finger. The Goertz cables, without the Match Links, have worked fine with many speaker-amp combinations in this system. Conversely, the GFA-7805 has worked superbly with several other cable-speaker combinations in both of my systems. My hunch is that this is one of those lurking but curable incompatibilities that a competent dealer should know about. Both Adcom and Alpha-Core were aware of the possibility; I assume that they've informed their dealers of it.—Kalman Rubinson