In practice, in my listening room, the Atlas lived up to its name. It was truly effortless in driving the Maggies at all listening levels, including levels at which I started worrying about blowing up the speakers. (But at least I never worried about blowing up the amps!) Throughout the entire review period, I never heard or felt any sense of strain from the Atlas Eclipse. In saying this, I must also give a tip of the hat to Magnepan. As an owner of all prior versions of the 20 series speaker, I used to have plenty of reasons to worry about blowing out a tweeter or even a midrange section. But the 20.7s are another matter, reliably playing to very high levels without sign of stress or failure (so far). I seriously doubt that any other full-range planar loudspeaker (except the new Magnepan 30.7) is capable of playing cleanly at the sustained high levels of which the 20.7 is capable.
Because it is a hybrid design, the Atlas is different from most other amplifiers. Through use of a tube input stage, Jim White (and other designers of hybrids) attempt to infuse the sound with the properties of the tube or tubes chosen to serve as input amplification. I am not qualified to delve into intricacies of amplifier design, but as I understand it, designers choose a hybrid topology when they want to obtain a tube-type sound with the benefits of a solid-state output stage, such as higher output and tighter control of the bass. Furthermore, a hybrid design can generally offer a higher power output, at a lower price, than similarly priced all-tube designs.
Not all hybrids, of course, are created equal. Years ago I tried some early hybrid amps made by Counterpoint. They had some nice qualities, but almost no control of the bass. But time and technology move on; the Atlas Eclipse is a state-of-the-art design with none of the flaws of those older amps. The gain stage of the Atlas uses one 6SN7 tube per channel. As this tube provides all of the voltage gain for the amplifier, it is fair to say that much of the sound of the amp is provided by this tube. Jim White informs me that as a zero-feedback design, the Atlas is unique in the world of hybrid amplifiers. He believes the lack of feedback is essential for maintaining air, space, and coherence. The Atlas’ inputs are sophisticated and flexible in that both single-ended and balanced are provided, as well as a separate direct input that bypasses all internal crossovers and switching for additional purity.
Both the Eclipse and the less expensive Signature versions of the Atlas utilize 16 bipolar output transistors per channel. Both versions also add a unique, but amazingly useful, wrinkle to connectivity. In addition to the full-range direct input, the Atlas has a separate set of inputs that allow the use of a built-in high-pass crossover. The crossover is adjustable and allows the user to roll off the low frequencies going to the loudspeaker driven by the Atlas. This capability is perfect for those who use subwoofers (or who own speakers such as Vandersteens with self-powered bass). The built-in crossover means there is no need to place an external crossover before the Atlas to remove the low bass frequencies before the Atlas sees the signal. This avoids any sonic degradation that may be added by the external crossover (including crossovers that may be built into subwoofers) and also eliminates the need to purchase additional connecting cables. Front panel controls on the amp allow 16 different crossover settings, all with a slope of 6dB/octave. This flexibility should allow the Atlas-driven main speaker to blend well with any subwoofer.
You may be thinking that since both the Signature and Eclipse versions contain so many of the same features, let alone the same essential circuit topology, can it be possible that the Eclipse sounds significantly different than the Signature? It can, and it does. Before examining the sound, it is important to note that the Eclipse differs from the Signature in many important respects. They both use the same number of output transistors, but for the Eclipse these transistors are carefully hand-matched across several parameters. Statistically, Jim informs me that 200–300 devices need to be measured before they obtain a matching set of 16 output transistors. He believes that this extreme matching leads to a reduction of distortion and a greater liquidity to the sound. This is all well and good, but of course the proof is in the listening.
The Eclipse also differs from its junior brethren in other important respects. First, each output transistor is accompanied by a dedicated reservoir capacitor (in addition to the large reservoir caps in the power supply). The aluminum casework has been redesigned to be more rigid and reduce resonances. The Eclipse also uses StealthCap coupling capacitors specifically designed for the Atlas by Peter Moncrieff. Changes have also been made to the power supply: The high-voltage transformer and choke are de-coupled from the chassis using specially designed isolation mounts. This reduces mechanical vibration introduced into the chassis, with the goal of increasing transparency. The power supply receives an additional upgrade in the form of lower-gauge winding made of the same material used in very high-end power cables.
It is apparent the Atlas Eclipse excels at connectivity and offers crossover features found almost nowhere else. It also contains many advanced design features. But these features would be wasted, in the final analysis, if the Eclipse did not offer sound reproduction noticeably superior to the Signature version of the amp. I am pleased to report that Jim’s efforts have succeeded in taking the performance of the Atlas to an entirely new level. Most importantly, the unflappable ease and power of the earlier Atlas versions have been retained, but with a new three-dimensionality and transparency unique to the Eclipse.