Australia's Halcro Has Some Explaining To Do
What has raised the eyebrows of the audio community, however, is a bold claim, repeated on the company website, that it is "the only amplifier in the world to challenge a theoretical limit." That limit is the specification for harmonic and IM distortion, which the company claims to have virtually eliminated via its circuitry, whose design is a closely-guarded secret. Halcro says that its dm 68 "produces less than 25 parts per billion harmonic distortion at 1kHz, and the worst result is about 600 parts per billion at 20kHz at full power, a frequency where most amplifiers exhibit very poor behaviour."
When asked how his company achieves such startling figures, founder Bruce Halcro Candy states that "it is extremely difficult to provide you with details without significantly disclosing the patents and intellectual property which are now hidden from public view." Candy does reveal that, about 10 years ago, he decided to try to eliminate the harsh sound that he heard in transistor amplifiers. "I think that I have succeeded," he says. "The result is now manifest in the Halcro units, which exhibit harmonic and IM distortion levels many hundreds of times lower than the typical lowest distortion amplifiers."
Candy explains that he has loved music all his life, and has very unusual hearing: "I can hear up to 23kHz in one ear and 21kHz in the other. When I was younger, I could hear much higher frequencies than these." Candy says that "unlike all my audiophile fashionable friends, I did not convert over to transistors for many years, even when valves were highly passé. As you can imagine, I was teased mercilessly about this idiosyncratic behaviour. However, I thought that the transistor sound was ghastly—very harsh. I owned a pair of Mullard 5-20 amps with those well-designed Partridge output transformers."
According to Candy, the Halcro specs, "which many of our competitors assume are absurd nonsense," have been twice independently verified, "once by a university commissioned by another audio magazine, and another time also by a university as a pre-requisite when the company won an award. As you can imagine, in order to achieve these ultra-low levels of distortion, all aspects of standard low distortion circuitry required considerable conceptual changes, about 12 fundamental conceptual changes in all," says Candy.
"Every stage of every circuit I have ever seen will increase distortion if swapped with the 'equivalent' stage in the Halcro amps. Hence there is not even a small bit of the Halcro circuitry that is at all similar to the conventional low-distortion high-quality circuits. It definitely would take the average amplifier designer quite some time to recognise the circuit as that of an amplifier if he were presented with no direct clue as to its purpose. I don't think that any designer would guess that the parts list, if seen in isolation, was of an amplifier."
Candy also takes issue with traditional distortion measuring techniques, saying that "they will clearly not suffice for the measurements attained by the Halcro amplifiers." As he explains, "It is often assumed that, if a product's distortion is less than the measured noise, then it is necessarily inaudible. As is well known, ears behave much more like spectrum analyzers for short periods than like time-domain analyzers. Thus, ears can detect tones or specific frequency spectra in noise even if the noise is far bigger in amplitude.
"It could be said that the skill of an acute audiophile ear resides in the ability to perceive false signals amongst desired signals. So, to what extent a good ear can hear below the amplifier noise when music is simultaneously produced is difficult to assess. However, it seems strange to me that distortion is measured to include noise, because one can definitely hear signals of amplitude substantially less than the noise measured in the distortion + noise measurements. The answer, I assume, is historical: Early THD analyzers simply measured THD+N, whereas it is now possible to measure arbitarily low noise floors, depending on the ability to notch out the fundamental and the frequency-point spacing in the FFT analysis."