With the distortion limit set at 3%, the amplifier delivered 76Wrms into an 8Ω load, a figure that increased to 134Wrms when the load was reduced to 4Ω. With a power increase factor of close to 1.8, we can conclude that the Six has substantial power reserves and should be able to drive loudspeakers whose impedance magnitude drops to low values.
As we also found when testing hARt’s power amplifier, this measurement does not fully reflect the capabilities of the power stage, which exhibits clear signs of clipping only at considerably higher power levels. Most of the distortion—which typically determines the power figure we measure—originates in the preamplifier stage. Put simply, this means that the Tune Six’s output stage can deliver considerably more power before it begins to distort on its own. We had the opportunity to confirm this during our measurements of the Tune Three, when we drove the amplifier into visible clipping , as you can see here. On this basis, the power rating published by the manufacturer appears more than realistic. The damping factor was calculated at around 33, which is a satisfactory figure, although it leaves some room for the load to have a minor influence on the overall frequency response.
Given what we already know about the circuit, the distortion measurements came as no surprise. The Tune Six consistently operated in the 1–2% range, with a slight tendency for distortion to increase with frequency, as would be expected from a circuit with relatively low feedback. This was true of both the THD and THD+N measurements, while the amplifier showed similar performance in the SMPTE IMD test. Interestingly, we found no major differences between the measurements taken with 8Ω and 4Ω loads. This suggests that the first, tube-based stage accounts for most of the measured distortion.
The amplifier proved to be quiet, with the signal-to-noise ratio exceeding -75dBr(A), referenced to an output of 1Wrms into 8Ω.
The frequency response into an 8Ω load is essentially flat, with gentle roll-offs at both ends of the spectrum: less than -3dB at 10Hz and less than -2dB at 60kHz. With a complex load, there are some small but visible deviations, although these do not exceed 0.5dB, in line with the measured damping factor. Channel matching is very good, with differences of less than 0.5dB.
The phase response is quite interesting. The phase difference between input and output remains within ±10° over a wide frequency range (100Hz–20kHz), suggesting low group delay and, consequently, good time-domain behavior. This supports hARt Lab’s design choices regarding feedback.

The output spectrum into an 8Ω load with a 1kHz signal contains a series of relatively strong harmonics, the second harmonic being the strongest at around -40dBr. Their levels decrease smoothly but relatively slowly, remaining above -80dBr until the seventh harmonic.
With a 4Ω load, the behavior of the lower-order harmonics remains unchanged, but the amplifier becomes somewhat more active at higher frequencies, although the levels there remain low, below -80dBr. It is also a positive sign that the amplifier’s noise level does not change significantly when the load is altered, showing only a slight tendency to increase at higher-order harmonics of 50Hz.

The intermodulation distortion spectrum into an 8Ω load follows the same pattern as the harmonic distortion spectra, with several components at relatively high levels. The third- and fifth-order components are at -50dBr and -60dBr, respectively, while the first-order component (1kHz) is at -40dBr.
With a 4Ω load, the noise level increases, accompanied by small changes in some of the higher-order components.

The noise spectrum indicates that the amplifier is fairly quiet. The power-supply noise measured at the output, with no signal at the input, was around 1mVrms. Mains-related components at 50/100Hz and higher frequencies are visible, but their levels are low, around or below -90dBr.
Despite being a complex component that incorporates a large display and a computer board, the high-frequency spectrum is remarkably calm. The corresponding plot, covering a 10MHz span centered at 5MHz with a 15kHz input signal, reveals nothing of particular significance.

The way distortion varies with frequency and output power holds no surprises. Distortion increases as the amplifier is asked to deliver more power, reaching 2% in the worst-case scenario: at 20kHz and one-third of the maximum power into 8Ω. At very low output levels, it remains below 0.4%.
With a 4Ω load, the behavior is similar, with distortion tending to increase at high frequencies, perhaps as a consequence of the low feedback level.

The harmonic distortion versus power plots show an amplifier with a smooth, gradual response. The “knee” that indicates the onset of overload is relatively indistinct, suggesting that the amplifier behaves gently as it approaches its limits.
With a square wave applied to the input, the response is smooth, with no problems in its transient behavior. The rise time was measured at 3.9μs and the slew rate at 4.3V/μs (at 10Vpp), comfortably exceeding the minimum required for this signal level.

Previous | Next | More Reviews