Amplification Reviews

While many lab tests were easily passed, the M300 did exhibit some odd quirks which suggested that there is still room for further development and improvement in future years.

8 ohm tap: Covering the basics first, the power testing commenced (after two hours' warm-up) with 8 ohm loads on the 8 ohm tap. Rated at a continuous 300W or 24.5dBW, this amplifier raised an extra 1dB heading toward an exact power of 400W into the precise load impedance (7.7 ohms)—some power indeed! In terms of a marginal clipping criterion, the power bandwidth was excellent, holding virtually within 1dB from below 10Hz to 20kHz, though the latter result has a severely relaxed distortion criterion. For -40dB, or 1% distortion, into 8 ohms, the output level required a reduction of reduction of 6dB to 19.7dBW.

With a 4 ohm loading on the 8 ohm tap, the amplifier fought to maintain level, dipping only to 23.2dBW in the midband—a true 400W into 4 ohms, though more loss was seen at the bandwidth extremes, eg: a fall to 18.8dB at 20kHz.

On the 8 ohm tap, a peak current capability of ±13.5A was noted, sourced from a constant output impedance with frequency of typically 0.25 ohms, rather lower than is usual for a tube model. The current capability suggested that kinder 8 ohm loads would be preferred, ideally not falling below 6 ohms, with a predominantly resistive impedance characteristic. The 8 ohm tap will drive a 4 ohm load quite well, but showed some strain. Peak power measurements showed a true 400W into 8 ohms, 420W into 4 ohms, and a steeper fall to 16.5dBW, 250W into a highly stressful 2 ohm impedance.

4 ohm tap: Moving to the 4 ohm tap, the output impedance reduced to a typical 0.2 ohms, the peak current lifting to ±16A, and the peak level into an 8 ohm load was held to 22.5dBW, 180W or so. The higher "strength" was reflected by the minor fall to 21.8dB on 4 ohms loading, while the 2 ohm loaded level was 18.1dBW, a 1.6dB improvement over the 8 ohm tap. Tapped at 2 ohms, ±18 amps peak was available, the output impedance lowered to typically 0.16 ohms, and the 8 ohm loaded output was still a substantial 20dBW or 100W into that impedance. A 4 ohm load was effortlessly handled, a drop of only 0.2dB at nearly 200W total. Finally, into a 2 ohm load the output was 18.4dBW, which represented only a small increment over the 4 ohm tap rating. For the sake of interest, the current delivery on the 1 ohm tap was checked out measuring ±24 amps.

The load-matching behavior is instructive, but the amplifier did not really address such problem speakers as the 1 ohm Apogee Scintilla. Low-impedance-ratio taps do provide the low output resistance and current necessary for such a load, but speakers as insensitive as this also require a good voltage swing to provide decent volumes. Current and voltage are both required. Transformer matching merely allows one to trade voltage for current, and vice versa. Take a Krell KMA-200 capable of running at a 24dBW level with a ±60A peak current capability: into a 1 ohm Scintilla, it will punch a true maximum power of nearly 2000W, while the 17A maximum current delivery of the M300 will limit the power into this load to just 300W.

Having checked several examples of the M300, I am aware of some variability with respect to linearity and the results for harmonic and intermodulation distortion. Frankly, it is unimpressive in these departments. Fig.3 shows the 20kHz waveform at 2/3rds full level, with slope changes akin to crossover non-linearity. These can still be seen at 1kHz as small glitches. The distortion results were just satisfactory in my opinion, at typically -50dB or 0.33%, worsening at high frequencies. Even at 0dBW, 1W, the 20kHz distortion was -35dB, or 2.2%.

Fig.3 Audio Research M300, 20kHz sinewave at 200W into 8 ohms.

At rated level the two-tone high-frequency intermodulation (19/20kHz mixed 1:1) read -45dB (fig.4), improving to -67dB at 0dBW (fig.5). At 1kHz the 0dBW distortion spectrum is given in fig.6, though other results suggested that this varied from sample to sample. Higher harmonic orders were present, unusual for a tube amp, and can be attributed to the slope non-linearity displayed.

Fig.4 Audio Research M300, HF intermodulation spectrum, DC-25kHz, 19+20kHz at 300W into 8 ohms (linear frequency scale).

Fig.5 Audio Research M300, HF intermodulation spectrum, DC-25kHz, 19+20kHz at 1W into 8 ohms (linear frequency scale).

Fig.6 Audio Research M300, spectrum of 1kHz sinewave, DC-10kHz, at 1W into 8 ohms (linear frequency scale).

Driven to clipping, some asymmetry was noticed together with some "hook" or slope latching (fig.7). By the way, this was noticeably improved over early production M300s which were not very good clippers. (However, you may well ask who in their right mind would be clipping a 300Wpc amplifier!)

Fig.7 Audio Research M300, 1kHz waveform at clipping.

The signal/noise ratio was fine, with low levels of supply hum. Tested for power modulation, in this case 150W into 4 ohms, 8 ohms tap, 35Hz fundamental, the result was very good, with the spectrogram (fig.8) showing simple harmonics with virtually no line components.

Fig.8 Audio Research M300, spectrum of 30Hz sinewave, DC-500Hz, at 150W into 4 ohms (linear frequency scale).

Channel balance was very close at typically 0.1dB, while the frequency response was very wide and improved over an earlier sample. The -0.5dB points were below 0.5Hz to 88kHz, at 1W, 8 ohms. For -3dB points, it extended from less than 0.3Hz to above 100kHz. A slight rise of 0.27dB was noted at 40kHz.

The input impedance was typically 60k ohms in parallel with 105pF and the amplifier was non-inverting of absolute phase. No DC offset was observed at the output (some cathode-coupled ARC amps can develop a small DC offset on the transformer secondary).

I need not mention channel separation in a monoblock such as this. Ultimately, separation will be controlled by the grounding and the source preamplifier.—Martin Colloms

Table 1: Audio Research M300, Measured Performance

Rated Power, 8 ohms: 300W, 24.5dBW (0dBW = 1W into 8 ohms)
4 ohms: 300W, 21.5dBW

Peak Current (via 1 ohm, 2ms pulse)
8 ohm tap: +13.7A -13.5A
4 ohm tap: +16A
2 ohm tap: +18A
1 ohm tap: +24A

DC Offset: None

Intermodulation Distortion:
19/20kHz, 1:1 ratio -47dB (rated power), -67dB (0dBW)
S/N ratio (full gain) Rel. 0dBW Rel. full rated level
22Hz-22kHz: 70dB 89dB
A-weighted: 77dB 97dB
Channel Balance: 0.11dB
Polarity: Non-inverting
Input Impedance: 60k ohms in parallel with l05pF
Input Sensivity: 50mV IHF, 0dBW
1.0V for program clip
Mechanical noise: Low hum; mild fan noise on "low"
Frequency response:
<0.5Hz-88kHz, +0dB, -0.5dB
<0.3Hz->100kHz, +0dB, -3dB
Slight (0.27dB) rise at 40kHz