Audio Note M2 Balanced Phono preamplifier Measurements

Sidebar 3: Measurements

Audio Note is an English company run by a Dane; I was surprised to see the M2's rear sporting a "Made in Canada" label.

I looked first at the phono stage. When I took its output from the tape output jacks, the voltage gain at 1kHz was 48.5dB, implying that the M2 will work well with moderate-output MC cartridges. Confirming this, the overload margin ref. 500µV at 1kHz was an excellent 25dB at low and midrange frequencies, dropping to a still good 20dB at 20kHz. Even though the phono stage's input impedance is specified at 47k ohms (I measured 53k at 1kHz), using a high-output MC or an MM cartridge will be on the verge of driving the circuit into overload.

The A-weighted signal/noise ratio, ref. 500µV input at 1kHz and taken with the input short-circuited, was a relatively good 62dB, though this worsened to 43dB with a wideband, unweighted measurement, due to the presence of very-low-frequency noise. The phono-stage frequency response rolled off slightly at the frequency extremes, being 1dB down at 23Hz and 15kHz (fig.1), which might have contributed to AD's feeling the preamp's balance was a little on the dark side. (Note also a 0.5dB mismatch between the channels in this graph.) The phono stage didn't invert signal polarity, at least from the tape outputs.

Fig.1 Audio Note M2 Balanced phono stage, RIAA error at 1mV input at 1kHz (0.5dB/vertical div., right channel dashed).

The line stage's maximum voltage gain—unbalanced input to unbalanced output—was significantly lower than specified, at 12.73dB rather than 17dB. This shortfall will have no practical consequences, as the M2 still has plenty enough gain to be capable of driving any power amplifier to its clipping point. The balanced output offered another 6dB of gain, as expected. The unity-gain setting of the volume control was 10:00 for the unbalanced output, 9:30 for the balanced.

The line stage inverted absolute polarity, which will be a factor in direct comparisons with other preamplifiers. The input impedance was a usefully high 63k ohms over most of the band, dropping to a still good 43k ohms at 20kHz. This means that the M2 will barely load down source components. The unbalanced output impedance was a low 200 ohms, rising a little at the audioband edges. The transformer-coupled balanced output's source impedance was somewhat higher, at 700 ohms, but this is still usefully low in absolute terms.

When I looked at the M2's line-stage frequency response, I was surprised by the fact that, while it was perfectly flat in the audioband, a series of ultrasonic peaks was also evident (fig.2). This graph was taken from the balanced output jacks, leading me to suspect that this behavior was due to the transformer. It could also be seen in the graph of the unbalanced output's response under the same conditions (fig.3, top pair of traces). Both of these graphs were taken with the volume control set to unity gain. The ultrasonic peaks progressively decreased in level as the volume control was rotated toward its maximum setting. Their levels decreased in a similar manner when the load impedance was reduced (fig.3, bottom traces). I'm not sure what the effect of this measured behavior would be on the Audio Note's sound quality, though it might well be subjectively benign, as suggested by AD's auditioning. But from an engineering viewpoint, I don't like to see it—it suggests that the line-stage circuitry might be only marginally stable.

Fig.2 Audio Note M2 Balanced line stage, balanced frequency response at 1V into 100k ohms with volume control at unity gain (2dB/vertical div., right channel dashed).

Fig.3 Audio Note M2 Balanced line stage, unbalanced frequency response at 1V into 100k ohms with volume control at unity gain (top) and into 600 ohms (bottom). (2dB/vertical div., right channel dashed.)

The M2 shares tubes between its channels, so it was not surprising to see only modest channel separation at high frequencies (fig.4), However, as can be seen from this graph, the coupling was worse in the L-R direction than in the R-L. The unweighted, wideband S/N ratio, ref. 1V output, was also relatively modest, at 67.3dB. This did improve to 85.2dB when A-weighted, again due to the elimination of some very-low-frequency noise as well as some low-level, 120Hz power-supply hum.

Fig.4 Audio Note M2 Balanced line stage, channel separation (10dB/vertical div., R-L dashed).

Fig.5 plots the THD+noise percentage in the M2's unbalanced output against output voltage into 100k, 10k, and 1k ohms. Below outputs of a few hundred millivolts, the measured figure is dominated by noise, but above that level the THD steadily rises due to an increasing amount of asymmetry in the signal waveform. True clipping doesn't happen until the preamp is putting out more than 10V. Commendably, the preamp's linearity is not significantly affected by the load impedance until the latter drops below 10k ohms. This can also be seen in the graph plotting the THD+N percentage against frequency at a 300mV level (fig.6). The 100k and 10k traces overlay one another and are respectably low in the midrange. A slight rise can be seen at the very top of the audioband, which might correlate with the ultrasonic peakiness seen in the frequency-response curves. More significant, the THD level rises with decreasing frequency throughout the bass region.

Fig.5 Audio Note M2 Balanced line stage, distortion (%) vs unbalanced 1kHz output voltage into (from bottom to top): 100k, 10k, 1k ohms.

Fig.6 Audio Note M2 Balanced line stage, unbalanced THD+N (%) vs frequency at 1V into (from bottom to top): 100k, 10k, and 1k ohms.

Fortunately, the spectrum of that distortion is heavily second-harmonic in nature (fig.7), the only exception being at low frequencies from the balanced output, where the third harmonic rises above the second (not shown). Note that the power-supply hum components can also be seen in this graph, though they are all below -80dB.

Fig.7 Audio Note M2 Balanced line stage, spectrum of 1kHz sinewave, DC-10kHz, at 1V unbalanced into 8k ohms (linear frequency scale).

The Audio Note did only okay in the final test: driving an equal mix of 19 and 20kHz tones at 1V into a fairly low impedance (fig.8). The 1kHz difference component lay at -50.6dB (0.29%). This graph was taken from the unbalanced jacks, by the way; the balanced performance was very similar. I should also note that with high-frequency test signals, the preamp's circuitry, or perhaps the output transformers, could be heard quietly "singing," even with the preamp's top cover in place.

Fig.8 Audio Note M2 Balanced, line-stage HF intermodulation spectrum, DC-24kHz, 19+20kHz at 1V unbalanced into 8k ohms (linear frequency scale).

Overall, the Audio Note M2 Balanced offers modest measured performance. This doesn't mean it will sound disappointing—AD was very impressed with its sound—but it does suggest a less-than-rigorous attitude to audio engineering on the part of the preamp's design team.—John Atkinson

Audio Note
US distributor: Triode & Co.
404 Hebron Avenue
Glastonbury, CT 06033
(860) 652-9691