Berning EA-2101 power amplifier Measurements part 2

Figs.7 and 8 provide another look at this from a different perspective. The THD+noise vs frequency curve in fig.7 shows the results for a constant 8 ohm load and 1W output with four different output taps—the standard 14.22 ohm and 8 ohm taps, and the 3.55 ohm and 0.89 ohm high-current-configured taps. While the 0.89 ohm tap gives the lowest distortion through the midband, it rapidly rises to meet, then exceed the distortion of the other taps at high frequencies. (Though this will not in itself be audible, the distortion harmonics lying above the audio band, it may well result in audible intermodulation products—see later.) And fig.8 (THD+noise vs level) shows why the 0.89 ohm tap may not be the best choice to drive 8 ohm loads: despite its lower level of distortion, the amplifier clips at less than 30W with this output transformer configuration. Figs.9 and 10 mirror figs.7 and 8, except that they were taken into a 4 ohm load.

Fig.7 Berning EA-2101, THD+N (%) vs frequency at 1W into 8 ohms (from bottom to top at 1kHz): 0.89 ohm, high-current tap; 3.55 ohm, high-current tap; 8 ohm tap; 14.22 ohm tap (right channel dashed).

Fig.8 Berning EA-2101, distortion (%) vs 1kHz continuous output power into 8 ohms (from bottom to top at 1kHz) from: 0.89 ohm, high-current tap; 3.55 ohm, high-current tap; 8 ohm tap; 14.22 ohm tap.

Fig.9 Berning EA-2101, THD+N (%) vs frequency at 2W into 4 ohms (from bottom to top at 1kHz): 0.89 ohm, high-current tap; 3.55 ohm, high-current tap; 8 ohm tap; 14.22 ohm tap (right channel dashed).

Fig.10 Berning EA-2101, distortion (%) vs 1kHz continuous output power into 4 ohms (from bottom to top at 1kHz) from: 0.89 ohm, high-current tap; 3.55 ohm, high-current tap; 8 ohm tap; 14.22 ohm tap.

Perhaps more important than the level of distortion is its harmonic spectrum (footnote 1). In fig.11 we see the Berning's spectral response to a 50Hz sinewave input with an output power of 67W into a 4 ohm load using the 3.55 ohm, high-current-configured output tap. The result at first glance does not look promising. While not particularly notable as modern amplifiers go, the highest-output harmonic is the third, at -46dB (about 0.5%). The second harmonic is at -47dB (0.45%), the remaining levels growing ever smaller until they dip below 0.1% at and above the 400Hz point. (The nomograph in fig.12, prepared by J. Gordon Holt, shows at a glance how a distortion level in dB below a reference level translates to the more traditional percentage figure.)

Fig.11 Berning EA-2101, 3.55 ohm tap, spectrum of 50Hz sinewave, DC-1kHz, at 67W into 4 ohms (linear frequency scale).

Fig.12 Chart for converting distortion levels in dB below fundamental into percentage of signal.

In contrast, the distortion with the same load but driven from the Berning's 8 ohm tap is considerably higher. This curve is not shown, but the second harmonic lies at -39dB (about 1.1%), the third at -36.4dB (about 1.5%). The readings are similarly higher for higher harmonics with this hookup, though they remain below 0.2% above 400Hz.

Fig.13 shows the intermodulation spectrum resulting from asking the Berning to drive an 8 ohm load with a combined 19+20kHz input signal from its 8 ohm output tap. (The resultant level was 67W.) Here the predominant artifacts are at 22kHz (-43dB or 0.7%), 21kHz (-46dB or 0.5%), 18kHz (-47dB or about 0.45%), 17kHz (-43dB or 0.7%), and 1kHz (-45dB or about 0.55%). Changing to a 4 ohm load from this same 8 ohm tap (with the same voltage output, graph not shown) results in an increase in distortion, though not at every point—1% at 22kHz, 0.5% at 21kHz, 0.45% at 18kHz, 1.1% at 17kHz, and 1% at 1kHz. As might be expected from fig.7, which reveals rapidly rising levels of distortion at high frequencies, one of the worst cases resulted from driving an 8 ohm load (at 67W) from the 3.55 ohm high-current tap. Here the distortion artifacts reached 3% at 22kHz, 5% at 21kHz, 5% at 18kHz, and 1.5% at 1kHz.

Fig.13 Berning EA-2101, 8 ohm tap, HF intermodulation spectrum, DC-22kHz, 19+20kHz at 67W into 8 ohms (linear frequency scale).

The lesson here, which will be repeated in the following data in a different form, is that the lowest distortion generally resulted from matching load impedance to output tap impedance. These generally high levels of distortion and the way in which they increase with level correlates with Dick Olsher's review finding that the Berning's midrange and lower treble acquired a slightly bright, wiry character that grew more apparent during loud passages.

The wide variety of possible loads and output taps makes the graphic presentation of further data on the maximum output levels cumbersome at best. I elected to condense the results into Table 3. The values presented here were read directly off the curves of THD+noise vs level, one channel driven. The knee of the distortion curve, as I've described in other reviews, is the point at which the distortion begins rising rapidly. It is also the point beyond which the output of the amplifier is of little use, although clipping is usually reached a bit above the knee. In this table you can read the power output and distortion reading at the knee, the distortion at mid-power (here specified for ease of comparison as half the output at the knee of the distortion curve), and the power output at 1% and 3% THD+noise. The distortion with certain combinations of output tap and load reaches 1% at quite a low power, but then rises only slightly more before reaching the knee of the curve.

Table 3: Distortion vs Power

Values read from graphs (one channel driven)
Output Tap (ohms)Knee (W/% THD+N) Mid-Power* % THD+NPower (W) @ 1% THD+NPower (W) @ 3% THD+N
14.22 ohms
16 ohm load150/
8 ohm load140/
4 ohm load29/
8 ohms
8 ohm load150/0.81.1160.0175
4 ohm load140/
2 ohm load27/
3.55 ohms (normal)
8 ohm load80/0.550.5383.090
4 ohm load130/0.720.85140.0160
2 ohm load150/
3.55 ohms (high-current)
8 ohm load83/0.520.688.095
4 ohm load140/0.750.9150.0170
2 ohm load140/
0.89 ohms (normal)
8 ohm load20/0.280.2225.026
4 ohm load36/0.320.3245.048
2 ohm load72/0.50.5375.081
0.89 ohms (high-current)
8 ohm load20/0.280.2226.027
4 ohm load39/0.350.3548.051
2 ohm load82/0.520.5885.092

* Power at ½ the power at the distortion curve "knee."

In addition, Table 4 gives the discretely measured readings of the Berning's clipping performance. Because of the amplifier's distortion characteristics, clipping here was taken as the 3% THD+noise level rather than our usually specified 1%.

Table 4: Clipping (3% THD+N at 1kHz)

Discrete Measurements
Output Tap (ohms)Both Channels Driven W (dBW)One Channel Driven W (dBW)AC Line Voltage
16 ohms  161.1 (25.1)112V
8 ohms81.1 (19.1)89.3 (19.5)173.7 (22.4)112V
4 ohms43.5 (13.4)49.2 (13.9)40.2 (13.0)111.5V
8 ohms137.3 (21.4)140.9 (21.5)167.4 (22.2)113V
4 ohms76.7 (15.8)84.8 (16.3)159.4 (19.0)112.5V
2 ohms  35.1 (9.4)113V
3.55 (normal)
8 ohms87.7 (19.4)88.4 (19.5)89.1 (19.5)114V
4 ohms139.6 (18.4)141.6 (18.5)145.2 (18.6)113.5V
2 ohms  158.4 (16.0)113V
3.55 (high-current)
8 ohms92.5 (19.7)92.9 (19.7)94.0 (19.7)116V
4 ohms153.4 (18.85)154.5 (18.9)159.9 (19.0)116V
2 ohms  178.1 (16.5)116V
0.89 (normal)
8 ohms25.8 (14.1)25.8 (14.1)25.9 (14.1)114V
4 ohms47.0 (13.7)47.2 (13.7)47.4 (13.75)114V
2 ohms  80.1 (13.0)114.5V
0.89 (high-current)
8 ohms26.8 (14.3)26.7 (14.3)26.9 (14.3)121V
4 ohms50.5 (14.0)50.5 (14.0)51.0 (14.0)121V
2 ohms  91.7 (13.6)121V

Two conclusions are clear from these tables. First, the 0.89 ohm output tap (either normal or high current) gives the lowest mid-power distortion but, ultimately, a THD-restricted maximum output power for any typical load. Second, it is generally advisable to match the output tap impedance as closely as possible to the load impedance, an observation made earlier. But use caution here—loudspeakers do not have constant impedances at all frequencies. The best match will still involve a certain degree of experimentation.

The Berning's unweighted wide-band signal/noise ratio measured -51.6dB in the unbalanced mode (referenced to 1W into 8 ohms). The result was essentially the same in the balanced mode (-51.4dB) (left channel; the right-channel S/N was just over 2dB worse in both cases). Fig.14 shows the EA-2101's output noise signature. Note the periodic pulses at 75kHz, most certainly residual noise from the switching power supply, which compromises the wide-band noise measurements.

Fig.14 Berning EA-2101, waveform of residual noise waveform with random components reduced in level by averaging (50µs time window).

The Berning's measured performance is acceptable but not inspiring. Its distortion, in particular, is higher than average for a modern amplifier in its price range, even a tube one, probably a result of the low-feedback design. There is a considerable difference of opinion on the amount of distortion which is audible (though more agreement that higher-order distortion is less tolerable than low-order), but it is probable that the Berning's distortion is at least a factor in the amplifier's sound.—Thomas J. Norton

Footnote 1: Stereophile's Test CD 2 carries a series of listening tests that compare the audibility of different distortion spectra at different levels.—John Atkinson
12430 McCrossin Lane
Potomac, MD 20854
(301) 926-3371