Nagra VPA monoblock power amplifier Measurements
The VPA was warmed up for approximately 15 minutes prior to any measurements being taken; according to Nagra, the amplifier should enter stable operation after this period. All measurements were made in balanced mode. The 8 ohm measurements were taken from the 8 ohm outputs, 4 and 2 ohm measurements from the 4 ohm outputs.
The Nagra's input impedance is too high (good) to measure accurately on our Audio Precision test set. The two voltages required to produce the measurement differed by mere thousandths of a volt. (With the slight variations in this range typical of tube amplifiers, the values actually overlapped.) Suffice it to say that this amplifier's input impedance should cause no matching problems with any competent preamplifier. Pin 2 of the balanced input is wired as positive.
The VPA's voltage gain into 8 ohms, from the 8 ohm tap, is a high 35.7dB. The measured output impedance was quite high, varying between 2.23 and 2.26 ohms depending on frequency and load impedance (measurements taken at the 8 ohm output). This suggests some potential frequency-response irregularities with typical loudspeaker loads.
The Nagra's DC offset was unmeasurable. Signal/Noise Ratio (unweighted ref. 1W into 8 ohms) measured 66.2dB with a bandwidth of 22Hz to 22kHz, 66.2dB from 10Hz to 500kHz, and 80dB A-weighted. (Most of the noise in the unweighted measurements was at low frequencies.)
Fig.1 shows the VPA's frequency response. The variations into our simulated real load indicate the sort of frequency-response variations this amplifier will produce into an actual loudspeaker, though the exact variations will depend on the particular loudspeaker. The 10kHz squarewave response is shown in fig.2—a very good result. The Nagra has a good rise time with little rounding of the waveform's leading edge. The 1kHz squarewave (not shown) is nearly perfect.
Fig.1 Nagra VPA, frequency response into (from top to bottom): 1W into 8 ohms, 8 ohm tap; 2W into 4 ohms, 4 ohm tap; and 2.828V into simulated loudspeaker load, 8 ohm tap (0.5dB/vertical div.).
Fig.2 Nagra VPA, small-signal 10kHz squarewave into 8 ohms, 8 ohm tap.
The low-power THD+noise vs frequency results shown in fig.3 are not exceptional. While the midband linearity is excellent, the amplifier becomes increasingly nonlinear at the frequency extremes, particularly below 100Hz—when you use no feedback, increased distortion is the consequence you must accept. The waveform of the distortion (1kHz) is shown in fig.4. At 1W the waveform was difficult to see in the noise, but at 5W a dominant second harmonic is apparent, as well as clear signs of higher-order distortion. Into 2 ohms (not shown), the dominant harmonic changes from second to third.
Fig.3 Nagra VPA, THD+noise vs frequency at (from top to bottom at 10kHz): 4W into 2 ohms, 4 ohm tap; 2W into 4 ohms, 4 ohm tap; 1W into 8 ohms, 8 ohm tap; and 2.83V into simulated loudspeaker load, 8 ohm tap.
Fig.4 Nagra VPA, 1kHz waveform at 5W into 4 ohms (top), 4 ohm tap, distortion and noise waveform with fundamental notched out (bottom, not to scale).
Fig.5 shows the Nagra's output spectrum, reproducing a 50Hz sinewave at 41W into 4 ohms. The distortion, again, is fairly high—a maximum of -43dB (about 0.7%) at 100Hz, with a regularly descending series of higher harmonics. Fig.6 shows the intermodulation distortion resulting from a combined 19+20kHz signal at 30.7W output into 4 ohms—a level just below visible signs of clipping with this test signal. The highest artifact here, at 18kHz, is -43.9dB, or about 0.7%. There is also approximately 0.5% distortion at 1kHz (-46.9dB). The results are not significantly different into 8 ohms (not shown).
Fig.5 Nagra VPA, spectrum of 50Hz sinewave, DC-1kHz, at 41W into 4 ohms, 4 ohm tap (linear frequency scale).
Fig.6 Nagra VPA, HF intermodulation spectrum, DC-22kHz, 19+20kHz at 30.7W into 4 ohms, 4 ohm tap (linear frequency scale).
The THD+noise percentage vs output power curves for 1kHz are shown in fig.7. The actual discrete-clipping-point measurements (1% THD+noise) are 59.9W (17.8dBW, 118V line) into an 8 ohm load, 61.2W (14.9dBW, 118V line) into 4 ohms, and 68.1W (12.3dBW, 118V line) into 2 ohms.
Fig.7 Nagra VPA, distortion (%) vs continuous output power into (from bottom to top at 10W): 8 ohms, 8 ohm tap; 4 ohms and 2 ohms, 4 ohm tap.
It wasn't possible to measure the VPA's toneburst power with the Miller Audio Research Amplifier Profiler; its floating output configuration was not happy with the test system's ground-referenced input.
The Nagra VPA's measured performance reflects its low-feedback design. In short, acceptable, but little more, I feel.—Thomas J. Norton