Prima Luna Prologue One integrated amplifier Measurements
I ran the Prologue One at one-third power into 8 ohms from the 8 ohm tap for 60 minutes—not so much to thermally stress it, as I do with solid-state designs, but to make sure it was up to its working temperature. (I didn't notice the pleasant smell, Art!) I performed a complete set of tests from both its 8 ohm and 4 ohm output transformer taps, but have described the results for both here only when relevant.
The input impedance was usefully high, at 95k ohms at 20Hz and 1kHz, this dropping inconsequentially to 76k ohms at 20kHz. The amplifier preserved absolute polarity—ie, was noninverting—from both sets of outputs. The voltage gain was a little hard to assess accurately, due to the presence of very-low-frequency noise, which caused the level reading to bounce a little. But it appeared that with the volume control set to its maximum, the gain from the 8 ohm tap into an 8 ohm load was 37.5dB; from the 4 ohm tap into 4 ohms, 36.3dB. Because the amplifier does not have preamp or tape outputs, I could not assess how much of this gain comes from the preamp stage and how much from the power amplifier circuitry.
The output impedance was very high, at 5.8 ohms from the 8 ohm tap and 3.1 ohms from the 4 ohm tap. (These figures were measured at 1kHz; the source impedance was a little higher at 20Hz, a little lower at 20kHz, but inconsequentially so.) These high source impedances guarantee maximum power delivery into matched loads: approximately 6 ohms in the case of the "8 ohm" output, 3 ohms in the case of the "4 ohm" output. However, the downside is that there is a very audible modification of the amplifier's frequency response due to the manner in which the loudspeaker's impedance varies with frequency. This is shown graphically by the top traces in figs.1 and 2, which were taken with the amplifier's 8 and 4 ohm taps, respectively, driving Stereophile's simulated loudspeaker load. The response modification is as high as ±2.5dB and ±1.6dB, and cannot be discounted when auditioning the amplifier.
Fig.1 Prima Luna Prologue One, 8 ohm tap, frequency response at 2.83V into (from top to bottom at 2kHz): simulated loudspeaker load, 8 ohms, 4 ohms, 2 ohms (2dB/vertical div., right channel dashed).
Fig.2 Prima Luna Prologue One, 4 ohm tap, frequency response at 2.83V into (from top to bottom at 2kHz): simulated loudspeaker load, 8 ohms, 4 ohms, 2 ohms (2dB/vertical div., right channel dashed).
These two graphs also show that the Prologue One's response into resistive loads is commendably flat across quite a wide bandwidth, particularly at low frequencies. At high frequencies, the rolloff didn't depend on the load to any great extent, which can be common in designs using an output transformer, but did depend on the setting of the volume control. It varied from –1dB at 30kHz at the extreme volume-control positions to –1dB at 20kHz in the unity-gain position (8:00). The Prima Luna's reproduction of a 10kHz squarewave (fig.3) shows some slowing of the waveform edges, as well as a faint hint of ringing. But not apparent in the graph, which was taken with a digital oscilloscope (footnote 1), is a very brief overshoot on each leading edge, which was visible with a 20MHz analog 'scope. It is possible, therefore, that the Prologue One's circuitry is not unconditionally stable; but if so, the parasitic instability appears to be in the MHz region and is probably inconsequential.
Fig.3 Prima Luna Prologue One, 8 ohm tap, small-signal 10kHz squarewave into 8 ohms.
The Prima Luna's channel separation was moderate, at better than 60dB below 2kHz, decreasing above that frequency at the usual 6dB/octave rate due to capacitive coupling between the channels. The separation at 20kHz was 47dB, which is merely adequate. Of perhaps more practical concern was leakage between adjacent pairs of inputs. Because the amplifier does not have a Mute button, I was using the input selector to switch to an unused input between tests. When I did so, however, I noticed that the oscilloscope was still showing some high-frequency activity. Inserting a shorting plug into the unused input didn't make any difference; the leakage response in both the shorted and open-circuit conditions is shown in fig.4. If you're playing a CD with, say, a switched-on tuner plugged into the next pair of inputs, you'll hear a quiet treble ghost of the tuner signal in the background. The solution is to turn unselected sources down (not off), but it's rare to see this behavior these days.
Fig.4 Prima Luna Prologue One, adjacent-input crosstalk (10dB/vertical div.).
With the input shorted and the volume control set to its maximum, the Prologue One's unweighted signal/noise ratio (ref. 1W into 8 ohms) was moderate, at 64.9dB. This increased to 75.9dB when A-weighted.
As expected from a design using a pair of EL34 output tubes for each channel, the Prologue One's maximum power is relatively modest. Fig.5 shows the percentage of THD+noise present in the amplifier's output while its 8 ohm output transformer tap drives increasing power levels into loads ranging from 2 to 16 ohms; fig.6 shows similar curves for the 4 ohm output tap. With the load matched to the nominal output tap, the 1% points are 21W from the 8 ohm tap (13.2dBW), and 24W from the 4 ohm tap (10.8dBW). It could be argued that using our usual definition of "clipping" as being 1% THD would be misleading, as the amplifier is not actually in hard clip at that point, but is really suffering from waveform triangulation and compression of the peaks. So, relaxing the definition to 3% THD, where actual clipping is visible on the oscilloscope screen, the 8 ohm tap gives out 42W into 8 ohms (16.2dBW) and 14W into 4 ohms (8.45dBW). The 4 ohm tap delivers 25W into 8 ohms (14dBW), 47W into 4 ohms (13.7dBW), and 17W into 2 ohms (6.3dBW).
Fig.5 Prima Luna Prologue One, 8 ohm tap, distortion (%) vs 1kHz continuous output power into (from bottom to top at 1W): 16 ohms, 8 ohms, 4 ohms, 2 ohms.
Fig.6 Prima Luna Prologue One, 4 ohm tap, distortion (%) vs 1kHz continuous output power into (from bottom to top at 1W): 16 ohms, 8 ohms, 4 ohms, 2 ohms.
I plotted the manner in which the THD+noise percentage changes with frequency at 5V output, a level at which the distortion products emerge from the noise floor. Figs.7 and 8 show the behavior from the 8 and 4 ohm taps, respectively. Though it increases to moderately high levels at high and low frequencies—reflecting the relatively modestly sized output transformers—and into loads that are much lower than the nominal value of the transformer tap, the midband THD is quite respectable. In fact, when you take into account the fact that the main distortion component is the subjectively benign second harmonic (fig.9), it is not surprising that the Prima Luna sounds more powerful than its measurements suggest.
Fig.7 Prima Luna Prologue One, 8 ohm tap, THD+N (%) vs frequency at 5V into (from bottom to top): 16 ohms, 8 ohms, 4 ohms, 2 ohms.
Fig.8 Prima Luna Prologue One, 4 ohm tap, THD+N (%) vs frequency at 5V into (from bottom to top): 16 ohms, 8 ohms, 4 ohms, 2 ohms.
Fig.9 Prima Luna Prologue One, 8 ohm tap, 1kHz waveform at 1W into 8 ohms (top), 0.3% THD+N; distortion and noise waveform with fundamental notched out (bottom, not to scale).
At low frequencies, the onset of saturation of the output transformer core adds odd-order harmonics to the amplifier's sonic signature (fig.10), but note that the 120Hz power-supply component is still 90dB down in this graph. Only when I measured the Prologue One's high-frequency intermodulation at a power just below visible waveform clipping on the oscilloscope screen did the amplifier begin to run out of steam, with power-supply components joining the intermodulation spuriae (fig.11). Even so, the level of intermodulation is not catastrophically bad, as it can be with some single-ended designs.
Fig.10 Prima Luna Prologue One, 8 ohm tap, spectrum of 50Hz sinewave, DC–1kHz, at 3W into 8 ohms (linear frequency scale).
Fig.11 Prima Luna Prologue One, 4 ohm tap, HF intermodulation spectrum, DC–24kHz, 19+20kHz at 25W peak into 4 ohms (linear frequency scale).
Considering its $1k price, the nicely made Prima Luna Prologue One offers relatively respectable measured performance. No, it will not rival a competent solid-state design on the test bench, but neither does it do anything to be ashamed of.—John Atkinson
Footnote 1: Some readers have asked why the usual Gibbs' Phenomenon "ringing" from the 'scope's digital antialiasing filter is not evident on my squarewave graphs. There is no "ringing" because the 'scope, a 1987-vintage 8-bit Heathkit, has no low-pass filtering on its input. I make sure that the picture is not obscured by aliasing by using a very high sampling frequency.—John Atkinson