Lindell AMPX power amplifier Measurements
Sidebar 3: Measurements
Given that he's an experienced record producer, it's not surprising that John Marks keeps an eye on the world of pro audio. In his December "The Fifth Element" column, he enthused over the performance of the Lindell AMPX power amplifier ($1599) from Sweden, which has a class-A output stage and offers 40W into 8 ohms. "The AMPX was unfailingly musical, sounding both powerful and revealing," JM wrote. Some aspects of its performance reminded him of Plinius's 8150 integrated amplifier: "a liveliness that was never undisciplined or 'electronic,' and its top-to-bottom coherence." "A great amp and an amazing bargain," he concluded.
John shipped the Lindell AMPX (serial no. 54046) to me so that I could look at its performance on the test bench. I performed a full set of measurements on the AMPX, using Stereophile's loan sample of the top-of-the-line Audio Precision SYS2722 system (see www.ap.com and the January 2008 "As We See It").
Before performing any measurements, I ran the AMPX for an hour at one-third its specified maximum power of 40W into 8 ohms. Thermally, this is the worst case for an amplifier with a class-B or -AB output stage. However, the more power you ask a class-A amplifier to deliver, the cooler it runs. Nevertheless, following the preconditioning, the top panel's temperature measured 104°F (39.9°C) and the heatsinks 114.3°F (45.8°C). The distortion remained at 0.028% throughout this period. During normal use, the heatsink temperature rose to 125.5°F (52°C), the top panel to 116°F (46.5°C). This amplifier will need to be used in a well-ventilated position.
The voltage gain at 1kHz was significantly lower than the norm, at just 20.1dB into 8 ohms. The AMPX preserved absolute polarity (ie, was non-inverting), and the input impedance was a constant 42.2k ohms across the audioband, this inconsequentially lower than the specified 47k ohms. The output impedance was a little higher than is usual for a solid-state design: 0.16 ohm at low and middle frequencies, rising to 0.21 ohm at the top of the audioband. This resulted in mild response variations of ±0.15dB with our standard simulated loudspeaker (fig.1, gray trace). The Lindell's response was well matched between channels, and flat within the audioband into 8 ohms (fig.1, blue and red traces) and 4 ohms (cyan, magenta), but began to roll off prematurely into 2 ohms (green). Into higher impedances, however, the shape of a 10kHz squarewave (fig.2) was well defined, with short risetimes but also with a small degree of overshoot and one cycle of critically damped ringing. This was also apparent with a 1kHz squarewave (fig.3).
Channel separation (not shown) was an even 70dB at low and middle frequencies, decreasing slightly to 60dB at 20kHz. The AMPX's unweighted, wideband signal/noise ratio, taken with the input short-circuited and referenced to 1W into 8 ohms, was modest, at 62.4dB left and 62.1dB right. These ratios improved to 76.5dB when the measurement bandwidth was restricted to 22Hz22kHz, and further improved, to 82.9dB, when A-weighted. The spectrum in fig.4 was taken with the amplifier driving a 1kHz tone at 25Wpc into 8 ohms. The primary noise components are 60Hz and its odd harmonics, presumably due to magnetic interference from the toroidal power transformer. The full-waverectified spuriae from the power supply at 120Hz and its harmonics are all 10dB lower in level.
Specified as having a maximum power delivery of 20Wpc into 8 ohms (16dBW), the AMPX easily exceeded that specification, clipping (defined as when the THD+noise reaches 1%) at 35Wpc into 8 ohms with both channels driven (15.4dBW, fig.5). The minimum distortion in this graph is very low, but is a little higher into 4 ohms, as seen in fig.6, which reveals that the AMPX clips at 60Wpc into 4 ohms with both channels driven (14.8dBW).
I didn't measure the Lindell's clipping power into 2 ohms, as fig.7 indicates that the amplifier is not comfortable driving 2 ohms. This graph plots the THD+noise percentage against frequency at 6.33V, equivalent to 5Wpc into 8 ohms, 10Wpc into 4 ohms, and 20Wpc into 2 ohms. The amplifier clearly offers low distortion into 8 ohms (blue and red traces) and 4 ohms (cyan, magenta), but becomes less linear into 2 ohms (gray). A rise in THD at higher frequencies is also evident in this graph, due to the circuit's limited open-loop gain-bandwidth product reducing the amount of available corrective negative feedback as the frequency rises.
Fortunately, the distortion is primarily the subjectively innocuous low orders in nature (fig.8). The third harmonic dominates, though the second harmonic is higher in the left channel (fig.9, blue trace) than the right (red). Even at a level just below visible clipping into 8 ohms on the oscilloscope screen, intermodulation distortion is relatively low in level (fig.10), with the second-order difference product at 1kHz lying at 70dB (0.03%), and the higher-order components at 18 and 21kHz lying at 66dB (0.05%).
I used to be a big fan of amplifiers with an output stage operating in class-A. However, I have come to realize that the class of output-stage operation is not necessarily the primary gauge of amplifier sound quality. Even so, Lindell's AMPX measures well, provided it is not used to drive speakers with impedances that drop significantly below 4 ohms.John Atkinson