Chord SPM 14000 Ultimate monoblock power amplifier Measurements
I preconditioned the big Chord SPM 14000 by running it at 172W into 8 ohms for an hour. (This is less than one third the specified power; see later.) The measured THD+noise percentage at the beginning of that period was 0.0108%. After 20 minutes, some high-order components appeared in the distortion waveform, still at a very low level, but these subsided at the 30-minute mark, with the final THD level at 0.015%, this almost pure third harmonic. The rear heatsinks were around 140°F (60°C) at the end of the hour, the side panels cooler than this.
The Chord's voltage gain was a high 29.75dB into 8 ohms, this the same for both unbalanced and balanced drive. Neither of the left-hand input jacks (LH as seen from the rear of the amplifier) inverted signal polarity; both of the right-hand jacks did so. (The LH XLR appears to be wired with pin 2 positive, the RH jack with pin 3 positive.) Given that the two RCA jacks are wired with opposite polarity, it might be thought that inserting a shorting plug into the unused jack would give the lowest possible noise with unbalanced drive. When I tried that, however, the amplifier turned itself off.
The unbalanced input impedance was a usefully high 52k ohms in the bass and midrange, this dropping slightly at 20kHz, to 35k ohms. The balanced input impedance was a constant 104k ohms across the audioband. The output impedance, including 6' of speaker cable and taken from one of the pairs of WBT binding posts, which I assumed were in parallel—the amp's terminal panel was devoid of any legends—was 0.23 ohm at all audio frequencies. This resulted in a moderate ±0.25dB of frequency-response modification with our standard dummy loudspeaker (fig.1, top trace at 2kHz). The Chord's small-signal bandwidth was wide, with a high-frequency –3dB point of 111kHz, correlating with the excellent 10kHz squarewave reproduction (fig.2). But as also can be seen from fig.1, the amplifier's bandwidth decreased with decreasing load impedance, the output being 0.7dB down at 20kHz into 2 ohms. This graph was taken using balanced drive; the unbalanced set of responses was identical.
Fig.1 Chord SPM 14000, balanced frequency response at 2.83V into (from top to bottom at 2kHz): simulated loudspeaker load, 8, 4, 2 ohms (1dB/vertical div., right channel dashed).
Fig.2 Chord SPM 14000, small-signal 10kHz squarewave into 8 ohms.
The level of the SPM 14000's noise floor depended on whether it was being driven by balanced or unbalanced signals. With balanced drive, I measured a moderate signal/noise ratio of 64.2dB, ref. 1W into 8 ohms. This was a wideband, unweighted figure; the A-weighted ratio was a good 79.3dB. However, with unbalanced drive, while the A-weighted ratio decreased only slightly, to 76.3dB, the unweighted, wideband figure was significantly worse than balanced, at 49.75dB. This was why I had assumed that the undriven, inverting RCA jack should have a shorting plug inserted, as some balanced amplifiers require the unused signal phase to be referenced to ground. But as I wrote earlier, the Chord went into protection when I tried this. As PB didn't remark on any issues with background noise, I assume he used the VTL TL-7.5 preamplifier's balanced outputs to feed the SPM 14000s.
To my surprise, though the Chord is definitely a powerhouse of an amplifier, it didn't quite meet its specified power at clipping (defined as 1% THD+N). Into 8 ohms, I measured 525W (27.2dBW), which is almost 3dB below the specified power of 1kW and just over 1dB less than what I had earlier understood to be the specified power, 700W (28.45dB). The shortfall is probably insignificant in my opinion, though I did check by repeating the measurement the next day with the amplifier cold rather than hot (the result was the same). It is also relevant to note that I don't hold my AC wall supply constant with a Variac for my power testing. With no signal, the wall voltage in my test lab was 125V; this dropped to a still high 121V with the amplifier clipping into 8 ohms, 119V with it clipping into 4 ohms, and 117V into 2 ohms. The 4 ohm clipping power was 980W (26.9dBW), the 2 ohm power 1450W (25.6dBW)—as I said, the amplifier is still a powerhouse!
The manner in which the distortion percentage changes with output power is shown in fig.3. The downward slopes of the traces below a few tens of watts in this graph indicate that the THD+N percentage is dominated by noise at these levels. Only above deliveries of 80W or so into 8 and 4 ohms does the Chord's true distortion start to rise above the background noise. Into 2 ohms, however, the THD rises above 30W, which suggests that the amplifier is working hard into this load. Even so, the THD remains well below 0.1% for almost all its working power range.
Fig.3 Chord SPM 14000, distortion (%)vs 1kHz continuous output power into (from bottom to top at 1W): 8, 4, 2 ohms.
Because of the very low level of true THD at small-signal levels, I plotted the Chord's THD+N against frequency at 20V output, equivalent to 50W into 8 ohms. The results are shown in fig.4. The 8 ohm trace in this graph correlates with the result in fig.3; however, both the 4 ohm and 2 ohm traces are significantly worse. I have no idea what these are due to, other than the fact that the amplifier was very hot for these tests, while it had cooled down considerably by the time I performed the low-impedance clipping measurements, which I do at the very end of the testing, in case the amp being tested breaks. The THD rises above the audioband into all three loads, presumably due to the decreasing gain margin at ultrasonic frequencies rendering negative feedback less effective. But again, the THD+N percentage is well below 0.01% in the audioband.
Fig.4 Chord SPM 14000, THD+N (%)vs frequency at 20V into (from bottom to top): 8, 4, 2 ohms.
Even at high levels into low impedances, that distortion is the subjectively benign third harmonic (fig.5). Looking at the spectrum of a high-power, low-frequency tone (fig.6), the third harmonic lies at –84dB (0.02%), with the second harmonic next highest in level below –90dB (0.01%). All other harmonics are below –100dB. Keeping the signal level the same but halving the load impedance gives the spectrum shown in fig.7: the second harmonic has risen by 6dB, the third by 12dB. Note also that a 120Hz component just touches the –100dB line. I couldn't eliminate this by changing the grounding between my Audio Precision analyzer and the amplifier. Nevertheless, it is negligible.
Fig.5 Chord SPM 14000, 1kHz waveform at 230W into 4 ohms (top), 0.035% THD+N; distortion and noise waveform with fundamental notched out (bottom, not to scale).
Fig.6 Chord SPM 14000, spectrum of 50Hz sinewave, DC–1kHz, at 235W into 8 ohms (linear frequency scale).
Fig.7 Chord SPM 14000, spectrum of 50Hz sinewave, DC–1kHz, at 450W into 4 ohms (linear frequency scale).
With its decreasing linearity at the top of the audioband, I wasn't surprised to see the Chord produce a 1kHz difference component at –70dB when driving a mix of 19 and 20kHz tones into 8 ohms close to visible clipping on the oscilloscope (fig.8). Halving the load impedance (not shown) didn't change the level of the 1kHz product, but did change the level of the higher-order products at 18kHz and 21kHz, to –76dB. Even so, this is very much a worst-case situation; this behavior shouldn't be an issue under normal circumstances.
Fig.8 Chord SPM 14000, HF intermodulation spectrum, DC–24kHz, 19+20kHz at 720W peak into 8 ohms (linear frequency scale).
The slight shortfall in its power delivery aside, the Chord SPM 14000 delivers excellent measured performance.—John Atkinson