Sidebar 3: Measurements
I performed a complete set of measurements on one of the two Goldmund Telos 2800 monoblocks sharing serial number T2_LN000023 with my Audio Precision SYS2722 system. I preconditioned the amplifier by operating it at 1/8 power into 8 ohms for 20 minutes. At the end of that time, the top panel's temperature was 89.8°F/32.1°C. The temperature of the hottest section of the heatsink on the rear panel was 108.9°F/42.8°C.
The amplifier preserved absolute polarity, ie, was noninverting, with both the single-ended and balanced inputs. (The XLR jack is wired with pin 2 positive, the IEC standard.) The balanced input impedance was 19k ohms at 20Hz and 1kHz, dropping inconsequentially to 16k ohms at 20kHz. As expected, the unbalanced input impedances were half those values.
Goldmund specifies the Telos 2800's gain as 35dB. With the Menu's Gain Adjust setting of –6dB, which was how the amplifier had been set when first powered up, the balanced input's gain into 8 ohms was 28.9dB. Setting the Gain Adjust to 0dB, both the balanced and single-ended inputs offered gains of 34.9dB into 8 ohms. Unusually, the Telos 2800 has a coaxial S/PDIF digital input. This locked to data sampled at rates up to 192kHz and preserved absolute polarity. With the Gain Adjust set to 0dB, a 1kHz tone at –20dBFS resulted in an output voltage of 11.62V into 8 ohms. This is 12.5dB below the output voltage at the specified maximum power of 300W into 8 ohms. Full-scale digital data should not be sent to the Goldmund's digital input unless the Gain Adjust is set to the minimum value of –9dB.
The Goldmund Telos 2800 offers high power even into low impedances and at high frequencies. While I was surprised by the presence of crossover distortion and the circuit's restricted open-loop bandwidth, the measured consequences of this behavior are extremely low in level. I can confidently say that they will not have audible consequences.—John Atkinson
Fig.1 Goldmund Telos 2800, balanced input, frequency response at 2.83V into 8 ohms (blue), 4 ohms (magenta), and 2 ohms (red), and unbalanced input into 8 ohms (gray) (1dB/vertical div.).
Fig.2 Goldmund Telos 2800, balanced input, small-signal 10kHz squarewave into 8 ohms.
Fig.3 Goldmund Telos 2800, unbalanced input, small-signal 10kHz squarewave into 8 ohms.
The Telos 2800's output impedance was very low, at 0.02 ohms at 20Hz and 1kHz, rising slightly to 0.045 ohms at 20kHz. The variation in the frequency response with our standard simulated loudspeaker (not shown) was therefore negligible. The response into resistive loads was flat in the audioband, rolling off above 20kHz, with the output into 8 ohms down by 3dB at 200kHz (fig.1, blue trace) and into 2 ohms at 125kHz (red trace). The Goldmund's reproduction of a 10kHz squarewave into 8 ohms (fig.2) featured very short risetimes and, commendably, no overshoot or ringing. Fig.2 and the blue, magenta, and red traces in fig.1 were taken with the balanced input. The single-ended input's response into 8 ohms (fig.1. gray trace) was flat up to 75kHz then started to rise, reaching +1dB at 200kHz, which resulted in a small amount of overshoot on the amplifier's reproduction of a 10kHz squarewave (fig.3).
Fig.4 Goldmund Telos 2800, balanced input, spectrum of 1kHz sinewave, DC–1kHz, at 1W into 8 ohms (linear frequency scale).
The unweighted, wideband signal/noise ratio (ref. 1W into 8 ohms), taken with the unbalanced input shorted to ground and its Gain Adjust set to 0dB, was 51.2dB. This ratio improved to 74.7dB when the measurement bandwidth was restricted to 22Hz–22kHz, and to 85.0dB when A-weighted. The S/N ratios with the balanced input shorted to ground and the Gain Adjust set to –6dB were 2–4dB greater. Spectral analysis of the low-frequency noisefloor while the Goldmund drove a 1kHz tone at 1W into 8 ohms revealed a fairly low random noisefloor, though spuriae at 60Hz and its odd-order harmonics were present (fig.4), presumably due to magnetic interference from the power transformer.
Fig.5 Goldmund Telos 2800, distortion (%) vs 1kHz continuous output power into 8 ohms.
Fig.6 Goldmund Telos 2800, distortion (%) vs 1kHz continuous output power into 4 ohms.
The Telos 2800's maximum power is specified as 300W into 8 ohms (24.77dB) and 530W into 4 ohms (24.23dBW). Stereophile defines an amplifier's clipping power as being when the THD+noise reaches 1%: With a 1kHz signal, the amplifier clipped at 350W into 8 ohms (24.84dBW, fig.5) and 530W into 4 ohms (24.23dB, fig.6). (The AC wall voltage had dropped from 118.3V with the amplifier idling to 116.3V when the amplifier clipped into 4 ohms.) The THD+noise percentage falls with increasing power up to 60W into both loads, which means that actual distortion lies beneath the noisefloor at low-to-moderate powers. The FTC's updated "Amplifier Rule" states that maximum power should also be assessed at frequencies other than 1kHz; therefore, I repeated the power test at 20kHz. The Telos 2800 clipped at the specified 300W into 8 ohms with this signal.
Fig.7 Goldmund Telos 2800, THD+N (%) vs frequency at 20V into: 8 ohms (blue) and 4 ohms (red).
Fig.7 shows how the THD+N percentage varied with frequency at 20V, which is equivalent to 50W into 8 ohms (blue trace) and 100W into 4 ohms (red). The THD+N percentage was extremely low in the bass and midrange but started to rise in the low treble, reaching 0.13% into 8 ohms and 0.2% into 4 ohms. This behavior will be due to the amplifier's circuit having a limited open-loop bandwidth, which means that less corrective negative feedback is available as the frequency increases.
Fig.8 Goldmund Telos 2800, 1kHz waveform at 50W into 8 ohms, 0.0092% THD+N (top); distortion and noise waveform with fundamental notched out (bottom, not to scale).
Fig.9 Goldmund Telos 2800, 1kHz waveform at 1W into 8 ohms, 0.034% THD+N (top); distortion and noise waveform with fundamental notched out (bottom, not to scale).
The distortion signature was predominantly the second harmonic (fig.8, bottom trace), but the distortion waveform has spikes at the sinewave's zero-crossing points indicating crossover distortion. Fig.8 was taken at 50W into 8 ohms; I repeated this test at 10W and 1W (fig.9). The crossover distortion spikes were present at these lower powers, too, which suggests that even though the level of this distortion is very low, the output stage bias current is lower than optimal.
Fig.10 Goldmund Telos 2800, spectrum of 50Hz sinewave, DC–1kHz, at 100W into 4 ohms (linear frequency scale).
Fig.11 Goldmund Telos 2800, HF intermodulation spectrum, DC–30kHz, 19+20kHz at 100W peak into 4 ohms (linear frequency scale).
Spectral analysis with a 50Hz tone at 100W into 4 ohms (fig.9) shows that the second harmonic lay at a very low –100dB (0.001%), joined by the fifth harmonic at the same low level and, unusually, the third harmonic of the 60Hz supply frequency. The second harmonic was higher with a 1kHz tone but still lay at a low –84dB (0.005%, not shown). With the increased distortion at high frequencies shown in fig.7, more intermodulation products were present with an equal mix of 19 and 20kHz tones than I usually find (fig.10), which was taken at 100W into 4 ohms. Nevertheless, the difference product at 1kHz lay close to –90dB (0.003%), with the higher-order products just 10dB higher in level.
