Conrad-Johnson MF-200 power amplifier Measurements
Sidebar 3: Measurements
Rated at 200W into 8 ohms, this is a power level of 23dBW referred to 0dB, 1W, 8 ohms. The MF-200 raised a solid 250W (24dBW) with an exceptionally good power bandwidth. No diminution was recorded at 20Hz, while at 20kHz the power loss was barely 0.1dB. Protected by modest supply-rail fuses of 5A, the extra current demanded at 20kHz into 4 ohms blew the fuse, while the results for 20Hz and 1kHz were 22.7dBW at this lower impedance, representing a true power of 373Wpc. This arduous test was made with both channels driven. In view of the fine result at 20Hz, the power-supply reservoir capacity was clearly quite generous.
On peak signals the 8 ohm power rose to 263W or 24.2dBW; these fine results were available regardless of the more demanding nature of the 50Hz supply used in the UK. One could not fault the '200 for load tolerance; it managed a peak level of 23dBW into 2 ohms, a true power of 800W. Backing this was the moderate output impedance of 0.15 ohms (rising to 0.25 ohms at 20kHz) and the peak current capacity of ±28.5A, hardly what one might expect from a low-feedback FET design.
Some evidence of the latter was forthcoming in the results for distortion, however, which were rather higher than those found in most high-end, solid-state power amplifiers. At rated power the distortion readings were satisfactory at 53 or 54dB, or 0.22%. Surprisingly, the results were very similar over the whole frequency range, and even for the two-tone (19:20kHz, 1:1) high-frequency intermodulation, where the 1kHz product measured at 53dB. Examination of the distortion spectra showed a fair balance of even and odd harmonics decaying with increasing frequencya tube-like character. At 1W, the distortion had improved by a factor of 10, or 20dB, roughly in proportion to the reduction in power level. Excepting a reading of 65dB or 0.56% at 20kHz, all distortion readings were fine at typically 70dB, or 0.02%.
I felt some further exploration was in order in the light of the dual nature of this amplifier's sound quality when judged at low and high power levels. Fig.1 shows four traces, the lower pair relating to the right-hand scale and plotting input level in dBV vs watts, for 4 ohm and 8 ohm loads. These reveal good power linearity to beyond 200W (this would be a straight line for a log scale in watts). More relevant are the upper two curves, which show distortion vs level again for 4 and 8 ohm loads. Between 10 and 5dBV, corresponding to an indicated power level of a few watts, there was a distinct change in this amplifier's linearity, exaggerated in the poorer upper curve, which is the result with the MF-200 loaded with 4 ohms. The other slope features were easily explained; for example, the rise at very low levels is simply due to residual system noise, while the abrupt slope change at 6dBV input denotes the onset of clipping. The 5dBV point, approximately 20dB below full power, marks a transition between two regimes, modified by loading and hence, for example, by impedance variations with frequency in a loudspeaker crossover.
Another useful test can reveal interference between power-supply components and the audio output, the so-called "supply modulation." Here the amplifier is stressed by the application of a 4-ohm load at two-thirds the rated 8-ohm output level, so as to draw substantial current. The input frequency is deliberately offset to 35Hz to allow visual separation of the 50Hz (UK) supply harmonics. Fig.2 shows this result where the natural harmonics of the 35Hz fundamental appear in progressively decaying order. A trace of 50Hz can be seen at 77dB down, but note that no higher supply harmonics100, 150, 200, and 250Hzare present even at the sensitive 95dB threshold seen for this measurement. No sidebands are evident either; I consider this a good result.
I investigated the transient and stability performance, rating the amplifier as unconditionally stable, with an obvious suitability for electrostatic speakers. The 1kHz squarewave performance (fig.3) for a difficult load8 ohms in parallel with 2µFshows that a fast risetime was maintained with strong damping of the usual resultant ringing. This graph also reveals the amplifier's excellent low-frequency response by the excellently flat tops. The amplifier was absolute phase-correct, DC coupled at the output, and of wide response, +0, 0.5dB 3.1Hz to 32kHz; +0, 3dB, 1.25Hz to 156kHz.
Channel separation was ample, typically 80dB, falling to 59dB by 20kHz. Balance held to 0.13dB tolerance, and the S/N ratio figures were satisfactory108dB ref. full power, 22Hz22kHzwith a very low hum content. The power transformer was also mechanically quiet; some hiss might just be audible on very sensitive speakers of over 93dB/W sensitivity.
The input loading was modest at 50k ohms, while a 2.3V input was required for full leveltoo low a sensitivity for direct connection of a passive line controller, even with CD players of 2V maximum output. A 0.7V sensitivity is ideally required for such applications. DC offset measured at the output was fine at less than 12mV, and showed negligible drift. Mild "crossover" distortion was evident at medium levels, but was of a beneficially low harmonic order thanks to the low-feedback design.
Taken overall, the test results were very good, particularly the high power available and the good load tolerance. A question remains concerning the "two-speed" nature of the tonal quality and its possible associations with the slope change in distortion at moderate output levels.Martin Colloms