Columns Retired Columns & Blogs |
McIntosh C1000 preamplifier system Measurements
Sidebar 3: Measurements
I performed full sets of measurements on both McIntosh preamplifiers, for both balanced and unbalanced operation, but the two units turned out to be electrically so similar that I've combined my comments on both into a single sidebar.
The phono stages, of course, are unbalanced only; I examined their behavior at each preamp's Processor Out jacks. The MM stages of the C1000P and C1000T are noninverting and offered a voltage gain of 40.1dB and an input impedance of 47k ohms across most of the audioband, the latter dropping inconsequentially to 42.5k ohms at 20kHz. Both MC stages were also noninverting and offered 60.1dB of voltage gain at 1kHz, with an input impedance of 30 ohms when set to "25" and 960 ohms when set to "1000."
The C1000P's RIAA error (fig.1) showed a very slight excess of energy in the upper bass and above the audioband, but was otherwise flat. The C1000T's RIAA error (fig.2) was very similar at audible frequencies, but featured an increasingly positive error at ultrasonic frequencies. This suggests that the tube preamp's phono-stage gain asymptotically approaches unity with rising frequency rather than, as I prefer to see, continuing to decrease. This behavior will not in itself be audible, but the late John Linsley-Hood suggested a quarter century ago that it might increase the audibility of record clicks, particularly with MC cartridges. Why would the two McIntosh preamps differ in this respect? I can surmise only that the more limited voltage gain available from tubes meant that a different circuit topology had to be used.
Fig.1 McIntosh C1000P, MM stage RIAA error (0.5dB/vertical div., right channel dashed).
Fig.2 McIntosh C1000T, MM stage RIAA error (0.5dB/vertical div., right channel dashed).
At normal levels, the phono-stages' THD+noise was around 0.017%, which is excellent. Both tubed and solid-state phono stages excelled in a potentially more important area: their immunity to overload. The MM and MC stages of both the C1000P and C1000T offered a superb 27dB or greater overload margin at all audio frequencies (ref. 5mV, MM, and 500µV, MC, both at 1kHz). The McIntosh preamps' phono stages are also extremely quiet—the unweighted audioband signal/noise ratios were 78.5dB (C1000P) and 69.4dB (C1000T) from the preamps' MM inputs, and only slightly lower from their MC inputs—meaning that, with their very high overload margins, they have commendably high dynamic range, especially the solid-state C1000P. One thing should be noted about the C1000T: When I was exploring the point at which its phono-stage output clipped at 20kHz, I experienced some latching; ie, once the circuit had clipped, it stayed clipped even after I'd reduced the input-signal level. However, as the output level was around 12V when this happened, it should have no practical consequences.
The C1000P's and C1000T's line stages offered a maximum of 15dB of voltage gain in both balanced and unbalanced operation, with 0dB on the meters (which come after the volume control) indicating when the preamp is reaching the limit of its dynamic range capability. The input impedance of both preamps was 16.5k ohms across most of the audioband for the unbalanced inputs, 25k ohms for the balanced, these decreasing slightly but inconsequentially at 20kHz. All inputs preserved absolute polarity (ie, were noninverting), the XLRs being wired with pin 2 hot.
Both preamplifier line stages offered the identical wide frequency response into 100k ohms (fig.3, top pair of traces), with a slightly early rolloff of low-bass frequencies into the punishing 600 ohm load (fig.3, bottom traces). The response was identical from both the balanced and unbalanced jacks and at all volume-control settings. Any crosstalk was buried under the noise (fig.4 shows the channel separation for the C1000T) and the noise floor was very low in level—particularly for the C1000P, which, with its input shorted but its volume control set to its maximum, gave an A-weighted S/N ratio (ref.1V output) of 116dB! The C1000T was not in that class but was still good, with an unweighted wideband measurement of 76dB improving to 88dB when A-weighted.
Fig.3 McIntosh C1000P, balanced frequency response at 1V into 100k ohms (top below 200Hz) and 600 ohms with volume control at "100" (0.5dB/vertical div., right channel dashed).
Fig.4 McIntosh C1000T, channel separation, balanced operation (bottom) and unbalanced (top). (10dB/vertical div., R–L dashed)
The C1000P and C1000T could both swing very high output voltages. Figs.5 and 6 show how the percentage of THD+noise present in the balanced outputs of the 'P and 'T, respectively, changed with output voltage into 100k ohms and 600 ohms. The solid-state preamp clipped (1% THD+N) at 32.5V into the higher impedance and at 18V into the low load, with the maximum unbalanced output half these figures, as expected. The tube preamp clipped a little earlier, at 20V into 100k ohms and 11.3V into 600 ohms, but the difference is academic—any power amplifier will be driven well into overload at these levels.
Fig.5 McIntosh C1000P, balanced distortion (%)vs 1kHz output level into (from bottom to top at 1V): 100k and 600 ohms.
Fig.6 McIntosh C1000T, balanced distortion (%)vs 1kHz output level into (from bottom to top at 1V): 100k and 600 ohms.
The downward slope below 10V of the traces in figs.5 and 6 suggests that actual distortion is buried under the already low noise at normal signal levels. To examine how the THD+N percentage changes with frequency, I therefore chose a high output level: 10V for the C1000P (fig.7) and 5V for the C1000T (fig.8). Into high impedances, the THD+N remains constant with frequency in the audioband for both the 'P and 'T, the former offering a very low 0.001% or less compared with the latter's still-low 0.016%. Into 600 ohms, however, these graphs show that the actual THD rises at both very low and very high frequencies, though not to any extent that might even begin to be audible.
Fig.7 McIntosh C1000P, balanced THD+N (%)vs frequency at 1V into (from bottom to top): 100k and 600 ohms (right channel dashed).
Fig.8 McIntosh C1000T, balanced THD+N (%)vs frequency at 1V into (from bottom to top): 100k and 600 ohms (right channel dashed).
At normal levels, spectral analysis reveals that the harmonic content is vanishingly low for both preamps: 0.0005% (true sum of the harmonics; fig.9, C1000P; fig.10, C1000T). The increase in THD at low frequencies and high levels into 600 ohms was due primarily to an increase in the subjectively innocuous third harmonic (fig.11), though at –88dB (0.006%) this won't bother any human listener. Intermodulation distortion was also essentially nonexistent at normal levels; I had to brutalize the preamps by driving them close to clipping into 600 ohms to get any significant evidence of high-order products (fig.12), and even then, the distortion products remained at or below 80dB (0.01%)!
Fig.9 McIntosh C1000P, unbalanced spectrum of 1kHz sinewave, DC–10kHz, at 1V into 8k ohms (linear frequency scale).
Fig.10 McIntosh C1000T, unbalanced spectrum of 1kHz sinewave, DC–10kHz, at 1V into 8k ohms (linear frequency scale).
Fig.11 McIntosh C1000P, balanced spectrum of 50Hz sinewave, DC–1kHz, at 10V into 600 ohms (linear frequency scale).
Fig.12 McIntosh C1000P, balanced HF intermodulation spectrum, DC–24kHz, 19+20kHz at 15V peak into 600 ohms (linear frequency scale).
The measured performance of these two McIntosh preamplifiers reveals some superb audio engineering, with no compromise evident for the tubed version apart from some slightly higher but still low noise levels and that ultrasonic RIAA error. It is very satisfying to measure such products, and, as Mikey found, equally satisfying to listen to them.—John Atkinson
- Log in or register to post comments