Sutherland Engineering 20/20 phono preamplifier Measurements
Sidebar 3: Measurements
I measured the Sutherland 20/20 with Stereophile's loan sample of the Audio Precision SYS2722 system (see the January 2008 "As We See It" and www.ap.com). Floating the unbalanced generator gave the lowest level of measured noise; although I tried connecting the chassis of the Audio Precision and the Sutherland with a separate wire, this didn't improve the signal/noise ratios.
But before I could test the Sutherland, I had to power it up, which had me scratching my head. Although each channel has its own 48V power supplya single-ended, switch-mode wall wartI could see nowhere to plug these into the 20/20 itself. Removing the top panel after undoing the four thumbscrews, I could see that each channel's module had a power-input jack at the other end from the input and output jacks. The gain and load impedances for each channel are set by printed-circuit-board jumpers by the input jack, and the circuitry is based on three high-performance chips from Texas Instruments: an INA163 low-noise, low-distortion instrumentation amplifier, a Burr-Brown OPA134 FET-input op-amp, and a Burr-Brown OPA627 op-amp, the last used as a DC servo amplifier with high-quality, 1% tolerance Rel-Caps used for EQ.
The 20/20 can be set to five different voltage gains, from 40 to 64dB in four steps of 6dB each. The measured gain was the same as that specified at all settings, within 0.5dB. The load impedance can be set to 47k ohms, for MM cartridges; or, for MC cartridges, to 1k, 475, 200, or 100 ohms. I measured 101, 201, 475, and 993 ohms across the audioband for the four lower settings; with the impedance set to 47k ohms, the impedance varied from 77k ohms at 20Hz to 41k ohms at 1kHz to 37k ohms at 20kHz. The output impedance was a low 198 ohms at all audio frequencies.
The 20/20's channels matched superbly well in both gain and frequency response, which can be seen from the plot of its departure from perfect RIAA equalization (fig.1). Other than a small, 0.2dB dip in the mid-treble, the RIAA error is nonexistent. However, the rise in response above 15kHz shows that Ron Sutherland wanted the 20/20's ultrasonic correction to conform to the so-called "Neumann fourth pole," which is intended to compensate for the cutter head's departure from a perfect preemphasis curve.1
Fig.1 Sutherland 20/20, response with RIAA correction (left channel blue, right red; 1dB/vertical div.).
The Sutherland's construction is completely dual-mono; crosstalk between the channels was unmeasurable. With the input shorted and the gain set to 40dB for MM cartridges, the 20/20's unweighted, wideband S/N ratio (ref.1kHz at 5mV input) was excellent, at 79.4dB. This improved to 87.9dB when A-weighted. In the worst case, with 64dB of gain, the unweighted, wideband S/N ratio was a still-good 64dB left and 68.2dB right, both figures referred to 1kHz at 500µV. Again, A-weighting the measurement gave the expected improvement, to 73dB left and 74dB right.
Overload margins with 40dB gain, ref. 1kHz at 5mV input, were excellent, at 25.4dB at low and midrange frequencies, and a still-good 17.4dB at 20kHz. With 64dB gain, and ref. 1kHz at 500µV, the margins were only slightly worse, at 22dB at low and midrange frequencies and 14.3dB at 20kHz. At 1% THD+noise, the output level was 9.5V at low and midrange frequencies, but 3.9V at 20kHz.
At more typical input levels, the distortion harmonics in the 20/20's output were all below the noise floor (fig.2), though some low-level enharmonic spuriae can be seen, particularly in the right channel. I have no idea where these come from, unless they are residual components from the switching power supply. (As the supply is single-ended, the audio circuitry floats midway between the two rails, with the signal ground apparently defined by a resistive ladder. You would have thought, therefore, that any spuriae riding on the supply rails would be rejected by the circuit's immunity to common-mode noise, as well as aggressively filtered by the multiple large capacitors used in each 20/20 circuit module.)
Fig.2 Sutherland 20/20, MM 40dB gain, spectrum of 50Hz sinewave, DC1kHz, at 1.1V into 600 ohms (left channel blue, right red; linear frequency scale).
To see any harmonics at all, I had to raise the input level at 1kHz to 66mV, which gave an output of 6V into 100k ohms (fig.3). Again the low-level enharmonic spuriae can be seen, but the two highest-level harmonics, the second and third, are still negligible at 118 and 121dB, respectively. This result was not made significantly worse by changing to the demanding 600 ohm load. Intermodulation distortion (not shown) was fairly low, with the 1kHz difference component resulting from a high-level mix of 19 and 20kHz tones (just below visible waveform clipping on the 'scope) lying at 76dB (0.015%).
Fig.3 Sutherland 20/20, MM 40dB gain, spectrum of 1kHz sinewave, DC10kHz, at 6V into 100k ohms (left channel blue, right red; linear frequency scale).
Like all of Ron Sutherland's designs, the 20/20 is well engineered, with low levels of noise. The presence of those enharmonic products did bother me a little, however, even if they lie at a very low level in absolute terms.John Atkinson
Footnote 1: See Keith Howard's "Cut and Thrust: RIAA LP Equalization," in the March 2009 Stereophile.