NAT Symmetrical line preamplifier Measurements
I used Stereophile's loan sample of the top-of-the-line Audio Precision SYS2722 system (see the January 2008 "As We See It" and www.ap.com) to examine the NAT Symmetrical preamplifier's measured behavior; for some tests, I also used my vintage Audio Precision System One Dual Domain.
The NAT Symmetrical line preamplifier is beautifully constructed, but I had some problems measuring its performance, as the right channel's balanced output was half that of the left's. I reseated all the tubes and ribbon cables, but this didn't fix things. The fact that the right channel's output was exactly half the left's suggests that a tube had failed, resulting in just the positive phase of the balanced signal being present on the output. Bob Reina hadn't noticed anything amiss with the channel balance in his auditioning, so the tube must have failed in shipping. I did examine each channel's performance, but in the main discuss only the properly functioning left channel, on the assumption that this is representative of the preamp's behavior as a whole. It wasn't possible to get another sample to me for measurement in time to meet this review's deadline, unfortunately, nor was it feasible to postpone publication. We will publish a Follow-Up in due course.
The maximum voltage gain for balanced input to balanced output was 18dB compared with the 15dB specified; for unbalanced input to unbalanced output, it was 13.6dB. The unity-gain setting of the volume control was three-quarters of full scale on the front-panel thermometer display. Both sets of inputs preserved absolute polarity (ie, were non-inverting), and the XLRs were wired with pin 2 hot. The input impedance was to specification at low and middle frequencies, at 97k ohms balanced and 51k ohms unbalanced, but dropped a little at 20kHz, to 96k and 42k ohms, respectively. This will have no sound-quality consequences, however.
The Symmetrical's output impedance was slightly higher than specified, but was still low at 197 ohms balanced and 91 ohms unbalanced. These figures were for the high and midrange frequencies; the output impedance did rise a little in the low bass, to 265 ohms balanced and 112 ohms unbalanced. This results in a slight degree of rolloff in the low bass, reaching 2dB at 10Hz (fig.1, red trace). Into 100k ohms (fig.1, blue trace), the response is flat down to below 10Hz, and, at the other end of the spectrum, doesn't reach 3dB until 160kHz. This graph was plotted with the volume control at its maximum; with the control at its unity-gain setting, the bandwidth increased slightly, to 3dB at 200kHz. The unbalanced bandwidth was also that wide, even with the volume control at its maximum.
Fig.1 NAT Symmetrical, volume control at maximum, left-channel balanced frequency response at 1V into 100k ohms (blue) and 600 ohms (red) (0.25dB/vertical div.).
Channel separation (not shown) was better than 100dB below 8kHz, which is superb, while the unweighted, wideband signal/noise ratio (ref. 1V, with the input shorted but the volume control at its maximum) was a reasonably good 74.5dB balanced and 77dB unbalanced. These figures improved to 82.6 and 76.1dB, respectively, when the measurement bandwidth was restricted to the audioband, and to 89.1 and 89.6dB when A-weighted. There was a little more 60Hz hum present in the unbalanced output than the balanced, which is why restricting the measurement bandwidth gave a greater improvement in S/N ratio with the balanced output than with the unbalanced.
Fig.2 shows how the THD+noise percentage in the balanced output changed with output voltage into 100k ohms (right) and 600 ohms (left). Though NAT claims on its website that the Symmetrical line stage is compatible with 600 ohm loads, you can see from fig.2 that the THD+N starts to rise above 0.1% at levels above 1.2V, though the preamp doesn't actually clip (defined as 1% THD) until 3.5V. Into 100k ohms, the distortion remains below the background noise until 1.5V or so, which implies a sensible gain architecture. The NAT preamp's balanced output doesn't clip into high impedances until 16V, which is well above any level it will be required to provide in practical use.
Fig.2 NAT Symmetrical, THD+N (%) vs balanced output voltage at 1kHz into (from bottom to top): 100k, 600 ohms.
I haven't shown the graph for unbalanced operationin this mode, the NAT Symmetrical clipped at just 2V into 100k ohms and at 1V into 600 ohms, with distortion not dropping below 0.1% until the output level was below 100mV. This strongly suggests that the unbalanced mode wasn't operating properly. As Bob Reina exclusively reviewed the Symmetrical from its balanced outputs, this damage may well have occurred before Stereophile was sent this sample (serial no.0906089) for review. Again, this is something I will return to in a Follow-Up, using another sample.
In balanced mode, not only is the NAT's distortion low at typical output levels, it primarily comprises the subjectively innocuous second and third harmonics, at 78dB (0.011%) and 88dB (0.034%), respectively (fig.3, left channel, blue trace). The red trace in fig.3 shows the spectrum of the right channel's output. The level is 6dB lower than the left's, as mentioned earlier, and while the distortion is correspondingly lower, harmonics of the AC power supply can be seen. Sticking with the better-performing left channel but dropping the load to 600 ohms increases the levels of both the second and third harmonics to around 64dB (0.06%), but the spectrum is commendably free from higher-order harmonics (fig.4).
Fig.3 NAT Symmetrical, balanced output, spectrum of 50Hz sinewave, DC1kHz, at 1V into 100k ohms (left channel blue, right red; linear frequency scale).
Fig.4 NAT Symmetrical, balanced output, spectrum of 50Hz sinewave, DC1kHz, at 1V into 600 ohms (left channel blue, linear frequency scale).
Finally, tested with the demanding combination of 19 and 20kHz tones into 100k ohms, the Symmetrical offered relatively low levels of intermodulation distortion (fig.5). Reducing the load impedance to 600 ohms increased both the difference and the higher-order distortion products to close to 60dB (0.1%), but this is still good performance in absolute terms.
Fig.5 NAT Symmetrical, balanced output, HF intermodulation spectrum, DC24kHz, 19+20kHz at 2V peak into 100k ohms (left channel blue, linear frequency scale).
Assuming that the better-performing left channel is typical of the NAT Symmetrical line stage's intrinsic performance, this tube preamp offers excellent measured performance. But we must wait for the second sample for it to get a totally clean bill of health.John Atkinson