- Written by John E. Johnson, Jr.
- Published on 09 September 2009
On the Bench
OK guys, here is where you are going to be amazed.
All distortion measurements were made within an 80 kHz bandwidth, both channels driven, 100 kOhms load, except where noted. For graphs 1, 2, 3, 7, and 8, the left channel is the yellow line and the right channel is the red line. I set the volume control to maximum and adjusted the input signal level to obtain specific output voltages.
At 1 kHz, the second and third harmonics were visible, but nothing beyond that. Distortion was a very low 0.002%.
With a combination input of 19 kHz and 20 kHz sine waves, the 1 kHz B-A was more than 100 dB below the fundamentals, and there were some harmonics in the out-of-band (beyond the audible band) region.
IMD was also very, very low.
THD+N vs. Frequency is shown below, both for 100 kOhm loads and 600 ohm loads. Notice that even with the torture of a 600 ohm load (for a typical power amp, the load would be 10 kOhms up to 100 kOhms), distortion did not go above 0.02% all the way out to 50 kHz. Within the audible band and a 100 kOhm load, distortion stayed below 0.005%. The slight upward trend, starting at about 2 kHz, gives the sound a slight snap to it. In fact, with a high-end preamplifier, this line might be flat, and some consumers don't like that, saying that the sound is "dry". In any case, this is good performance for any preamplifier, let alone one that costs $399.
The THD+N vs. Output curves show that the USP-1 delivered 10 volts rms into a 100 kOhm load at clipping (1% THD+N). It output 9 volts at 0.003% distortion. Using a 600 ohm load, the output was 3.4 volts at 0.002% distortion, before rising to clip at 3.8 volts.
The measured frequency response was 10 Hz - 100 kHz, - 2 dB. The interesting thing here was that with a 600 ohm load, the frequency response at the extreme high end was slightly better. First time I have seen this happen.
I measured the distortion and noise in the phono stage by inputting a 1 kHz sine wave at 1 mv for the MC test and 4 mv for the MM test. The results are shown below. You can see that the noise level dropped considerably with the MM setting. That is because there is no extra gain stage in that circuit. MC cartridges have very low output, except for a few noteworthy exceptions like the Sumiko line, and when you try to amplify 1 mv all the way up to 20 volts (50 watts into an 8 ohm speaker), noise is a huge factor. In fact, here, the larger THD+N value for the MC setting is actually due to it simply having more noise, rather than more distortion than the MM setting. In any case, there were no noticeable harmonic peaks.
I also measured the frequency response in the split outputs, setting the control dials all the way counter-clockwise. You can see that the rolloff was complementary between the high-pass output and the low-pass output, which is where one would set it for use with a subwoofer. The output of the low-pass is lower than the high-pass output because the high-pass output goes to a power amplifier input, while the low-pass output goes to the input stage of a subwoofer amplifier, which is a preamplifier. I would have liked to see about 1 volt output from the low-pass output though. Having such a low output means you won't have much room to work with on the subwoofer volume control.