Sidebar 3: Measurements
I measured the Belleson Brilliance phono preamplifier with my Audio Precision SYS2722 system, checking some of my findings with the magazine's APx500 analyzer. I adjusted the preamp's settings with the app on the supplied Android tablet, which was connected via the preamp's own Wi-Fi network. To minimize noise, I connected a wire from the grounding terminal on the preamplifier chassis's rear panel to the analyzer's ground. I performed the testing primarily with one of the single-ended inputs, as I had understood from TF that that was mostly how he had auditioned the Brilliance. To do so, I pressed in the recessed pushbutton on the rear panel as instructed in the manual, which grounds the shields of the input connectors. I repeated some of the testing with a balanced input, resetting the pushbutton to the recommended position (footnote 1). I also followed the manual's warning not to connect the balanced and unbalanced outputs at the same time.
While the balanced inputs preserved absolute polarity from both the balanced and single-ended outputs, the unbalanced inputs inverted polarity. The gain from the unbalanced outputs was very close to the numeric values. In MM mode, the gain with both single-ended and balanced inputs was 29.6dB with the gain set to "30dB," 35.6dB with it set to "36dB," 43.6dB with it set to "44dB," and 49.6dB with it set to "50dB." In MC mode, the gains were 49.5dB ("50dB"), 55.5dB ("56dB"), 63.5dB ("64dB"), and 69.5dB ("70dB"). The gains from the balanced outputs were exactly 6dB higher than these values.
Peculiarly, the unbalanced input impedances were close to half the values set with the Android app: 23k ohms with the impedance set to "47k," 153 ohms with it set to "300," 57 ohms with it set to "160," and 11.3 ohms with it set to "20," for example, all values consistent from 20Hz to 20kHz. (The capacitance was set to 0pF for these measurements.) Perhaps the single-ended/balanced pushbutton was miswired. The balanced input impedances were close to the indicated values. The preamp's source impedances were usefully low, at 3.5 ohms, single-ended, and 6.6 ohms, balanced, both values consistent from 20Hz to 20kHz.
Footnote 1: Noting that pin 1 of the signal generator's balanced XLR output jack needed to be floated with respect to ground. If it wasn't floated the Brilliance's output was afflicted by supply-related spuriae.
Fig.1 Belleson Brilliance, response with RIAA correction into 100k ohms (left channel blue, right red) (1dB/vertical div.).
The error in the Brilliance's RIAA equalization (fig.1, blue and red traces) was impressively low, with superb channel matching. Channel separation (not shown) was 80–90dB in both directions across the audioband.
Fig.2 Belleson Brilliance, MM mode, gain "30," spectrum, DC–1kHz, of output ref. 5mV input (left channel blue, right red, linear frequency scale, 20dB/vertical div.).
With the MM gain set to "50dB," the Belleson preamp's unweighted, wideband S/N ratio, measured at the unbalanced outputs with the single-ended input shorted to ground and ref. 1kHz at 5mV, was a very good 78.5dB (average of both channels). Restricting the measurement bandwidth to 22Hz–22kHz increased the ratio to 81.25dB, while the A-weighted ratio was a superb 86.4dB. The audioband and A-weighted ratios were identical with the MM gain set to "30dB," though the wideband ratio was 10dB lower. In MC mode, the ratios at the highest and lowest gain settings ref. 1kHz at 500µV were very similar to each other. Though they were approximately 20dB lower than in MM mode, this is still a low level of noise. Spectral analysis of the Brilliance's low-frequency noisefloor with the MM gain set to "30dB" and ref. 1kHz at 5mV (fig.2) revealed that both the random noise and any supply-related spuriae were low in level.
The Brilliance's overload margins in MC mode, with the gain set to "50dB," ref. 1kHz at the standard MC level of 500µV, were excellent, at 37.7dB at 20Hz and 1kHz, and still 28dB at 20kHz. With the MC gain set to "70dB," the margins were lower, at 16dB across the audioband, but this is still good performance. The margins in MM mode with the gain set to "50dB" and ref. 1kHz at 5mV were 16.1dB at 20Hz and 1kHz but 8.5dB at 20kHz. With the MM gain set to "30dB," the overload margins in the bass and midrange were higher, at 22.1dB, but still 8.5dB at the top of the audioband.
Fig.3 Belleson Brilliance, MM mode, gain "44," spectrum of 1kHz sinewave, DC–10kHz, into 100k ohms for 5mV input (left channel blue, right red, linear frequency scale, 20dB/vertical div.).
Fig.4 Belleson Brilliance, MM mode, gain "44," spectrum of 1kHz sinewave, DC–10kHz, into 100k ohms for 35mV input (left channel blue, right red, linear frequency scale, 20dB/vertical div.).
Fig.5 Belleson Brilliance, MM mode, gain "44," HF intermodulation spectrum, DC–30kHz, 19+20kHz into 100k ohms for 50mV peak input (left channel blue, right red, linear frequency scale, 20dB/vertical div.).
The Brilliance offered very low distortion. Fig.3 shows the spectrum of the preamplifier's output reproducing an input signal of 1kHz at 5mV with the gain set to "44." The only distortion harmonics that can be seen above the noisefloor are the second, at –104dB (0.0006%), and the third, at –96dB (0.0015%). These very low levels didn't increase significantly when I reduced the load impedance to the current-demanding 600 ohms and were similar at different gain settings. Distortion rose at higher input levels, but even with an input voltage 6dB below the point where the Brilliance in MM mode overloaded, the second harmonic lay at –89dB (0.03%) and the third harmonic at –60dB (0.1%) (fig.4). Intermodulation distortion with an equal mix of 19kHz and 20kHz tones at a peak input level 8.6dB below the MM clipping voltage at 20kHz was relatively low (fig.5). The difference product at 1kHz lay at –70dB (0.03%), the higher-order products at –60dB (0.02%).
Other than the limited overload margin at the top of the audioband in MM mode and that peculiar error in the unbalanced MM load impedances, the Belleson Brilliance's measured performance was excellent, with extremely accurate RIAA equalization and very low noise and harmonic distortion. I was impressed by the Android app's ease of use; I just wish the setting of that recessed pushbutton on the rear panel was easier to see.—John Atkinson
Footnote 1: Noting that pin 1 of the signal generator's balanced XLR output jack needed to be floated with respect to ground. If it wasn't floated the Brilliance's output was afflicted by supply-related spuriae.
