I measured the Primare PRE35 with my Audio Precision SYS2722 system, repeating some tests with the magazine's higher-performance APx500 system. Looking first at the analog inputs, the balanced and single-ended line input impedances are specified, respectively, at 30k ohms and 15k ohms. I measured, for the balanced inputs, 29.6k ohms from 20Hz to 20kHz, and for the unbalanced inputs, 14.8k ohms at 20Hz and 1kHz and 13k ohms at 20kHz.
The maximum gain at the balanced outputs was 16.7dB for the balanced inputs and 22.75dB for the single-ended inputs. There are two sets of single-ended outputs, one labeled "Pre1," the other "Line"; I assumed the latter to be a fixed-level passthrough. The maximum gain at the Pre1 output was 16.7dB for the unbalanced inputs and 10.7dB for the balanced inputs. The gain at the fixed-level Line output was close to unity, at –0.1dB. The output impedances were very close to the specified figures, at a relatively low 376 ohms balanced and a lower 94 ohms unbalanced.
All three of the Primare's analog outputs preserved absolute polarity with analog input signals, and the analog input's frequency response featured a very wide bandwidth, flat to 100kHz into both 100k ohms and 600 ohms (fig.1). This graph was taken with a balanced input and output and with the volume control set to its maximum. Both the response and the superb channel matching were identical with the unbalanced input and output and at lower settings of the volume control.
Channel separation (not shown) was superb, at >120dB in both directions below 3kHz, decreasing to a still excellent 101dB at the top of the audioband. The low-frequency noisefloor (fig.2) was free from any power supply–related spuriae, and random noise was very low in level. Measured with an unbalanced input shorted to ground and the volume control set to its maximum, the Primare's unweighted, wideband signal/noise ratio was a high 78dB in both channels ref. 1V into 100k ohms, improving to 91dB when the measurement bandwidth was restricted to the audioband. Switching an A-weighting filter into circuit further improved the ratio to 93.75dB.
Fig.3 plots the THD+noise percentage against the balanced output level into 100k ohms. The distortion lies below the noisefloor up to 8V, with clipping occurring at 15V. Into the punishing 600 ohm load, the Primare's balanced outputs clipped just above 6V, which is still higher than the level the PRE35 will be required to deliver in practice. The single-ended outputs clipped at 7V. Fig.4 shows how the THD+N of the balanced outputs varies with frequency at 5V. The percentage is very low and doesn't change with frequency into 100k ohms (blue and red traces). Into 600 ohms (green, gray), the THD+N is as low as it was into 100k ohms below 2.5kHz then has a slight rise in the top three octaves. Even so, it remains very low, below 0.0035%.
The only harmonic visible above the noisefloor in the spectrum of the PRE35's balanced output driving 100k ohms was the third, at just –120dB (0.0001%, not shown). Into a difficult 600 ohms, the level of this harmonic rose by just 3dB (fig.5). Intermodulation distortion with an equal mix of 19 and 20kHz tones, the combintion peaking at 2V into 600 ohms, was vanishingly low in level (fig.6).
With the PRE35 connected to my network, Roon recognized it as an AirPlay Device. However, I did all the testing of the Primare's optional digital inputs with the coaxial and optical S/PDIF inputs. (These locked to datastreams sampled at all rates up to 192kHz.) I repeated some of the tests with the USB input, which Apple's AudioMIDI utility revealed accepted 16- and 32-bit integer data sampled at all rates from 44.1kHz to 768kHz. Apple's USB Prober app identified the PRE35 as "Primare_DM35_v1.06" from "Primare," and the USB port operated in the optimal isochronous asynchronous mode.
The preamplifier's outputs preserved absolute polarity from all digital inputs. The maximum level from the PRE35's analog outputs with 1kHz data at –12dBFS was 7.73V from the balanced output and 3.85V from the single-ended Pre1 output, both voltages about 6dB below the clipping point. The level with data at –12dBFS was 559mV from the single-ended Line output. To avoid clipping the main outputs with full-scale data, I performed all the testing of the PRE35's digital inputs from the Line output with the volume control turned down.
Fig.7 shows the PRE35's impulse response with 44.1kHz data. It is typical of a linear-phase reconstruction filter, with equal amounts of ringing before and after the single, full-scale sample. With 44.1kHz white-noise data, the filter rolled off rapidly above the audioband (fig.8, magenta and red traces), reaching full stop-band attenuation at 24kHz. With a 19.1kHz tone at 0dBFS (cyan, blue), the aliased image at 25kHz is suppressed by 100dB. The harmonics associated with the 19.1kHz tone are all low in level, with the second harmonic lying at –86dB (0.005%) and the third 10dB lower than that.
The Primare's frequency response with data sampled at 44.1, 96, and 192kHz is shown in fig.9. The responses all follow the same shape, flat almost to 20kHz with then a slow rolloff broken by a sharp drop in level just below half of each sample rate. An increase in bit depth from 16 to 24 with dithered data representing a 1kHz tone at –90dBFS (fig.10) dropped the PRE35's noisefloor by 20dB, which implies a high DAC resolution of between 19 and 20 bits. This was confirmed by the fact that the linearity error at –116dBFS was <0.1dB. While some enharmonic products can be seen with the 24-bit traces in fig.10, these are too low in level to have audible consequences. When I played undithered data representing a tone at exactly –90.31dBFS, the waveform was symmetrical, with the three DC voltage levels described by the data cleanly resolved (fig.11). Repeating the measurement with undithered 24-bit data gave a well-formed, noise-free sinewave (not shown).
As shown in fig.8, the PRE35's digital inputs featured a low level of harmonic distortion, though higher than with the analog inputs. With a full-scale 50Hz tone, the second and third were the only harmonics visible above the noisefloor (fig.12), both lying around –85dB (0.005%). The intermodulation distortion products with an equal mix of 19 and 20kHz tones, the signal peaking at 0dBFS, were all low in level, at close to –100dB (0.001%, fig.13), and the aliased images of the primary tones were suppressed by 96dB.
Fig.14 shows the spectrum of the PRE35's output when it was fed high-level, optical, undithered, 16-bit J-Test data. The odd-order harmonics of the undithered low-frequency, LSB-level squarewave mostly lie close to the correct levels, indicated by the sloping green line. However, the sideband pair closest to the frequency of the high-level tone at one-quarter the sample rate was boosted in level. This was also the case with 24-bit J-Test data. Jitter rejection via the USB port, however, was superb (fig.15).
Primare's PRE35 did very well on the test bench, offering very low distortion and noise for both analog and digital input signals.—John Atkinson
