Solid State Preamp Reviews

I ran a conventional set of measurements on this preamplifier. This was not so much to find a cause behind any facet of each unit's sound quality, but to see if there were any particular aspects which would lead to system incompatibility problems. At this price level, I would be surprised by any electronic component that didn't depart from neutrality, so careful system matching is even more essential with the Parasound preamplifier than with expensive models.

Fig.1 shows the overall frequency response of the third sample of the P/FET-900 measured via the phono inputs. Noticeable is a very slight degree of exaggeration in the low bass, reaching a maximum of 0.3dB at 30Hz, below which it starts to roll off, if anything a little too gently, considering that it is likely the P/FET-900 will be used with inexpensive turntables. (The first sample had slightly more low-frequency boost.) There is a touch of lift at the top of the audio band, while the HF response extends too high, in my opinion, not reaching its –3dB point until above 100kHz, which, as with the Rotel, might make this preamp more prone than usual to picking up RF with MM cartridges (though the use of discrete FETs rather than op-amps in its phono stage, as well as the overall lower gain required, may well help here).

Measuring the line-level inputs, the response was flat in the audio band, with –3dB points at 4Hz and at more than 100kHz. As explained above, switching in the tone-control stage inverted signal polarity. The tone controls are specified as having a ±10dB action at 85Hz and 12kHz. Looking at the edge-of-band behavior, this translates to a maximum 10dB boost and 12.5dB cut at 20Hz and 20kHz. I have to admit that I am no big fan of loudness controls, it being impossible to establish a true baseline level where the response should be flat. However, the Parasound's loudness contour seems well-designed, with the maximum amount of tonal modification, appropriate for very low listening levels, shown in fig.2. The loudness control becomes inoperative when the volume control is past the halfway point.

The S/N ratio via the (shorted) MM input was a little worse than the Rotel, but very good, nevertheless, at 68dB unweighted and 76dB A-weighted. Crosstalk via the line-level inputs was buried in the noise at 20Hz and 1kHz, lying at –52dB at 20kHz. Via the phono input, separation was 64dB, 67dB, and 49dB at 20Hz, 1kHz, and 20kHz, respectively.

As the volume control is detented, I checked its action. The steps range from 1dB at the very top of the range to 6dB or more at the very bottom, with less than 1dB/step around the 12 noon position, and 1.5–2dB between 12 and 9 o'clock. The two channels tracked each other to within a 0.5dB accuracy over most of the range, this worsening to a 1dB tolerance or worse with the volume control set at 10 o'clock or lower. This is good at this price level.

Looking at input overload levels, the maximum output-stage swing at 1kHz is ±22V peak-peak, this given by 3V RMS into the line inputs with the volume control wide open. One concern with CMOS switches is that, unlike mechanical devices, they cannot swing more than a certain voltage. Feeding the test signal into the CD input and looking at the tape output, the P/FET-900's switches didn't overload until asked to swing more than 10.5V RMS. One anomaly was that hard-clipping the CMOS switches at 20kHz (12V RMS input) caused a drastic latch-up with a correspondent 15V DC-level shift. Fortunately, this is extremely unlikely to be triggered in use (I'd almost say impossible). Looking at the signal available from the tape output, the phono input appeared to overload at 95mV RMS at 20Hz, 170mV RMS at 1kHz, and 765mV RMS at 20kHz, though it actually appeared to be the CMOS switches that were clipping rather than the phono circuitry. This amount of headroom is excellent, and this Parasound shouldn't ever be overdriven even by very-high-output MM cartridges.

I measured the output impedance as 700 ohms—the P/FET-900 should have no problems driving reasonably long cables.—John Atkinson