Integrated Amp Reviews

I performed a full set of measurements on the Woo WA5 using my Audio Precision SYS2722 system (see the January 2008 "As We See It") and one of the RCA inputs. The voltage gain into 8 ohms from the speaker terminals was 31dB with the gain switch in the H position, just 3.8dB with it in the L position. From the unbalanced headphone jack set to High Power, the gain into 300 ohms was a very high 44.5dB at the H setting, 17.1dB at L. Switched to Low Power, the respective headphone-output gains were 31.1 and 17.1dB. Both the speaker and headphone outputs inverted absolute polarity.

TheWA5's input impedance was usefully high, at 83k ohms at 20Hz and 1kHz, 60k ohms at 20kHz. The output impedance from the headphone jack was 38 ohms at 1kHz at High Impedance, 16 ohms at Low Impedance—both a little on the high side. The output impedance from the speaker terminals was also high, at 3.2 ohms across the audioband, which gave rise to variations of ±1.9dB in the frequency response with our standard simulated loudspeaker (fig.1, gray trace). However, this graph also reveals that the response into a resistive load was impressively flat between 20Hz and 20kHz, and that the extended low-frequency response resulted in accurate reproduction of a 1kHz squarewave—ie, a wave with flat tops and bottoms (fig.2). The WA5 has excellent output transformers. With a 10kHz squarewave at the speaker outputs (fig.3) there was a slight amount of ringing visible, suggesting the presence of one or two well-suppressed ultrasonic resonances. Fig.4 reveals that these resonances are much more greatly developed from the headphone output with a load representative of Sennheiser headphones, but are very high in frequency, at 90 and 180kHz. Again, the WA5's output transformers are impressively engineered.

Channel separation was okay, at 60dB in both directions across the audioband, but the amplifier's noise floor, measured at the speaker terminals, was contaminated with low-level power-supply–related hum components (fig.5). The unweighted, wideband signal/noise ratio, ref. 2.83V into 8 ohms and taken with the inputs shorted to ground but the volume control set to its maximum, was therefore disappointing, at 72dB in the left channel and 73.6dB in the right. A-weighted, those ratios respectively improved to 83.5 and 84.25dB.

Claimed to deliver 10Wpc into 8 ohms (10dBW) at 3% distortion, the WA5 didn't meet that specification with the Psvane 300B. Fig.6 reveals that, into 8 ohms, the amplifier reached 1% THD (our usual definition of clipping) at just 618mW, and at 3% THD+noise delivered just 3W (4.8dBW). Even at 10% THD+N, the power is only 8.5Wpc (9.3dBW). Note also the rapid rise in THD as the power increases above a couple of hundred milliwatts—this amplifier's output stage has what could be kindly referred to as a "bent" transfer function. The WA5 actually offers better linearity into 4 ohms (fig.7), with 4.8Wpc available at 3% THD+N and 8Wpc at 10% THD+N, these powers respectively equivalent to 3.8 and 6dBW.

Looking at how the percentage of THD+N varied with frequency at 2V, which is equivalent to 500mW into 8 ohms or 1W into 4 ohms, it appears that the right channel (fig.8, magenta trace) is more linear than the left into 4 ohms (cyan), though the channels behave very similarly into 8 ohms (blue, red). At 3W, which turned out to be a relatively high power into 8 ohms for this amplifier, the residual distortion (fig.9, red trace) is predominantly the third harmonic. But if you look in fig.9 at the waveform of the fundamental tone (blue trace), you can see that the tops of the waveform are more rounded than the bottoms. This waveform asymmetry, due to the bent transfer function, suggests that there will also be even-order harmonic distortion. At lower powers and frequencies (fig.10), the second harmonic is dominant, though the third harmonic rises to an equal level into 4 ohms (fig.11).

This commentary applies to the speaker outputs: The distortion signature from the headphone outputs was very similar, though with a slightly higher level of random noise. Finally, even with the WA5's better linearity into 4 ohms than 8, it didn't do well on the demanding high-frequency intermodulation test (fig.12). Yes, the right channel (red trace) is still better than the left (blue), but the difference product at 1kHz is high, at –48dB left (0.4%) and –56dB right (0.15%).

It is difficult to sum up the Woo Audio WA5's measured performance. Looked at in traditional terms, it didn't do well on the test bench: Its low power output and high distortion will not be too much of an issue with headphones but would seem to rule the WA5 out of contention for driving loudspeakers. But, of course, the bent transfer function responsible for those failings is due to the designer's decision to use a single-ended output stage topology. Considered in that light, the WA5 performed better than I would have expected.—John Atkinson