Obviously I was anxious to see whether the improved performance would also show up in my measurements. Letís start off with the THD (total harmonic distortion) versus frequency plot as thatís by far the most interesting one. The measurement bandwidth was set to 40kHz with both plots taken at 10W. The measurement bandwidth is different from that used in my first article, so Iím including fresh measurements made on the Reference 8, made at the same time for best comparison. The red plot shows the Reference 8, the blue plot the Reference 9. The Reference 8ís THD keeps rising all the way up to 20kHz indicating that its distortion performance is fully dominated by the second harmonic (it fits inside the 40kHz measurement bandwidth up to an input frequency of 20kHz). The blue plot drops off at 16kHz, indicating that the second harmonic problem is corrected in the Reference 9. At the bottom end, the rise in distortion is gone too. Most likely the addition of two large buffer caps on the power supply are responsible for the improvement.
The second plot is THD versus power. Whatís obvious is that the Reference 9 (blue) has quite a bit more muscle than the Reference 8 (red), as is clear from the blue plot. We get about 250W at 4 ohms, due to the beefier power supply. At the bottom end we see a marked decrease in noise. Quite puzzling, however, is the large amount of distortion at really low levels (20mW). This could be caused by a negative dead-time setting for the HIP4080 MOSFET driver; fortunately it did result in lower THD scores at higher output power.
The third plot is an IMD, intermodulation distortion plot with two test tones, one at 18.5KHz, and the other at 19.5KHz, for a total of 10W. Again we see a surprise even product between the two test tones, indicating the circuit is still asymmetrical in operation. In comparison to the Reference 8 the Reference 9 doesnít fare much better here.
Overall the Reference 9 is indeed a better amplifier than the Reference 8, it however isnít a small miracle as NuForce chooses to pitch it; it is a mediocre performer at best. Upon looking at the measurements thereís a clear improvement, although thereís quite some room for improvement still. To be honest Iíve seen quite a bit of common mistakes that, once corrected, will up the performance of the NuForce amplifiers considerably. In hindsight the Reference 9 is what the Reference 8 should have been, as it really isnít much more than a tweaked and upgraded Reference 8.
To give credit where credit is due however they do deserve a closer look if youíre in the market for a class-D amplifier. The Reference 9 has the potential to be a better amplifier than many class-D amplifiers that donít have post-filter feedback, use a triangle-wave based modulator, or all digital designs (which are flawed by design). In terms of quality of execution and attention to detail however, it falls well short when compared to more mature designs.
Model: Reference 9.02 class-D amplifier