Obviously one can look at the NuForce website to see how the insides of these amplifiers look like, I however wanted to see for myself. This would also give me the opportunity to further investigate how these Reference 8 amplifiers work and how they’ve been constructed. A closer look at the components used would give me a far better impression on what decisions the manufacturer made in their design process and whether any corners have been cut along the way, or whether they have a strong grasp of the underlying technology.
The NuForce Reference 8 amplifier, take notice of the comment on the cable sleeving.
As mentioned in the introduction many amplifiers that are sold as mono-blocks feature a heavy chassis and a beefy power supply. I was surprised, although disappointed more accurately describes my feelings, to find that the Reference 8 uses a run-of-the-mill, off-the-shelf, 48-Volts/100-Watt industrial switch mode power supply. This doesn’t just limit the output significantly as it cuts the power if that rating is surpassed, but is a genuine low-budget approach. Any amplifier which has HiFi aspirations should have a decent power supply, that’s the first basic requirement. This simple switch mode power supply however is far from it, as it exhibits pour power and load regulation as well as a significant amount of noise and other RF garbage on its output.
The low-cost switch mode power supply used in the NuForce Reference 8 amplifier.
But there’s more, after a closer inspection of the design it became quickly apparent that it uses the well known Harris Semiconductor HIP4080 MOSFET driver. This is an integrated circuit (IC) which has become the de-facto standard for class-D car amplifiers used in a full-bridge configuration. This means that the Reference 8 actually features two amplifiers; one amplifier is sourcing the signal and the other one is sinking it, summing their outputs at the loudspeaker, which quadruples its output power versus a half-bridge configuration. This also explains why the Reference 8 runs off of a single supply voltage, as the HIP4080 is able to run off of a single supply voltage. The downside is that there’s half the supply voltage at the output. Hence shorting either one of the outputs to the chassis or GND is not recommended as it’ll cause the amplifier to draw excessive currents and might even render it defective, we didn’t dare to try out this hypothesis though.
The connectors as featured on the amplifier, take notice of the petite wiring used for the signal.
Much to my surprise the full-bridge didn’t make full use of one of the benefits of a full-bridge amplifier; the ability to cancel out all even harmonics by tying a feedback loop around both the sourcing and sinking side of the full-bridge. NuForce only opted to tie a feedback loop around one half of the bridge basically hoping that the other half will play along without being kept in check by a feedback mechanism. This more or less convinces me that NuForce has quite a few things to learn still about amplifier design; this amplifier is starting to look more and more like a high-school project rather than an amplifier designed and built by a reputable manufacturer.
NuForce uses a filter on the amplifier to filter out noise and HF garbage from the power supply.
What’s also clear is that NuForce hasn’t really found a novel way of incorporating the output filter into the feedback loop. Their method of incorporating the output filter in a single global feedback loop by relying heavily on the derivative term is neither novel nor unique, as it is commonly used by other amplifiers that rely on Philips’ UcD technology. If I’m not mistaken it could very well be that the claims for this patent quite neatly cover the NuForce circuit. For NuForce to suggest on their web site that UcD doesn't use the derivative term is telling in this respect. Whether these similartites are intentional or not I’m guessing Philips’ IP department might want to check into these.
