Nuforce is probably one of the few class-D manufacturer with its own patent (US patent granted in 2007) and R&D. And we uses not only our own class-d design but also other chips on the market. Since we started selling our first product in 2005, we have expanded into many categories and performance level, so I can say that our experience and road map reflects where class-D is going.
As many of you have recognized, with advance in technology, each generation of class-d gets better. When we started to win over many audiophiles with our power amp in 2005, after more than 8 years, with thousands of high-end customers worldwide, there is no question that the best class-d implementation can match the best solid state or tube amp. On the other hand, commercial reality (pricing pressure) and market requirement also resulted in many class-d power amp (mostly chips) being designed for low cost, low power requirement.
Nuforce vision is to design the best sounding audio for a consumer lifestyle, whether it is for portable, entry level home use, or high-end audio. So how do we decide what goes into high-end audio and how far can we push the performance limit for a $299 (Dia) or $549 all digital power amp (DDA-100, to be released in late June) ?
To be honest, the performance difference between high-end audio and entry level priced audio is getting smaller. For example, the soon to be released DDA-100, at $549 (USD or Euro), offers 4 digital inputs (1 USB, 1 coaxial, 2 optical) and 50W@8ohm x 2 speaker outputs with high-end audiophile-grade sound quality. On the other extreme end, our Ref 18 cost more than 15X. In any industry, extreme performance and limited market demand resulted in high price (Ferrari cost 20X the price of entry level car). I will use these two products to illustrate where we see class-d going in the coming years.
Why does Ref 18 or 20 cost so much more? These are the factors:
1. R&D time. A new generation of high-end class-d amp takes years to develop, not months. A typical "analog" class-D design converts input audio signal into small pulses with varying width, amplifies the pulses, then convert them into audio. A good class-D design implements the conversion at very high frequency (Nuforce is perhaps the only implementation running at 400kHz) with real time feedback and error correction, utilizing high quality parts. At low power, it is possible to implement the entire design in a chip, therefore reducing cost and optimizing performance. At high power (> 100W) things started to get costly and complicated. Therefore it is exponentially more difficult to make a high-end audiophile-grade class-D amp to sound as good as a solid state amp. There are many things to optimize. The bottom line is class-d implementation is a lot more complex than solid state or tube.
To increase the power of Ref 18 (Ref 20 is not going to sound much better than Ref 18), we originally thought that adding a bigger power supply with some minor adjustment on the amp would be enough, after some initial prototyping. We were wrong. Because this new amp has to be just as perfect as the Ref 18. So round after round of optimization and reliability testing. And the schedule keep slipping. Ref 20 is simply designed for people who really need more power than Ref 18.
2. Performance requirement - audiophiles pick on everything. High-end class-D such as Ref 18 is becoming near flawless (on our website review section, you can read 3 generations of reviews from V1 to V3). Therefore to develop V4, we can not do any incremental improvement on the same architecture. We have every intention to keep advancing high-end class D, but V4 is going to be a multi-year effort. Even if we think it is possible, there is no point in redesigning a good architecture just to end up with incremental sound improvement. With our Ref 18 near flawless using V3 technology, V4 has to do more than just sounding better. So we will make an attempt with V4 not only sounding better than current V3, but also more reliable and with adjustable sound characteristic.
We plan to go where no class-D has gone before
3. Build quality - lower price product such as DDA-100 have to cut cost whereas Ref 18 uses big and heavier chassis, with lots of shielding inside.
We have to amortize the R&D cost over a smaller number of units, that added to the cost of high-end products as well.
How can we offer a high-end integrated amp such as DDA-100 for so little money and how far have we pushed the performance?
I would dare to say that DDA-100 sounded better than our V1 generation power amp which retails for more than $2000 8 years ago. Today, we can achieve the same performance at about 1/4 the price 8 years ago. DDA-100 uses an all digital chip so that brings the cost down significantly. Surprisingly, the performance is incredibly good after we did some optimization to the system design. The limitation with such a design is that we don't have much room to modify it or increase the power. We have to start from the ground up again if we are going to offer a higher power DDA-100.
The bottom line for high-end Class D is that it has to keep pushing the limit, and fewer and fewer high-end companies are able to do so (either because they lack the R&D capability or are dependent on module supplier such as icepower).