The benefits of an additional DSP
System designers would prefer to integrate any processing into the main system. Processing in the phone chipset will generally indeed give the smallest, most power efficient, and cheapest solution. However, accurate feedback is the key to successful speaker boost, and it needs low latency and high bandwidth.
Comprehensive speaker protection actually requires multiple input points - just knowing current and voltage isn't enough. Furthermore, interrupts and system integration issues can become a major hassle. Multiple sensing points are needed to optimize the amplified signal. The processing must also optimize the performance of both amplifier and DC/DC converter.
To properly integrate this system into the central processor, all these signals would need to be converted and fed in to the chipset and all the controls properly taken out. A separate DSP can handle all these interactions automatically and can run continually even when the central processing shuts down.
The TFA9887 for speaker boost and protection
NXP's TFA9887 (Figure 4) is offered as the first IC to dramatically boost output while fully protecting the speaker. It has an embedded CoolFlux DSP, Class-D amplifier with integrated current sensing, and intelligent DC/DC boost converter.
Figure 4. Block diagram of TFA9887.
The IC holds a software model of the speaker, and automatically adapts to any changes over the speaker's lifetime including ageing, enclosure damage, blocked speaker ports, or whatever the world can throw at it. Better sound quality can also be traded against even smaller speakers and back volumes, giving smaller end products.
To confirm its performance, we compared the SPL of a speaker driven by the TFA9887 with a popular unmodified smart phone which uses a software compressor to enhance the volume. The test used identical test files and an identical phone (so identical speaker and enclosure).
Figure 5 shows more than 6 dB SPL increase in output volume. Optimized for bandwidth, bass output is increased by around 10 dB SPL - a huge improvement for a simple replacement in a state-of-the-art reference phone!
Figure 5. Volume and bass levels from the TFA9887 are noticeably improved.
Figure 6. Despite delivering around 2.5 W peak into a 0.5 W speaker (and over 5 W into a 4-ohm speaker), the excursion for music and speech clearly remains well within specifications.
This leap in performance illustrates an important design trend. The days of stand-alone amplifiers and converters designed in isolation have gone. The performance of phones and other portable devices have seen so many refinements that components must be treated as part of a bigger system. Each part of the system must sense and interact with the real world for best possible system performance.
So, audio systems must monitor the performance of the acoustics and adjust for the best user experience (Figure 6). Here as elsewhere, there is a clear trend to producing systems that measure and interact with the real world.
About the author
Shawn Scarlett is the marketing director for NXP’s mobile audio group, with a long history specializing in audio semiconductors including positions at Analog Devices and National Semiconductor, as well as start-ups such as Tripath and GTRonix. He has a B.S. in Electrical Engineering from the University of Arizona as well as an MBA from Santa Clara University. Before moving into semiconductors, he developed his audio skills as a professional sound engineer with the I.A.T.S.E working on sound reinforcement for major touring shows.