Description
The CS47L15 is a highly integrated, low-power audio hub for smartphones, tablets, and other portable audio devices
including wearable technology. It combines an advanced DSP feature set with a flexible, high-performance audio hub
codec. The CS47L15 combines a programmable DSP core with a variety of power-efficient fixed-function audio
processors. An SPI master interface is provided, for autonomous boot-up and configuration using an external non-volatile
memory—enabling the CS47L15 to be used independently of a host processor.
The DSP core supports advanced audio processing functions such as wideband noise reduction, acoustic-echo
cancelation (AEC), speech enhancement, karaoke, and many more. Low-power analog and digital interfaces provide
flexible support for always-on voice applications and speaker-protection algorithms implemented on the programmable
DSP core. The DSP core is integrated within a fully flexible, all-digital mixing and routing engine with sample-rate
converters, for wide use-case flexibility. Support for third-party DSP programming provides far-reaching opportunities for
product differentiation.
Three digital audio interfaces are provided, each supporting a wide range of standard audio sample rates and serial
interface formats. Automatic sample-rate detection enables seamless wideband/narrowband voice-call handover. The
DACs and output paths provide full support for high definition audio throughout the entire signal chain.
The stereo headphone driver provides ground-referenced output, with noise levels as low as 0.45 VRMS for hi-fi quality
line or headphone output. The CS47L15 also features a mono bridge-tied load (BTL) earpiece output, mono 2.5-W Class D
speaker driver, two channels of stereo PDM output, and an IEC-60958-3–compatible S/PDIF transmitter. A signal
generator for controlling haptics devices is included; vibe actuators can connect directly to the Class D speaker output, or
via an external driver on the PDM output interface.
The CS47L15 supports up to five analog inputs, and up to four PDM digital inputs. As many as four analog microphone
connections can be supported; a separate analog input channel is provided for use in speaker-protection applications.
Microphone activity detection with interrupt is available. A smart accessory interface supports most standard 3.5-mm
accessories. Impedance sensing and measurement is provided for external accessory and push-button detection
(Android™ headset specification compliant).
The CS47L15 supports SPI™ and I2C interface modes for control-register access. The CS47L15 can also be configured
as SPI master, enabling autonomous boot-up and configuration without dependency on a host processor. Two integrated
FLLs support a wide range of system-clock frequencies. The device is powered from 1.8- and 1.2-V supplies. Separate
MICVDD input can be supported, for microphone operation above 1.8 V. An additional supply is required for the Class D
speaker drivers (typically direct connection to 4.2-V battery). The power, clocking, and output driver architectures are
designed to maximize battery life in voice, music, and standby modes. Low-power (25 W) Sleep Mode is supported, with
configurable wake-up events.

Features
• 150 MIPS, 150 MMAC audio-signal processor
— Low-power, always-on voice trigger capability
— Speaker protection algorithm support
— Event loggers with time-stamp and interrupt functions
• Programmable wideband audio processing
— Transmit-path noise reduction and echo cancelation
• Integrated multichannel 24-bit hi-fi audio hub codec
— 98-dB signal-to-noise ratio (SNR) mic input (48 kHz)
— 127-dB SNR headphone playback (48 kHz)
— Low-power analog input modes
• Up to four analog or four digital microphone (DMIC)
inputs
— Speaker-monitoring input path (analog or digital)
• Stereo headphone/earpiece/line output driver: 30 mW
into 32- load at 0.1% total harmonic distortion + noise
(THD+N)
• Earpiece, speaker, and digital (pulse-density modulation,
PDM) output interfaces
— Two-way stereo PDM interface
• Three full digital-audio interfaces
— Standard sample rates from 8 to 192 kHz
— Multichannel support on AIF1 and AIF2
• Self-boot capability from external non-volatile memory
• Flexible clocking, derived from MCLKn or AIFn
• Low-power frequency-locked loops (FLLs) support
reference clocks down to 32 kHz
• Advanced accessory detection functions
• Configurable functions on up to 15 general-purpose
input/output (GPIO) pins
• Small WLCSP package, 0.4-mm ball array

Applications
• Smartphones, tablets, and wearable technology
— Karaoke algorithm support