NXP Semiconductors
TFA9896
Boosted audio system with adaptive sound maximizer and speaker protection
TFA9896 All information provided in this document is subject to legal disclaimers. © NXP Semiconductors N.V. 2016. All rights reserved.
Product short data sheet Rev. 1 — 22 December 2016
COMPANY PUBLIC 6 / 22
8 Functional description
The TFA9896 is a highly efficient mono Bridge Tied Load (BTL) class-D audio amplifier
with a sophisticated SpeakerBoost protection algorithm. Figure 1 is a block diagram of
the TFA9896.
The device contains two TDM input/output channels. The number of slots and number
of bits per slot can be configured for each TDM channel and the channel can be
configured as input or output. Typically, one TDM channel is configured as a standard
I2S input while the other TDM channel is used for stereo sync, where gain information is
transferred between the devices.
It is also possible to output current sense and voltage sense information on the TDM
interface, which can be processed by the audio host.
The SpeakerBoost protection algorithm, running on a CoolFlux Digital Signal Processor
(DSP) core, maximizes the acoustical output of the speaker while limiting membrane
excursion and voice coil temperature to a safe level. The mechanical protection
implemented guarantees that speaker membrane excursion never exceeds its rated limit,
to an accuracy of 10 %. Thermal protection guarantees that the voice coil temperature
never exceeds its rated limit, to an accuracy of ±10 °C. Furthermore, advanced signal
processing ensures the audio quality remains acceptable at all times.
The protection algorithm implements an adaptive loudspeaker model that is used to
predict the extent of membrane excursion. The model is continuously updated to ensure
that the protection scheme remains effective even when speaker parameter values
change or the acoustic enclosure is modified.
The SpeakerBoost protection algorithm boosts the output sound pressure level within
given mechanical, thermal and quality limits. An optional Bandwidth extension mode
extends the low frequency response up to a predefined limit before maximizing the output
level. This mode is suitable for listening to high quality music in quiet environments.
The frequency response of the TFA9896 can be modified via ten fully programmable
cascaded second-order biquad filters. The first two biquads are processed with 48-bit
double precision; biquads 3 to 8 are processed with 24-bit single precision.
A battery supply safeguard mechanism can be configured to reduce the gain at low
battery voltage levels to limit battery current.
The output volume can be controlled by the SpeakerBoost protection algorithm or by
the host application (external). In the latter case, the boost features of the SpeakerBoost
protection algorithm must be disabled to avoid neutralizing external volume control.
The SpeakerBoost protection algorithm output is converted into two pulse width
modulated (PWM) signals which are then injected into the class-D audio amplifier. The 3-
level PWM scheme supports filterless speaker drive.
The adaptive DC-to-DC converter boosts the battery supply voltage in line with the output
of the SpeakerBoost protection algorithm. It switches to Follower mode (V
BST
= V
BAT
; no
boost) when the audio output voltage is lower than the battery voltage.
