This application note describes how to interface the STR7xx SPI peripheral with an audio
device (Codec, ADC, DAC, filter...) using the I2S protocol via an external interface consisting
of a low cost small CPLD. The design is referenced to the I2S protocol specification, so any
I2S device is able to interface with the STR7xx and STR91x MCU through this bridge.
To avoid using extra hardware, an I2S solution could be implemented entirely by firmware
but this would put a huge load on the CPU. The SPI to I2S bridge solution proposed in this
application note is made with a very low cost hardware (about 1$) using a CPLD with less
than 32 macrocells and a minimum of firmware overhead. Only one Timer (one Output
Compare), 6 GPIOs and the DMA peripheral are used. The DMA available on STR75x and
STR91x allows the CPU to be totally free for other tasks, and on STR71x the CPU load is
3.9% at 32 MHz MCLK frequency. The footprint is less than 5 Kbytes in Thumb mode.
The example CPLD described in this application note was built for the STR71x and STR750
microcontrollers but can be easily tailored to the STR91x or STM32.
I2S (IC to IC Sound) is an audio data transfer standard using a three-line bus for serial and
synchronous data transmission.
Data are transmitted on the SD line (Serial Data) in Little Endian format (MSB first). Data
length is not limited (usually 16/20/24/32/64 bits). Data are synchronized by the SCK (Serial
Clock) rising or falling edge for the transmitter, and falling edge only for the receiver. Refer to
Figure 1.
Data represent stereo digital sound, so each sample contains two words, the right channel
sample and the left channel sample. Instead of using two data channels, muxing is
performed by transmitting each word in half a sampling period and that allows, by doubling
the sampling rate, to transmit two words per period.
A control signal WS (Word Select) is then used to determine if the word being sent is the
right or the left one. This signal also determines the beginning and the end of the data: there
is no need to fix the data length. Receiver and transmitter data lengths can therefore be
different, as well as the right and left data lengths.
WS is synchronized either on the rising or the falling edge of SCK and precedes the MSB by
one SCK period in order to have enough time for storing and shifting operations.
As in the SPI protocol, there must be a master and a slave. The master controls and
provides the SCK clock and the WS signal, while the slave only sends or receives data. The
master can be the receiver or the transmitter or a third element (Controller). Refer to
Figure 2.
Figure 1.I2S Philips protocol waveforms 16/32-bit
CK
WS
SD
Transmission
MSB
Reception
16-bit or 32-bit
MSB
LSB
Left channel Right channel
3/15
General description AN2682 - Application note
Figure 2.I2S protocol signal description and configuration
SCK
TransmitterReceiver
WS
SD
Transmitter = MasterReceiver = Master
ReceiverTransmitter
Controller
SCK
Transmitter
WS
SD
Controller = Master
Receiver
SCK
WS
SD
1.2 SPI protocol
The SPI is chosen to implement this solution because its protocol is the most similar to the
I2S protocol.
It uses four pins:
●Two pins (MOSI and MISO) to transmit and receive serial data
●One pin (SCLK) for data clocking (either rising or falling edge modes are available)
●One pin (nSS) to choose between master and slave modes.
When nSS is tied to Gnd, the SPI peripheral functions in Slave mode: it receives the data
clock (SCLK) from the master which controls how and when data are transmitted or
received.
When nSS is tied to Vcc, the SPI peripheral functions in master mode and SCLK is then an
output.
MISO is the Master Input in master mode and the Slave Output in slave mode.
MOSI is the Master Output in master mode and the Slave Input in slave mode.
There are two main differences between the I2S and SPI protocols:
●The I2S clock is continuous while the SPI clock is generated only during word
transmission and is stopped afterwards.
●The data format in SPI protocol is fixed to 16 or 8 bits only, while the data in I2S can
have any length.
Due to these two differences, the SPI peripheral cannot be configured in master mode in
this case. This is due to the fact that in SPI master mode, data are transmitted with a delay
between the words. Besides, many difficulties would arise when other tasks are performed
by the microcontroller. So the solution uses the SPI in slave mode.
4/15
AN2682 - Application noteCPLD design
2 CPLD design
The CLPD design is based on the MAX3000A 32 macrocell device shown in Figure 3.
Figure 3.MAX3000A CPLD pin description
The CPLD bridge is intended to function as the master for both SPI and I2S devices. Since
the SPI peripheral is bidirectional, two I2S devices can be connected at once, as long as
they have the same SCK clock (some I2S devices are also bidirectional).
When a single device is interfaced to the microcontroller, the unused direction logic can
simply be ignored and has no effect on the interface. The CPLD bridge connections with the
STR7/9 MCU and the I2S audio device(s) are illustrated in Figure 4.
Figure 4.SPI to I2S CPLD bridge connections
AF
GPIO/AF
GPIO/AF
STR7/9
MCU
PWM OCMPA
GPIO
GPIO
MOSI
MISO
SCLK
CLK
Start
Conf
CPLD
interface
SDo
SDi
SCK
WSo
WSi
Audio
I2S
device(s)
The Conf pin is used only in configuration mode. In communication mode, it can be used for
any other application purpose.
5/15
Loading...
+ 10 hidden pages
You need points to download manuals.
1 point = 1 manual.
You can buy points or you can get point for every manual you upload.