The UDA1384 is a single-chip consisting of 4 plus 1 Analog-to-Digital Converters (ADC)
and 6 Digital-to-Analog Converters (DAC) with signal processing features employing
bitstream conversion techniques. The multichannel configuration makes the device
eminently suitable for use in digital audio equipment which incorporates surround feature.
The UDA1384 supports conventional 2 channels per line data transfer conformable to the
I2S-bus format with word lengths of up to 24 bits, the MSB-justified format with word
lengths of up to 24 bits and the LSB-justified format with word lengths of 16 bits, 20 bits
and 24 bits, as well as 4 channels to 6 channels per line transfer mode. The device also
supports a combination of the MSB-justified output format and the LSB-justified input
format. The UDA1384 has special sound processing features in the Direct Stream Digital
(DSD) playback mode, de-emphasis, volume and mute which can be controlled via the
L3-bus or I2C-bus interface.
2.Features
2.1 General
■ 2.7 V to 3.6 V power supply
■ 5 V tolerant digital inputs
■ 24-bit data path
■ Selectable control: via L3-bus or I2C-bus microcontroller interface
■ Supports sample frequency ranges for:
◆ Audio ADC: fs = 16 kHz to 100 kHz
◆ Voice ADC: fs = 7 kHz to 50 kHz
◆ Audio DAC: fs = 16 kHz to 200 kHz
■ Separate power control for ADC and DAC
■ ADC plus integrated high-pass filter to cancel DC offset
■ Integrated digital filter plus DAC
■ Slave mode only applications
■ Easy application
Philips Semiconductors
2.2 Multiple format data interface
■ Audio interface supports standard I2S-bus, MSB-justified, LSB-justified and two
multichannel formats
■ Voice interface supports I2S-bus and mono channel formats
2.3 Digital sound processing
■ Control via L3-bus or I2C-bus:
◆ Channel independent digital logarithmic volume
◆ Digital de-emphasis for fs = 32 kHz, 44.1 kHz, 48 kHz or 96 kHz
◆ Soft or quick mute
◆ Output signal polarity control
2.4 Advanced audio configuration
■ Inputs:
◆ 4 single-ended audio inputs (2 × stereo) with programmable gain amplifiers
◆ 1 single-ended voice input
■ Outputs:
◆ 6 differential audio outputs (3 × stereo)
■ DSD mode to support stereo DSD playback
■ High linearity, wide dynamic range and low distortion
■ DAC digital filter with selectable sharp or soft roll-off
UDA1384
Multichannel audio coder-decoder
3.Applications
■ Excellently suitable for multichannel home audio-video application
4.Quick reference data
Table 1:Quick reference data
V
= V
DDD
(pins V
SymbolParameterConditionsMinTypMaxUnit
Supplies
V
V
V
I
DDA(AD)
I
DDA(DA)
I
DDD
DDA(AD)
); unless otherwise specified.
SS
DDA(AD)
DDA(DA)
DDD
= V
ADC analog supply
voltage
DAC analog supply
voltage
digital supply voltage2.73.33.6V
ADC analog supply
current
DAC analog supply
current
digital supply current f
DDA(DA)
= 3.3 V; T
= 25°C; RL = 22 kΩ; all voltages referenced to ground
Product data sheetRev. 02 — 17 January 20055 of 55
Philips Semiconductors
UDA1384
Multichannel audio coder-decoder
Table 3:Pin description
…continued
SymbolPinTypeDescription
DATAV16DOvoice data output
BCKV17DISvoice bit clock input
WSV18DIOvoice word select input or output
SYSCLK19DISsystem clock input: 256f
, 384fs, 512fs or 768f
s
s
MCMODE20DIL3-bus L3MODE input or I2C-bus DAC mute control input
2
MCCLK21DISL3-bus L3CLOCK input or I
MCDATA22IICL3-bus L3DATA input and output or I
C-bus SCL input
2
C-bus SDA input and
output
WSDA23DIDAC word select input
BCKDA24DISDAC bit clock input
DATADA125DIDAC channel 1 and channel 2 data input
DATADA226DIDAC channel 3 and channel 4 data input
DATADA327DIDAC channel 5 and channel 6 data input
V
Product data sheetRev. 02 — 17 January 20056 of 55
Philips Semiconductors
UDA1384
Multichannel audio coder-decoder
Table 4:Pin types
TypeDescription
DISdigital Schmitt-triggered input
DOdigital output
DSdigital supply
IICinput and open-drain output for I
8.Functional description
8.1 System clock
The UDA1384 operates in slave mode only; this means that in all applications the system
must provide either the system clock (the bit clock for the voice ADC) or the word clock.
The audio ADC part, the voice ADC part and the DAC part can operate at different
sampling frequencies (DAC-WS and ADC-WS modes) as well as a common frequency
(SYSCLK, WSDA and DSD modes).
The voice ADC part supports a sampling frequency up to 50 kHz and the audio ADC
supports a sampling frequency up to 100 kHz. The DAC sampling frequency range is
extended up to 200 kHz with the range above 100 kHz being supported through 192 kHz
sampling mode, which halves the oversampling ratio of SYSCLK and internal clocks.
…continued
2
C-bus
The mode of operation of the audio and voice channels can be set via the L3-bus or
I2C-bus microcontroller interface and are summarized in Table 5 and Table 6.
When applied, the system clock must be locked in frequency to the corresponding digital
interface clocks.
The voice ADC part can either receive or generate the WSV signal as shown in Table 6.
Product data sheetRev. 02 — 17 January 20057 of 55
, 384fs, 512fs or 768fsSYSCLK 256fs, 384fs, 512fs or 768f
s
SYSCLK 128fs, 192fs, 256fs or 384fs;
192 kHz sampling mode
, 384fs, 512fs or 768fsWSDA1f
s
s
s
, 64fs, 128fs or 256f
s
, 64fs, 128fs or 256f
s
SYSCLK 256fs, 384fs, 512fs or 768f
SYSCLK 128fs, 192fs, 256fs or 384fs;
WSDA1f
s
s
s
192 kHz sampling mode
s
input
output
s
s
Philips Semiconductors
8.2 Audio analog-to-digital converter (audio ADC)
The audio analog-to-digital front-end of the UDA1384 consists of 4-channel single-ended
ADCs with programmable gain stage (from 0 dB to 24 dB with 3 dB steps), controlled via
the microcontroller interface.Using the PGA feature,it is possible to accept an input signal
of 900 mV (RMS) or 1.8 V (RMS) if an external resistor of 10 kΩ is used in series. The
schematic of audio ADC front-end is shown in Figure 3.
Fig 3. Schematic of audio ADC front-end
input signal
2 V (RMS)
10 kΩ
VINL,
VINR
Multichannel audio coder-decoder
10 kΩ (0 dB setting)
10 kΩ
V
ref
V
DDA
mgu582
= 3.3 V
UDA1384
ADC
8.3 Voice Analog-to-Digital Converter (voice ADC)
The voice analog-to-digital front-end of the UDA1384 consists of a single-channel
single-ended ADC with a fixed gain (26 dB) Low Noise Amplifier (LNA). Together with the
digital variable gain amplification stage, the voice ADC provides optimal processing and
reproduction of the microphone signal. The supported sampling frequency range is from
7 kHz to 50 kHz. Power-down of the LNA and the ADC can be controlled separately.
8.4 Decimation filter of audio ADC
The decimation from 64fs is performed in two stages. The first stage realizes
characteristics with a decimation factor of 8. The second stage consists of three half-band
filters, each decimating by a factor of 2. The filter characteristics are shown in Table 7.
The voice ADC decimation filter is realized with the combination of a Finite Impulse
Response (FIR) filter and Infinite Impulse Response (IIR) filter forshorter group delay. The
filter characteristics are shown in Table 8. During the power-on sequence, the output of
the ADC is hard muted for a certain period. This hard-mute time can be chosen between
1024 samples and 2048 samples.
The digital interpolation filter interpolates from 1fsto 128fs (or to 64fs in the 192 kHz
sampling mode) by cascading FIR filters, and has two sets of filter coefficients for sharp
and slow roll-off as given in Table 9 and Table 10.
The 3rd-order noise shaper operates at either 128fs or 64fs (in the 192 kHz sampling
mode), and converts the 24-bit input signal into a 5-bit signal stream. The noise shaper
shifts in-band quantization noise to frequencies well above the audio band. This noise
shaping technique enables high signal-to-noise ratios to be achieved.
8.8 Digital mixer
The UDA1384 has 6 digital mixers inside the interpolator (see Figure 4). The ADC signals
can be mixed with the I2S-bus input signals. The mixing of the ADC signals can be
selected by the bits MIX[1:0].
Product data sheetRev. 02 — 17 January 20059 of 55
Philips Semiconductors
]
MIX[1:0
from ADC
ch1
ch2
ch3
ch4
2
from I
S-bus
mixer input
MIXER
VOLUME
VOLUMEDE-EMPHASISMUTE
MIXER
MUTE
UDA1384
Multichannel audio coder-decoder
INTERPOLATION
+
1f
s
FILTER
DAC1
DATADA1
DATADA2
DATADA3
]
DIS[1:0
]
ICS[1:0
Fig 4. Block diagram of DAC mixer
8.9 Audio digital-to-analog converters
The audio digital-to-analog front-end of the UDA1384 consists of 6-channel differential
SDACs: an SDAC is a multi-bit DAC based upon switched resistors. To minimize data
dependent modulation effects, a Dynamic Element Matching (DEM) algorithm scrambler
circuit and DC current compensation circuit are implemented with the SDAC.
same as above
same as above
same as above
same as above
same as above
DAC2
DAC3
DAC4
DAC5
DAC6
mgw786
8.10 Power-on reset
The UDA1384 has an internal power-on reset circuit which initializes the device (see
Figure 5). All the digital sound processing features and the system controlling features are
set to their default values in the L3-bus and the I2C-bus modes.
The reset time (see Figure 6) is determined by an external capacitor which is connected
between pin V
When V
DDA(AD)
During the reset time, the system clock should be running.
Product data sheetRev. 02 — 17 January 200513 of 55
Philips Semiconductors
8.12 Voice digital interface
The following voice formats can be selected via the microcontroller interface:
2
• I
S-bus format with data word length of up to 20 bits. The left and the right channels
contain the same data.
• Mono channel format with data word length of up to 20 bits.
The formats are illustrated in Figure 9.
UDA1384
Multichannel audio coder-decoder
WS
BCK
DATA
WS
BCK
DATA
MSB B2
MSB B2
LEFT
Fig 9. Voice digital interface formats
8.13 DSD mode
The UDA1384 can receive 2.8224 MHz DSD signals and generate 88.2 kHz multibit PCM
signals as well as analog signal outputs. The configuration of the UDA1384 in the DSD
mode is shown in Figure 10.
Product data sheetRev. 02 — 17 January 200514 of 55
Philips Semiconductors
8.14 Microcontroller interface mode
The microcontroller interface mode can be selected as shown in Table 11:
• L3-bus mode when pin I2C_L3 = LOW
2
• I
C-bus mode when pin I2C_L3 = HIGH
Table 11: Pin function in the L3-bus or I2C-bus mode
PinLevel on pin I2C_L3
MCCLKL3CLOCKSCL
MCDATAL3DATASDA
MCMODEL3MODEQMUTE
Table 12: QMUTE
Signal QMUTEFunction
LOWno muting
HIGHmuting
UDA1384
Multichannel audio coder-decoder
LOWHIGH
L3-bus mode signalI2C-bus mode signal
All the features are accessible with the I2C-bus interface protocol as with the L3-bus
interface protocol.
The detailed description of the device operation in the L3-bus mode and I2C-bus mode is
given in Section 9 and Section 10, respectively.
9.L3-bus interface
9.1 General
The UDA1384 has an L3-bus microcontroller interface and all the digital sound processing
features and various system settings can be controlled by a microcontroller.
The exchange of data and control information between the microcontroller and the
UDA1384 is LSB first and is accomplished through a serial hardware L3-bus interface
comprising the following pins:
• MCCLK: clock line with signal L3CLOCK
• MCDATA: data line with signal L3DATA
• MCMODE: mode line with signal L3MODE
The L3-bus format has two modes of operation:
• Address mode
• Data transfer mode
The address mode is used to select a device for a subsequent data transfer. The address
mode is characterized by signal L3MODE = LOW and a burst of 8 pulses for signal
L3CLOCK, accompanied by 8 bits (see Figure 11).
Product data sheetRev. 02 — 17 January 200515 of 55
Philips Semiconductors
The data transfer mode is characterized by signal L3MODE = HIGH and is used to
transfer one or more bytes representing a register address, instruction or data.
Basically, two types of data transfers can be defined:
• Write action: data transfer to the device
• Read action: data transfer from the device.
9.2 Device addressing
The device address consists of one byte with:
• Data Operating Mode (DOM) bits 0 and 1 representing the type of data transfer (see
Table 13)
• Address bits 2 to 7 representing a 6-bit device address. The address of the UDA1384
is 01 0100 (bits 2 to 7).
Table 13: Selection of data transfer
DOMTransfer
UDA1384
Multichannel audio coder-decoder
Bit 1Bit 0
00not used
01not used
10write data or prepare read
11read data
9.3 Register addressing
After sending the device address (including DOM bits), indicating whether the information
is to be read or written, one data byte is sent using bit 0 to indicate whether the
information will be read or written and bits 1 to 7 for the destination register address.
Basically, there are 3 methods for register addressing:
1. Addressing for write data: bit 0 is logic 0 indicating a write action to the destination
register, followed by bits 1 to 7 indicating the register address (see Figure 11).
2. Addressing for prepare read: bit is logic 1, indicating that data will be read from the
register (see Figure 12).
3. Addressing for data read action. Here, the device returns a register address prior to
sending data from that register. When bit 0 is logic 0, the register address is valid;
when bit 0 is logic 1, the register address is invalid (see Figure 12).