AD AD1896YRS

a
FEATURES
Automatically Senses Sample Frequencies
No Programming Required
Attenuates Sample Clock Jitter
3.3 V–5 V Input and 3.3 V Core Supply Voltages
Accepts 16-/18-/20-/24-Bit Data
Up to 192 kHz Sample Rate
Input/Output Sample Ratios from 7.75:1 to 1:8
Bypass Mode
Multiple AD1896 TDM Daisy-Chain Mode
Multiple AD1896 Matched-Phase Mode
142 dB Signal-to-Noise and Dynamic Range
(A-Weighted, 20 Hz–20 kHz BW)
Up to –133 dB THD + N
Linear Phase FIR Filter
Hardware Controllable Soft Mute
Supports 256 fS, 512 fS, or 768 fS Master
Mode Clock
Flexible 3-Wire Serial Data Port with Left-Justified,
I2S, Right-Justified (16-,18-, 20-, 24-Bits), and
TDM Serial Port Modes
Master/Slave Input and Output Modes
28-Lead SSOP Plastic Package
APPLICATIONS
Home Theater Systems, Studio Digital Mixers,
Automotive Audio Systems, DVD, Set-Top Boxes,
Digital Audio Effects Processors, Studio-to-Transmitter
Links, Digital Audio Broadcast Equipment,
DigitalTape Varispeed Applications
PRODUCT OVERVIEW
The AD1896 is a 24-bit, high performance, single-chip, secondgeneration asynchronous sample rate converter. Based on Analog
Devices experience with its first asynchronous sample rate
converter, the AD1890, the AD1896 offers improved performance
and additional features. This improved performance includes a
THD + N range of –117 dB to –133 dB depending on the sample
rate and input frequency, 142 dB (A-Weighted) dynamic range,
192 kHz sampling frequencies for both input and output sample
rates, improved jitter rejection, and 1:8 upsampling and 7.75:1
downsampling ratios. Additional features include more serial
formats, a bypass mode, better interfacing to digital signal processors, and a matched-phase mode.
The AD1896 has a 3-wire interface for the serial input and
output ports that supports left-justified, I2S, and right-justified
(16-, 18-, 20-, 24-bit) modes. Additionally, the serial output
*Patents pending.
192 kHz Stereo Asynchronous
Sample Rate Converter
AD1896*
FUNCTIONAL BLOCK DIAGRAM
RESET
GRPDLYS
VDD_IO VDD_CORE
AD1896
MUTE_I
SDATA_I
SCLK_I
LRCLK_I
SMODE_IN_0
SMODE_IN_1
SMODE_IN_2
FSOUT
FIFO
SDATA_O
SCLK_O
LRCLK_O
FSIN
SERIAL
INPUT
TDM_IN
DIGITAL
PLL
BYPASS
FIR
FILTER
SERIAL
OUTPUT
SMODE_O_0
SMODE_O_1
MUTE_O
CLOCK DIVIDER
MCLK_I
MSMODE_0
MCLK_O
ROM
WLNGTH_O_0
WLNGTH_O_1
MSMODE_2
MSMODE_1
port supports TDM mode for daisy-chaining multiple AD1896s to
a digital signal processor. The serial output data is dithered down
to 20, 18, or 16 bits when 20-, 18-, or 16-bit output data is selected. The AD1896 sample rate converts the data from the
serial input port to the sample rate of the serial output port. The
sample rate at the serial input port can be asynchronous with
respect to the output sample rate of the output serial port. The
master clock to the AD1896, MCLK, can be asynchronous to
both the serial input and output ports.
MCLK can be generated either off-chip or on-chip by the AD1896
master clock oscillator. Since MCLK can be asynchronous to the
input or output serial ports, a crystal can be used to generate
MCLK internally to reduce noise and EMI emissions on the
board. When MCLK is synchronous to either the output or input
serial port, the AD1896 can be configured in a master mode where
MCLK is divided down and used to generate the left/right
and bit clocks for the serial port that is synchronous to MCLK.
The AD1896 supports master modes of 256 ¥ fS, 512 ¥ fS,
and 768 ¥ fS for both input and output serial ports.
Conceptually, the AD1896 interpolates the serial input data by
a rate of 220 and samples the interpolated data stream by the
output sample rate. In practice, a 64-tap FIR filter with 220
polyphases, a FIFO, a digital servo loop that measures the time
difference between the input and output samples within 5 ps,
and a digital circuit to track the sample rate ratio are used to
perform the interpolation and output sampling. Refer to the
Theory of Operation section. The digital servo loop and sample
rate ratio circuit automatically track the input and output
sample rates.
(Continued on Page 17)
REV. A
Information furnished by Analog Devices is believed to be accurate and
reliable. However, no responsibility is assumed by Analog Devices for its
use, nor for any infringements of patents or other rights of third parties that
may result from its use. No license is granted by implication or otherwise
under any patent or patent rights of Analog Devices. Trademarks and
registered trademarks are the property of their respective companies.
One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A.
Tel: 781/329-4700
www.analog.com
Fax: 781/326-8703
© 2003 Analog Devices, Inc. All rights reserved.
AD1896–SPECIFICATIONS
TEST CONDITIONS, UNLESS OTHERWISE NOTED.
Supply Voltages
VDD_CORE
VDD_IO
Ambient Temperature
Input Clock
Input Signal
Measurement Bandwidth
Word Width
Load Capacitance
Input Voltage High
Input Voltage Low
3.3 V
5.0 V or 3.3 V
25°C
30.0 MHz
1.000 kHz, 0 dBFS
20 to fS_OUT/2 Hz
24 Bits
50 pF
2.4 V
0.8 V
Specifications subject to change without notice.
DIGITAL PERFORMANCE (VDD_CORE = 3.3 V 5%, VDD_IO = 5.0 V 10%)
Parameter
Min
Typ
Max
Resolution
Sample Rate @ MCLK_I = 30 MHz
Sample Rate (@ Other Master Clocks)1
Sample Rate Ratios
Upsampling
Downsampling (Short GRPDLYS)
Downsampling (Long GRPDLYS)
Dynamic Range2
(20 Hz to fS_OUT/2, 1 kHz, –60 dBFS Input) A-Weighted
Worst-Case (192 kHz:48 kHz)
44.1 kHz:48 kHz
48 kHz:44.1 kHz
48 kHz:96 kHz
44.1 kHz:192 kHz
96 kHz:48 kHz
192 kHz:32 kHz
(20 Hz to fS_OUT/2, 1 kHz, –60 dBFS Input) No Filter
Worst-Case (192 kHz:48 kHz)
44.1 kHz:48 kHz
48 kHz:44.1 kHz
48 kHz:96 kHz
44.1 kHz:192 kHz
96 kHz:48 kHz
192 kHz:32 kHz
Total Harmonic Distortion + Noise2
(20 Hz to fS_OUT/2, 1 kHz, 0 dBFS Input) No Filter
Worst-Case (32 kHz:48 kHz)3
44.1 kHz:48 kHz
48 kHz:44.1 kHz
48 kHz:96 kHz
44.1 kHz:192 kHz
96 kHz:48 kHz
192 kHz:32 kHz
Interchannel Gain Mismatch
Interchannel Phase Deviation
Mute Attenuation (24 Bits Word Width) (A-Weighted)
24
6
215
MCLK_I/5000 ≤ fS < MCLK_I/138
Unit
Bits
kHz
kHz
1:8
7.75:1
7.0:1
132
142
141
142
141.5
140
140
dB
dB
dB
dB
dB
dB
dB
139
139
139
137
137
138
dB
dB
dB
dB
dB
dB
dB
–123
–124
–120
–123
–132
–133
0.0
0.0
–144
dB
dB
dB
dB
dB
dB
dB
dB
Degrees
dB
132
–117
NOTES
1
Lower sampling rates than given by this formula are possible, but the jitter rejection will decrease.
2
Refer to the Typical Performance Characteristics section for DNR and THD + N numbers over wide range of input and output sample rates.
3
For any other sample rate ratio, the minimum THD + N will be better than –117 dB. Please refer to detailed performance plots.
Specifications subject to change without notice.
–2–
REV. A
AD1896
DIGITAL TIMING (–40C < TA < +105C, VDD_CORE = 3.3 V 5%, VDD_IO = 5.0 V 10%)
Parameter1
tMCLKI
fMCLK
tMPWH
tMPWL
Min
MCLK_I Period
MCLK_I Frequency
MCLK_I Pulsewidth High
MCLK_I Pulsewidth Low
Typ
33.3
Output Serial Port Timing
tTDMS
TDM_IN Setup to SCLK_O Falling Edge
TDM_IN Hold from SCLK_O Falling Edge
tTDMH
SDATA_O Propagation Delay from SCLK_O, LRCLK_O
tDOPD
tDOH
SDATA_O Hold from SCLK_O
LRCLK_O Setup to SCLK_O (TDM Mode Only)
tLROS
LRCLK_O Hold from SCLK_O (TDM Mode Only)
tLROH
tSOH
SCLK_O Pulsewidth High
SCLK_O Pulsewidth Low
tSOL
RESET Pulsewidth Low
tRSTL
Propagation Delay from MCLK_I Rising Edge to SCLK_O Rising Edge
(Serial Output Port MASTER)
Propagation Delay from MCLK_I Rising Edge to LRCLK_O Rising Edge
(Serial Output Port MASTER)
9
12
8
8
8
8
3
ns
ns
ns
ns
ns
–3–
12
ns
12
ns
3
3
20
3
5
3
10
5
200
NOTES
1
Refer to Timing Diagrams section.
2
The maximum possible sample rate is: FSMAX = fMCLK /138.
3
fMCLK of up to 34 MHz is possible under the following conditions: 0∞C < TA < 70∞C, 45/55 or better MCLK_I duty cycle.
Specifications subject to change without notice.
Unit
ns
MHz
ns
ns
30.02, 3
Input Serial Port Timing
tLRIS
LRCLK_I Setup to SCLK_I
SCLK_I Pulsewidth High
tSIH
tSIL
SCLK_I Pulsewidth Low
SDATA_I Setup to SCLK_I Rising Edge
tDIS
SDATA_I Hold from SCLK_I Rising Edge
tDIH
Propagation Delay from MCLK_I Rising Edge to SCLK_I Rising Edge
(Serial Input Port MASTER)
Propagation Delay from MCLK_I Rising Edge to LRCLK_I Rising Edge
(Serial Input Port MASTER)
REV. A
Max
ns
ns
ns
ns
ns
ns
ns
ns
ns
12
ns
12
ns
AD1896
TIMING DIAGRAMS
MCLK I
LRCLK_I
t SIH
t LRIS
SCLK I
t DIS
RESET
t SIL
t RSTL
SDATA I
Figure 2. RESET Timing
t DIH
LRCLK O
tSOH
t MPWH
SCLK O
tSOL
tDOPD
SDATA O
tDOH
t MPWL
tLROS
Figure 3. MCLK_I Timing
LRCLK O
tLROH
SCLK O
tTDMS
TDM IN
tTDMH
Figure 1. Input and Output Serial Port Timing (SCLK I/O,
LRCLK I/O, SDATA I/O, TDM_IN)
–4–
REV. A
AD1896
DIGITAL FILTERS (VDD_CORE = 3.3 V 5%, VDD_IO = 5.0 V 10%)
Parameter
Pass-Band
Pass-Band Ripple
Transition Band
Stop-Band
Stop-Band Attenuation
Group Delay
Min
Typ
0.4535 fS_OUT
0.5465 fS_OUT
Max
Unit
0.4535 fS_OUT
± 0.016
0.5465 fS_OUT
Hz
dB
Hz
Hz
dB
Max
Unit
0.8
+2
–2
+150
–150
10
V
mA
mA
mA
mA
pF
V
V
mA
mA
–125
Refer to the Group Delay Equations section.
Specifications subject to change without notice.
DIGITAL I/O CHARACTERISTICS (VDD_CORE = 3.3 V 5%, VDD_IO = 5.0 V Parameter
Min
Input Voltage High (VIH)
Input Voltage Low (VIL)
Input Leakage (IIH @ VIH = 5 V)1
Input Leakage (IIL @ VIL = 0 V)1
Input Leakage (IIH @ VIH = 5 V)2
Input Leakage (IIL @ VIL = 0 V)2
Input Capacitance
Output Voltage High (VOH @ IOH = –4 mA)
Output Voltage Low (VOL @ IOL = +4 mA)
Output Source Current High (IOH)
Output Sink Current Low (IOL)
2.4
10%)
Typ
VDD_CORE – 0.5
5
VDD_CORE – 0.4
0.2
0.5
–4
+4
NOTES
1
All input pins except GRPDLYS.
2
GRPDLYS pin only.
Specifications subject to change without notice.
POWER SUPPLIES
Parameter
Supply Voltage
VDD_CORE
VDD_IO*
Active Supply Current
I_CORE_ACTIVE
48 kHz:48 kHz
96 kHz:96 kHz
192 kHz:192 kHz
I_IO_ACTIVE
Power-Down Supply Current: (All Clocks Stopped)
I_CORE_PWRDN
I_IO_PWRDN
Min
Typ
Max
Unit
3.135
VDD_CORE
3.3
3.3/5.0
3.465
5.5
V
V
20
26
43
2
mA
mA
mA
mA
0.5
10
mA
mA
*For 3.3 V tolerant inputs, VDD_IO supply should be set to 3.3 V; however, VDD_CORE supply voltage should not exceed VDD_IO.
Specifications subject to change without notice.
REV. A
–5–
AD1896
POWER SUPPLIES (VDD_CORE = 3.3 V 5%, VDD_IO = 5.0 V 10%)
Parameter
Min
Typ
Total Active Power Dissipation
48 kHz:48 kHz
96 kHz:96 kHz
192 kHz:192 kHz
Total Power-Down Dissipation: (RESET LO)
Max
Unit
65
85
132
2
mW
mW
mW
mW
Specifications subject to change without notice.
TEMPERATURE RANGE
Parameter
Min
Specifications Guaranteed
Functionality Guaranteed
Storage
Thermal Resistance, qJA (Junction to Ambient)
–40
–55
Typ
Max
Unit
+105
+150
∞C
∞C
∞C
∞C/W
Min
Max
Unit
–0.3
–0.3
+3.6
+6.0
V
V
DGND – 0.3
–40
± 10
VDD_IO + 0.3
+105
mA
V
∞C
25
109
Specifications subject to change without notice.
ABSOLUTE MAXIMUM RATINGS*
Parameter
Power Supplies
VDD_CORE
VDD_IO
Digital Inputs
Input Current
Input Voltage
Ambient Temperature (Operating)
*Stresses greater than those listed under Absolute Maximum Ratings may cause permanent damage to the device. This is a stress rating only; functional operation of the
device at these or any other conditions above those indicated in the operational section of this specification is not implied. Exposure to absolute maximum rating conditions
for extended periods may affect device reliability.
ORDERING GUIDE
Model
Temperature Range
Package Description
Package Option
AD1896AYRS
AD1896AYRSRL
–40∞C to +105∞C
–40∞C to +105∞C
28-Lead SSOP
28-Lead SSOP
RS-28
RS-28 on 13" Reel
CAUTION
ESD (electrostatic discharge) sensitive device. Electrostatic charges as high as 4000 V readily
accumulate on the human body and test equipment and can discharge without detection. Although
the AD1896 features proprietary ESD protection circuitry, permanent damage may occur on
devices subjected to high-energy electrostatic discharges. Therefore, proper ESD precautions are
recommended to avoid performance degradation or loss of functionality.
–6–
WARNING!
ESD SENSITIVE DEVICE
REV. A
AD1896
PIN CONFIGURATION
GRPDLYS 1
28 MMODE_2
MCLK_IN 2
27 MMODE_1
MCLK_OUT 3
SDATA_I 4
SCLK_I 5
LRCLK_I 6
26 MMODE_0
AD1896
25 SCLK_O
TOP VIEW
24 LRCLK_O
(NOT TO SCALE)
23 SDATA_O
VDD_IO 7
22 VDD_CORE
21 DGND
DGND 8
20 TDM_IN
BYPASS 9
SMODE_IN_0 10
19 SMODE_OUT_0
SMODE_IN_1 11
18 SMODE_OUT_1
SMODE_IN_2 12
17 WLNGTH_OUT_0
RESET 13
16 WLNGTH_OUT_1
15 MUTE_OUT
MUTE_IN 14
PIN FUNCTION DESCRIPTIONS
Pin No.
IN/OUT
Mnemonic
Description
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
IN
IN
OUT
IN
IN/OUT
IN/OUT
IN
IN
IN
IN
IN
IN
IN
IN
OUT
IN
IN
IN
IN
IN
IN
IN
OUT
IN/OUT
IN/OUT
IN
IN
IN
GRPDLYS
MCLK_IN
MCLK_OUT
SDATA_I
SCLK_I
LRCLK_I
VDD_IO
DGND
BYPASS
SMODE_IN_0
SMODE_IN_1
SMODE_IN_2
RESET
MUTE_IN
MUTE_OUT
WLNGTH_OUT_1
WLNGTH_OUT_0
SMODE_OUT_1
SMODE_OUT_0
TDM_IN
DGND
VDD_CORE
SDATA_O
LRCLK_O
SCLK_O
MMODE_0
MMODE_1
MMODE_2
Group Delay High = Short, Low = Long
Master Clock or Crystal Input
Master Clock Output or Crystal Output
Input Serial Data (at Input Sample Rate)
Master/Slave Input Serial Bit Clock
Master/Slave Input Left/Right Clock
3.3 V/5 V Input/Output Digital Supply Pin
Digital Ground Pin
ASRC Bypass Mode, Active High
Input Port Serial Interface Mode Select Pin 0
Input Port Serial Interface Mode Select Pin 1
Input Port Serial Interface Mode Select Pin 2
Reset Pin, Active Low
Mute Input Pin—Active High Normally Connected to MUTE_OUT
Output Mute Control, Active High
Hardware Selectable Output Wordlength—Select Pin 1
Hardware Selectable Output Wordlength—Select Pin 0
Output Port Serial Interface Mode Select Pin 1
Output Port Serial Interface Mode Select Pin 0
Serial Data Input* (Only for Daisy-Chain Mode). Ground when not used.
Digital Ground Pin
3.3 V Digital Supply Pin
Output Serial Data (at Output Sample Rate)
Master/Slave Output Left/Right Clock
Master/Slave Output Serial Bit Clock
Master/Slave Clock Ratio Mode Select Pin 0
Master/Slave Clock Ratio Mode Select Pin 1
Master/Slave Clock Ratio Mode Select Pin 2
*Also used to input matched-phase mode data.
REV. A
–7–
AD1896–Typical Performance Characteristics
0
0
–20
–20
–40
–40
–60
–60
–80
dBFS
dBFS
–80
–100
–120
–120
–140
–140
–160
–160
–180
–180
–200
–200
2.5
5.0
7.5
10.0 12.5 15.0
FREQUENCY – kHz
17.5
20.0
10
22.5
TPC 1. Wideband FFT Plot (16k Points) 0 dBFS 1 kHz Tone,
48 kHz:48 kHz (Asynchronous)
0
0
–20
–20
–40
–40
–60
–60
dBFS
dBFS
30
40
50
60
FREQUENCY – kHz
70
80
90
–80
–100
–100
–120
–120
–140
–140
–160
–160
–180
–180
–200
20
TPC 4. Wideband FFT Plot (16k Points) 44.1 kHz:192 kHz,
0 dBFS 1 kHz Tone
–80
2.5
5.0
7.5
10.0 12.5 15.0
FREQUENCY – kHz
17.5
20.0
–200
22.5
2.5
TPC 2. Wideband FFT Plot (16k Points) 0 dBFS 1 kHz Tone,
44.1 kHz:48 kHz (Asynchronous)
0
0
–20
–20
–40
–40
–60
–60
–80
–80
–100
–120
–140
–140
–160
–160
–180
–180
–200
10
15
20
25
30
FREQUENCY – kHz
35
40
7.5
10.0
12.5
15.0
FREQUENCY – kHz
17.5
20.0
–100
–120
5
5.0
TPC 5. Wideband FFT Plot (16k Points) 48 kHz:44.1 kHz,
0 dBFS 1 kHz Tone
dBFS
dBFS
–100
–200
45
2.5
5.0
7.5
10.0
12.5
15.0
17.5
20.0
22.5
FREQUENCY – kHz
TPC 3. Wideband FFT Plot (16k Points) 48 kHz:96 kHz,
0 dBFS 1 kHz Tone
TPC 6. Wideband FFT Plot (16k Points) 96 kHz:48 kHz,
0 dBFS 1 kHz Tone
–8–
REV. A
AD1896
0
–50
–60
–20
–70
–40
–80
–90
–60
–100
–110
dBFS
dBFS
–80
–100
–120
–120
–130
–140
–150
–140
–160
–160
–170
–180
–180
–190
–200
2.5
5.0
7.5
10.0 12.5 15.0
FREQUENCY – kHz
17.5
20.0
–200
22.5
–50
–50
–60
–60
–70
–80
–70
–90
–80
–90
–100
–100
–110
–110
dBFS
dBFS
15
20
25
30
35
40
45
TPC 10. Wideband FFT Plot (16k Points) 48 kHz:96 kHz,
–60 dBFS 1 kHz Tone
–120
–130
–140
–120
–130
–140
–150
–150
–160
–160
–170
–170
–180
–180
–190
–190
–200
–200
2.5
5.0
7.5
10.0 12.5 15.0
FREQUENCY – kHz
17.5
20.0
22.5
TPC 8. Wideband FFT Plot (16k Points) –60 dBFS 1 kHz
Tone, 48 kHz:48 kHz (Asynchronous)
10
20
30
40
50
60
FREQUENCY – kHz
70
80
90
TPC 11. Wideband FFT Plot (16k Points) 44.1 kHz:192 kHz,
–60 dBFS 1 kHz Tone
–50
–60
–50
–70
–70
–80
–90
–80
–100
–110
–100
–60
–90
–110
dBFS
dBFS
10
FREQUENCY – kHz
TPC 7. Wideband FFT Plot (16k Points) 192 kHz:48 kHz,
0 dBFS 1 kHz Tone
–120
–130
–120
–140
–130
–140
–150
–150
–160
–160
–170
–170
–180
–180
–190
–190
–200
–200
2.5
5.0
7.5
10.0 12.5 15.0
FREQUENCY – kHz
17.5
20.0
2.5
22.5
5.0
7.5
10.0
12.5
15.0
17.5
20.0
FREQUENCY – kHz
TPC 9. Wideband FFT Plot (16k Points) 44.1 kHz:48 kHz,
–60 dBFS 1 kHz Tone
REV. A
5
TPC 12. Wideband FFT Plot (16k Points) 48 kHz:44.1 kHz,
–60 dBFS 1 kHz Tone
–9–
AD1896
–50
0
–20
–40
–100
–60
–110
–120
–80
dBFS
dBFS
–60
–70
–80
–90
–130
–140
–100
–150
–160
–120
–170
–140
–180
–190
–160
–200
2.5
5.0
7.5
10.0 12.5 15.0
FREQUENCY – kHz
17.5
20.0
–180
22.5
2.5
TPC 13. Wideband FFT Plot (16k Points) 96 kHz:48 kHz,
–60 dBFS 1 kHz Tone
5.0
7.5
10.0 12.5 15.0
FREQUENCY – kHz
17.5
20.0
22.5
TPC 16. IMD, 10 kHz and 11 kHz 0 dBFS Tone
96 kHz:48 kHz
0
–50
–60
–20
–70
–80
–40
–90
–60
dBFS
dBFS
–100
–110
–120
–130
–140
–80
–100
–120
–150
–160
–140
–170
–180
–190
–200
–160
–180
2.5
5.0
7.5
10.0 12.5 15.0
FREQUENCY – kHz
17.5
20.0
22.5
2.5
TPC 14. Wideband FFT Plot (16k Points) 192 kHz:48 kHz,
–60 dBFS 1 kHz Tone
7.5
10.0
12.5 15.0
FREQUENCY – kHz
17.5
20.0
TPC 17. IMD, 10 kHz and 11 kHz 0 dBFS Tone
48 kHz:44.1 kHz
0
0
–20
–20
–40
–40
–60
–60
–80
–80
dBFS
dBFS
5.0
–100
–100
–120
–120
–140
–140
–160
–160
–180
–180
2.5
5.0
7.5
10.0 12.5 15.0
FREQUENCY – kHz
17.5
20.0
–200
22.5
2.5
TPC 15. IMD, 10 kHz and 11 kHz 0 dBFS Tone
44:1 kHz:48 kHz
5.0
7.5
10.0 12.5 15.0
FREQUENCY – kHz
17.5
20.0
22.5
TPC 18. Wideband FFT Plot (16k Points) 44.1 kHz:48 kHz,
0 dBFS 20 kHz Tone
–10–
REV. A
0
0
–20
–20
–40
–40
–60
–60
–80
–80
dBFS
dBFS
AD1896
–100
–120
–120
–140
–140
–160
–160
–180
–180
–200
–200
10
20
30
40
50
60
FREQUENCY – kHz
70
80
90
5
TPC 19. Wideband FFT Plot (16k Points) 192 kHz:192 kHz,
0 dBFS 80 kHz Tone
10
15
20
25
30
FREQUENCY – kHz
35
40
45
TPC 22. Wideband FFT Plot (16k Points) 48 kHz:96 kHz,
0 dBFS 20 kHz Tone
0
0
–20
–20
–40
–40
–60
–60
–80
dBFS
–80
dBFS
–100
–100
–100
–120
–120
–140
–140
–160
–160
–180
–180
–200
–200
2.5
5.0
7.5
10.0 12.5 15.0 17.5
FREQUENCY – kHz
20.0
2.5
22.5
TPC 20. Wideband FFT Plot (16k Points) 48 kHz:48 kHz,
0 dBFS 20 kHz Tone
5.0
7.5
10.0 12.5 15.0
FREQUENCY – kHz
17.5
20.0
22.5
TPC 23. Wideband FFT Plot (16k Points) 96 kHz:48 kHz,
0 dBFS 20 kHz Tone
0
–119
–20
–121
–40
–123
THD+N – dBFS
–60
dBFS
–80
–100
–120
–125
–127
–129
–140
–131
–160
–133
–180
–200
–135
2.5
5.0
7.5
10.0
12.5
15.0
FREQUENCY – kHz
17.5
20.0
TPC 21. Wideband FFT Plot (16k Points) 48 kHz:44:1 kHz,
0 dBFS 20 kHz Tone
REV. A
30
55
80
105
130
155
180
OUTPUT SAMPLE RATE – kHz
TPC 24. THD + N vs. Output Sample Rate, fS_IN = 192 kHz,
0 dBFS 1 kHz Tone
–11–
–119
–119
–121
–121
–123
–123
THD+N – dBFS
THD+N – dBFS
AD1896
–125
–127
–129
–125
–127
–129
–131
–131
–133
–133
–135
–135
30
55
80
105
130
155
30
180
55
OUTPUT SAMPLE RATE – kHz
TPC 25. THD + N vs. Output Sample Rate, fS_IN = 48 kHz,
0 dBFS 1 kHz Tone
80
105
130
155
OUTPUT SAMPLE RATE – kHz
180
TPC 28. THD + N vs. Output Sample Rate, fS_IN = 96 kHz,
0 dBFS 1 kHz Tone
–130
–119
–131
–132
–123
–133
–125
–134
DNR – dBFS
THD+N – dBFS
–121
–127
–129
–135
–136
–137
–131
–138
–133
–135
30
–139
55
80
105
130
155
–140
30
180
55
OUTPUT SAMPLE RATE – kHz
TPC 26. THD + N vs. Output Sample Rate, fS_IN = 44.1 kHz,
0 dBFS 1 kHz Tone
80
105
130
155
OUTPUT SAMPLE RATE – kHz
180
TPC 29. DNR vs. Output Sample Rate, fS_IN = 192 kHz,
–60 dBFS 1 kHz Tone
–135
–119
–136
–137
–123
–138
–125
–139
DNR – dBFS
THD+N – dBFS
–121
–127
–129
–140
–141
–142
–131
–143
–133
–135
30
–144
55
80
105
130
155
OUTPUT SAMPLE RATE – kHz
–145
30
180
TPC 27. THD + N vs. Output Sample Rate, fS_IN = 32 kHz,
0 dBFS 1 kHz Tone
55
80
105
130
155
OUTPUT SAMPLE RATE – kHz
180
TPC 30. DNR vs. Output Sample Rate, fS_IN = 32 kHz,
–60 dBFS 1 kHz Tone
–12–
REV. A
AD1896
–135
–130
–131
–136
–132
DNR – dBFS
DNR – dBFS
–133
–134
–135
–136
–137
–138
–137
–139
–138
–139
–140
30
–140
55
80
105
130
155
30
180
55
80
105
130
155
180
OUTPUT SAMPLE RATE – kHz
OUTPUT SAMPLE RATE – kHz
TPC 34. DNR vs. Output Sample Rate, fS_IN = 44.1 kHz,
–60 dBFS 1 kHz Tone
TPC 31. DNR vs. Output Sample Rate, fS_IN = 96 kHz,
–60 dBFS 1 kHz Tone
0.00
0
–0.01
–20
–0.02
–40
–0.03
192kHz:96kHz
–0.04
192kHz:48kHz
dBFS
dBFS
–60
–80
–0.05
–0.06
–0.07
–100 192kHz:32kHz
–0.08
192kHz:48kHz
–120
–0.09
–140
–0.10
0
10
20
30
40
FREQUENCY – kHz
50
60
0
4
–137
3
LINEARITY ERROR – dBr
5
–136
DNR – dBFS
–138
–139
–140
–141
–142
14
16
18
20
22
24
–1
–2
–4
–145
155
–5
–140
180
OUTPUT SAMPLE RATE – kHz
–120
–100
–80
–60
–40
INPUT LEVEL – dBFS
–20
TPC 36. Linearity Error, 48 kHz:48 kHz, 0 dBFS to
–140 dBFS Input, 200 Hz Tone
TPC 33. DNR vs. Output Sample Rate, fS_IN = 48 kHz,
–60 dBFS 1 kHz Tone
REV. A
12
1
–144
130
10
0
–3
105
8
2
–143
80
6
TPC 35. Pass-Band Ripple, 192 kHz:48 kHz
–135
55
4
FREQUENCY – kHz
TPC 32. Digital Filter Frequency Response
30
2
–13–
0
5
5
4
4
3
3
LINEARITY ERROR – dBr
LINEARITY ERROR – dBr
AD1896
2
1
0
–1
–2
2
1
0
–1
–2
–3
–3
–4
–4
–5
–140
–5
–120
–100
–80
–60
–40
–20
–140
0
–120
–100
INPUT LEVEL – dBFS
5
4
3
3
LINEARITY ERROR – dBr
LINEARITY ERROR – dBr
5
4
2
1
0
–1
–2
2
1
0
–1
–2
–3
–3
–4
–4
–5
–5
–120
–100
–80
–60
–40
–20
–140
0
–120
–100
INPUT LEVEL – dBFS
5
4
4
3
3
LINEARITY ERROR – dBr
LINEARITY ERROR – dBr
5
2
1
0
–1
–2
0
–1
–2
–3
–4
–5
–80
–60
–40
INPUT LEVEL – dBFS
–20
0
1
–4
–100
–20
2
–3
–120
–80
–60
–40
INPUT LEVEL – dBFS
TPC 41. Linearity Error, 44.1 kHz:192 kHz, 0 dBFS to
–140 dBFS Input, 200 Hz Tone
TPC 38. Linearity Error, 96 kHz:48 kHz, 0 dBFS to
–140 dBFS Input, 200 Hz Tone
–140
0
–20
TPC 40. Linearity Error, 48 kHz:96 kHz, 0 dBFS to
–140 dBFS Input, 200 Hz Tone
TPC 37. Linearity Error, 48 kHz:44.1 kHz, 0 dBFS to
–140 dBFS Input, 200 Hz Tone
–140
–80
–60
–40
INPUT LEVEL – dBFS
–5
0
–140
–120
–100
–80
–60
–40
INPUT LEVEL – dBFS
–20
0
TPC 42. Linearity Error, 192 kHz:44:1 kHz, 0 dBFS to
–140 dBFS Input, 200 Hz Tone
TPC 39. Linearity Error, 44.1 kHz:48 kHz, 0 dBFS to
–140 dBFS Input, 200 Hz Tone
–14–
REV. A
–110
–110
–115
–115
–120
–120
–125
–125
–130
–130
–135
–135
–140
–140
dBr
dBr
AD1896
–145
–150
–155
–155
–160
–160
–165
–165
–170
–170
–175
–175
–180
–140
–120
–100
–80
–60
–40
INPUT LEVEL – dBFS
–20
–180
–140
0
–110
–115
–115
–120
–120
–125
–125
–130
–130
–135
–135
–140
–140
dBr
dBr
–110
–145
–150
–155
–155
–160
–160
–165
–165
–170
–170
–175
–175
–120
–100
–80
–60
–40
INPUT LEVEL – dBFS
–20
–180
–140
0
–110
–115
–115
–120
–120
–125
–125
–130
–130
–135
–135
–140
–140
dBr
–110
–145
–150
–155
–155
–160
–160
–165
–165
–170
–170
–175
–175
–100
–80
–60
–40
INPUT LEVEL – dBFS
–20
–180
–140
0
0
–120
–100
–80
–60
–40
INPUT LEVEL – dBFS
–20
0
–120
–100
–80
–60
–40
INPUT LEVEL – dBFS
–20
0
TPC 48. THD + N vs. Input Amplitude, 192 kHz:48 kHz,
1 kHz Tone
TPC 45. THD + N vs. Input Amplitude, 44.1 kHz:48 kHz,
1 kHz Tone
REV. A
–20
–145
–150
–120
–80
–60
–40
INPUT LEVEL – dBFS
TPC 47. THD + N vs. Input Amplitude, 44.1 kHz:192 kHz,
1 kHz Tone
TPC 44. THD + N vs. Input Amplitude, 96 kHz:48 kHz,
1 kHz Tone
–180
–140
–100
–145
–150
–180
–140
–120
TPC 46. THD + N vs. Input Amplitude, 48 kHz:96 kHz,
1 kHz Tone
TPC 43. THD + N vs. Input Amplitude, 48 kHz:44.1 kHz,
1 kHz Tone
dBr
–145
–150
–15–
AD1896
–110
–115
–115
–120
–120
–125
–125
–130
–130
–135
–135
–140
–140
dBr
dBr
–110
–145
–150
–155
–155
–160
–160
–165
–165
–170
–170
–175
–175
–180
–180
2.5
5.0
7.5
10.0
12.5
FREQUENCY – kHz
15.0
17.5
20.0
2.5
TPC 49. THD + N vs. Frequency Input, 48 kHz:44.1 kHz,
0 dBFS
5.0
7.5
10.0
12.5
FREQUENCY – kHz
15.0
17.5
20.0
TPC 51. THD + N vs. Frequency Input, 48 kHz:96 kHz,
0 dBFS
–110
–110
–115
–115
–120
–120
–125
–125
–130
–130
–135
–135
–140
–140
dBr
dBr
–145
–150
–145
–150
–145
–150
–155
–155
–160
–160
–165
–165
–170
–170
–175
–175
–180
–180
2.5
5.0
7.5
10.0
12.5
FREQUENCY – kHz
15.0
17.5
20.0
2.5
5.0
7.5
10.0
12.5
FREQUENCY – kHz
15.0
17.5
20.0
TPC 52. THD + N vs. Frequency Input, 96 kHz:48 kHz,
0 dBFS
TPC 50. THD + N vs. Frequency Input, 44.1 kHz:48 kHz,
0 dBFS
–16–
REV. A
AD1896
(Continued from Page 1)
The digital servo loop measures the time difference between
the input and output sample rates within 5 ps. This is necessary
in order to select the correct polyphase filter coefficient. The
digital servo loop has excellent jitter rejection for both input and
output sample rates as well as the master clock. The jitter rejection begins at less than 1 Hz. This requires a long settling
time whenever RESET is deasserted or when the input or
output sample rate changes. To reduce the settling time, upon
deassertion of RESET or a change in a sample rate, the digital
servo loop enters the fast settling mode. When the digital servo
loop has adequately settled in the fast mode, it switches into the
normal or slow settling mode and continues to settle until the
time difference measurement between input and output sample
rates is within 5 ps. During fast mode, the MUTE_OUT signal
is asserted high. Normally, the MUTE_OUT is connected to the
MUTE_IN pin. The MUTE_IN signal is used to softly mute
the AD1896 upon assertion and softly unmute the AD1896
when it is deasserted.
The sample rate ratio circuit is used to scale the filter length of
the FIR filter for decimation. Hysteresis in measuring the
sample rate ratio is used to avoid oscillations in the scaling of
the filter length, which would cause distortion on the output.
REV. A
However, when multiple AD1896s are used with the same serial
input port clock and the same serial output port clock, the hysteresis causes different group delays between multiple AD1896s.
A phase-matching mode feature was added to the AD1896 to
address this problem. In phase-matching mode, one AD1896,
the master, transmits its sample rate ratio to the other AD1896s,
the slaves, so that the group delay between the multiple AD1896s
remains the same.
The group delay of the AD1896 can be adjusted for short or
long delay. An address offset is added to the write pointer of the
FIFO in the sample rate converter. This offset is set to 16 for
short delay and 64 for long delay. In long delay, the group delay
is effectively increased by 48 input sample clocks.
The sample rate converter of the AD1896 can be bypassed
altogether using the bypass mode. In bypass mode, the AD1896’s
serial input data is directly passed to the serial output port without any dithering. This is useful for passing through nonaudio
data or when the input and output sample rates are synchronous
to one another and the sample rate ratio is exactly 1 to 1.
The AD1896 is a 3.3 V, 5 V input tolerant part and is available
in a 28-lead SSOP package. The AD1896 is 5 V input-tolerant
only when the VDD_IO supply pin is supplied with 5 V.
–17–
AD1896
ASRC FUNCTIONAL OVERVIEW
THEORY OF OPERATION
Asynchronous sample rate conversion is converting data from
one clock source at some sample rate to another clock source at
the same or a different sample rate. The simplest approach to an
asynchronous sample rate conversion is the use of a zero-order
hold between the two samplers shown in Figure 4. In an asynchronous system, T2 is never equal to T1 nor is the ratio between
T2 and T1 rational. As a result, samples at fS_OUT will be repeated
or dropped producing an error in the resampling process. The
frequency domain shows the wide side lobes that result from
this error when the sampling of fS_OUT is convolved with the
attenuated images from the sin(x)/x nature of the zero-order
hold. The images at fS_IN, dc signal images, of the zero-order
hold are infinitely attenuated. Since the ratio of T2 to T1 is an
irrational number, the error resulting from the resampling at
fS_OUT can never be eliminated. However, the error can be significantly reduced through interpolation of the input data at
fS_IN. The AD1896 is conceptually interpolated by a factor of 220.
IN
ZERO-ORDER
HOLD
fS_IN = 1/T1
between each fS_IN sample and convolving this interpolated
signal with a digital low-pass filter to suppress the images. In the
time domain, it can be seen that fS_OUT selects the closest fS_IN ¥ 220
sample from the zero-order hold as opposed to the nearest fS_IN
sample in the case of no interpolation. This significantly reduces
the resampling error.
IN
INTERPOLATE
BY N
LOW-PASS
FILTER
OUT
ZERO-ORDER
HOLD
fS_IN
fS_OUT
TIME DOMAIN OF fS_IN SAMPLES
TIME DOMAIN OUTPUT OF THE LOW-PASS FILTER
OUT
fS_OUT = 1/T2
TIME DOMAIN OF fS_OUT RESAMPLING
ORIGINAL SIGNAL
SAMPLED AT fS_IN
TIME DOMAIN OF THE ZERO-ORDER HOLD OUTPUT
Figure 5. Time Domain of the Interpolation and
Resampling
SIN(X)/X OF ZERO-ORDER HOLD
SPECTRUM OF ZERO-ORDER HOLD OUTPUT
SPECTRUM OF fS_OUT SAMPLING
2 fS_OUT
fS_OUT
FREQUENCY RESPONSE OF fS_OUT CONVOLVED WITH ZERO-ORDER
HOLD SPECTRUM
Figure 4. Zero-Order Hold Being Used by fS_OUT to
Resample Data from fS_IN
THE CONCEPTUAL HIGH INTERPOLATION MODEL
Interpolation of the input data by a factor of 220 involves placing
(220 – 1) samples between each fS_IN sample. Figure 5 shows
both the time domain and the frequency domain of interpolation
by a factor of 220. Conceptually, interpolation by 220 would
involve the steps of zero-stuffing (220 – 1) number of samples
In the frequency domain shown in Figure 6, the interpolation
expands the frequency axis of the zero-order hold. The images
from the interpolation can be sufficiently attenuated by a good
low-pass filter. The images from the zero-order hold are now
pushed by a factor of 220 closer to the infinite attenuation point
of the zero-order hold, which is fS_IN ¥ 220. The images at the
zero-order hold are the determining factor for the fidelity of the
output at fS_OUT. The worst-case images can be computed from
the zero-order hold frequency response, maximum image =
sin (p ¥ F/fS_INTERP)/(p ¥ F/fS_INTERP). F is the frequency of the
worst-case image that would be 220 ¥ fS_IN ± fS_IN/2 , and
fS_INTERP is fS_IN ¥ 220.
The following worst-case images would appear for fS_IN =
192 kHz:
–18–
Image at fS_INTERP – 96 kHz = –125.1 dB
Image at fS_INTERP + 96 kHz = –125.1 dB
REV. A
AD1896
IN
INTERPOLATE
BY N
LOW-PASS
FILTER
ZERO-ORDER
HOLD
fS_IN
OUT
fS_OUT
FREQUENCY DOMAIN OF SAMPLES AT fS_IN
the output samples is less than the Nyquist frequency of the
input samples. To move the cutoff frequency of the antialiasing
filter, the coefficients are dynamically altered and the length
of the convolution is increased by a factor of (fS_IN/fS_OUT).
This technique is supported by the Fourier transform property
that if f(t) is F(w), then f(k ¥ t) is F(w/k). Thus, the range of
decimation is simply limited by the size of the RAM.
fS_IN
THE SAMPLE RATE CONVERTER ARCHITECTURE
FREQUENCY DOMAIN OF THE INTERPOLATION
220 fS_IN
SIN(X)/X OF ZERO-ORDER HOLD
FREQUENCY DOMAIN OF fS_OUT RESAMPLING
FREQUENCY DOMAIN AFTER
RESAMPLING
220 fS_IN
The architecture of the sample rate converter is shown in
Figure 7. The sample rate converter’s FIFO block adjusts the
left and right input samples and stores them for the FIR filter’s
convolution cycle. The fS_IN counter provides the write address
to the FIFO block and the ramp input to the digital servo
loop. The ROM stores the coefficients for the FIR filter convolution and performs a high order interpolation between the
stored coefficients. The sample rate ratio block measures the
sample rate for dynamically altering the ROM coefficients and
scaling of the FIR filter length as well as the input data. The
digital servo loop automatically tracks the fS_IN and fS_OUT
sample rates and provides the RAM and ROM start addresses
for the start of the FIR filter convolution.
220 fS_IN
RIGHT DATA IN
LEFT DATA IN
Figure 6. Frequency Domain of the Interpolation and
Resampling
FIFO
The difficulty with the above approach is that the correct interpolated sample needs to be selected upon the arrival of fS_OUT.
Since there are 220 possible convolutions per fS_OUT period, the
arrival of the fS_OUT clock must be measured with an accuracy
of 1/201.3 GHz = 4.96 ps. Measuring the fS_OUT period with a
clock of 201.3 GHz frequency is clearly impossible; instead,
several coarse measurements of the fS_OUT clock period are made
and averaged over time.
Another difficulty with the above approach is the number of
coefficients required. Since there are 220 possible convolutions
with a 64-tap FIR filter, there needs to be 220 polyphase coefficients for each tap, which requires a total of 226 coefficients. To
reduce the amount of coefficients in ROM, the AD1896 stores a
small subset of coefficients and performs a high order interpolation between the stored coefficients. So far the above approach
works for the case of fS_OUT > fS_IN. However, in the case when
the output sample rate, fS_OUT, is less than the input sample
rate, fS_IN, the ROM starting address, input data, and the length
of the convolution must be scaled. As the input sample rate
rises over the output sample rate, the antialiasing filter’s cutoff
frequency has to be lowered because the Nyquist frequency of
REV. A
HIGH
ROM B
ORDER
ROM C
INTERP
ROM D
HARDWARE MODEL
The output rate of the low-pass filter of Figure 5 would be the
interpolation rate, 220 ¥ 192000 kHz = 201.3 GHz. Sampling at
a rate of 201.3 GHz is clearly impractical, not to mention the
number of taps required to calculate each interpolated sample.
However, since interpolation by 220 involves zero-stuffing 220– 1
samples between each fS_IN sample, most of the multiplies in
the low-pass FIR filter are by zero. A further reduction can be
realized by the fact that since only one interpolated sample is
taken at the output at the fS_OUT rate, only one convolution
needs to be performed per fS_OUT period instead of 220 convolutions. A 64-tap FIR filter for each fS_OUT sample is sufficient
to suppress the images caused by the interpolation.
ROM A
fS_IN
COUNTER
DIGITAL
SERVO LOOP
SAMPLE RATE RATIO
FIR FILTER
fS_IN
fS_OUT
SAMPLE RATE
RATIO
L/R DATA OUT
EXTERNAL
RATIO
Figure 7. Architecture of the Sample Rate Converter
The FIFO receives the left and right input data and adjusts the
amplitude of the data for both the soft muting of the sample rate
converter and the scaling of the input data by the sample rate
ratio before storing the samples in the RAM. The input data is
scaled by the sample rate ratio because as the FIR filter length
of the convolution increases, so does the amplitude of the
convolution output. To keep the output of the FIR filter from
saturating, the input data is scaled down by multiplying it by
(fS_OUT/fS_IN) when fS_OUT < fS_IN. The FIFO also scales the input
data for muting and unmuting of the AD1896.
The RAM in the FIFO is 512 words deep for both left and right
channels. An offset to the write address provided by the fS_IN
counter is added to prevent the RAM read pointer from ever
overlapping the write address. The offset is selectable by the
GRPDLYS, group delay select, signal. A small offset, 16, is
added to the write address pointer when GRPDLYS is high,
and a large offset, 64, is added to the write address pointer when
GRPDLYS is low. Increasing the offset of the write address pointer
is useful for applications when small changes in the sample rate
ratio between fS_IN and fS_OUT are expected. The maximum decimation rate can be calculated from the RAM word depth and
GRPDLYS as (512 – 16)/64 taps = 7.75 for short group delay and
(512 – 64)/64 taps = 7 for long group delay.
–19–
AD1896
10
0
–10
–20
–30
–40
SLOW MODE
–50
FAST MODE
–60
–70
–80
–90
–100
–110
–120
–130
–140
–150
–160
–170
–180
–190
–200
–210
–220
0.01
0.1
1
10
100
1e3
1e4
1e5
FREQUENCY – Hz
Figure 8. Frequency Response of the Digital Servo Loop. fS_IN Is the X-Axis, fS_OUT = 192 kHz, Master Clock
Frequency Is 30 MHz.
The digital servo loop is essentially a ramp filter that provides
the initial pointer to the address in RAM and ROM for the start
of the FIR convolution. The RAM pointer is the integer output
of the ramp filter while the ROM is the fractional part. The
digital servo loop must be able to provide excellent rejection of
jitter on the fS_IN and fS_OUT clocks as well as measure the arrival
of the fS_OUT clock within 4.97 ps. The digital servo loop will
also divide the fractional part of the ramp output by the ratio of
fS_IN/fS_OUT for the case when fS_IN > fS_OUT, to dynamically alter
the ROM coefficients.
The digital servo loop is implemented with a multirate filter. To
settle the digital servo loop filter quicker upon start-up or a change
in the sample rate, a “fast mode” was added to the filter. When
the digital servo loop starts up or the sample rate is changed, the
digital servo loop kicks into “fast mode” to adjust and settle
on the new sample rate. Upon sensing the digital servo loop
settling down to some reasonable value, the digital servo loop
will kick into “normal” or “slow mode.” During “fast mode”
the MUTE_OUT signal of the sample rate converter is asserted
to let the user know that they should mute the sample rate
converter to avoid any clicks or pops. The frequency response of
the digital servo loop for “fast mode” and “slow mode” are
shown in Figure 8.
The FIR filter is a 64-tap filter in the case of fS_OUT ≥ fS_IN and is
(fS_IN/fS_OUT) ¥ 64 taps for the case when fS_IN > fS_OUT. The FIR
filter performs its convolution by loading in the starting address
of the RAM address pointer and the ROM address pointer
from the digital servo loop at the start of the fS_OUT period.
The FIR filter then steps through the RAM by decrementing its
address by 1 for each tap, and the ROM pointer increments its
address by the (fS_OUT/fS_IN) ¥ 220 ratio for fS_IN > fS_OUT or 220
for fS_OUT ≥ fS_IN. Once the ROM address rolls over, the convolution is completed. The convolution is performed for both
the left and right channels, and the multiply accumulate circuit
used for the convolution is shared between the channels.
The fS_IN/fS_OUT sample rate ratio circuit is used to dynamically
alter the coefficients in the ROM for the case when fS_IN >
fS_OUT. The ratio is calculated by comparing the output of an
fS_OUT counter to the output of an fS_IN counter. If fS_OUT >
fS_IN, the ratio is held at one. If fS_IN > fS_OUT, the sample rate
ratio is updated if it is different by more than two fS_OUT periods
from the previous fS_OUT to fS_IN comparison. This is done to
provide some hysteresis to prevent the filter length from oscillating and causing distortion.
–20–
REV. A
AD1896
However, the hysteresis of the fS_OUT/fS_IN ratio circuit can cause
phase mismatching between two AD1896s operating with the
same input clock and the same output clock. Since the hysteresis requires a difference of more than two fS_OUT periods for the
fS_OUT/fS_IN ratio to be updated, two AD1896s may have differences in their ratios from 0 to 4 fS_OUT period counts. The
fS_OUT/fS_IN ratio adjusts the filter length of the AD1896, which
corresponds directly with the group delay. Thus, the magnitude
in the phase difference will depend upon the resolution of the
fS_OUT and fS_IN counters. The greater the resolution of the
counters, the smaller the phase difference error will be.
The fS_IN and fS_OUT counters of the AD1896 have three bits of
extra resolution over the AD1890, which reduces the phase
mismatch error by a factor of 8. However, an additional feature
was added to the AD1896 to eliminate the phase mismatching
completely. One AD1896 can set the fS_OUT/fS_IN ratio of other
AD1896s by transmitting its fS_OUT/fS_IN ratio through the
serial output port.
OPERATING FEATURES
RESET and Power-Down
When RESET is asserted low, the AD1896 will turn off the
master clock input to the AD1896, MCLK_I, initialize all of its
internal registers to their default values, and three-state all of the
I/O pins. While RESET is active low, the AD1896 is consuming
minimum power. For the lowest possible power consumption
while RESET is active low, all of the input pins to the AD1896
should be static.
When RESET is deasserted, the AD1896 begins its initialization
routine where all locations in the FIFO are initialized to zero,
MUTE_OUT is asserted high, and any I/O pins configured as
outputs are enabled. When RESET is deasserted, the master
serial port clock pins SCLK_I/O and LRCLK_I/O become
active after 1024 MCLK-I cycles. The mute control counter,
which controls the soft mute attenuation of the input samples,
is initialized to maximum attenuation, –144 dB (see the Mute
Control section).
When asserting RESET and deasserting RESET, the RESET
should be held low for a minimum of five MCLK_I cycles.
During power-up, the RESET should be held low until the
power supplies have stabilized. It is recommended that the
AD1896 be reset when changing modes.
Power Supply and Voltage Reference
The AD1896 is designed for 3 V operation with 5 V input tolerance on the input pins. VDD_CORE is the 3 V supply that is
used to power the core logic of the AD1896 and to drive the
output pins. VDD_IO is used to set the input voltage tolerance
of the input pins. In order for the input pins to be 5 V input
tolerant, VDD_IO must be connected to a 5 V supply. If the
input pins do not have to be 5 V input tolerant, then
VDD_IO can be connected to VDD_CORE. VDD_IO should
never be less than VDD_CORE. VDD_CORE and VDD_IO
should be bypassed with 100 nF ceramic chip capacitors, as
close to the pins as possible, to minimize power supply and
ground bounce caused by inductance in the traces. A bulk aluminium electrolytic capacitor of 47 mF should also be provided
on the same PC board as the AD1896.
REV. A
Digital Filter Group Delay
The group delay of the digital filter may be selected by the logic
pin GRPDLYS. As mentioned in the Theory of Operation section,
this pin is particularly useful in varispeed applications. The
GRPDLYS pin has an internal pull-up resistor of approximately
33 kW to VDD_CORE. When GRPDLYS is high, the filter group
delay will be short and is given by the equation:
GDS =
16
32
+
seconds for fS_OUT > fS_IN
fS_IN
fS_IN
GDS =
16
Ê 32 ˆ Ê fS_IN ˆ
+Á
˜ ¥Á
˜ seconds for fS_OUT < fS_IN
fS_IN Ë fS_IN ¯ Ë fS_OUT ¯
For short filter group delay, the GRPDLYS pin can be left open.
When GRPDLYS is low, the group delay of the filter will be
long and is given by the equation:
GDL =
GDL =
64
32
+
seconds for fS_OUT > fS_IN
fS_IN
fS_IN
64
fS_IN
Ê 32 ˆ Ê fS_IN ˆ
+Á
˜ ¥Á
˜ seconds for fS_OUT < fS_IN
Ë fS_IN ¯ Ë fS_OUT ¯
NOTE: For the long group delay mode, the decimation ratio is
limited to less than 7:1.
Mute Control
When the MUTE_IN pin is asserted high, the MUTE_IN control
will perform a soft mute by linearly decreasing the input data to
the AD1896 FIFO to zero, –144 dB attenuation. When MUTE_IN
is deasserted low, the MUTE_IN control will linearly decrease
the attenuation of the input data to 0 dB. A 12-bit counter,
clocked by LRCLK_I, is used to control the mute attenuation.
Therefore, the time it will take from the assertion of MUTE_IN
to –144 dB full mute attenuation is 4096/LRCLK_I seconds.
Likewise, the time it will take to reach 0 dB mute attenuation from
the deassertion of MUTE_IN is 4096/LRCLK_I seconds.
Upon RESET, or a change in the sample rate between LRCLK_I
and LRCLK_O, the MUTE_OUT pin will be asserted high. The
MUTE_OUT pin will remain asserted high until the digital
servo loop’s internal fast settling mode has completed. When
the digital servo loop has switched to slow settling mode, the
MUTE_OUT pin will deassert. While MUTE_OUT is asserted,
the MUTE_IN pin should be asserted as well to prevent any
major distortion in the audio output samples.
Master Clock
A digital clock connected to the MCLK_I pin or a fundamental
or third overtone crystal connected between MCLK_I and
MCLK_O can be used to generate the master clock, MCLK_I.
The MCLK_I pin can be 5 V input tolerant just like any of the
other AD1896 input pins. A fundamental mode crystal can be
inserted between MCLK_I and MCLK_O for master clock
frequency generation up to 27 MHz. For master clock frequency generation with a crystal beyond 27 MHz, it is
recommended that the user use a third overtone crystal and to
add an LC filter at the output of MCLK_O to filter out the
fundamental, do not notch filter the fundamental. Please refer
to your quartz crystal supplier for values for external capacitors and inductor components.
–21–
AD1896
Table I. Serial Data Input Port Mode
SMODE_IN_[0:2]
AD1896
MCLK_I
MCLK_O
R
C1
C2
Figure 9a. Fundamental-Mode Circuit Configuration
AD1896
MCLK_I
MCLK_O
R
1nF
C1
C2
L1
Figure 9b. Third-Overtone Circuit Configuration
There are, of course, maximum and minimum operating frequencies for the AD1896 master clock. The maximum master
clock frequency at which the AD1896 is guaranteed to operate is
30 MHz. A frequency of 30 MHz is more than sufficient to
sample rate convert sampling frequencies of 192 kHz + 12%.
The minimum required frequency for the master clock generation
for the AD1896 depends upon the input and output sample
rates. The master clock has to be at least 138 times greater than
the maximum input or output sample rate.
Serial Data Ports—Data Format
The serial data input port mode is set by the logic levels on the
SMODE_IN_0/SMODE_IN_1/SMODE_IN_2 pins. The serial
data input port modes available are left justified, I2S, and right
justified (RJ), 16, 18, 20, or 24 bits as defined in Table I.
Interface Format
2
1
0
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
Left Justified
I 2S
Undefined
Undefined
Right Justified, 16 Bits
Right Justified, 18 Bits
Right Justified, 20 Bits
Right Justified, 24 Bits
The serial data output port mode is set by the logic levels on the
SMODE_OUT_0/SMODE_OUT_1 and WLNGTH_OUT_0/
WLNGTH_OUT_1 pins. The serial mode can be changed to
left justified, I2S, right justified, or TDM as defined in the following table. The output word width can be set by using the
WLNGTH_OUT_0/WLNGTH_OUT_1 pins as shown in
Table III. When the output word width is less than 24 bits, dither
is added to the truncated bits. The right justified serial data out
mode assumes 64 SCLK_O cycles per frame, divided evenly for
left and right. Please note that 8 bits of each 32-bit subframe are
used for transmitting matched-phase mode data. Please refer to
Figure 14. The AD1896 also supports 16-bit, 32-clock packed
input and output serial data in LJ and I2S format.
Table II. Serial Data Output Port Mode
SMODE_OUT_[0:1]
1
0
0
0
1
1
0
1
0
1
Interface Format
Left Justified (LJ)
I 2S
TDM Mode
Right Justified (RJ)
Table III. Word Width
WLNGTH_OUT_[0:1]
1
0
0
0
1
1
0
1
0
1
Word Width
24 Bits
20 Bits
18 Bits
16 Bits
The following timing diagrams show the serial mode formats.
–22–
REV. A
AD1896
RIGHT CHANNEL
LEFT CHANNEL
LRCLK
SCLK
MSB
LSB
MSB
MSB
SDATA
LSB
LEFT-JUSTIFIED MODE – 16 BITS TO 24 BITS PER CHANNEL
LRCLK
LEFT CHANNEL
RIGHT CHANNEL
SCLK
MSB
SDATA
LSB
MSB
MSB
LSB
I2S MODE – 16 BITS TO 24 BITS PER CHANNEL
RIGHT CHANNEL
LEFT CHANNEL
LRCLK
SCLK
MSB
SDATA
LSB
MSB
LSB
RIGHT-JUSTIFIED MODE – SELECT NUMBER OF BITS PER CHANNEL
LRCLK
SCLK
MSB
SDATA
LSB
MSB
LSB
TDM MODE – 16 BITS TO 24 BITS PER CHANNEL
1/fs
NOTES
1 LRCLK NORMALLY OPERATES AT ASSOCIATIVE INPUT OR OUTPUT SAMPLE FREQUENCY (f ).
s
2 SCLK FREQUENCY IS NORMALLY 64 LRCLK EXCEPT FOR TDM MODE WHICH IS N 64 f ,
s
WHERE N = NUMBER OF STEREO CHANNELS IN THE TDM CHAIN, IN MASTER MODE N = 4.
3 PLEASE NOTE THAT 8 BITS OF EACH 32-BIT SUBFRAME ARE USED FOR TRANSMITTING
MATCHED-PHASE MODE DATA. PLEASE REFER TO FIGURE 14.
Figure 10. Input/Output Serial Data Formats
TDM MODE APPLICATION
In TDM mode, several AD1896s can be daisy-chained together
and connected to the serial input port of a SHARC DSP. The
AD1896 contains a 64-bit parallel load shift register. When the
LRCLK_O pulse arrives, each AD1896 parallel loads its left and
right data into the 64-bit shift register. The input to the shift
register is connected to TDM_IN, while the output is connected
to SDATA_O. By connecting the SDATA_O to the TDM_IN
of the next AD1896, a large shift register is created, which is
clocked by SCLK_O.
The number of AD1896s that can be daisy-chained together is
limited by the maximum frequency of SCLK_O, which is about
25 MHz. For example, if the output sample rate, fS, is 48 kHz,
up to eight AD1896s could be connected since 512 ¥ fS is less
than 25 MHz. In master/TDM mode, the number of AD1896s
that can be daisy-chained is fixed to four.
LRCLK
SCLK
AD1896
AD1896
SDATA_O
DR0
LRCLK_O
LRCLK_O
LRCLK_O
RFS0
SCLK_O
SCLK_O
SCLK_O
SDATA_O
TDM_IN
SHARC
DSP
AD1896
SDATA_O
TDM_IN
PHASE-MASTER
TDM_IN
SLAVE-1
RCLK0
SLAVE-n
M2
M1
M0
M2
M1
M0
M2
M1
M0
0
0
0
0
0
0
0
0
0
STANDARD MODE
1
0
0
1
0
0
MATCHED-PHASE MODE
Figure 11. Daisy-Chain Configuration for TDM Mode (All AD1896s Being Clock-Slaves)
REV. A
–23–
AD1896
AD1896
AD1896
SHARC
DSP
AD1896
SDATA_O
DR0
LRCLK_O
LRCLK_O
LRCLK_O
RFS0
SCLK_O
SCLK_O
SCLK_O
SDATA_O
TDM_IN
SDATA_O
TDM_IN
CLOCK-MASTER
AND
PHASE-MASTER
TDM_IN
RCLK0
SLAVE-n
SLAVE-1
M2
M1
M0
M2
M1
M0
M2
M1
M0
0
1
1
0
0
0
0
0
0
STANDARD MODE
1
0
0
1
0
0
MATCHED-PHASE MODE
Figure 12. Daisy-Chain Configuration for TDM Mode (First AD1896 Being Clock-Master)
MATCHED-PHASE MODE (NON-TDM MODE) APPLICATION
LRCLKI (fS_IN)
SCLKI
SDOm
SDO1
SDO2
AD1896
AD1896
AD1896
AD1896
PHASE-MASTER
TDM_IN
SLAVE1
TDM_IN
SLAVE2
TDM_IN
SLAVEn
TDM_IN
SDATA_I SDATA_O
LRCLK_I LRCLK_O
SCLK_O
SCLK_I
SDATA_I SDATA_O
LRCLK_I LRCLK_O
SCLK_O
SCLK_I
SDATA_I SDATA_O
LRCLK_I LRCLK_O
SCLK_O
SCLK_I
SDATA_I SDATA_O
LRCLK_I LRCLK_O
SCLK_O
SCLK_I
MCLK
MCLK
MCLK
MCLK
RESET
RESET
RESET
RESET
M2 M1 M0
0
0
M2 M1 M0
1
0
0
M2 M1 M0
1
0
0
0
SDOn
M2 M1 M0
1
0
0
LRCLKO (fS_OUT)
SCLKO (64fS_OUT)
MCLK
RESET
Figure 13. Typical Configuration for Matched-Phase Mode Operation
Serial Data Port Master Clock Modes
Table IV. Serial Data Port Clock Modes
Either of the AD1896 serial ports can be configured as a master
serial data port. However, only one serial port can be a master
while the other has to be a slave. In master mode, the AD1896
requires a 256 ¥ fS, 512 ¥ fS, or 768 ¥ fS master clock (MCLK_I).
For a maximum master clock frequency of 30 MHz, the maximum sample rate is limited to 96 kHz. In slave mode, sample
rates up to 192 kHz can be handled.
When either of the serial ports is operated in master mode, the
master clock is divided down to derive the associated left/
right subframe clock (LRCLK) and serial bit clock (SCLK).
The master clock frequency can be selected for 256, 512, or 768
times the input or output sample rate. Both the input and output serial ports will support master mode LRCLK and SCLK
generation for all serial modes, left justified, I2S, right justified, and
TDM for the output serial port.
MMODE_0/
MMODE_1/
MMODE_2
2
1
0
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
–24–
Interface Format
Both serial ports are in slave mode.
Output serial port is master with 768 ¥ fS_OUT.
Output serial port is master with 512 ¥ fS_OUT.
Output serial port is master with 256 ¥ fS_OUT.
Matched-phase Mode
Input serial port is master with 768 ¥ fS_IN.
Input serial port is master with 512 ¥ fS_IN.
Input serial port is master with 256 ¥ fS_IN.
REV. A
AD1896
configured as the master, while the rest of the AD1896s in the
chain would be configured as slaves with their MMODE_2,
MMODE_1, and MMODE_0 pins set to 100, respectively.
Matched-Phase Mode
The matched-phase mode is the mode discussed in the Theory of
Operation section that eliminates the phase mismatch between
multiple AD1896s. The master AD1896 device transmits
its fS_OUT/fS_IN ratio through the SDATA_O pin to the slave
AD1896’s TDM_IN pins. The slave AD1896s receive the
transmitted fS_OUT/fS_IN ratio and use the transmitted fS_OUT/
fS_IN ratio instead of their own internally derived fS_OUT/fS_IN
ratio. The master device can have both its serial ports in slave
mode as depicted or either one in master mode. The slave
AD1896s must have their MMODE_2, MMODE_1, and
MMODE_0 pins set to 100, respectively. LRCLK_I and
LRCLK_O may be asynchronous with respect to each other in
this mode. Another requirement of the matched-phase mode is
that there must be 32 SCLK_O cycles per subframe. The
AD1896 will support the matched-phase mode for all serial
output data formats, left justified, I2S, right justified, and
TDM. In the case of TDM, the AD1896 shown in the TDM
mode operation figure with its TDM_IN tied to ground would be
AUDIO DATA LEFT CHANNEL, 24 BITS
Please note that in the left-justified, I2S, and TDM modes,
the lower eight bits of each channel subframe are used to
transmit the matched-phase data. In right-justified mode, the
upper eight bits are used to transmit the matched-phase data.
This is shown in Figures 14a and 14b.
Bypass Mode
When the BYPASS pin is asserted high, the input data bypasses
the sample rate converter and is sent directly to the serial output
port. Dithering of the output data when the word length is set
to less than 24 bits is disabled. This mode is ideal when the
input and output sample rates are the same and LRCLK_I and
LRCLK_O are synchronous with respect to each other. This
mode can also be used for passing through non-AUDIO data
since no processing is performed on the input data in this mode.
MATCHED-PHASE
DATA, 8 BITS
AUDIO DATA RIGHT CHANNEL, 24 BITS
MATCHED-PHASE
DATA, 8 BITS
Figure 14a. Matched-Phase Data Transmission (Left-Justified, I 2S, and TDM Mode)
MATCHED-PHASE
DATA, 8 BITS
AUDIO DATA LEFT CHANNEL,
16 BITS – 24 BITS
MATCHED-PHASE
DATA, 8 BITS
AUDIO DATA RIGHT CHANNEL,
16 BITS – 24 BITS
Figure 14b. Matched-Phase Data Transmission (Right-Justified Mode)
REV. A
–25–
AD1896
OUTLINE DIMENSIONS
28-Lead Shrink Small Outline Package [SSOP]
(RS-28)
Dimensions shown in millimeters
10.50
10.20
9.90
28
15
5.60
5.30
5.00
8.20
7.80
7.40
14
1
1.85
1.75
1.65
2.00 MAX
0.10
COPLANARITY
0.25
0.09
0.05
MIN
0.65
BSC
0.38
0.22
SEATING
PLANE
8ⴗ
4ⴗ
0ⴗ
0.95
0.75
0.55
COMPLIANT TO JEDEC STANDARDS MO-150AH
–26–
REV. A
AD1896
Revision History
Location
Page
3/03—Data Sheet changed from REV. 0 to REV. A.
Edits to DIGITAL PERFORMANCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Edits to DIGITAL TIMING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Edits to ORDERING GUIDE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Edits to RESETand Power-Down section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Edits to Figures 9a and 9b . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Edits to Serial Data Ports—Data Format section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Edits to Figure 13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Update to OUTLINE DIMENSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
REV. A
–27–
–28–
PRINTED IN U.S.A.
C02403–0–3/03(A)