SYSTEM DIAGRAM

AD1893

EXAMPLE

FREQUENCIES:

DAT 48kHz OR

CD 44.1kHz OR

BROADCAST 32kHz

EXAMPLE

FREQUENCIES:

DAT 48kHz OR

CD 44.1kHz OR

BROADCAST 32kHz

INPUT SERIAL DATA

OUTPUT SERIAL DATA

INPUT SAMPLE CLOCK

OUTPUT SAMPLE CLOCK

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
which may result from its use. No license is granted by implication or
otherwise under any patent or patent rights of Analog Devices.

FEATURES
Low Cost
LQFP and PDIP Packages
3 V Supply Performance Specified--Very Low Power
Automatically Senses Sample Frequencies--No

Programming Required

Rejects Sample Clock Jitter
Accommodates Dynamically Changing Asynchronous

Sample Clocks

8 kHz to 56 kHz Sample Clock Frequency Range
Approximately 1:2 to 2:1 Ratio Between Sample

Clocks

96 dB THD+N at 1 kHz
96 dB Dynamic Range
Optimal Clock Tracking Control--Slow/Fast Settling

Modes

Linear Phase in All Modes
Automatic Output Mute
Flexible Four-Wire Serial Interfaces with Right-Justified

Mode

Power-Down Mode
On-Chip Oscillator

APPLICATIONS
Consumer CD-R, DAT, DCC, MD and 8 mm Video Tape

Recorders Including Portables

Digital Audio Communication/Network Systems
Computer Multimedia Systems

PRODUCT OVERVIEW

The AD1893 SamplePort is a fully digital, stereo Asynchronous
Sample Rate Converter (ASRC) that solves sample rate interfacing
and compatibility problems in digital audio equipment. Concep-
tually, this converter interpolates the input data up to a very high
internal sample rate with a time resolution of 300 ps, then deci-
mates down to the desired output sample rate. The AD1893 is
intended for 16-bit low cost, non-varispeed applications where low
voltage, low power (i.e., battery-powered) operation is required.
Refer to the AD1890/AD1891 data sheet for other products in the
SamplePort family. This device is asynchronous because the fre-
quency and phase relationships between the input and output
sample clocks (both are inputs to the AD1893 ASRC) are arbitrary
and need not be related by a simple integer ratio. There is no need
to explicitly select or program the input and output sample clock
frequencies, as the AD1893 automatically senses the relationship
between the two clocks. The input and output sample clock fre-
quencies can nominally range from 8 kHz to 56 kHz, and the ratio
between them can vary from approximately 1:2 to 2:1.

SamplePort is a registered trademark of Analog Devices, Inc.

AD1893

Low Cost SamplePort

16-Bit Stereo Asynchronous

Sample Rate Converter

The AD1893 uses multirate digital signal processing techniques
to construct an output sample stream from the input sample
stream. The input word width is 4 to 16 bits for the AD1893.
Shorter input words are automatically zero-filled in the LSBs.
The output word width is 24 bits. The user can receive as many
of the output bits as desired. Internal arithmetic is performed
with 22-bit coefficients and 27-bit accumulation. The digital
samples are processed with unity gain.

The input and output control signals allow for considerable
flexibility for interfacing to a variety of DSP chips, AES/EBU
receivers and transmitters and for I

S compatible devices. Input

and output data can be independently right- or left- (with or
without a one bit clock delay) justified to the left/right clock
edge. In the right-justified mode, the MSB is delayed 16-bit
clock periods from the left/right clock edge transition. Input and
output data can also be independently justified to the word
clock rising edge. The data justification options are encoded on
two mode pins for both the input port and the output port. The
bit clocks can also be independently configured for rising edge
active or falling edge active operation.

The AD1893 SamplePort ASRC has on-chip digital coefficients
that correspond to a highly oversampled 0 Hz to 20 kHz low-
pass filter with a flat passband, a very narrow transition band,
and a high degree of stopband attenuation. A subset of these
filter coefficients are dynamically chosen on the basis of the
filtered ratio between the input sample clock (LR_I) and the
output sample clock (LR_O), and these coefficients are then
used in an FIR convolver to perform the sample rate conversion.
Refer to the Theory of Operation section of this data sheet for a
more thorough functional description. The low-pass filter has
been designed so that full 20 kHz bandwidth is maintained
when the input and output sample clock frequencies are as low
as 44.1 kHz. If the output sample rate drops below the input
sample rate, the bandwidth of the input signal is automatically

(continued on Page 4)

One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A.
Tel: 781/329-4700

World Wide Web Site: http://www.analog.com

Fax: 781/326-8703

© Analog Devices, Inc., 1998

AD1893SPECIFICATIONS

REV. A

TEST CONDITIONS UNLESS OTHERWISE NOTED

Supply Voltage

+3.0

Ambient Temperature

Crystal Frequency

MHz

Load Capacitance

100

All minimums and maximums tested except as noted.

PERFORMANCE

(Guaranteed for V

= +3.3 V to +5.0 V

10%)

Min

Max

Units

Dynamic Range (20 Hz to 20 kHz, 60 dB Input)

Total Harmonic Distortion + Noise

(20 Hz to 20 kHz, Full-Scale Input, F

SOUT

SIN

Between 0.51 and 1.99)

(1 kHz Full-Scale Input, F

SOUT

SIN

Between 0.7 and 1.4)

(10 kHz Full-Scale Input, F

SOUT

SIN

Between 0.7 and 1.4)

Interchannel Phase Deviation

Degrees

Input and Output Sample Clock Jitter

(For

1 dB Degradation in THD+N with 10 kHz Full-Scale Input, Slow-Settling Mode)

DIGITAL INPUTS

(Guaranteed for V

= +3.0 V to +5.0 V

10%)

Min

Max

Units

2.0

+3.0 V)

0.8

(+2.7 V

< +3.0 V)

0.7

@ V

= +5.0 V, All Pins Except XTAL_I

@ V

= +5.0 V, XTAL_I Pin

@ V

= 0 V, All Pins Except XTAL_I

@ V

= 0 V, XTAL_I Pin

@ I

= 4 mA (V

+3.0 V)

2.4

@ I

= 4 mA (+2.7 V

< +3.0 V)

2.2

@ I

= 4 mA

0.4

Input Capacitance

DIGITAL TIMING

(Guaranteed for V

= +3.0 V to +5.0 V

10%) See Figures 26 through 28.

Min

Max

Units

CRYSTAL

Crystal Period

62.5

125

CRYSTAL

Crystal Frequency (1/t

CRYSTAL

)

MHz

PWL

Crystal LO Pulsewidth

PWH

Crystal HI Pulsewidth

LRI

LR_I Frequency with 16 MHz Crystal

kHz

RPWL

RESET LO Pulsewidth

125

RESET Setup to Crystal Falling

BCLK

BCLK_I/O Period

120

BCLK

BCLK_I/O Frequency (l/t

BCLK

)

8.33

MHz

BPWL

BCLK_I/O LO Pulsewidth

BPWH

BCLK_I/O HI Pulsewidth

WSI

WCLK_I Setup to BCLK_I

WSO

WCLK_O Setup to BCLK_O

LRSI

LR_I Setup to BCLK_I

LRSO

LR_O Setup to BCLK_O

Data Setup to BCLK_I

Data Hold from BCLK_I

DPD

Data Propagation Delay from BCLK_O

DOH

Data Output Hold from BCLK_O

AD1893

DIGITAL FILTER CHARACTERISTICS

Min

Max

Units

Passband Ripple (0 kHz to 20 kHz)

0.01

Transition Band

4.1

kHz

Stopband Attenuation

110

Group Delay (LR_I = 50 kHz)

700

3000

POWER

SIN

= 48 kHz, F

SOUT

= 44.1 kHz)

Min

Typ

Max

Units

Supplies

Voltage, V

2.7

5.5

Operational Current, I

= +5.0 V)

Operational Current, I

= +3.0 V)

Power-Down Current, I

= +5.0 V)

1.5

2.5

Power-Down Current, I

= +3.0 V)

0.5

1.0

Dissipation

Operation (V

= +5.0 V)

150

200

Operation (V

= +3.0 V)

Power-Down (V

= +5.0 V)

7.5

12.5

Power-Down (V

= +3.0 V)

1.5

3.0

TEMPERATURE RANGE

Min

Max

Units

Specifications Guaranteed

+70

Operational Guaranteed

+85

Storage

+100

ABSOLUTE MAXIMUM RATINGS

Min

Max

Units

to GND

0.3

7.0

DC Input Voltage

0.3

+ 0.3

Latch-Up Trigger Current

1000

+1000

Soldering

+300

sec

NOTES

Guaranteed, Not Tested

Valid only when F

SOUT

SIN

(i.e., upsampling), F

SIN

= 44.1 kHz.

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.

Specifications subject to change without notice.

ORDERING GUIDE

Model

Temperature Range

Package Descriptions

Package Options

AD1893JN

C to +70

Plastic DIP

N-28

AD1893JST

C to +70

LQFP

ST-44

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 AD1893 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.

REV. A

WARNING!

ESD SENSITIVE DEVICE

AD1893

REV. A

(continued from Page 1)

PRODUCT OVERVIEW (Continued)
limited to avoid alias distortion on the output signal. The
AD1893 dynamically alters the low-pass filter cutoff frequency
smoothly and slowly, so that real-time variations in the sample
rate ratio are possible without degradation of the audio quality.

The AD1893 has a pin selectable slow- or fast-settling mode.
This mode determines how quickly the ASRC adapts to a
change in either the input sample clock frequency (F

SIN

) or the

output sample clock frequency (F

SOUT

). In the slow-settling

mode, the control loop which computes the ratio between F

SIN

and F

SOUT

settles in approximately 800 ms and begins to reject

jitter above 3 Hz. The slow-settling mode offers the best signal
quality and the greatest jitter rejection. In the fast-settling mode,
the control loop settles in approximately 200 ms and begins to
reject jitter above 12 Hz. The fast-settling mode allows rapid,
real time sample rate changes to be tracked without error, at the
expense of some narrowband noise modulation products on the
output signal.

The AD1893 features short group delay processing. This feature
relates to the depth of the First-In, First-Out (FIFO) memory,
which buffers the input data samples before they are processed
by the FIR convolver. In the AD1893, the group delay is
approximately 700

s. If the read and write pointers that

manage the FIFO cross (indicating underflow or overflow), the
AD1893 asserts the mute output (MUTE_O) pin HI for 128
output clock cycles. If MUTE_O is connected to the mute input
(MUTE_I) pin, as it normally should be, the serial output will
be muted (i.e., all bits zero) during this transient event.

The AD1893 includes an on-chip oscillator that only requires
the user provide an external crystal. By removing the need for
an external oscillator, the AD1893 lowers the total cost of own-
ership to the end user. The AD1893 also includes a power-
down mode, which is invoked with the PWRDWN pin.
Asserting this control signal HI will place the AD1893 into a
very low power dissipation in active and standby condition.

The AD1893 is fabricated in a 0.8

m single poly, double metal

CMOS process and are packaged in a 0.6" wide 28-lead plastic
DIP and a 10 mm by 10 mm body size 44-lead LQFP. The
AD1893 operates from a +3 V to +5 V power supply over the
temperature range of 0

C to +70

DEFINITIONS
Dynamic Range

The ratio of a near full-scale input signal to the integrated noise
in the passband (0 kHz to

20 kHz), expressed in decibels (dB).

Dynamic range is measured with a 60 dB input signal and
"60 dB" arithmetically added to the result.

Total Harmonic Distortion + Noise
Total Harmonic Distortion plus Noise (THD+N) is defined as
the ratio of the square root of the sum of the squares of the values
of the harmonics and noise to the rms value of a sinusoidal
input signal. It is usually expressed in percent (%) or decibels.

Interchannel Phase Deviation
Difference in input sampling times between stereo channels,
expressed as a phase difference in degrees between 1 kHz inputs.

Group Delay
Intuitively, the time interval required for a full-level input pulse
to appear at the converter's output, at full level, expressed in
milliseconds (ms). More precisely, the derivative of radian
phase with respect to radian frequency at a given frequency.

Transport Delay
The time interval between when an impulse is applied to the
converter's input and when the output starts to be affected by
this impulse, expressed in milliseconds (ms). Transport delay is
independent of frequency.

AD1893

REV. A

AD1893 PIN LIST

Serial Input Interface

Pin Name DIP

LQFP I/O

Description

DATA_I

Serial input, MSB first, containing two channels of 4 to 16 bits of twos-complement data per
channel.

BCLK_I

Bit clock input for input data. Need not run continuously; may be gated or used in a burst fashion.

WCLK_I

Word clock input for input data. This input is rising edge sensitive. (Not required in LR input
data clock triggered modes.)

LR_I

Left/right clock input for input data. Must run continuously.

Serial Output Interface

Pin Name DIP

LQFP I/O

Description

DATA_O

Serial output, MSB first, containing two channels of 4- to 24-bits of twos-complement data per
channel.

BCLK_O

Bit clock input for output data. Need not run continuously; may be gated or used in a burst
fashion.

WCLK_O

Word clock input for output data. This input is rising edge sensitive. (Not required in LR output
data clock triggered modes.)

LR_O

Left/right clock input for output data. Must run continuously.

Input Control Signals

Pin Name DIP

LQFP

I/O Description

BKPOL_I

Bit clock polarity. LO: Normal mode. Input data is sampled on rising edges of BCLK_I. HI:
Inverted mode. Input data is sampled on falling edges of BCLK_I.

MODE0_I 11

Serial mode zero control for input port.

MODE1_I 12

Serial mode one control for input port.

MODE0_I

MODE1_I

Left-justified, no MSB delay, LR_I clock triggered.

Left-justified, MSB delay, LR_I clock triggered.

Right-justified, MSB delayed 16 bit clock periods from LR_I transition.

WCLK_I triggered, no MSB delay.

PIN CONFIGURATIONS

DIP

SERIAL IN

SERIAL OUT

COEF ROM

MULT

FIFO

CLOCK

TRACKING

ACCUM

XTAL_I

DATA_I

BCLK_I

WCLK_I

GND

BKPOL_I

MODE0_I

MODE1_I

GND

RESET

SETSLW

PWRDWN

BCLK_O

WCLK_O

DATA_O

GND

BKPOL_O

MODE0_O

MODE1_O

MUTE_O

MUTE_I

AD1893

NC = NO CONNECT

XTAL_O

LQFP

NC = NO CONNECT

WCLK_O

GND

DATA_O

BCLK_I

GND

WCLK_I

DATA_I

XTAL_O

SETSLW

XTAL_I

MODE1_I

GND

MUTE_I

PWRDWN

BCLK_O

MUTE_O

MODE1_O

BKPOL_O

MODE0_O

BKPOL_I

MODE0_I

39 38

AD1893

21 22

12 13

15 16 17

35 34

RESET

SERIAL IN

MULT

FIFO

ACCUM

CLOCK

TRACKING

SERIAL OUT

COEF ROM

AD1893

REV. A

Output Control Signals

Pin Name DIP

LQFP

I/O

Description

BKPOL_O 19

Bit clock polarity. LO: Normal mode. Output data is valid on rising edges of BCLK_O, changed
on falling. HI: Inverted mode. Output data is valid on falling edges of BCLK_O, changed on
rising.

MODE0_O 18

Serial mode zero control for output port.

MODE1_O 17

Serial mode one control for output port.

MODE0_O

MODE1_O

Left-justified, no MSB delay, LR_O clock triggered.

Left-justified, MSB delay, LR_O clock triggered.

Right-justified, MSB delayed 16 bit clock periods from LR_O transition.

WCLK_O triggered, no MSB delay.

Miscellaneous

Pin Name DIP

LQFP

I/O

Description

XTAL_O

Crystal output. Connect to one side of nominal 16 MHz crystal for sampling frequencies
(F

word rates) from 8 kHz to 56 kHz.

XTAL_I

Crystal input. Connect to other side of nominal 16 MHz crystal for sampling frequencies

word rates) from 8 kHz to 56 kHz. Use this input to overdrive the on-chip oscillator

with an external clock source.

RESET

Active LO reset. Set HI for normal chip operation.

MUTE_O

Mute output. HI indicates that data is not currently valid due to read and write FIFO

memory pointer overlap. LO indicates normal operation.

MUTE_I

Mute input. HI mutes the serial output to zeros (midscale). Normally connected to

MUTE_O. Reset LO for normal operation.

SETLSLW 28

Settle slowly to changes in sample rates. HI: Slow-settling mode (

800 ms). Less sensitive

to sample clock jitter. LO: Fast-settling mode (

200 ms). Some narrow-band noise

modulation may result from jitter on the LR clocks. This signal may be asynchronous with
respect to the crystal frequency, and dynamically changed, but is normally pulled up or
pulled down on a static basis.

PWRDWN 27

Power-down input. Set HI for inactive, low power dissipation state. Reset LO for normal

operation.

9, 20

1, 5, 8, 11,

No connect. Reserved. Do not connect.

12, 15, 17,
19, 22, 23,
26, 29, 33,
34, 37, 39,
41, 44

Power Supply Connections

Pin Name

DIP

LQFP

I/O

Description

7, 22

6, 28

Positive digital voltage supply.

GND

8, 14, 21

7, 16, 27

Digital ground. Pin 14 (DIP) and Pin 16 (LQFP) need not be decoupled.

AD1893

REV. A

THEORY OF OPERATION

There are at least two logically equivalent methods of explaining
the concept of asynchronous sample rate conversion: the high
speed interpolation/decimation model and the polyphase filter
bank model. Using the AD1893 SamplePort does not require
understanding either model. This section is included for those
who wish a deeper understanding of its operation.

Interpolation/Decimation Model

In the high speed interpolation/decimation model, illustrated in
Figure 1, the sampled data input signal (Plot A in Figure 1) is
interpolated at some ratio (IRATIO) by inserting IRATIO-1
zero valued samples between each of the original input signal
samples (Plot B in Figure 1). The frequency domain charac-
teristics of the input signal are unaltered by this operation, ex-
cept that the zero-padded sequence is considered to be sampled
at a frequency which is the product of original sampling fre-
quency multiplied by IRATIO.

The zero-padded values are fed into a digital FIR low-pass filter
(Plot C in Figure 1) to smooth or integrate the sequence, and
limit the bandwidth of the filter output to 20 kHz. The inter-
polated output signal has been quantized to a much finer time
scale than the original sequence. The interpolated sequence is

INPUT

SIGNAL

OUTPUT
SIGNAL

ZERO STUFF

INTERPOLATION

FIR LOW-

PASS

FILTER

ZERO ORDER

HOLD

REGISTER

RESAMPLING

DECIMATION

TIME

AMP

Figure 1. Interpolation/Decimation Model--Time Domain View

then passed to a zero-order hold functional block (physically
implemented as a register, Plot D in Figure 1) and then asyn-
chronously resampled at the output sample frequency (Plot E in
Figure 1). This resampling can be thought of as a decimation
operation since only a very few samples out of the great many
interpolated samples are retained. The output values represent
the "nearest" values, in a temporal sense, produced by the inter-
polation operation. There is always some error in the output
sample amplitude due to the fact that the output sampling switch
does not close at a time that exactly corresponds to a point on the
fine time scale of the interpolated sequence. However, this error
can be made arbitrarily small by using a very large interpolation
ratio. The AD1893 SamplePort ASRC uses an equivalent IRATIO
of 65,536 to provide 16-bit accuracy (

96 dB THD+N) across

the 0 kHz to 20 kHz audio band.

The number of FIR filter taps and associated coefficients is
approximately 4 million. The equivalent FIR filter convolution
frequency (or "upsample" frequency) is 3.2768 GHz, and the
fine time scale has resolution of about 300 ps. Various propri-
etary efficiencies are exploited in the AD1893 ASRC to reduce
the complexity and throughput requirements of the hardware
implied by this interpolation/decimation model.

AD1893

REV. A

Polyphase Filter Bank Model

Although less intuitively understandable than the interpolation/
decimation model, the polyphase filter bank model is useful to
explore because it more accurately portrays the operation of the
actual AD1893 SamplePort hardware. In the polyphase filter
bank model, the stored FIR filter coefficients are thought of as
the impulse response of a highly oversampled 0 kHz to 20 kHz
low-pass prototype filter, as shown in Figure 2. If this low-pass
filter is oversampled by a factor of N, then it can be conceptu-
ally decomposed into N different "subfilters," each filter consist-
ing of a different subset of the original set of impulse response
samples. If the temporal position of each of the subfilters is
maintained, then they can be summed to recreate the original
oversampled impulse response. Since the original impulse re-
sponse is highly oversampled, the more sparsely sampled
subfilters still individually meet the Nyquist criterion (i.e., they

AMP

TIME

OVERSAMPLED
LOW-PASS FILTER
IMPULSE RESPONSE

DECOMPOSED INTO
FOUR SUBFILTERS

PHASE

0 Deg

180

270

1/4F

1/2F

3/4F

FREQ

AMP

1/4F

1/2F

3/4F

1/4F

1/2F

3/4F

1/4F

1/2F

3/4F

1/4F

1/2F

3/4F

Figure 2. Four Polyphase Subfilters in the Time and Frequency Domains

are adequately sampled). The baseband magnitude and phase
responses of the subfilters are identical. The out-of-band (i.e.,
alias) regions of the subfilters however have phase responses
which are shifted relative to one another, in a manner that
causes them to cancel when they are summed.

The subfilter coefficients are then aligned to the left, as shown
in Figure 3, so that the first coefficient of each subfilter is
aligned to the first point on a coarse time scale. (This conceptual
step accounts for how the hardware implementation is able to
operate at the slower rate corresponding to the coarse time
scale.) Each subfilter has been shifted in time by a different
amount, and though they still share identical magnitude re-
sponses, they now have in-band phase responses which have
fractionally different slopes (i.e., group delays).

AD1893

REV. A

AMP

TIME

SUBFILTER COEFFICIENTS
ALIGNED TO THE LEFT

SIN

= 1/F

SIN

FREQ

SIN

PHASE

DELAY = NOMINAL

DELAY = NOMINAL .25/F

SIN

DELAY = NOMINAL .5/F

SIN

DELAY = NOMINAL .75/F

SIN

Figure 3. Four Polyphase Subfilters Realigned to Coarse Time Grid

POLYPHASE FILTER 1

POLYPHASE FILTER 2

POLYPHASE FILTER 3

POLYPHASE FILTER 4

POLYPHASE FILTER 5

POLYPHASE FILTER 6

POLYPHASE FILTER 7

INPUT

SIGNAL

POLYPHASE FILTER N-1

POLYPHASE FILTER N

SELECT

N TO 1

MUX

OUTPUT

SIGNAL

SAMPLE

CLOCK

TRACKING

CIRCUIT

Figure 4. Polyphase Filter Bank Model--Conceptual Block
Diagram

The full set of subfilters can be considered to form a parallel
bank of "polyphase" filters which have decrementing, linear
phase group delays. All of the polyphase filters conceptually
process the input signal simultaneously, as illustrated in Figure
4, at the input sample rate.

Asynchronous sample rate conversion under the polyphase filter
bank model is accomplished by selecting the output of a particu-
lar polyphase filter on the basis of the temporal relationship
between the input sample clock and the output sample clock
events. Figure 5 shows the desired filter group delay as a func-
tion of the relative time difference between the current output
sample clock and the last input sample clock. If an output
sample is requested late in the input sample period, then a short
filter delay is required, and if an output sample is requested
early in the input sample period, then a long filter delay is re-
quired. This nonintuitive result arises from the fact that FIR
filters always produce some delay, so that selecting a filter with
shorter delay moves the interpolated sample closer to the newest
input sample.

AD1893

REV. A

OUTPUT SEQUENCE

PAST

FUTURE

AMPLITUDE

SHORT

DELAY

LONG
DELAY

SMALL
OFFSET

LARGE
OFFSET

OFFSET INTO DENSE FIR FILTER COEFFICIENT ARRAY

TO ACCESS REQUIRED POLYPHASE FILTER

REQUIRED FILTER GROUP DELAY TO

COMPUTE REQUESTED OUTPUT SAMPLE

INPUT SEQUENCE

Figure 5. Input and Output Clock Event Relationship

A short delay corresponds to a large offset into the dense FIR
filter coefficient array, and a long delay corresponds to a small
offset. Note that because the output sample clock can arrive at
any arbitrary time with respect to the input sample clock, the
selection of a polyphase filter with which to convolve the input
sequence occurs on every output sample clock event. Occasion-
ally the FIFO which holds the input sequence in the FIR con-
volver is either not incremented, or incremented by two between
output sample clocks (see periods A and B in Figure 5); this
happens more often when the input and output sample clock
frequencies are dissimilar than when they are close together.
However, in this situation, an appropriate polyphase filter is
selected to process the input signal, and thus an accurate output
sample is computed. Input and output samples are not skipped
or repeated (unless the input FIFO underflows or overflows), as
is the case in some other sample rate converter implementations.

To obtain an accurate conversion, a large number of polyphase
filters are needed. The AD1893 SamplePort uses the equivalent
of 65,536 polyphase filters to achieve its high quality distortion
and dynamic range specifications.

Sample Clock Tracking
It should be clear that, in either model, the correct computation
of the ratio between the input sample rate (as determined from
the left/right input clock, LR_I) and the output sample rate (as
determined from the left/right output clock, LR_O) is critical to
the quality of the output data stream. It is straightforward to
compute this ratio if the sample rates are fixed and synchronous;
the challenge is to accurately track dynamically varying and
asynchronous sample rates, as well as to account for jitter.

The AD1893 SamplePort solves this problem by embedding the
ratio computation circuit within a digital servo control loop, as
shown in Figure 6. This control loop includes special provisions
to allow for the accurate tracking of dynamically changing
sample rates. The outputs of the control loop are the starting
read addresses for the input data FIFO and the filter coefficient
ROM. These start addresses are used by the FIFO and ROM
address generators, as shown in Figure 6.

The input data FIFO write address is generated by a counter
which is clocked by the input sample clock (i.e., LR_I). It is very
important that the FIFO read address and the FIFO write ad-
dress do not cross, as this means that the FIFO has either
underflowed or overflowed. This consideration affects the
choice of settling time of the control loop. When a step change
in the sample rate occurs, the relative positions of the read and
write addresses will change while the loop is settling. A fast
settling loop will act to keep the FIFO read and write addresses
separated better than a slow settling loop. The AD1893 includes
a user selectable pin (SETLSLW) to set the loop settling time
that essentially changes the coefficients of the digital servo con-
trol loop filter. The state of the SETLSLW pin can be changed
on-the-fly but is normally set and forgotten.

FIFO WRITE

ADDRESS

GENERATOR

FIFO

SERIAL DATA

INPUT UNIT

FIFO READ

ADDRESS

GENERATOR

ACCUMULATOR

SERIAL DATA
OUTPUT UNIT

SAMPLE CLOCK RATIO

SERVO CONTROL LOOP

POLYPHASE

COEFFICIENT

ROM

ROM ADDRESS

GENERATOR

POLYPHASE FILTER

SELECTOR

DATA_I

WCLK_O

BCLK_O

WCLK_I

BCLK_I

FIR CONVOLVER

DATA_O

START
ADDRESS

FREQUENCY

RESPONSE

COMPRESSION

SOUT

< F

SIN

)

SOUT

)

Figure 6. Functional Block Diagram

AD1893

REV. A

Sample Clock Jitter Rejection
The loop filter settling time also affects the ability of the
AD1893 ASRC to reject sample clock jitter, since the control
loop effectively computes a time weighted average or "esti-
mated" new output of many past input and output clock events.
This first order low pass filtering of the sample clock ratio pro-
vides the AD1893 with its jitter rejection characteristic. In the
slow settling mode, the AD1893 attenuates jitter frequencies
higher than 3 Hz (

800 ms for the control loop to settle to an

18-bit "pure" sine wave), and thus rejects all but the most se-
vere sample clock jitter; performance is essentially limited only
by the FIR filter. In the fast settling mode, the ASRC attenuates
jitter components above 12 Hz (

200 ms for the control loop to

settle). Due to the effects of on-chip synchronization of the
sample clocks to the 16 MHz (62.5 ns) crystal master clock,
sample clock jitter must be a large percentage of the crystal
period (>10 ns) before performance degrades in either the slow
or fast settling modes. Note that since both past input and past
output clocks are used to compute the filtered "current" internal
output clock request, jitter on both the input sample clock and
the output sample clock is rejected equally. In summary: the fast-
settling mode is best for applications when the sample rates will
be dynamically altered (e.g., varispeed situations), while the
slow-settling mode provides the most sample clock jitter rejection.

Clock jitter can be modeled as a frequency modulation process.
Figure 7 shows one such model, where a noise source combined
with a sine wave source modulates the "carrier" frequency gen-
erated by a voltage controlled oscillator.

NOISE SOURCE

VCO

DIGITAL
OUT

ANALOG IN

ADC

VOLTAGE

SOURCE

NOISE
WAVEFORM

SINE

WAVE

Figure 7. Clock Jitter Modeled as a Modulated VCO

If the jittered output of the VCO is used to clock an analog-to-
digital converter, the digital output of the ADC will be contami-
nated by the presence of jitter. If the noise source is spectrally
flat (i.e., "white" jitter), an FFT of the ADC digital output
would show a spectrum with a uniform noise floor that is

elevated compared to the spectrum with the noise source turned
off. If the noise source has distinct frequency components (i.e.,
"correlated" jitter), then an FFT of the ADC digital output
would show symmetrical sidebands around the ADC input
signal, at amplitudes and frequencies determined by frequency
modulation theory. One notable result is that the level of the
noise or the sidebands is proportional to the slope of the input
signal, i.e., the worst case occurs at the highest frequency full-
scale input (a full-scale 20 kHz sinusoid).

The AD1893 applies rejection to these jitter frequency compo-
nents referenced to the input signal. In other words, if a 5 kHz
digital sinusoid is applied to the ASRC, depending on the set-
tling mode selected, the ASRC will attenuate sample clock jitter
at either 3 Hz above and below 5 kHz (slow settling) or 12 Hz
above and below 5 kHz (fast settling). The rolloff is 6 dB per
octave. As an example, suppose there was correlated jitter
present on the input sample clock with a 1 kHz component,
associated with the same 5 kHz sinusoidal input data. This
would produce sidebands at 4 kHz and 6 kHz, 3 kHz and
7 kHz, etc., with amplitudes that decrease as they move away
from the input signal frequency. For the slow-settling mode
case, 1 kHz represents more than nine octaves (relative to
3 Hz), so the first two sideband pairs would be attenuated by
more than 54 dB. For the fast-settling mode case, 1 kHz repre-
sents more than seven octaves (relative to 12 Hz), so that the
first two sideband pairs would be attenuated by more than 42 dB.
The second and higher sideband pairs are attenuated even more
because they are spaced further from the input signal frequency.

Group Delay Modes

The other parameter that determines the likelihood of FIFO
input overflow or output underflow is the FIFO depth. This
FIFO induced group delay is better termed transport delay,
since it is frequency independent, and should be kept conceptu-
ally distinct from the notion of group delay as used in the poly-
phase filter bank model. The total group delay of the AD1893
equals the FIFO transport delay plus the FIR (polyphase) filter
group delay.

In the AD1893, the FIFO read and write pointers are separated
by five memory locations (

100

s equivalent transport delay at

a 50 kHz sample rate). This is added to the FIR filter delay
(64 taps divided by 2) for a total nominal group delay in short
mode of

700

This delay is deterministic and constant except when F

SOUT

drops below F

SIN

which causes the number of FIR filter taps to

increase (see Cutoff Frequency Modification section). If the
FIFO read and write addresses cross, the MUTE_O signal will
be asserted. Note that under all conditions, both the highly
oversampled low-pass prototype and the polyphase subfilters of
the AD1893 ASRC possess a linear phase response.

AD1893

REV. A

Cutoff Frequency Modification

The final important operating concept of the ASRC is the modi-
fication of the filter cutoff frequency when the output sample
rate (F

SOUT

) drops below the input sample rate (F

SIN

), i.e.,

during downsampling operation. The AD1893 automatically
reduces the polyphase filter cutoff frequency under this condi-
tion. This lowering of the cutoff frequency (i.e., the reduction of
the input signal bandwidth) is required to avoid alias distortion.
The AD1893 SamplePort takes advantage of the scaling prop-
erty of the Fourier transform which can be stated as follows: if
the Fourier transform of f(t) is F(w), then the Fourier transform
of f(k

t) is F(w/k). This property can be used to linearly com-

press the frequency response of the filter, simply by multiplying
the coefficient ROM addresses (shown in Figure 6) by the ratio
of F

SOUT

to F

SIN

whenever F

SOUT

is less than F

SIN

. This scaling

property works without spectral distortion because the time scale
of the interpolated signal is so dense (300 ps resolution) with
respect to the cutoff frequency that the discrete-time representa-
tion is a close approximation to the continuous time function.

The cutoff frequency (3 dB down) of the FIR filter during
downsampling is given by the following relation:

Downsampling Cutoff Frequency = (F

SOUT

/44.1 kHz)

20 kHz

The AD1893 frequency response compression circuit includes a
first order low-pass filter to smooth the filter cutoff frequency
selection during dynamic sample rate conditions. This allows
the ASRC to avoid objectionable clicking sounds that would
otherwise be imposed on the output while the loop settles to a
new sample rate ratio. Hysteresis is also applied to the filter
selection with approximately 300 Hz of cutoff frequency "noise
margin," which limits the available selection of cutoff frequen-
cies to those falling on an approximately 300 Hz frequency grid.
Thus if a particular sample frequency ratio was reached by slid-
ing the output sample frequency up, it is possible that a filter
will be chosen with a cutoff frequency that could differ by as
much as 300 Hz from the filter chosen when the same sample
frequency ratio was reached by sliding the output sample fre-
quency down. This is necessary to ensure that the filter selection
is stable even with severely jittered input sample clocks.

Note that when the filter cutoff frequency is reduced, the transi-
tion band of the filter becomes narrower since the scaling prop-
erty affects all filter characteristics. The number of FIR filter
taps necessarily increases because there are now a smaller num-
ber of longer length polyphase filters. Nominally, when F

SOUT

greater than F

SIN

, the number of taps is 64. When F

SOUT

is less

than F

SIN

, the number of taps linearly increase to a maximum

of 128 when the ratio of F

SOUT

to F

SIN

equals 1:2. The number

of filter taps as a function of sample clock ratio is illustrated in
Figure 8. The natural consequence of this increase in filter taps
is an increase in group delay.

0.5

1.0

1.5

2.0

NUMBER OF FILTER TAPS

128

SOUT

SIN

UPSAMPLING

DOWN-

SAMPLING

Figure 8. Number of Filter Taps as a Function of F

SOUT

SlN

When the AD1893 output sample frequency is higher than the
input sample frequency (i.e., upsampling operation), the cutoff
frequency of the FIR polyphase filter can be greater than 20 kHz.
The cutoff frequency of the FIR filter during upsampling is
given by the following relation:

Upsampling Cutoff Frequency = (F

SIN

/44.1 kHz)

20 kHz

Noise and Distortion Phenomena

There are three noise/distortion phenomena that limit the per-
formance of the AD1893 ASRC. First, there is broadband,
Gaussian noise that results from polyphase filter selection
quantization. Even though the AD1893 has a large number of
polyphase filters (the equivalent of 65,536) from which to
choose, the selection is not infinite. Second, there is narrowband
noise that results from the nonideal synchronization of the
sample clocks to the 16 MHz system clock, which leads to a
nonideal computation of the sample clock ratio, which leads
to a nonideal polyphase filter selection. This noise source is
narrowband because the digital servo control loop averages the
polyphase filter selection, leading to a strong correlation be-
tween selections from output to output. In slow mode, the selec-
tion of polyphase filters is completely unaffected by the clock
synchronization. In fast mode, some narrowband noise modula-
tion may be observed with very long FFT measurements. This
situation is analogous to the behavior of a phase locked loop
when presented with a noisy or jittered input. Third, there are
distortion components that are due to the noninfinite stopband
rejection of the low-pass filter response. Noninfinite stopband
rejection means that some amount of out-of-band spectral en-
ergy will alias into the baseband. The AD1893 performance
specifications include the effects of these phenomena.

Note that Figures 16 through 18 are shown with full-scale input
signals. The distortion and noise components will scale with the
input signal amplitude. In other words, if the input signal is
attenuated by 20 dB, the distortion and noise components will
also be attenuated by 20 dB. This dependency holds until the
effects of the 16-bit input quantization are reached.

AD1893

REV. A

OPERATING FEATURES
Serial Input/Output Ports

The AD1893 uses the frequency of the left/right input clock
(LR_ I) and the left/right output clock (LR_O) signals to deter-
mine the sample rate ratio, and therefore these signals must run
continuously and transition twice per sample period. (The LR_I
clock frequency is equivalent to F

SIN

and the LR_O clock fre-

quency is equivalent to F

SOUT

.) The other clocks (WCLK_I,

WCLK_O, BCLK_I, BCLK_O) are edge sensitive and may be
used in a gated or burst mode (i.e., a stream of pulses during
data transmission or reception followed by periods of inactivity).
The word clocks and the output bit clock are used only to write
data into or read data out of the serial ports; only the left/right
clocks are used in the internal DSP blocks. The input bit clock
is used to sample the input left/right clock. It is important that
the left/right clocks are "clean" with monotonic rising and falling
edge transitions and no excessive overshoot or undershoot which
could cause false triggering on the AD1893.

The AD1893's flexible serial input and output ports consume
and produce data in twos-complement, MSB-first format. The
left channel data field always precedes the right channel data
field; the current channel being consumed or produced is indi-
cated by the state of the left/right clock (LR_I and LR_O). A left
channel field, right channel field pair is called a frame. The
input data field consists of 4 to 16 bits. The output data field
consists of 4 to 24 bits. The input signals are specified to TTL
logic levels, and the outputs swing to full CMOS logic levels.
The ports are configured by pin selections.

Serial I/O Port Modes

The AD1893 has pin-selectable bit clock polarity for the input
and output ports. In "normal" mode (BKPOL_I or BKPOL_O
LO) the data is valid on the rising edge. In the "inverted" mode
(BKPOL_I or BKPOL_O HI) the data is valid on the falling
edge. Both modes are shown in Figures 23 and 24.

The AD1893 uses two multiplexed input pins to control the mode
configuration of the input and output serial ports. MODE0_I
and MODE1_I control the input serial port, and MODE0_O
and MODE1_O control the output serial port as follows:

MODE0_I

MODE1_I

Serial Input Port Mode

Left-justified, no MSB delay, LR_I clock
triggered.

Left-justified, MSB delay, LR_I clock
triggered.

Right-justified, MSB delayed 16-bit clock
periods from LR_I clock transition, LR_I
clock triggered.

Word clock triggered, no MSB delay.

MODE0_O MODE1_O Serial Output Port Mode

Left-justified, no MSB delay, LR_O clock
triggered.

Left-justified, MSB delay, LR_O clock
triggered.

Right-justified, MSB delayed 16-bit clock
periods from LR_O clock transition, LR_O
clock triggered.

Word clock triggered, no MSB delay.

The MSB delay is useful for I

S format compatibility and for

ease of interfacing to some DSP processors.

The AD1893 SamplePort serial ports operate in either the word
clock (WCLK_I, WCLK_O) triggered mode or left/right clock
(LR_I, LR_O) triggered mode. These modes can be utilized
independently for the input and output ports. In the word clock
triggered mode, as shown in Figure 23, after the left/right clock
is valid, the appearance of the MSB of data is synchronous with
the rising edge of the word clock. Note that the word clock is
rising edge sensitive, and can fall anytime after it is sampled HI
by the bit clock. In the left-justified left/right clock triggered
modes, as shown in Figure 24, the appearance of the MSB of
data is synchronous with the rising edge of the left/right clock
for the left channel and the falling edge of left/right clock for the
right channel. The MSB is delayed by one bit clock after the
left/right clock if the MSB delay mode is selected. In the right-
justified left/right clock triggered mode, as shown in Figure 25,
the MSB is delayed 16 bit clock periods from a left/right clock
edge, so that when there are 64 bit clock periods per frame, the
LSB is right-justified to a left/right clock edge. The word clock
is not required in the left/right clock triggered modes, and
should be tied either HI or LO. Figure 24 shows the bit clock in
the optional gated or burst mode; the bit clock is inactive be-
tween data fields, and can take either the HI state or the LO
state while inactive.

Note that there is no requirement for a delay between the left
channel data and the right channel data. The left/right clocks
and the word clocks can transition immediately after the LSB of
the data, so that the MSB of the subsequent channel appears
without any timing delay. The AD1893 is therefore capable of a
32-bit frame mode, in which both 16-bit channels are packed
into a 32-bit clock period. More generally, there is no particular
requirement for when the left/right clock falls (i.e., there is no
left/right clock duty cycle or pulsewidth specification), provided
that the left/right clock frequency equals the intended sample
frequency, and there are sufficient bit clock periods to clock in
or out the intended number of data bits.

On-Chip Oscillator

The AD1893 includes an on-chip oscillator so that the user
need only supply an external quartz crystal or ceramic resonator.
The crystal or the resonator should be tied to the XTAL_I and
XTAL_O pins of the AD1893. An external crystal oscillator can
be used to overdrive the AD1893 on-chip oscillator. The exter-
nal clock source should be applied to the XTAL_I pin, and the
XTAL_O pin should be left unconnected.

AD1893

XTAL_I

XTAL_O

AD1893

XTAL_I

XTAL_O

16MHz CRYSTAL CONNECTION

16MHz OSCILLATOR CONNECTION

16MHz

2064pF

Figure 9. Crystal and Oscillator Connections

AD1893

REV. A

Some applications using multiple AD1893s may desire to use
the same master clock frequency for all the SamplePorts, sup-
plied by a single crystal. The crystal output can be buffered with
a 74HCXX gate and distributed to the other XTAL_I inputs, as
shown in Figure 10.

AD1893

XTAL_I

XTAL_O

16MHz

20pF

15pF

74HC DEVICE

TO XTAL_I
INPUTS

Figure 10. Buffered 16 MHz Crystal Connection

Power-Down Mode

The AD1893 includes a power-down control input pin
PWRDWN. This control signal is active HI, and puts the
AD1893 in an inactive state with very low power dissipation.
The PWRDWN pin should be connected LO when normal
operation of the AD1893 is desired.

Control Signals

The SETLSLW, BKPOL_I, BKPOL_O, MODE0_I, MODE1_I,
MODE0_O and MODE1_O inputs are asynchronous signals in
that they need obey no particular timing relation to the crystal
frequency or the sample clocks. Ordinarily, these pins are hard-
wired or connected to an I/O register for microprocessor control.
The only timing requirement on these pins is that the control
signals are stable and valid before the first serial input data bit
(i.e., the MSB) is presented to the AD1893.

Reset

Figure 27 shows the reset timing for the AD1893 SamplePort. A
crystal (or resonator) must be connected to the AD1893 when
RESET is asserted, and the bit clocks, the word clocks and the
left/right clocks may also be running. When the AD1893 comes
out of reset, it defaults to a F

SIN

to F

SOUT

ratio of 1:1. The filter

pipeline is not cleared. However, the mute output goes HI for at
least 128 cycles, adequate to allow the pipeline to clear. If F

SIN

differs significantly from F

SOUT

, then the AD1893 sample clock

servo control loop also has to settle. While settling, the mute
output will be HI. After the external system resets the AD1893,
it should wait until the mute output goes LO before clocking in
serial data.

There is no requirement for using the RESET pin at power-up
or when the input or output sample rate changes. If it is not
used, the AD1893 will settle to the sample clocks supplied within

200 ms in fast-settling mode or within

800 ms in slow-settling

mode.

APPLICATION ISSUES
Dither

Due to the large output word length, no redithering of the
AD1893 output is necessary. This assumes that the input is
properly dithered and the user retains the same or greater num-
ber of output bits as there are input bits. The AD1893 output
bit stream may thus be used directly as the input to downstream
digital audio processors, storage media or output devices.

If the AD1893 is to be used to dramatically downsample (i.e.,
output sample frequency is much lower than input sample fre-
quency), the input should be sufficiently dithered to account for
the limiting of the input signal bandwidth (which reduces the
rms level of the input dither). No dither is internally used or
applied to the audio data in the AD1893 SamplePort.

Decoupling and PCB Layout

The AD1893 ASRC has two power and two ground connections to
minimize output switching noise and ground bounce. (Pins 14
[DIP] and 16 [LQFP] are actually control inputs, and should be
tied LO, but need not be decoupled.) The DIP version places
the power and ground pins at the center of the device to optimize
switching performance. The AD1893 should be decoupled
with two high quality 0.1

F or 0.01

F ceramic capacitors

(preferably surface mount chip capacitors, due to their low in-
ductance), one between each V

/GND pair. Best practice PCB

layout and interconnect guidelines should be followed. This may
include terminating the bit clocks or the left/right clocks if exces-
sive overshoot or undershoot is evident and avoiding parallel
PCB traces to minimize digital crosstalk between clocks and
control lines. Note that DIP and LQFP sockets reduce elec-
trical performance due to the additional inductance they
impose; sockets should therefore be used only when required.

Master Clock

Using a 16 MHz crystal, the nominal range of sample frequencies
that the AD1893 accepts is from 8 kHz to 56 kHz. Other
sample frequency ranges are possible by linearly scaling the
crystal frequency. For example, a 12 MHz crystal would yield a
sample frequency range of 6 kHz to 42 kHz. The approximate
relative upper bound sample frequency is the crystal frequency
divided by 286; the approximate relative lower bound sample
frequency is the crystal frequency divided by 2000. The audio
performance will not degrade if the sample frequencies are kept
within these bounds. The AD1893 SamplePort is production
tested at 16 MHz. Note that due to finite register length con-
straints, there is a minimum input sample frequency (LR_I).
The allowable input and output sample frequency ranges for
crystal frequencies of 16 MHz and 12 MHz are shown in Figures
11 and 12.

SIN

 kHz

SOUT

 kHz

SIN

SOUT

= 2/1

SIN

SOUT

= 1/1

SIN

SOUT

= 1/2

8kHz

DOWNSAMPLING

UPSAMPLING

56kHz

Figure 11. Allowable Input and Output Sample Frequencies
F

CRYSTAL

= 16 MHz Case

AD1893

REV. A

SIN

 kHz

SOUT

 kHz

SIN

SOUT

= 2/1

SIN

SOUT

= 1/1

SIN

SOUT

= 1/2

6kHz

UP-

SAMPLING

DOWN-

SAMPLING

42kHz

Figure 12. Allowable Input and Output Sample Frequencies
F

CRYSTAL

= 12 MHz Case

Multiple SamplePort Synchronization and Performance
Degradation

Multiple parallel AD1893 SamplePorts may be used in a single
system. Multiple AD1893s can be "synchronized" by simply
sharing the same reset and buffered crystal connections (see
Figure 10), and ensuring that all the SamplePorts leave the reset
state on the same crystal falling edge. No other provision is
necessary since the different AD1893s will process samples
identically if they are presented with the same input and output
clocks (neglecting the effect of excessive clock skew on the PCB,
as well process variations between ASRCs which could cause
different devices to trigger at slightly different times on excess-
ively slow rising or falling clock edges).

It is also likely that several AD1893s could end up in a serial
cascade arrangement, either in a single system design or as the
result of two or more systems, each using a single AD1893 in
the signal path. The audio signal quality will be degraded with
each pass through a SamplePort, though to a very minor degree.
The THD+N performance will degrade by 3 dB with every
doubling of the number of passes through an ASRC. For ex-
ample, the AD1893 THD+N specification of 94 dB will rise to
91 dB if the signal makes two passes through an ASRC. The
overall system THD+N specification will rise to 88 dB with
four passes, and so on.

Clipping

Under certain rare input conditions, it is possible for the
AD1893 to produce a clipped output sample. This situation is
best comprehended by employing the interpolation/decimation
model. If two consecutive samples happened to have full-scale
amplitudes (representing the peak of a full-scale sine wave, for
example), the interpolated sample (or samples) between these
two samples might have an amplitude greater than full scale. As
this is not possible, the AD1893 will compute a full-scale ampli-
tude for the interpolated sample or samples (see Figure 13).
Clipping can also arise due to the pre-echo and post-echo Gibbs
phenomena of the FIR filter, when presented with a full-scale
step input. The result of this erroneous or clipped output
sample may be measured as an extremely small decrease in
headroom for transient signals.

Sample Rate Conversion Ratio Range

The AD1893 does not support exact 1:2 (or 2:1) sample rate
conversion. The SamplePort will convert to within several hertz
of the 1:2 range, but will mute before reaching the exact 1:2
ratio. Thus the AD1893 will not support applications where the
input and output sample clocks are derived from a common
source but differ by a divide-by-two. When the ratio between
the input and output sample clock reaches to within 1% of 1:2
or 2:1, the THD+N performance may degrade by several deci-
bels, due to the wraparound of the internal read/write memory
location pointers.

CORRECTLY INTERPOLATED SAMPLE

CLIPPED INTERPOLATED SAMPLE

FULL-SCALE

AMPLITUDE

TIME

Figure 13. Nipped Output Sample

Options for Sample Rate Conversion over a Wider Range

There are systems requiring sample rate conversion over a range
that is wider than the 1:2 or 2:1 range provided by a single
AD1893, such as for "scrubbing" in digital audio editors. There
are at least two options in this situation. The first is to use a
programmable DSP chip to perform simple integer ratio inter-
polation or decimation, and then use the AD1893 when this
intermediate output sample frequency is within the approximate
1:2 or 2:1 range of the final desired output sample frequency. The
second is to use multiple AD1893 devices cascaded in series to
achieve the required sample rate range.

"Almost Synchronous" Operation

It is possible to apply input and output sample frequencies
which are very close (within a few hertz) or in fact synchronous
(LR_I and LR_O tied together). There is no performance pen-
alty when using the AD1893 in "almost synchronous" applica-
tions. Indeed, there is a very slight performance benefit when
the input and output sample clocks are synchronous since the
alias distortion components which arise from the noninfinite
stopband attenuation of the FIR filter will pile up exactly on top
of the sinusoidal frequency components of the input signal, and
will thus be masked.

It has been empirically observed that during almost synchronous
operation, certain AES/EBU receivers, when used to generate
the input bit clock (BCLK_I) using a 128 times F

bit clock

frequency, can suffer sympathetic phase lock to the output bit
clock (BCLK_O) when the output bit clock is also operated at a
128 times F

rate. In other words, due to a noise pickup

mechanism in the analog phase lock loop portion of these AES/
EBU receivers, the lock frequency is pulled to match the fre-
quency of the output bit clock. The system can suffer inter-
mittent bursts of audible distortion when this sympathetic lock
phenomenon occurs. Analog Devices recommends avoiding the
use of a 128 times F

output bit clock frequency if almost syn-

chronous application is intended. The use of a 64 times F

output bit clock rate is recommended.

AD1893

REV. A

System Mute

The mute function applies to both right and left channels on the
AD1893. The user can include a system-specific output mute
signal, while retaining the automatic mute feature of the AD1893
by using the circuit shown in Figure 14.

AD1893

EXTERNAL SYSTEM MUTE
ACTIVE HI

MUTE_O

MUTE_I

Figure 14. External Mute Circuit

Performance Graphs

100

20k

10k

FREQUENCY Hz

60

80

70

90

110

100

130

120

140

160

150

dBFS

Figure 15. Dynamic Range from 20 Hz to 20 kHz,
60 dBFS, 48 kHz Input Sample Frequency, 44.1 kHz
Output Sample Frequency, 16k-Point FFT, BH4 Window

dBFS

40

140

20

80

60

120

100

20k

10k

FREQUENCY Hz

Figure 16. 1 kHz Tone at 0 dBFS, 48 kHz Input Sample
Frequency, 44.1 kHz Output Sample Frequency,
16k-Point FFT, BH4 Window

dBFS

40

140

20

80

60

120

100

20k

10k

FREQUENCY Hz

Figure 17. 15 kHz Tone at 0 dBFS, 48 kHz Input Sample
Frequency, 44.1 kHz Output Sample Frequency,
16k-Point FFT, BH4 Window

100

20k

10k

FREQUENCY Hz

120

80

100

110

90

95

105

115

85

dBFS

Figure 18. THD+N vs. Frequency, 48 kHz Input Sample
Frequency, 44.1 kHz Output Sample Frequency, Full-Scale
Input Signal

AD1893

REV. A

90

100

10

20

30

40

50

60

70

80

1kHz

AMPLITUDE dBFS

dBFS

96

100

98

94

95

97

99

92

93

90

91

20kHz

Figure 19. THD+N vs. Input Amplitude, 44.1 kHz Input
Sample Frequency, 48 kHz Output Sample Frequency,
1 kHz and 20 kHz Tones

25k

30k

40k

44.1k

35k

FREQUENCY kHz

10

dBFS

Figure 20. Digital Filter Signal Transfer Function, 10 kHz
to 20 kHz, 44.1 kHz Input Sample Frequency, 44.1, 40, 35,
30 and 25 kHz Output Sample Frequencies

140

80

120

100

20

60

40

dBFS

20k

18k

16k

14k

12k

10k

FREQUENCY Hz

Figure 21. Twintone, 10 kHz and 11 kHz, 44.1 kHz Input
Sample Frequency, 48 kHz Output Sample Frequency,
16k-Point FFT, BH4 Window

140

10k

80

120

100

20

60

40

dBFS

FREQUENCY Hz

Figure 22. 5 kHz Tone at 0 dBFS with 100 ns p-p Bino-
minal Jitter on L/R Clocks, Fast Settling Mode, 48 kHz
Input Sample Frequency, 44.1 kHz Output Sample
Frequency, 16k-Point FFT, BH4 Window

AD1893

REV. A

DATA IN/OUT

BCLK_I, BCLK_O

NORMAL MODE

BCLK_I, BCLK_O

INVERTED MODE

_I, L

INPUT

WCLK_I, WCLK_O

INPUT

MSB1 MSB2 MSB3

LEFT DATA

RIGHT DATA

LSB+1

LSB

MSB3

MSB1 MSB2

LSB

MSB

Figure 23. Serial Data Input and Output Timing, Word Clock Triggered Mode

BCLK_I, BCLK_O

NORMAL MODE

BCLK_I, BCLK_O

INVERTED MODE

INPUT

DATA IN/OUT

NO MSB DELAY MODE

MSB2 MSB3

LEFT DATA

RIGHT DATA

LSB+1

LSB

MSB3

MSB2

LSB

DATA IN/OUT

MSB DELAY MODE

MSB1 MSB2

LEFT DATA

LSB

LSB+1

LSB+2

MSB1 MSB2

RIGHT DATA

LSB+1

LSB

_I, L

INPUT

MSB

MSB1

MSB

MSB1

MSB

Figure 24. Serial Data Input and Output Timing, Left-Justified Left/Right Clock Triggered Modes

RIGHT DATA

MSB3

MSB

MSB1 MSB2

MSB2 MSB3

MSB

MSB1

LSB+1

LSB

LSB+1

BCLK_I, BCLK_O

NORMAL MODE

BCLK_I, BCLK_O

INVERTED MODE

INPUT

_I, L

INPUT

LEFT DATA

LSB1

16-BIT CLOCK PERIODS

DATA IN/OUT

16-BIT CLOCK PERIODS

LSB

Figure 25. Serial Data Input and Output Timing, Right-Justified Left / Right Clock Triggered Mode

Ð­Ð»ÐµÐºÑ‚Ñ€Ð¾Ð½Ð½Ñ‹Ð¹ ÐºÐ¾Ð¼Ð¿Ð¾Ð½ÐµÐ½Ñ‚: AD1893

ÐÐ»ÐµÐºÑ‚Ñ€Ð¾Ð½Ð½Ñ‹Ð¹ ÐºÐ¾Ð¼Ð¿Ð¾Ð½ÐµÐ½Ñ‚: AD1893