ChipFind - документация

Электронный компонент: CS5317-KP

Скачать:  PDF   ZIP

Document Outline

1
Copyright
Cirrus Logic, Inc. 1997
(All Rights Reserved)
Cirrus Logic, Inc.
Crystal Semiconductor Products Division
P.O. Box 17847, Austin, Texas 78760
(512) 445 7222 FAX: (512) 445 7581
http://www.crystal.com
CS5317
16-Bit, 20 kHz Oversampling A/D Converter
Features
l
Complete Voiceband DSP Front-End
- 16-Bit A/D Converter
- Internal Track & Hold Amplifier
- On-Chip Voltage Reference
- Linear-Phase Digital Filter
l
On-Chip PLL for Simplified Output Phase
Locking in Modem Applications
l
84 dB Dynamic Range
l
80 dB Total Harmonic Distortion
l
Output Word Rates up to 20 kHz
l
DSP-Compatible Serial Interface
l
Low Power Dissipation: 220 mW
Description
The CS5317 is an ideal analog front-end for voiceband
signal processing applications such as high-perfor-
mance modems, passive sonar, and voice recognition
systems. It includes a 16-bit A/D converter with an inter-
nal track & hold amplifier, a voltage reference, and a
linear-phase digital filter.
An on-chip phase-lock loop (PLL) circuit simplifies the
CS5317's use in applications where the output word rate
must be locked to an external sampling signal.
The CS5317 uses delta-sigma modulation to achieve
16-bit output word rates up to 20 kHz. The delta-sigma
technique utilizes oversampling followed by a digital fil-
tering and decimation process. The combination of
oversampling and digital filtering greatly eases antialias
requirements. Thus, the CS5317 offers 84 dB dynamic
range and 80 dB THD and signal bandwidths up to
10 kHz at a fraction of the cost of hybrid and discrete
solutions.
The CS5317's advanced CMOS construction provides
low power consumption of 220 mW and the inherent re-
liability of monolithic devices.
ORDERING INFORMATION
See page 20.
I
5()%8)
9ROWDJH
5HIHUHQFH
$,1
1&
3//&ORFN *HQHUDWRU
&RPSDUDWRU
WK 2UGHU
'HFLPDWLRQ
)LOWHU
6WDWH
%XIIHUV
9$
9$
$*1'
9'
9'
'*1'
9&2,1
3+'7
567
02'(
&/.,1
'2(
&/.287
'$7$
'287
MAR `95
DS27F4
ANALOG CHARACTERISTICS
(T
A
= T
MIN
- T
MAX
; VA+, VD+ = 5V
10%; VA-, VD- = -5V
10%;
CLKIN = 4.9152 MHz in CLKOR mode; 1kHz Input Sinewave; with 1.2 k
, .01
F antialiasing filter.)
Parameter*
Min
Typ
Max
Units
Specified Temperature Range
0 to 70
C
Resolution
16
-
-
Bits
Dynamic Performance
Dynamic Rnage
(Note 1)
78
84
-
dB
Total Harmonic Distortion
72
80
-
dB
Signal to Intermodulation Distorition
-
84
-
dB
dc Accuracy
Differential Nonlinearity
(Note2)
-
0.4
-
LSB
Positive Full-Scale Error
-
150
-
mV
Positive Full-Scale Drift
-
500
-
V/C
Bipolar Offset Error
-
10
-
mV
Bipolar Offset Drift
-
50
-
V/C
Filter Characteristics
Absolute Group Delay
(Note 3)
78.125
-
-
s
Passband Frequency
(Note 4)
-
5
-
kHz
Input Characteristics
AC Input Impedance
(1kHz)
-
80
-
k
Analog Input Full Scale Signal Level
2.75
-
-
V
Power Supplies
Power Dissipation
(Note5)
-
220
300
mW
Notes:
1. Measured over the full 0 to 9.6kHz band with a -20dB input and extrapolated to full-scale. Since this
includes energy in the stopband above 5kHz, additional post-filtering at the CS5317's output can
typically achieve 88dB dynamic range by improving rejection above 5kHz. This can be increased to
90dB by bandlimiting the output to 2.5kHz.
2. No missing codes is guaranteed by design.
3. Group delay is constant with respect to input analog frequency; that is, the digital FIR filter has
linear phase. Group delay is determined by the formula D
grp
= 384/CLKIN in CLKOR mode, or
192/CLKOUT in any mode.
4. The digital filter's frequency response scales with the master clock. Its -3dB point is determined by
f
-3dB
= CLKIN/977.3 in CLKOR mode, or CLKOUT/488.65 in any mode.
5. All outputs unloaded. All inputs CMOS levels.

* Refer to the
Parameter Definitions
section after the Pin Description section.
CS5317
2
DS27F4
ANALOG CHARACTERISTICS
(continued)
Parameter
Min
Typ
Max
Units
Power Supply Rejection
VA+
(Note 6)
VA-
VD+
VD-
-
-
-
-
60
45
60
55
-
-
-
-
dB
dB
dB
dB
Specified Temperature Range
0 to 70
C
Phase-Lock Loop Characteristics
VCO Gain Constant, Ko
(Note 7)
-4
-10
-30
Mrad/Vs
VCO Operating Frequency
1.28
-
5.12
MHz
Phase Detector Gain Control, Kd
-3
-8
-12
A/rad
Phase Detector Prop. Delay
(Note 8)
-
50
100
ns
Notes:
6. With 300mV p-p, 1kHz ripple applied to each supply separately.
7. Over 1.28 MHz to 5.12 MHz VCO output range, where VCO frequency = 2 * CLKOUT.
8. Delay from an input edge to the phase detector to a response at the PHDT output pin.
DIGITAL CHARACTERISTICS
(T
A
= T
MIN
- T
MAX
; VA+, VD+ = 5V
10%; VA-, VD- = -5V
10%)
All measurements performed under static conditions.
Parameter
Symbol
Min
Typ
Max
Units
High-Level Input Voltage
V
IH
2.0
-
-
V
Low-Level Input Voltage
V
IL
-
-
0.8
V
High-Level Output Voltage
(Note 9)
V
OH
(VD+)-1.0V
-
-
V
Low-Level Output Voltage
IOUT = 1.6mA
V
OL
-
-
0.4
V
Input Leakage Current
I
in
-
-
10
A
3-State Leakage Current
I
OZ
-
-
10
A
Digital Output Pin Capacitance
C
out
-
9
-
pF
Note:
9. I
out
=-100
A. This specification guarantees the ability to drive one TTL load (V
OH
=2.4V @ I
out
=-40
A.).
RECOMMENDED OPERATING CONDITIONS
(DGND, AGND = 0V, see Note 10.)
Parameter
Symbol
Min
Typ
Max
Units
DC Power Supplies:
Positive Digital
Negative Digital
Positive Analog
Negative Analog
VD+
VD-
VA+
VA-
4.5
-4.5
4.5
-4.5
5.0
-5.0
5.0
-5.0
5.5
-5.5
5.5
-5.5
V
V
V
V
Master Clock Frequency
f
clk
0.01
-
5.12
MHz
Note:
10. All voltages with respect to ground.
Specifications are subject to change without notice.
CS5317
DS27F4
3
SWITCHING CHARACTERISTICS
(T
A
= T
MIN
-T
MAX
; C
L
=50 pF; VD+ = 5V
10%; VD- = -5V
10%)
Parameter
Symbol
Min
Typ
Max
Units
Master Clock Frequency:
CLKIN
CLKG1 Mode
CLKG2 Mode
CLKOR Mode
f
clkg1
f
clkg2
f
clkor
-
-
-
-
-
-
20
10
5.12
kHz
kHz
MHz
Output Word Rate:
DOUT
f
dout
-
-
20
kHz
Rise Times:
Any Digital Input
Any Digital Output
t
risein
t
riseout
-
-
20
15
1000
20
ns
ns
Fall Times:
Any Digital Input
Any Digital Output
t
fallin
t
fallout
-
-
20
15
1000
20
ns
ns
CLKIN Duty Cycle
CLKG1 and CKLG2 Modes
Pulse Width Low
Pulse Width High
CLKOR Mode
Pulse Width Low
Pulse Width High
t
pwl1
t
pwh1
t
pwl1
t
pwh1
200
200
45
45
-
-
-
-
-
-
-
-
ns
ns
ns
ns
RST Pulse Width Low
t
pwr
400
-
-
ns
Set Up Times:
RST High to CLKIN High
CLKIN High to RST High
t
su1
t
su2
40
40
-
-
-
-
ns
ns
Propagation Delays:
DOE Falling to Data Valid
CLKIN Rising to DOUT Falling
(Note 11)
DOE Rising to Hi-Z Output
CLKOUT Rising to DOUT Falling
CLKOUT Rising to DOUT Rising
CLKOUT Rising to Data Valid
CLKIN Rising to CLKOUT Falling
(Note 12)
CLKIN Rising to CLKOUT Rising
(Note 12)
t
phl1
t
phl2
t
plh1
t
plh2
t
plh3
t
plh4
t
plh5
t
plh6
-
-
-
-
-
-
-
-
-
1
-
-
-
-
-
-
150
-
80
60
60
100
200
200
ns
CLKOUT
cycles
ns
ns
ns
ns
ns
Notes: 11. CLKIN only pertains to CLKG1 and CLKG2 modes.
12. Only valid in CLKOR mode.
ABSOLUTE MAXIMUM RATINGS
(DGND, AGND = 0V, all voltages with repect to groung)
Parameter
Symbol
Min
Max
Units
DC Power Supplies:
Positive Digital
Negative Digital
Positive Analog
Negative Analog
VD+
VD-
VA+
VA-
-0.3
0.3
-0.3
0.3
(VA+) + 0.3
-6.0
6.0
-6.0
V
V
V
V
Input Current, Any Pin Except Supplies
(Note 13)
I
in
-
10
mA
Analog Input Voltage (AIN and VREF pins)
V
INA
(VA-) - 0.3
(VA+) + 0.3
V
Digital Input Voltage
V
IND
-0.3
(VD+) + 0.3
V
Ambient Operating Temperature
T
A
-55
125
C
Storage Temperature
T
stg
-65
150
C
Notes: 13. Transient currents up to 100mA will not cause SCR latch-up.
WARNING:Operating this device at or beyond these extremes may result in permanent damage to the device.
Normal operation of the part is not guaranteed at these extremes.
CS5317
4
DS27F4
2.0 V
0.8 V
t
fallin
t risein
2.4 V
0.4 V
triseout
t fallout
Rise and Fall Times
DATA
CLKOUT
(Note 15)
CLKIN
DOUT
DOE
CLKIN
t pwh1
t pwl1
CLKIN
CLKOUT
(Note 14)
t plh5
t su1
t su2
t pwr
t plh5
t su1
t su2
t plh6
RST
DOUT
t plh1
t plh2
t plh3
t phl2
t phl1
14
1
0
15
(MSB)
(Note 16)
t plh4
Notes: 14. CLKIN only pertains to CLKG1 and CLKG2 modes.
15. If DOE is brought high during serial data transfer, CLKOUT, DOUT, and DATA will immediately
3-state and the rest of the serial data is lost.
16. RST must be held high except in the clock override (CLKOR) mode where it can be used to align
the phases of all internal clocks.
CLKIN Timing
Serial Output Timing
Reset Timing
CS5317
DS27F4
5
GENERAL DESCRIPTION
The CS5317 functions as a complete data conver-
sion subsystem for voiceband signal processing.
The A/D converter, sample/hold, voltage refer-
ence, and much of the antialiasing filtering are
performed on-chip. The CS5317's serial interface
offers its 16-bit, 2's complement output in a for-
mat which easily interfaces with
industry-standard micro's and DSP's.
The CS5317 also includes a phase-locked loop
that simplifies the converter's application in sys-
tems which require sampling to be locked to an
external signal source. The CS5317 continuously
samples its analog input at a rate set by an exter-
nal clock source. On-chip digital filtering, an
integral part of the delta-sigma ADC, processes
the data and updates the 16-bit output register at
up to 20 kHz. The CS5317 can be read at any rate
up to 20 kHz.
The CS5317 is a CS5316 with an on-chip sam-
pling clock generator. As such, it replaces the
CS5316 and should be considered for all new de-
signs. In addition, a CS5316 look-alike mode is
included, allowing a CS5317 to be dropped into a
CS5316 socket.
THEORY OF OPERATION
The CS5317 utilizes the delta-sigma technique of
executing low-cost, high-resolution A/D conver-
sions. A delta-sigma A/D converter consists of
two basic blocks: an analog modulator and a digi-
tal filter.
Conversion
The analog modulator consists of a 1-bit A/D
converter (that is, a comparator) embedded in an
analog negative feedback loop with high open-
loop gain. The modulator samples and converts
the analog input at a rate well above the band-
width of interest (2.5 MHz for the CS5317). The
modulator's 1-bit output conveys information in
the form of duty cycle. The digital filter then
processes the 1-bit signal and extracts a high
resolution output at a much lower rate (that is,
16-bits at a 20 kHz word rate with a 5 kHz input
bandwidth).
An elementary example of a delta-sigma A/D
converter is a conventional voltage-to-frequency
converter and counter. The VFC's 1-bit output
conveys information in the form of frequency (or
duty-cycle), which is then filtered (averaged) by
the counter for higher resolution. In comparison,
the CS5317 uses a more sophisticated multi-order
modulator and more powerful FIR filtering to ex-
tract higher word rates, much lower noise, and
more useful system-level filtering.
Filtering
At the system level, the CS5317's digital filter
can be modeled exactly like an analog filter with
a few minor differences. First, digital filtering re-
sides behind the A/D conversion and can thus
reject noise injected during the conversion proc-
ess (i.e. power supply ripple, voltage reference
noise, or noise in the ADC itself). Analog filtering
cannot.
Also, since digital filtering resides behind the
A/D converter, noise riding unfiltered on a near-
full-scale input could potentially saturate the
ADC. In contrast, analog filtering removes the
noise before it ever reaches the converter. To ad-
dress this issue, the CS5317's analog modulator
and digital filter reserve headroom such that the
device can process signals with 100mV "excur-
sions" above full-scale and still output accurately
converted and filtered data. Filtered input signals
above full-scale still result in an output of all
ones.
An Application Note called "Delta Sigma Over-
view" contains more details on delta-sigma
conversion and digital filtering.
CS5317
6
DS27F4
SYSTEM DESIGN WITH THE CS5317
Like a tracking ADC, the CS5317 continuously
samples and converts, always tracking the analog
input signal and updating its output register at a
20 kHz rate. The device can be read at any rate to
create any system-level sampling rate desired up
to 20kHz.
Clocking
Oversampling is a critical function in delta-sigma
A/D conversion. Although system-level output
sample rates typically remain between 7kHz and
20kHz in voiceband applications, the CS5317 ac-
tually samples and converts the analog input at
rates up to 2.56 MHz. This internal sampling rate
is typically set by a master clock which is on the
order of several megahertz. See Table1 for a com-
plete description of the clock relationships in the
various CS5317 operating modes.
Some systems such as echo-canceling modems,
though, require the output sampling rate to be
locked to a sampling signal which is 20 kHz or
below. For this reason the CS5317 includes an
on-chip phase-lock loop (PLL) which can gener-
ate its requisite 5.12 MHz master clock from a
20 kHz sampling signal.
The CS5317 features two modes of operation
which utilize the internal PLL. The first, termed
Clock Generation 1 (CLKG1), accepts a sam-
pling clock up to 20 kHz at the CLKIN pin and
internally generates the requisite 5.12 MHz clock.
The CS5317 then processes samples updating its
output register at the rate defined at CLKIN, typi-
cally 20 kHz. For a 20 kHz clock input the digital
filter's 3 dB corner is set at 5.239 kHz, so CLKG1
provides a factor of 2X oversampling at the sys-
tem level
(20 kHz is twice the minimum possible
sampling frequency needed to reconstruct a 5
kHz input). The CLKG1 mode is initiated by ty-
ing the MODE input to +5V.
+5V
Analog
Supply
0.1
F
1
2
10
CS5317
11
AIN
Signal
Conditioning
14
4
10
VA-
AGND
DGND
VD-
0.1
15
12
REFBUF
VA+
VD+
7
9
Clock
Source
Source
Signal
Analog
-5V
Analog
Supply
6
5
Serial
Data
Interface
8
3
16
uP or DSP
Control
(clock gen. mode / CLKG2)
VD- (clock override mode / CLKOR)
VD+ (clock gen. mode / CLKG1)
17
PHDT
F
VCOIN
18
VA+
10
25 nF
25 k
2.75V
10
F
10
F
10
F
10
F
0.1
F
CLKIN
MODE
CLKOUT
DATA
DOE
RST
DOUT
0.1
F 0.1
F
Figure 1. System Connection Diagram with Example PLL Components
CS5317
DS27F4
7
The second PLL mode is termed Clock Genera-
tion 2
(CLKG2) which generates its 5.12 MHz
clock from a 10 kHz external sampling signal.
Again, output samples are available at the system
sampling rate set by CLKIN, typically 10 kHz.
For the full-rated 10 kHz clock CLKG2 still sets
the filter's 3 dB point at 5 kHz. Therefore,
CLKG2 provides no oversampling beyond the
Nyquist requirement at the system level
(10 kHz : 5 kHz) and its internal digital filter pro-
vides little anti-aliasing value. The CLKG2 mode
is initiated by grounding the MODE pin.
The CS5317 features a third operating mode
called Clock Override (CLKOR). Initiated by ty-
ing the MODE pin to -5V, CLKOR allows the
5.12 MHz master clock to be driven directly into
the CLKIN pin. The CS5317 then processes sam-
ples updating its output register at f
clkin
/256.
Since all clocking is generated internally, the
CLKOR mode includes a Reset capability which
allows the output samples of multiple CS5317's
to be synchronized.
The CS5317 also has a CS5316 compatible
mode, selected by tying RST low, and using
MODE (pin 7) as the FSYNC pin. See the
CS5316 data sheet for detailed timing informa-
tion.
Analog Design Considerations
DC Characteristics
The CS5317 was designed for signal processing.
Its analog modulator uses CMOS amplifiers re-
sulting in offset and gain errors which drift over
temperature. If the CS5317 is being considered
for low-frequency (< 10 Hz) measurement appli-
cations, Crystal Semiconductor recommends the
CS5501, a low-cost, d.c. accurate, delta-sigma
ADC featuring excellent 60 Hz rejection and a
system-level calibration capability.
The Analog Input Range and Coding Format
The input range of the CS5317 is nominally
3V,
with
250 mV possible gain error. Because of
this gain error, analog input levels should be kept
below
2.75V. The converter's serial output ap-
pears MSB-first in 2's complement format.
Antialiasing Considerations
In applying the CS5317, aliasing occurs during
both the initial sampling of the analog input at f
s
in
(~2.5 MHz) and during the digital decimation
process to the 16-bit output sample rate, f
s
out
.
Mode
Symbol
Mode
Pin
RESET
Output Word
Rate Provides
System-level 2X
Oversampling
CLKIN
(kHz)
CLKOUT
f
sin
(MHz)
DOUT
f
sout
(kHz)
F
(kHz)
t
dcD
*
(ns)
Clock
Gen. 2
CLKG2
CLKG2
CLKG2
0V
HIGH
NO
7.2
9.6
10.0 (max)
1.8432
2.4576
2.56
7.2
9.6
10.0
14.4
19.2
20.0
542.5
406.9
390.6
Clock
Gen. 1
CLKG1
CLKG1
CLKG1
+5V
HIGH
YES
14.4
19.2
20.0 (max)
1.8432
2.4576
2.56
14.4
19.2
20.0
14.4
19.2
20.0
542.5
406.9
390.6
Clock
Override
CLKOR
CLKOR
CLKOR
-5V
SYNC
YES
3686.4
4915.2
5120.0 (max)
1.8432
2.4576
2.56
14.4
19.2
20.0
14.4
19.2
20.0
N/A
N/A
N/A
CS5316
CS5316
FSYNC
LOW
YES
5120.0 (max)
2.56
20.0
20.0
N/A
* t
dcD
- Delay from CLKIN rising to DOUT falling = 1 CLKOUT cycle
Table 1. Mode Comparisons
CS5317
8
DS27F4
Initial Sampling
The CS5317 samples the analog input, AIN, at
one-half the master clock frequency (~2.5 MHz
max). The input sampling frequency, f
s
in
, appears
at CLKOUT regardless of whether the master
clock is generated on-chip (CLKG1 and CLKG2
modes) or driven directly into the CS5317
(CLKOR mode). The digital filter then processes
the input signal at the input sample rate.
Like any sampled-data filter, though, the digital
filter's passband spectrum repeats around integer
multiples of the sample rate, f
s
in
. That is, when
the CS5317 is operating at its full-rated speed any
noise within
5 kHz bands around 2.5 MHz, 5
MHz, 7.5 MHz, etc. will pass unfiltered and alias
into the baseband. Such noise can only be filtered
by analog filtering before the signal is sampled.
Since the signal is heavily oversampled (2.5
MHz : 5 kHz, or 500 : 1), a single-pole passive
RC filter can be used as shown in Figure 2.
0
F
2F
3F
dc
5 kH
z
10 k
H
z
20 k
H
z
40
k
H
z
60 kH
z
.25 F .5 F
0
-40
-80
-2.74 dB
f
sin
(
128
T
)
128sin
(
T
)
3
=
Magnitude where: T = 1/
sin
sin
= input sampling frequency = CLKOUT frequency for all modes
= CLKIN/2 in CLKOR mode
= CLKIN*128 in CLKG1 mode
= CLKIN*256 in CLKG2 mode
F =
sin
/128 for all modes
= input frequency
sout
=
sin
/128 = output data rate for CLKOR & CLKG1 = F
sout
=
sin
/256 = output data rate for CLKG2 = F/2
Examples:
For
sin
= 2.56 MHz at
= 5 kHz: Magnitude is -2.74 dB
For
sin
= 2.56 MHz at
= 10 kHz: Magnitude is -11.8 dB
Figure 3. CS5317 Low-Pass Filter Response
Mag
H(e
j
)
(dB)
1.2 K
0.01
F
AIN
Input
Signal
Note: Any nonlinearities contributed by this filter
will be encoded as distortion by the CS5317.
Therefore a low distortion, high frequency ca-
pacitor such as COG-ceramic is recommended.
Figure 2. Anti-alias Filter
CS5317
DS27F4
9
Decimation
Aliasing effects due to decimation are identical in
the CLKOR and CLKG1 modes. Aliasing is dif-
ferent in the CLKG2 mode due to the difference
in output sample rates (10 kHz vs. 20 kHz) and
thus will be discussed separately.
Aliasing in the CLKOR and CLKG1 Modes
The delta-sigma modulator output is fed into the
digital low-pass filter at the input sampling rate,
f
s
in
. The filter's frequency response is shown in
Figure 3. In the process of filtering the digitized
signal the filter decimates the sampling rate by
128 (that is, f
s
out
= f
s
in
/128). In its most elemen-
tary form, decimation simply involves ignoring -
or selectively reading - a fraction of the available
samples.
In the process of decimation the output of the
digital filter is effectively resampled at f
s
out
, the
output word rate, which has aliasing implications.
Residual signals after filtering at multiples of f
s
out
will alias into the baseband. For example, an in-
put tone at 28 kHz will be attenuated by 39.9 dB.
If f
s
out
= 20 kHz, the residual tone will alias into
the baseband and appear at 8 kHz in the output
spectrum.
If the input signal contains a large amount of out-
of-band energy, additional analog and/or digital
antialias filtering may be required. If digital post-
filtering is used to augment the CS5317's
rejection above f
s
out
/4 (that is, above 5 kHz), the
filtering will also reject residual quantization
noise from the modulator. This will typically in-
crease the converter's dynamic range to 88 dB.
Further bandlimiting the digital output to f
s
out
/8
(2.5 kHz at full speed) will typically increase dy-
namic range to 90 dB.
Aliasing in the CLKG2 Mode
Aliasing effects in the CLKG2 mode can be mod-
eled exactly as those in the CLKG1 mode with
the output decimated by two (from 20 kHz to 10
kHz). This is most easily achieved by ignoring
every other sample. In the CLKG2 mode the ratio
of the output sampling rate to the filter's -3 dB
point is two, with no oversampling beyond the
demands of the Nyquist criterion. Without the
ability to roll-off substantially before f
s
out
/2, the
on-chip digital filter's antialiasing value is dimin-
ished.
The CLKG2 mode should therefore be used only
when the output data rate must be minimized due
to communication and/or storage reasons. In ad-
dition, adequate analog filtering must be provided
prior to the A/D converter.
Digital Design Considerations
The CS5317 presents its 16-bit serial output
MSB-first in 2's complement format. The con-
verter's serial interface was designed to easily
interface to a wide variety of micro's and DSP's.
Appendix A offers several hardware interfaces to
industry-standard processors.
CLKOUT
DATA
13
12 11
10
9
8
7
6
5
4
3
2
1
0
MSB
(sign bit)
LSB
15 14
15
14
f out
DOUT
Figure 4. Data Output
CS5317
10
DS27F4
Data Output Characteristics & Coding Format
As shown in Figure 4, the CS5317 outputs its 16-
bit data word in a serial burst. The data appears at
the DATA pin on the rising edge of the same
CLKOUT cycle in which DOUT falls. Data
changes on the rising edge of CLKOUT, and can
be latched on the falling edge. The CLKOUT rate
is set by the CLKIN input (f
clkin
/2 in the CLKOR
mode; f
clkin
*128 in the CLKG1 mode; and
f
clkin
*256 in the CLKG2 mode). DOUT returns
high after the last bit is transmitted. After trans-
mitting the sixteen data bits, DATA will remain
high until DOUT falls again, initiating the next
data output cycle.
A 3-state capability is available for bus-oriented
applications. The 3-state control input is termed
Data Output Enable, DOE, and is asynchronous
with respect to the rest of the CS5317. If DOE is
taken high at any time, even during a data burst,
the DATA, DOUT and CLKOUT pins go to a
high impedance state. Any data which would be
output while DOE is high is lost.
Power Supplies
Since the A/D converter's output is digitally fil-
tered in the CS5317, the device is more forgiving
and requires less attention than conventional 16-
bit A/D converters to grounding and layout
arrangements. Still, care must be taken at the de-
sign and layout stages to apply the device
properly. The CS5317 provides separate analog
and digital power supply connections to isolate
digital noise from its analog circuitry. Each sup-
ply pin should be decoupled to its respective
ground, AGND or DGND. Decoupling should be
accomplished with 0.1
F ceramic capacitors. If
significant low frequency noise is present in the
supplies, 10
F tantalum capacitors are recom-
mended in parallel with the 0.1
F capacitors.
The positive digital power supply of the CS5317
must never exceed the positive analog supply by
more than a diode drop or the chip could be per-
manently damaged. If the two supplies are de-
rived from separate sources, care must be taken
that the analog supply comes up first at power-up.
Figure 1 shows a decoupling scheme which al-
lows the CS5317 to be powered from a single set
of
5V rails. The digital supplies are derived
from the analog supplies through 10
resistors
to prevent the analog supply from dropping be-
low the digital supply.
PLL Characteristics
A phase-locked loop is included on the CS5317
and is used to generate the requisite high fre-
quency A/D sampling clock. A functional
diagram of the PLL is shown in Figure 5. The
PLL consists of a phase detector, a filter, a VCO
(voltage-controlled oscillator), and a counter/di-
vider. The phase detector inputs are CLKIN (
1
)
and a sub-multiple of the VCO output signal (
2
).
The inputs to the phase detector are positive-edge
triggered and therefore the duty cycle of the
CLKIN signal is not significant. With this type of
phase detector, the lock range of the PLL is equal
to the capture range and is independent of the low
pass filter. The output of the phase detector is in-
put to an external low pass filter. The filter
characteristics are used to determine the transient
response of the loop. The output voltage from the
filter functions as the input control voltage to the
VCO. The output of the VCO is then divided in
frequency to provide an input to the phase detec-
tor. The clock divider ratio is a function of the
PLL mode which has been selected.
Phase Detector Gain (Kd)
A properly designed and operating phase-locked
loop can be described using steady state linear
analysis. Once in frequency lock, any phase dif-
ference between the two inputs to the phase
detector cause a current output from the detector
during the phase error. While either the +50
A or
the -50
A current source may be turned on, the
average current flow is:
CS5317
DS27F4
11
i
out
avg
=
Kd
(
1
-
2
)
(-
50
A
/
2
)
(
1
-
2
)
where
1
is the phase of IN1,
2
is the phase of
IN2 and Kd is the phase detector gain. The factor
2
comes from averaging the current over a full
CLKIN cycle. Kd is in units of micro-am-
peres/radian
.
VCO Gain (Ko)
The output frequency from the VCO ranges from
1.28 MHz to 5.12 MHz. The frequency is a func-
tion of the control voltage input to the VCO. The
VCO has a negative gain factor, meaning that as
the control voltage increases more positively the
output frequency decreases. The gain factor units
are Megaradians per Volt per Second. This is
equivalent to 2
Megahertz per volt. Changes in
output frequency are given by:
vco
= Ko
VCO
in
[Ko is typ. -10Mrad/Vs.]
Counter/Divider Ratio
The CS5317 PLL multiplies the CLKIN rate by
an integer value. To set the multiplication rate, a
counter/divider chain is used to divide the VCO
output frequency to develop a clock whose fre-
quency is compared to the CLKIN frequency in
the phase detector. The binary counter/divider ra-
tio sets the ratio of the VCO frequency to the
CLKIN frequency. As illustrated in Figure 5, the
VCO output is always divided by two to yield the
CLKOUT signal which is identical in frequency
to the delta-sigma modulator sampling clock.
The CLKOUT signal is then further divided by
either 128 in the CLKG1 mode or by 256 in the
CLKG2 mode. When the divide by two stage is
included, the divider ratio (N) for the PLL in the
CLKG1 mode is effectively 256. In the CLKG2
mode the divider ratio (N) is 512.
Loop Transfer Function
As the phase-locked loop is a closed loop system,
an equation can be determined which describes its
closed loop response. Using the gain factors for
the phase detector and the VCO, the filter ar-
rangement and the counter/divider constant N,
analysis will yield the following equation which
describes the transfer function of the PLL:
2
1
=
KoKdR
N
s
+
KoKd
NC
s
2
+
KoKdR
N
s
+
KoKd
NC
This equation may be rewritten such that its ele-
ments correspond with the following
Phase/Frequency
Detect Logic
IN1
IN2
DOWN
UP
CLKIN
1
2
+5V
-5V
VA+
VCO
128
2
2
Conversion Output Rate -
Same Frequency as DOUT
Internal Sync for Digital Filter
Delta-Sigma
Sampling Clock
CLKG1
CLKG2
CLKOR
CLKOR
VCOIN
External RC
Kd = -8
A/rad
K0 = -10 Mrad/V.s
R
C
50
A
50
A
(CLKOUT)
PHDT
C2
Figure 5. PLL Functional Diagram
CS5317
12
DS27F4
characteristic form in which the damping factor,
, and the natural frequency,
n
, are evident:
2
1
=
2
n
s
+
n
2
s
2
+
2
n
s
+
n
2
Both the natur al frequency and the damping fac-
tor are particularly important in determining the
transient response of the phase-locked loop when
subjected to a step input of phase or frequency. A
family of curves are illustrated in Figure 6 that
indicate the overshoot and stability of the loop as
a function of the damping factor. Each response is
plotted as a function of the normalized time,
n
t.
For a given
and lock time, t, the
n
required
can be determined. Alternatively, phase lock con-
trol loop bandwidth may be a specified parameter.
In some systems it may be desirable to reduce the
-3dB bandwidth of the PLL control loop to re-
duce the effects of jitter in the phase of the input
clock. The 3 dB bandwidth of the PLL control
loop is defined by the following equation:
3dB
=
n

2
2
+
1
+
(
2
2
+
1
)
2
+
1
The equations used to describe the PLL and the
3 dB bandwidth are valid only if the frequency of
CLKIN is approximately 20 times greater than
the 3 dB corner frequency of the control loop.
Filter Components
Using the equations which describe the transfer
function of the PLL system, the following exter-
nal filter component equations can be determined:
C
=
KoKd
N
n
2
R
=
2
n
N
KoKd
The gain factors (Ko, Kd) are specified in the
Analog Characteristics table. In the event the sys-
tem calls for very low bandwidth, hence a
corresponding reduction in loop gain, the phase
detector gain factor Kd can be reduced. A large
series resistor (R1) can be inserted between the
output of the detector and the filter. Then the
50
A current sources will saturate to the supplies
and yield the following gain factor:
Kd
-
5V
2
R
1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
1.3
0
1
2
3
4
5
6
7
8
9
10
0.1
1
10
-10
-9
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
Figure 6a.
2
Unit Step Response
Figure 6b. Second Order PLL Frequency Response
n
t.
2
normalized to
1
/
n
20 log(
2
/
1
)
= 0.5
= 0.6
= 0.7
= 0.8
= 0.9
= 1.0
= 1.5
= 2.0
= 3.0
= 10.0
= 10.0
= 0.5
= 0.5
= 10
CS5317
DS27F4
13
In some applications additional filtering may be
useful to eliminate any jitter associated with the
discrete current pulses from the phase detector.
In this case a capacitor whose value is no more
than 0.1 C can be placed across the RC filter net-
work (C
2
in Figure 5).
Filter Design Example
The following is a step by step example of how to
derive the loop filter components. The CS5317
A/D sampling clock is to be derived from a 9600
Hz clock source. The application requires the sig-
nal passband of the CS5317 to be 4 kHz. The
on-chip digital filter of the CS5317 has a 3 dB
passband of CLKOUT/488.65 (see Note 4 in the
data sheet specifications tables). The 4 kHz pass-
band requirement dictates that the sample clock
(CLKOUT) of the CS5317 be a minimum of
4000 X 488.65 = 1.954 MHz. This requires the
VCO to run at 3.908 MHz. The 3.908 MHz rate is
407 times greater than the 9600 Hz PLL input
clock. Therefore the CS5317 must be set up in
mode CLKG2 with N = 512. If the CLKG1 mode
were used (N = 256), too narrow of a signal band-
width through the A/D would result.
Once the operating mode has been determined
from the system requirements, a value for the
damping factor must be chosen. Figure 6 illus-
trates the dynamic aspects of the system with a
given damping factor. Damping factor is gener-
ally chosen to be between 0.5 and 2.0. The choice
of 0.5 will result in an overshoot of 30 % to a step
response whereas the choice of 2.0 will result in
an overshoot of less than 5 %. For example pur-
poses, let us use a damping factor of 1.0.
So, let us begin with the following variables :
Ko = - 10 Mradians/volt.sec
Kd = - 8
A/radian
N = 512
= 1.0
To calculate values for the resistor R and capaci-
tor C of the filter, we must first derive a value for
n
. Using the general rule that the sample clock
should be at least 20 times higher frequency than
the 3dB bandwidth of the PLL control loop:
CLKIN
20
3dB
where CLKIN = 9600 Hz = 2
9600 radians/sec.
So:
3dB
= 2
9600/20 = 3016 radians/sec.
Knowing
3dB
and the damping factor of 1.0, we
can calculate the natural frequency,
n
, of the
control loop:
n
=
3dB

2
2
+
1
+

(
2
2
+
1
)
2
+
1
n
=
3016

2
(
1
)
2
+
1
+

(
2
(
1
)
2
+
1
)
2
+
1
n
=
1215 1
/
sec
Once the natural frequency,
n
, is determined,
values for R and C for the loop filter can be cal-
culated:
R = 2
n
N/KoKd
R = 2(1)(1215 1/s) 512/(-10Mrad/v.s.)(-8
A/rad)
R = 15552 v/A = 15.55 k
. Use R = 15 k
.
C = KoKd/N
n
2
C = (-10 Mrad/v.s)(- 8
A/rad)/512 (1215 1/s)
2
C = 105.8 x 10
-9
A.s/v = 105 nF. Use 0.1
F.
The above example assumed typical values for
Ko and Kd. Your application may require a worst
case analysis which includes the minimum or
maximum values. Table 2 shows some other ex-
ample situations and R and C values.
CS5317
14
DS27F4
CS5317 PERFORMANCE
The CS5317 features 100% tested dynamic per-
formance. The following section is included to
illustrate the test method used for the CS5317.
FFT Tests and Windowing
The CS5317 is tested using Fast Fourier Trans-
form (FFT) techniques to analyze the converter's
dynamic performance. A pure sine wave is ap-
plied to the CS5317 and a "time record" of 1024
samples is captured and processed. The FFT algo-
rithm analyzes the spectral content of the digital
waveform and distributes its energy among 512
"frequency bins". Assuming an ideal sinewave,
distribution of energy in bins outside of the fun-
damental and dc can only be due to quantization
effects and errors in the CS5317.
If sampling is not synchronized to the input sine-
wave it is highly unlikely that the time record will
contain an exact integer number of periods of the
input signal. However, the FFT assumes that the
signal is periodic, and will calculate the spectrum
of a signal that appears to have large discontinui-
ties, thereby yielding a severely distorted
spectrum. To avoid this problem, the time record
is multiplied by a window function prior to per-
forming the FFT. The window function smoothly
forces the endpoints of the time record to zero,
removing the discontinuities. The effect of the
"window" in the frequency domain is to convo-
lute the spectrum of the window with that of the
actual input.
The quality of the window used for harmonic
analysis is typically judged by its highest side-
lobe level. The Blackman-Harris window used to
test the CS5317 has a maximum side-lobe level
of -92 dB.
Figure 7 shows an FFT plot of a typical CS5317
with a 1 kHz sinewave input generated by an "ul-
tra-pure" sine wave generator and the output
multiplied by a Blackman-Harris window. Arti-
facts of windowing are discarded from the
signal-to-noise calculation using the assumption
that quantization noise is white. All FFT plots in
this data sheet were derived by averaging the FFT
results from ten time records. This filters the
spectral variability that can arise from capturing
finite time records, without disturbing the total
energy outside the fundamental. All harmonics
and the -92 dB side-lobes from the Blackman-
Harris window are therefore clearly visible in the
plots.
CLKIN
(Hz)
Mode
N
CLKOUT
(MHz)
3dB
n
R * (k
)
C * (nF)
7200
CLKG2
512
1.8432
1.0
2262
911
11.6
187
9600
CLKG2
512
2.4576
1.0
3016
1215
15.5
106
14400
CLKG1
256
1.8432
1.0
4524
1822
11.6
94
19200
CLKG1
256
2.4576
1.0
6032
2430
15.5
52
* The values for R and C are as calculated using the described method. Component tolerances have
not been allowed for. Notice that Ko and Kd can vary over a wide range, so using tight tolerances
for R and C is not justified. Use the nearest conveniently available value.
Table 2 Example PLL Loop Filter R and C values
Figure 7. CS5317 Dynamic Performance
Signal
Amplitude
Relative to
Full Scale
dc
Input Frequency
0dB
-20dB
-40dB
-60dB
-80dB
-100dB
-120dB
9.6 kHz
1 kHz
Sampling Rate: 19.2 kHz
Full Scale:
S/(N+D): 81.39 dB
+ 2.75 V
_
CS5317
DS27F4
15
Full - scale signal - to - noise - plus - distortion
[S/(N+D)] is calculated as the ratio of the rms
power of the fundamental to the sum of the rms
power of the FFT's other frequency bins, which
include both noise and distortion. For the
CS5317, signal-to-noise-plus-distortion is shown
to be better than 81 dB for an input frequency
range of 0 to 9.6 kHz (fs/2).
Harmonic distortion characteristics of the CS5317
are excellent at 80 dB full scale signal to THD
(typical), as are intermodulation distortion char-
acteristics, shown in Figure 8. Intermodulation
distortion results from the modulation distortion
of two or more input frequencies by a non-linear
transfer function.
DNL Test
Figure 9 shows a plot of the typical differential
non-linearity (DNL) of the CS5317. This test is
done by taking a large number of conversion re-
sults, and counting the occurrences of each code.
A perfect A/D converter would have all codes of
equal size and therefore equal numbers of occur-
rences. In the DNL test, a code with the average
number of occurrences is considered ideal and
plotted as DNL = 0 LSB. A code with more or
less occurrences than average will appear as a
DNL of greater than or less than zero. A missing
code has zero occurrences, and will appear as a
DNL of -1 LSB.
The plot below illustrates the typical DNL per-
formance of the CS5317, and clearly shows the
part easily achieves no missing codes.
Signal
Amplitude
Relative to
Full Scale
dc
Input Frequency (Hz)
S/I.M.D.: 84.7 dB
0dB
-20dB
-40dB
-60dB
-80dB
-100dB
-120dB
800
1450
9600
Figure 9. CS5317 DNL Plot
Figure 8. CS5317 Intermodulation Distortion
0
65,535
Codes
32,768
DNL (LSB)
+1
0
-1
+1/2
-1/2
Schematic & Layout Review Service
Confirm Optimum
Schematic & Layout
Before Building Your Board.
Confirm Optimum
Schematic & Layout
Before Building Your Board.
For Our Free Review Service
Call Applications Engineering.
For Our Free Review Service
Call Applications Engineering.
C a l l : ( 5 1 2 ) 4 4 5 - 7 2 2 2
CS5317
16
DS27F4
PIN DESCRIPTIONS
(Pin numbers refer to the 18-pin DIP package)
18 pin DIP Pinout
20 pin SOIC pinout
Power Supplies
VD+ - Positive Digital Power, PIN 2.
Positive digital supply voltage. Nominally 5 volts.
VD- - Negative Digital Power, PIN 10.
Negative digital supply voltage. Nominally -5 volts.
DGND - Digital Ground, PIN 4.
Digital ground reference.
VA+ - Positive Analog Power, PIN 1.
Positive analog supply voltage. Nominally 5 volts.
VA- - Negative Analog Power, PIN 14.
Negative analog supply voltage. Nominally -5 volts.
AGND - Analog Ground, PIN 15.
Analog ground reference.
PLL/Clock Generator
CLKIN - Clock Input, PIN 9.
Clock input for both clock generation modes and the clock override mode (see MODE).
18
17
1
2
3
4
5
6
7
8
11
10
9
16
15
14
13
12
POSITIVE ANALOG POWER
VA+
VCOIN
VCO INPUT
POSITIVE DIGITAL POWER
VD+
PHDT
PHASE DETECT
DATA OUTPUT ENABLE
DOE
RST
RESET
DIGITAL GROUND
DGND
AGND
ANALOG GROUND
SERIAL CLOCK OUTPUT CLKOUT
VA-
NEGATIVE ANALOG POWER
SERIAL DATA OUTPUT
DATA
NC
NO CONNECT
CLOCKING MODE SELECT
MODE
REFBUF
POSITIVE REFERENCE BUFFER
DATA OUTPUT READY
DOUT
AIN
ANALOG INPUT
CLOCK INPUT
CLKIN
VD-
NEGATIVE DIGITAL POWER
1
2
3
4
5
6
7
8
10
9
18
17
11
16
15
14
13
12
19
20
POSITIVE ANALOG POWER
VA+
VCOIN
VCO INPUT
POSITIVE DIGITAL POWER
VD+
PHDT
PHASE DETECT
DATA OUTPUT ENABLE
DOE
RST
RESET
DIGITAL GROUND
DGND
AGND
ANALOG GROUND
NO CONNECT
NC
NC
NO CONNECT
SERIAL CLOCK OUTPUT CLKOUT
NC
NO CONNECT
SERIAL DATA OUTPUT
DATA
VA-
NEGATIVE ANALOG POWER
CLOCKING MODE SELECT
MODE
REFBUF
POSITIVE REFERENCE BUFFER
DATA OUTPUT READY
DOUT
AIN
ANALOG INPUT
CLOCK INPUT
CLKIN
VD-
NEGATIVE DIGITAL POWER
CS5317
DS27F4
17
MODE - Mode Set, PIN 7.
Determines the internal clocking mode utilized by the CS5317. Connect to +5V to select
CLKG1 mode. Connect to DGND to select CLKG2 mode. Connect to -5V to select CLKOR
mode. This pin becomes equivalent to FSYNC in the CSZ5316 compatible mode.
VCOIN - VCO Input, PIN 18.
This pin is typically connected to PHDT. A capacitor and resistor in series connected between
VA+ and this pin sets the filter response of the on-chip phase locked loop.
PHDT - Phase Detect, PIN 17.
This pin is typically connected to VCOIN. A capacitor and resistor in series connected between
VA+ and this pin sets the filter response of the on-chip phase locked loop.
Inputs
AIN - Analog Input, PIN 11.
DOE - Data Output Enable, PIN 3.
Three-state control for serial output interface. When low, DATA, DOUT, and CLKOUT are
active. When high, they are in a high impedance state.
RST - Sample Clock Reset, PIN 16.
Sets phase of CLKOUT. Functions only in the clock override mode, CLKOR. Used to
synchronize the output samples of multiple CS5317's. Must be kept high in CLKG1 or CLKG2
modes. Also, tying this pin low, with MODE not tied to - 5V, will place the CS5317 into
CSZ5316 compatible mode.
Outputs
DOUT - Data Output Flag, PIN 8.
The falling edge indicates the start of serial data output on the DATA pin. The rising edge
indicates the end of serial data output.
DATA - Data Output, PIN 6.
Serial data output pin. Converted data is clocked out on this pin by the rising edge of
CLKOUT. Data is sent MSB first in two's complement format.
CLKOUT - Data Output Clock, PIN 5.
Serial data output clock. Data is clocked out on the rising edge of this pin. The falling edge
should be used to latch data. Since CLKOUT is a free running clock, DOUT can be used to
indicate valid data.
REFBUF - Positive Voltage Reference Noise Buffer, PIN 12.
Used to attenuate noise on the internal positive voltage reference. Must be connected to the
analog ground through a 0.1
F ceramic capacitor.
CS5317
18
DS27F4
PARAMETER DEFINITIONS
Resolution - The number of different output codes possible. Expressed as N, where 2
N
is the number
of available output codes.
Dynamic Range - The ratio of the largest allowable input signal to the noise floor.
Total Harmonic Distortion - The ratio of the rms sum of all harmonics to the rms value of the largest
allowable input signal. Units in dB's.
Signal to Intermodulation Distortion - The ratio of the rms sum of two input signals to the rms sum
of all discernible intermodulation and harmonic distortion products.
Linearity Error - The deviation of a code from a straight line passing through the endpoints of the
transfer function after zero- and full-scale errors have been accounted for. "Zero-scale" is a
point 1/2 LSB below the first code transition and "full-scale" is a point 1/2 LSB beyond the
code transition to all ones. The deviation is measured from the middle of each particular code.
Units in %FS.
Differential Nonlinearity - The deviation of a code's width from the ideal width. Units in LSB's.
Positive Full Scale Error - The deviation of the last code transition from the ideal, (VREF - 3/2 LSB).
Units in mV.
Positive Full Scale Drift - The drift in effective, positive, full-scale input voltage with temperature.
Negative Full Scale Error - The deviation of the first code transition from the ideal, (-VREF + 1/2
LSB). Units in mV.
Negative Full Scale Drift - The drift in effective, negative, full-scale input voltage with temperature.
Bipolar Offset - The deviation of the mid-scale transition from the ideal. The ideal is defined as the
middle transition lying on a straight line between actual positive full-scale and actual negative
full-scale.
Bipolar Offset Drift - The drift in the bipolar offset error with temperature.
Absolute Group Delay - The delay through the filter section of the part.
Passband Frequency - The upper -3 dB frequency of the CS5317.
CS5317
DS27F4
19
ORDERING GUIDE
Model Number
Temperature Range
Package
CS5317-KP
0 to 70
C
18 Pin Plastic DIP
CS5317-KS
0 to 70
C
20 Pin Plastic SOIC
CS5317
20
DS27F4
APPENDIX A
APPLICATIONS
Figure A1 shows one method of converting the
serial output of the CS5317 into 16-bit, parallel
words. The associated timing is also shown.
S2
A
S1
P
A
P
B
P
C
P
D
P
E
P
F
P
G
P
H
D0
D1
D2
D3
D4
D5
D6
D7
7
4
HCT2
9
9
Q
H
P
A
P
B
P
C
P
D
P
E
P
F
P
G
P
H
D8
D9
D10
D11
D12
D13
D14
D15
74
H
C
T
2
9
9
CS5317
DATA
DOUT
CLKOUT
CS
8
8
16
Data
Bus
+5V
OE2
+5V
OE2
S2
A
S1
OE1
OE1
INT
INT
Only needed for level sensitive interrupt driven systems.
D
Q
Q
SET
7
4
HCT7
4
15
13
14
0
2
1
INT Cleared when data read
(CS goes low)
(MSB)
DATA
INT
CLKOUT
DOUT
Figure A1. CS5317-to-Parallel Data Bus Interface
CS5317
DS27F4
21
Figure A2 shows the interconnection and timing
details for connecting a CS5317 to a NEC
PD7730 DSP chip.
Figure A3 shows the interconnection and timing
details for connecting a CS5317 to a Motorola
DSP 56000.
Status Register (SR)
Meaning
External Clock
16 bit data
MSB First
Setting
0
1, 0
0
Bit
9
7, 6
3
Mnemonic
SCI
SDLI
SIF
PD77230
SICK
SI
SIEN
CS5317
DATA
CLKOUT
DOUT
15
13
14
DATA
CLKOUT
DOUT
0
2
1
(MSB)
Figure A2. CS5317-to-NEC
PD77230 Serial Interface
CS5317
DATA
CLKOUT
DOUT
DSP56000
SCK, SC0
SRD
SC2, SC1
SSI Control Reg. A
CRA (X:FFEC)
WL1 = 1
WL0 = 0
16 bits
ASYNC
SC0
SC1
-
0
-
0
0
0
0
SYNC
SCK
SC2
0
1
-
0
0
0
-
CLKOUT
DOUT
GCK
SYN
FSL
SCKD
SCD2
SCD1
SCD0
PINS
SSI Contr
o
l Reg. B
CRB (X:FFED)
15
13
14
DATA
CLKOUT
DOUT
0
2
1
(MSB)
Figure A3. CS5317-to-Motorola DSP56000 Serial Interface
CS5317
22
DS27F4
Figure A4 shows the interconnection and timing
details for connecting a CS5317 to a WE DSP16
DSP chip.
Figure A5 shows the interconnection and timing
details for connecting a CS5317 with TMS32020
and TMS320C25 DSP chips.
CS5317
DATA
CLKOUT
DOUT
DSP16
ICK
DI
ILD
Q
D
74
DATA
d
Serial I/O Control Register (SIOC)
Meaning
MSB input first
ILD is an input
ICK is an input
16 bit input data
Value
1
0
0
0
Field
MSB
ILD
ICK
ILEN
CLKOUT
DOUT
DATA d
15
13
14
2
1
0
(MSB)
Figure A4. CS5317-to-WE DSP16 Serial Interface
TMS32020
TMS320C25
CLKR
DR
FSR
CS5317
DATA
CLKOUT
DOUT
TMS32020 Status Register (ST1):
F0 = 0 (16 bit data)
TMS320C25 Status Register (ST1):
FSM = 1 (Frame Sync used)
F0 = 0 (16 bit data)
15
13
14
DATA
CLKOUT
DOUT
0
2
1
(MSB)
Figure A5. CS5317-to-TMS32020/TMS320C25 Serial Interface
CS5317
DS27F4
23
Notes
25
Copyright
Cirrus Logic, Inc. 1998
(All Rights Reserved)
Cirrus Logic, Inc.
Crystal Semiconductor Products Division
P.O. Box 17847, Austin, Texas 78760
(512) 445 7222 FAX: (512) 445 7581
http://www.crystal.com
CDB5317
Evaluation Board for CS5317
Features
l
Easy to Use Digital Interface
Parallel 16 Bits With Clock
Serial Output With Clock
l
Multiple Operating Modes
Including Two PLL Modes
l
IDC Header used to access Parallel Data,
Serial Data, and Clock Input and Output
Description
The CDB5317 Evaluation Board is designed to allow the
user to quickly evaluate performance of the CS5317 Del-
ta-Sigma Analog-to-Digital Converter. All that is required
to use this board is an external power supply, a signal
source, a clock source, and an ability to read either serial
or parallel 16bitdata words.
ORDERING INFORMATION
CDB5317
Evaluation Board
I
CLKIN
CS5317
CONVERTER
SERIAL TO PARALLEL
CLKOUT
DATA
D0-D15
IDC HEADER
AIN
DACK
+5V
-5V
GND
CLKIN
DRDY
V
L
MAR `95
DS27DB3
GENERAL DESCRIPTION
The CDB5317 Evaluation Board is a stand-alone
environment for easy lab evaluation of the
CS5317 Delta-Sigma Analog-to-Digital Converter.
Included on the board is a serial-to-parallel con-
verter. The user can access output data in either
parallel or serial form. When supplied with the
necessary +5 V and -5 V power supplies, a
CLKIN signal, and an analog signal source, the
CDB5317 will provide converted data at the 40
pin header.
SUGGESTED EVALUATION METHOD
An efficient evaluation of the CS5317 using the
CDB5317 may be accomplished as described be-
low.
The following equipment will be required for the
evaluation:
The CDB5317 Evaluation Board.
A power supply capable of supplying +5V and
-5V.
A clock source as the CLKIN signal of the
CS5317.
A spectrally pure sine wave generator such as
the Krohn-Hite Model 4400A "Ultra-Low Distor-
tion Oscillator".
A PC equipped with a digital data acquisition
board such as the Metrabyte Model PIO12 "24 Bit
Parallel Digital I/O Interface".
A software routine to collect the data and per-
form a Fast Fourier Transform (FFT).
The evaluation board includes filter components
for the on-chip phase locked loop. The compo-
nents are adequate for testing if the CLKIN signal
has little or no phase-jitter. If the CDB5317 board
is being tested as part of a system which generates
a CLKIN which contains jitter, the PLL filter
components may need to be optimized for your
system (see the CS5317 data sheet).
Set-up for evaluation is straightforward. First de-
cide the operating mode and place the jumper on
the board for the proper selection. Then decide
whether the filter components for the phase
locked loop are adequate or whether they should
be changed for your evaluation. The PLL will
lock on a steady clock input with the filter as it is.
Connect the necessary 5 V (CMOS compatible)
CLKIN signal for the application. Use the sine-
wave generator to supply the analog signal to the
CDB5317. Apply the analog input and CLKIN
signals only when the evaluation board is pow-
ered up. Converted data will then appear at the
header on the CDB5317. The header should be
connected to the digital data acquisition board in
the PC through an IDC 40 pin connector and ca-
ble. The software routine should collect the data
from the CDB5317 and run a standard 1024 point
Fast Fourier Transform (FFT). Such an analysis
results in a plot similar to Figure 1. This plot re-
sulted from using a 1kHz input signal and a
Blackman-Harris window for the FFT.
The signal to noise and signal to total harmonic
distortion characteristics of the CS5317 may be
easily measured in this way. The signal to total
harmonic distortion value for a particular input is
the ratio of the RMS value of the input signal and
the sum of the RMS values of the harmonics
shown in the diagram. The dynamic range of the
CS5317 can be measured by reducing the input
Figure 1. FFT Plot Example
Signal
Amplitude
Relative to
Full Scale
dc
Input Frequency
0dB
-20dB
-40dB
-60dB
-80dB
-100dB
-120dB
9.6 kHz
1 kHz
Sampling Rate: 19.2 kHz
Full Scale:
S/(N+D): 81.39 dB
2.75 V
CDB5317
26
DS27DB3
amplitude so that distortion products become neg-
ligible. This allows an accurate measurement of
the noise floor.
More complex analysis such as intermodulation
distortion measurements can be accomplished
with the addition of another sine-wave generator.
CIRCUIT DESCRIPTION
Figure 2 illustrates the CS5317 A/D converter IC
circuit connections. The chip operates off of
5V.
These voltages are supplied from a power source
external to the evaluation board. Binding posts
are supplied on the board to connect the +5, -5,
and ground power lines. A good quality low rip-
p le, low no ise su pp ly will giv e the best
performance. The +5 V supply can also be used
for VL and should be connected between the VL
board jack and the power supply, as opposed to
connecting the VL jack straight to the +5V jack.
The +5V jack is the positive power source for the
CS5317 IC whereas the VL jack supplies power
to all the digital ICs. Care should be taken that
noise is not coupled between VL and +5V; how-
ever, supply noise is generally not a problem with
the CS5317 since the on-chip decimation filter
will remove any interference outside of its pass-
band. The +5 and -5 V supply lines are filtered on
U6
9
11
CLKIN
GND
-5V
AIN
TP10
17
18
R11
15
12
R8
10
C13
C12
CLKG1
CLKG2
CLKOR
CLKIN (fig.6)
P2
10
F
TP9
C18
C16
10
F
D2
6.8v
TP7
+
TP8
R1*
R2
200
51
10k
CS5317
+5V
D3
C19
C17
R10
6.8v
0.22
F
R7
10
VA+
1
VD+
2
C7
0.1
F
C6
10
F
R5
10 k
CLKOUT
DATA
DOUT
(fig. 3)
(fig. 3)
(fig. 3)
R9
10 k
R6
10 k
7
16
4
3
5
6
8
0.1
F
C15
0.1
F
0.1
F
C14
TP6
D1
C4
C5
0.1
F
VL
VL
10
F
VA-
14
VD-
10
C8
0.1
F
C9
10
F
* Remove for logic gate CLKIN source
DGND
CLKOUT
DATA
MODE
RST
DOE
DOUT
AIN
CLKIN
AGND
REFBUF
PHDT
VCOIN
+
1.2k
0.01
F
NPO
Figure 2. Analog-to-Digital Converter
CDB5317
DS27DB3
27
the board and then connected to the V
A
+ and V
A
-
supply pins of the chip. The +5 V and -5V are
then connected by means of ten
resistors to the
V
D
+ and V
D
- pins respectively. Capacitive filter-
ing is provided on all supply pins of the chip. In
addition there is a 0.1
F filter capacitor con-
nected from the REFBUF pin of the chip to the
V
A
- supply pin.
To properly operate, the CS5317 chip requires an
external (5 V CMOS compatible) clock. A BNC
connector labeled CLKIN is provided to connect
the off-board clock signal to the board. The
CLKIN signal is also available on the 40 pin
header connector. The CLKIN signal is one input
to the phase detector of the on-chip phase locked
loop of the CS5317.
Header connector P2 (see Figure 2) is provided to
allow mode selection for the CS5317 chip. The
mode selection works together with the CLKIN
signal to set the sample rate and the output word
rate of the CS5317. See the CS5317 data sheet
for details on mode selection. Two of the avail-
able modes (CLKG1 and CLKG2) utilize the
on-chip phase locked loop to step up the CLKIN
frequency to obtain the necessary sample rate
clock for the A/D converter. Another mode (the
CLKOR mode) does not use the on-chip PLL but
instead drives the sample function directly. The
1
3
9
11
13
2
4
8
10
12
U5
UNUSED GATES
D
S
2
GND R
74HC74
CL
VL
14
4
C10
0.1
F
D
C11
VL
R4
10k
0.1
F
6
U5
5
5
6
Q
Q
14
C1
0.1
F
V
L
3
1
13
12
2
7
VL
(fig. 2)
DOUT
7
3
5
6
U1
4
CLKOUT
(fig. 2)
11
10
9
8
DRDY
DACK
(fig. 6)
(fig. 6)
(fig. 6)
DOUT2
(fig. 6)
CLKOUT
(fig. 6)
CLKOUT2
DATA
DATA 1
(fig. 6)
(fig. 6)
DATA
(fig. 2)
TP3
TP2
U4
14
7
D
8
10
VL
VL
Q
12
R
U4
CL
VL
11
VL
Q
9
S
13
Figure 3. Buffers and Parallel Handshake Flip-Flop
CDB5317
28
DS27DB3
two modes which use the phase locked loop will
require appropriate low pass filter components on
the Evaluation Board. The low pass filter compo-
nents help determine the PLL control loop
response, including its bandwidth and stability
and therefore directly affect the transient response
of the PLL control loop. Appropriate filter compo-
nents should be installed if a particular dynamic
response to changes of the CLKIN signal is de-
sired.
The filter components which are installed on the
board have been chosen for the following parame-
ters: MODE: CLKG2; CLKIN: 7,200; N=512;
damping factor: 1.0; Control loop -3 dB band-
width: 2262 radians/second. These parameters
yield R as 10 k
and C as 0.22
F for the filter
components.
The analog signal to be digitized is input to the
AIN BNC connector. The digital output words
from the CS5317 are buffered by HEX inverters
as shown in Figure 3. The buffered versions of
the CLKOUT and DATA signals are available on
the header connector P1 in Figure 6. The serial
data signals out of the CS5317 are illustrated in
Figure 4. If remote control of the DOE line is
desired, the trace on the PC Board can be opened
and a wire connection can be soldered to the DOE
input line. Remote control of the RST line of the
CS5317 is also available if desired.
Figures 5 and 6 illustrate the serial to parallel shift
registers including timing information. The DATA
output signal from the CS5317 is input to the data
input of the shift register. An inverted version of
the CLKOUT signal is used to clock the DATA
into the shift registers. The two 8-bit shift register
ICs also include output latches. The rising edge
Note: For a complete description of serial timing see the CS5317 Data Sheet
CLKOUT
DATA
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
DOUT
Figure 4 Serial Data Timing
DATA
Serial Data
Shifting Out
Parallel Data (D0-D15) Valid
Parallel Data Valid
DRDY
DACK
DOUT
Serial Data
Shifting Out
Figure 5. Parallel Data Timing
CDB5317
DS27DB3
29
of the DOUT signal from the CS5317 is used to
latch the data once it is input to the shift registers.
The rising edge of DOUT is also used to toggle
the DRDY flip flop (see Figure 3). The flip flop
is used to signal a remote device whenever new
data is latched into the output registers. The
DRDY flip flop is reset whenever DACK occurs.
A component layout of the CDB5317 board is il-
lustrated in Figure 7.
13
13
U2
D8
D9
D10
D11
D12
D13
D14
D15
DATA
CLKOUT
7
6
5
4
3
2
1
15
QH
QG
QF
QE
QD
QC
QB
QA
GND
12
11
14
LATCH CLK
SHIFT CLK
DIN
74HC595
RST
CLKIN
P1
DACK
D0
D1
D2
D3
D4
D5
D6
D7
DRDY
7
6
5
4
3
2
1
15
QH
QG
QF
QE
QD
QC
QB
QA
U3
GND
12
11
14
74HC595
8
9
TP4
TP5
10k
VL
10 16
0.1
F
C3
VL
10 16
0.1
F
C2
CLKOUT2
(fig. 3)
(fig. 2)
CLKIN
RST
LATCH CLK
SHIFT CLK
DIN
DOUT
DATA1
(fig. 3)
(fig. 2)
DATA
(fig. 3)
(fig. 3)
DOUT2
CLKOUT
DRDY
(fig. 3)
(fig. 3)
DACK
OE
8
R3
Figure 6.
CDB5317
30
DS27DB3
Figure 7. Bird's Eye View
CDB5317
DS27DB3
31