Fine delay adjust on 2 LVDS/CM OS outputs
4-wire or 3-wire serial control port
Space-saving 64-lead LFCSP
APPLICATIONS
Low jitter, low phase noise clock distribution
Clocking high speed ADCs, DACs, DDS, DDC, DUC, MxFEs
High performance wireless transceivers
High performance instrumentation
Broadband infrastructure
Dividers, Delay Adjust, Eight Outputs
AD9510
FUNCTIONAL BLOCK DIAGRAM
CPRSET
PLL
REF
CHARGE
PUMP
PLL
SETTINGS
LVPECL
LVPECL
LVPECL
LVPECL
LVDS/CMOS
LVDS/CMOS
∆
T
LVDS/CMOS
∆
T
LVDS/CMOS
CP
STATUS
CLK2
CLK2B
OUT0
OUT0B
OUT1
OUT1B
OUT2
OUT2B
OUT3
OUT3B
OUT4
OUT4B
OUT5
OUT5B
OUT6
OUT6B
OUT7
OUT7B
05046-001
REFIN
REFINB
FUNCTION
CLK1
CLK1B
SCLK
SDIO
SDO
CSB
GNDVSVCP
RSET
SYNCB,
RESETB
PDB
SERIAL
CONTROL
PORT
DISTRIBUTION
REF
R DIVIDER
N DIVIDER
PROGRAMMABLE
AD9510
FREQUENCY
DETECTOR
DIVIDERS AND
PHASE ADJUST
/1, /2, /3... /31, /32
/1, /2, /3... /31, /32
/1, /2, /3... /31, /32
/1, /2, /3... /31, /32
/1, /2, /3... /31, /32
/1, /2, /3... /31, /32
/1, /2, /3... /31, /32
/1, /2, /3... /31, /32
PHASE
Figure 1.
GENERAL DESCRIPTION
The AD9510 provides a multi-output clock distribution
function along with an on-chip PLL core. The design
emphasizes low jitter and phase noise in order to maximize data
converter performance. Other applications with demanding
phase noise and jitter requirements also benefit from this part.
The PLL section consists of a programmable reference divider
(R); a low noise phase frequency detector (PFD); a precision
charge pump (CP); and a programmable feedback divider (N).
By connecting an external VCXO or VCO to the CLK2/CLK2B
pins, frequencies up to 1.5 GHz may be synchronized to the
input reference.
There are eight independent clock outputs. Four outputs are
LVPECL, and four are selectable as either LVDS or CMOS
levels. The LVPECL and LVDS outputs operate to 800 MHz,
and the CMOS outputs operate to 250 MHz.
Rev. PrB
Information furnished by Analog Devices is believed to be accurate and reliable.
However, no responsibility is assumed by Analog Devices for its use, nor for any
infringements of patents or other rights of third parties that may result from its use.
Specifications subject to change without notice. No license is granted by implication
or otherwise under any patent or patent rights of Analog Devices. Trademarks and
registered trademarks are the property of their respective owners.
Each output has a programmable divider that may be bypassed
or set to divide by any integer up to 32. The phase of one clock
output relative to another clock output may be varied by means
of a divider phase select function that serves as a coarse timing
adjustment. Two of the LVDS/CMOS outputs also feature
programmable delay elements with full-scale ranges up to 10 ns
of delay. This fine tuning delay block has 5-bit resolution, giving
32 possible delays from which to choose for each full-scale
setting.
The AD9510 is ideally suited for data converter clocking
applications where maximum converter performance is
achieved by encode signals with subpicosecond jitter.
The AD9510 is available in a 64-lead LFCSP and may be
operated from a single 3.3 V supply. An external VCO that
requires an extended voltage range may be accommodated by
connecting the charge pump supply (VCP) to 5.5 V. The
temperature range is −40°C to +85°C.
Phase Frequency Detector Input Frequency 80 MHz Antibacklash pulse width 0Dh<1:0>= 00b.
Phase Frequency Detector Input Frequency --- MHz Antibacklash pulse width 0Dh<1:0>= 01b.
Phase Frequency Detector Input Frequency --- MHz Antibacklash pulse width 0Dh<1:0>= 10b.
Antibacklash Pulse Width 1.3 ns 0Dh<1:0>= 00b.
Antibacklash Pulse Width 2.9 ns 0Dh<1:0>= 01b.
Antibacklash Pulse Width 6.0 ns 0Dh<1:0>= 10b.
CHARGE PUMP (CP)
ICP Sink/Source Programmable.
High Value 5 mA
Low Value 625 µA
Absolute Accuracy 2.5 % VCP = VS/2.
CPR
Range 2.7/10 kΩ
SET
ICP Three-State Leakage 1 nA
Sink-and-Source Current Matching 2 % 0.5 V < CP < VCP − 0.5 V.
ICP vs. VCP 1.5 % 0.5 V < CP < VCP − 0.5 V.
ICP vs. Temperature 2 % CP = VS/2.
In-Band Noise of the Charge Pump/
Phase Frequency Detector (In-Band
Means Within the LBW of the PLL)
@ 50 kHz PFD Frequency −172 dBc/Hz
@ 2 MHz PFD Frequency −156 dBc/Hz
@ 10 MHz PFD Frequency −149 dBc/Hz
@ 50 MHz PFD Frequency −142 dBc/Hz
= 4.12 kΩ, CPR
SET
= 5.1 kΩ, unless otherwise noted.
SET
When dc-coupled, REFINB capacitively
bypassed to RF ground
CLK2 is electrically identical to CLK1, the
distribution only input (see Clock Inputs)
can be used as differential or single-ended
inputs.
Frequencies > 800 MHz require a minimum
divide-by-2 (see the Distribution Section)
When dc-coupled, CLK2B capacitively
bypassed to RF ground.
The synthesizer phase noise floor is
estimated by measuring the in-band
phase noise at the output of the VCO and
subtracting 20logN (where N is the
N divider value).
1
.
Rev. PrB | Page 4 of 52
Page 5
Preliminary Technical Data AD9510
Parameter Min Typ Max Unit Test Conditions/Comments
PLL Figure of Merit
−219 +
10 × log (f
PRESCALER
Prescaler Input Frequency
P = 2 DM (2/3) 500 MHz
P = 4 DM (4/5) 750 MHz
P = 8 DM (8/9) 1500 MHz
P = 16 DM (16/17) 1500 MHz
P = 32 DM (32/33) 1500 MHz
Prescaler Output Frequency 300 MHz
PLL DIGITAL LOCK DETECT WINDOW
Required to Lock (Coincidence of Edges) Selected by Register ODh.
Low Range 3.5 ns <5> = 1.
High Range 9.5 ns <5> = 0.
To Unlock After Lock (Hysteresis) Selected by Register ODh.
Low Range 7 ns <5> = 1.
High Range 15 ns <5> = 0.
REFIN to CLK2 Delay 500 ps
1
REFIN and REFINB self-bias points are offset slightly to avoid chatter on an open input condition.
2
Example: −219 + 10 × log(f
) + 20 × log(N) should give the values for the in-band noise at the VCO output.
PFD
dBc/Hz
)
PFD
Approximation of the PFD/CP phase noise
floor (in the flat region) inside the PLL loop
bandwidth. When running closed loop this
phase noise is gained up by 20 × log(N)
2
.
Signal available at STATUS pin
when selected by 08h<5:2>.
CLOCK INPUTS
Table 2.
Parameter Min Typ Max Unit Test Conditions/Comments
CLOCK INPUTS – CLK1, CLK2
Input Frequency 1.5 GHz
Input Sensitivity, Differential --- 200 mV
Input Common-Mode Voltage , VCM --- 1.6 --- V Self-biased; enables ac coupling
Input Single-Ended Sensitivity --- VCM ± 100 mV
CLK1 and CLK2 are electrically
identical; can be used as differential
or single-ended inputs
Frequencies > 800 MHz require a
minimum divide-by-2, see the
Distribution Section
When dc-coupled, B input capacitively
bypassed to RF ground
Rev. PrB | Page 5 of 52
Page 6
AD9510 Preliminary Technical Data
CLOCK OUTPUTS
Table 3.
Parameter Min Typ Max Unit Test Conditions/Comments
LVPECL CLOCK OUTPUTS Termination = 50 Ω to VS − 2 V
OUT0, OUT1, OUT2, OUT3; Differential Output level 3Ch (3Dh) (3Eh) (3Fh)<3:2>= 10b
Output Frequency 800 MHz
Output High Voltage (VOH) VS − 1.2 VS − 0.8 V @ dc
Output Low Voltage (VOL) VS − 1.8 VS − 1.6 V @ dc
Output Differential Voltage (VOD) --- 800 --- mV @ dc
Isolation LVPECL-to-LVPECL Output --- dB Typical worst case, desired out to one other out1
Isolation LVDS-to-LVPECL Output --- dB Typical worst case, desired out to one other out 1
Isolation CMOS-to-LVPECL Output --- dB Typical worst case, desired out to one other out1
Output Frequency 800 MHz
Differential Output Voltage (VOD) --- 350 --- mV
Delta VOD --- 5 --- mV
Output Offset Voltage (VOS) --- 1.25 --- V
Delta VOS --- 5 --- mV
Short-Circuit Current (ISA, ISB) --- 13 --- mA Output shorted to GND
Isolation LVDS to LVDS --- dB Typical worst case, desired out to one other out1
Isolation LVPECL to LVDS --- dB Typical worst case, desired out to one other out1
Isolation CMOS to LVDS --- dB Typical worst case, desired out to one other out1
CMOS CLOCK OUTPUTS B outputs are inverted; termination = open
OUT4, OUT5, OUT6, OUT7; Single Ended
Output Frequency 250 MHz 5 pF load
Output Voltage High (VOH) --- 2.7 V
Output Voltage Low (VOL) 0.4 --- V
Isolation CMOS to CMOS --- dB Typical worst case, desired out to one other out1
Isolation LVPECL to CMOS --- dB Typical worst case, desired out to one other out1
Isolation LVDS to CMOS --- dB Typical worst case, desired out to one other out1
1
Desired output is 100 MHz and 50 MHz on one other output; isolation is level of 50 MHz signal referred to the 100 MHz signal on the desired output. Results shown are
typical worst case of isolation from a single output of indicated type.
Output Level 40h (41h) (42h) (43h)<2:1>= 01b
3.5 mA termination current
TIMING CHARACTERISTICS
Table 4.
Parameter Min Typ Max Unit Test Conditions/Comments
LVPECL
Termination = 50 Ω to V
Output level 3Ch (3Dh) (3Eh) (3Fh)<3:2> = 10b
Output Rise Time, tRP 125 --- ps 20% to 80%
Output Fall Time, tFP 125 --- ps 80% to 20%
Defined as the worst-case difference between any two similar delay paths within a single device operating at the same voltage and temperature.
2
Defined as the absolute worst-case difference between any two delay paths on any two devices operating at the same voltage and temperature. Part-to-part skew is
the total skew difference; pin-to-pin skew + part-to-part skew.
3
Incremental delay. Does not include propagation delay.
Distribution Section only; does not include PLL or external VCO/VCXO.
1
Distribution Section only;
does not include PLL or external VCO/VCXO
Calculated from SNR of ADC method;
FC = 100 MHz with AIN = 170 MHz
Distribution Section only;
does not include PLL or external VCO/VCXO
Calculated from SNR of ADC method;
= 100 MHz with AIN = 170 MHz
F
C
Distribution Section only;
does not include PLL or external VCO/VCXO
Calculated from SNR of ADC method;
= 100 MHz with AIN = 170 MHz
F
C
PLL AND DISTRIBUTION PHASE NOISE AND SPURIOUS
Table 7.
Parameter Min Typ Max Unit Test Conditions/Comments
PHASE NOISE AND SPURIOUS
Setup No.1
245.76 MHz VCXO, F
= 1.2288 MHz; R = 25, N = 200
PFD
245.76 MHz Output Divide by 1
Phase Noise @100 kHz Offset --- dBc/Hz
Spurious --- dBc First and second harmonics of F
61.44 MHz Output Divide by 4
Phase Noise @100 kHz Offset --- dBc/Hz
Spurious --- dBc First and second harmonics of F
Setup No. 2
245.76 MHz VCXO, F
= 30.72 MHz; R = 1, N = 8
PFD
245.76 MHz Output Divide by 1
Phase Noise @100 kHz Offset --- dBc/Hz
Spurious --- dBc First and second harmonics of F
61.44 MHz Output Divide by 4
Phase Noise @100 kHz Offset --- dBc/Hz
Spurious --- dBc First and second harmonics of F
Depends on VCO/VCXO selection.
Characterization ongoing.
Measured at LVPECL clock outputs;
ABP = 6 ns; I
= 5 mA; Ref = 30.72 MHz
CP
Measured at LVPECL clock outputs;
ABP = 6 ns; I
= 5 mA; Ref = 30.72 MHz
CP
PFD
PFD
PFD
PFD
Rev. PrB | Page 11 of 52
Page 12
AD9510 Preliminary Technical Data
SERIAL CONTROL PORT
Table 8.
Parameter Min Typ Max Unit Test Conditions/Comments
CSB, SCLK (INPUTS)
Input Logic 1 Voltage --- V
Input Logic 0 Voltage --- V
Input Logic 1 Current --- µA
Input Logic 0 Current --- µA
Input Capacitance --- pF
SDIO (WHEN INPUT)
Input Logic 1 Voltage --- V
Input Logic 0 Voltage --- V
Input Logic 1 Current --- µA
Input Logic 0 Current --- µA
Input Capacitance --- pF
SDIO, SDO (OUTPUTS)
Output Logic 1 Voltage --- V
Output Logic 0 Voltage --- V
TIMING
Clock Rate (SCLK, 1/t
Pulse-Width High, t
Pulse-Width Low, t
SDIO and CSB to SCLK Setup, tDS --- ns
SCLK to SDIO Hold, tDH --- ns
SCLK to Valid SDIO and SDO, tDV --- ns
) 25 MHz
SCLK
16 24 ns
PWH
16 24 ns
PWL
CSB and SCLK have 30 kΩ
internal pull-down resistors
FUNCTION PIN
Table 9.
Parameter Min Typ Max Unit Test Conditions/Comments
INPUT CHARACTERISTICS
Logic 1 Voltage --- V
Logic 0 Voltage --- V
Logic 1 Current --- µA
Logic 0 Current --- µA
Capacitance --- pF
RESET TIMING
Pulse-Width Low --- ns
SYNC TIMING
Pulse-Width Low 1.5 Clock cycles
Setup Time --- ps
Hold Time --- ps
Sync single chip; CLK1 or CL2,
whichever is being used for distribution
Sync multichip; Write CLK1 or CLK2,
whichever is being used for distribution
Sync multichip; Write CLK1 or CLK2,
whichever is being used for distribution
Rev. PrB | Page 12 of 52
Page 13
Preliminary Technical Data AD9510
STATUS PIN
Table 10.
Parameter Min Typ Max Unit Test Conditions/Comments
OUTPUT CHARACTERISTICS
Output Voltage High (VOH) --- mV
Output Voltage Low (VOL) --- mV
MAXIMUM TOGGLE RATE 100 MHz
ANALOG LOCK DETECT
Capacitance 3 pF
POWER
Table 11.
Parameter Min Typ Max Unit Test Conditions/Comments
POWER-UP DEFAULT MODE POWER DISSIPATION 650 --- mW
MAXIMUM POWER DISSIPATION 1050 --- mW
Full Sleep Power-Down --- ---
Power-Down (PDB) --- ---
POWER DELTA
CLK1, CLK2 Power-Down --- --- mW
Divider, DIV 2 − 32 to Bypass --- --- mW
LVPECL Output Power-Down
Safe Power-Down (PD2) 56 --- mW
Total Power-Down (PD3) 58 --- mW PD3 mode; use only with no load resistors connected.
LVDS Output Power-Down 33 46 mW
CMOS Output Power-Down 24 38 mW
Delay Block Bypass --- --- mW
Delay Block Power-Down --- --- mW Versus delay block bypass.
PLL Section Power-Down 40 --- mW Versus PLL in loop with … (conditions)
Applies when PLL mux is set to any divider or counter output,
or PFD up/down pulse. Also applies in analog lock detect mode.
Usually debug mode only. Beware that spurs may couple
to output when this pin is toggling.
On-chip capacitance; used to calculate RC time
constant for analog lock detect readback. Use pull-up resistor.
Power-up default state; does not
include power dissipated in output
load resistors.
All functions enabled, all outputs on
and terminated, maximum clock rates,
and frequencies. Does not include
power dissipated in load resistors.
(Pick these conditions.)
Maximum sleep is entered by setting 0Ah<1:0> = 01b
and 58h<4>= 1b. This powers off the PLL BG and the
distribution BG references. Does not include power
dissipated in terminations.
Set FUNCTION pin for PDB operation by setting
58h<6:5> = 11b. Pull PDB low. Does not include
power dissipated in terminations.
PD2 mode (safe) power-down is required
when load resistors are connected. Delta
does not include dissipation in load resistors.
Versus delay block operation at 10 ns fs
with maximum delay; output clocking at 25 MHz.
Rev. PrB | Page 13 of 52
Page 14
AD9510 Preliminary Technical Data
C
TIMING DIAGRAMS
t
CLK1
LK1
t
PECL
t
LVDS
t
CMOS
Figure 2. CLK1/CLK1B to Clock Output Timing, DIV = 1 Mode
05046-002
Figure 4. LVPECL Fall Time
Figure 3. LVPECL Rise Time
Figure 5. LVDS Timing
Rev. PrB | Page 14 of 52
Page 15
Preliminary Technical Data AD9510
ABSOLUTE MAXIMUM RATINGS
Table 12.
With
Respect
Parameter or Pin
VS GND −0.3 +3.6 V
VCP GND −0.3 +6 V
VCP V
REFIN, REFINB V
RSET GND V
CPRSET GND V
CLK1, CLK1B, CLK2, CLK2B V
CLK1 CLK1B V
CLK2 CLK2B V
SCLK, SDIO, SDO, CSB GND V
Outputs 0, 1, 2, 3 V
Outputs 4, 5, 6, 7 V
FUNCTION V
STATUS V
Junction Temperature 150 °C
Storage Temperature −65 +150 °C
Lead Temperature (10 sec) 300 °C
to
−0.3 V
S
Min Max Unit
Stresses above 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
sections of this specification is not implied. Exposure to
absolute maximum ratings for extended periods may affect
device reliability.
THERMAL CHARACTERISTICS1
Thermal Resistance
64-Lead LFCSP
= 24°C/W
θ
JA
1
Thermal impedance measurements were taken on a 4-layer board in still air,
in accordance with EIA/JESD51-7.
ESD 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 this product 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. PrB | Page 15 of 52
Page 16
AD9510 Preliminary Technical Data
T
PIN CONFIGURATION AND FUNCTION DESCRIPTIONS
VS
CPRSE
GND
RSETVSVS
OUT0
OUT0BVSGND
OUT1
OUT1BVSVS
GND
646362616059585756555453525150
GND
49
REFIN
REFINB
GND
VS
VCP
CP
GND
GND
VS
CLK2
CLK2B
GND
VS
CLK1
CLK1B
FUNCTION
PIN 1
1
INDICATOR
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
171819202122232425262728293031
STATUS
SCLK
SDO
SDIO
AD9510
TOP VIEW
(Not to Scale)
VS
CSB
GND
OUT7B
VS
OUT7
GND
OUT3
OUT3B
VS
VS
32
GND
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
VS
OUT4
OUT4B
VS
VS
OUT5
OUT5B
VS
VS
OUT6
OUT6B
VS
VS
OUT2
OUT2B
VS
05046-003
Figure 6. 64-Lead LFCSP Pin Configuration
Note that the exposed paddle on this package is an electrical connection as well as a thermal enhancement. For the device to
function properly, the paddle must be attached to ground, GND.
11 CLK2B Complementary Clock Input Used in Conjunction with CLK2.
14 CLK1 Clock Input That Drives Distribution Section of the Chip.
15 CLK1B Complementary Clock Input Used in Conjunction with CLK1.
16 FUNCTION Multipurpose Input May Be Programmed as a Reset (RESETB), Sync (SYNCB), or Power-Down (PDB) Pin.
17 STATUS Output Used to Monitor PLL Status and Sync Status.
18 SCLK Serial Data Clock.
19 SDIO Serial Data I/O.
20 SDO Serial Data Output.
21 CSB Serial Port Chip Select.
24 OUT7B Complementary LVDS/Inverted CMOS Output.
25 OUT7 LVDS/CMOS Output.
28 OUT3B Complementary LVPECL Output.
29 OUT3 LVPECL Output.
34 OUT2B Complementary LVPECL Output.
35 OUT2 LVPECL Output.
38 OUT6B Complementary LVDS/Inverted CMOS Output. OUT6 includes a delay block.
39 OUT6 LVDS/CMOS Output. OUT6 includes a delay block.
42 OUT5B Complementary LVDS/Inverted CMOS Output. OUT5 includes a delay block.
43 OUT5 LVDS/CMOS Output. OUT5 includes a delay block.
46 OUT4B Complementary LVDS/Inverted CMOS Output.
47 OUT4 LVDS/CMOS Output.
53 OUT1B Complementary LVPECL Output.
54 OUT1 LVPECL Output.
57 OUT0B Complementary LVPECL Output.
58 OUT0 LVPECL Output.
61 RSET Current Set Resistor to Ground. Nominal value = 4.12 kΩ.
63 CPRSET Charge Pump Current Set Resistor to Ground. Nominal value = 5.1 kΩ.
GND Ground.
VS Power Supply (3.3 V).
Charge Pump Power Supply. It should be greater than or equal to VS. VCP may be set as high as 5.5 V for
VCOs requiring extended tuning range.
Clock Input Used to Connect External VCO/VCXO to Feedback Divider, N. CLK2 also drives the distribution
section of the chip and may be used as a generic clock input when PLL is not used.
Note that the exposed paddle on this package is an electrical connection as well as a thermal enhancement. For the device to
function properly, the paddle must be attached to ground, GND.
Rev. PrB | Page 17 of 52
Page 18
AD9510 Preliminary Technical Data
TERMINOLOGY
Phase Jitter and Phase Noise
An ideal sine wave can be thought of as having a continuous
and even progression of phase with time from 0 to 360 degrees
for each cycle. Actual signals, however, display a certain amount
of variation from ideal phase progression over time. This
phenomenon is called phase jitter. Although there are many
causes that can contribute to phase jitter, one major component
is due to random noise which is characterized statistically as
being Gaussian (normal) in distribution.
This phase jitter leads to a spreading out of the energy of the
sine wave in the frequency domain, producing a continuous
power spectrum. This power spectrum is usually reported as a
series of values whose units are dBc/Hz at a given offset in
frequency from the sine wave (carrier). The value is a ratio
(expressed in dB) of the power contained within a 1 Hz
bandwidth with respect to the power at the carrier frequency.
For each measurement, the offset from the carrier frequency is
also given.
It is also meaningful to integrate the total power contained
within some interval of offset frequencies (for example, 10 kHz
to 10 MHz). This is called the integrated phase noise over that
frequency offset interval and can be readily related to the time
jitter due to the phase noise within that offset frequency
interval.
Phase noise has a detrimental effect on the performance of
ADCs, DACs, and RF mixers. It lowers the achievable dynamic
range of the converters and mixers, although they are affected
in somewhat different ways.
Time Jitter
Phase noise is a frequency domain phenomenon. In the time
domain, the same effect is exhibited as time jitter. When
observing a sine wave, the time of successive zero crossings is
seen to vary. In the case of a square wave, the time jitter is seen
as a displacement of the edges from their ideal (regular) times
of occurrence. In both cases, the variations in timing from the
ideal are the time jitter. Since these variations are random in
nature, the time jitter is specified in units of seconds root mean
square (rms) or 1 sigma of the Gaussian distribution.
Time jitter that occurs on a sampling clock for a DAC or an
ADC decreases the SNR and dynamic range of the converter. A
sampling clock with the lowest possible jitter provides the
highest performance from a given converter.
Additive Phase Noise
It is the amount of phase noise that is attributable to the device
or subsystem being measured. The phase noise of any external
oscillators or clock sources has been subtracted. This makes it
possible to predict the degree to which the device impacts the
total system phase noise when used in conjunction with the
various oscillators and clock sources, each of which contribute
their own phase noise to the total. In many cases, the phase
noise of one element dominates the system phase noise.
Additive Time Jitter
It is the amount of time jitter that is attributable to the device or
subsystem being measured. The time jitter of any external
oscillators or clock sources has been subtracted. This makes it
possible to predict the degree to which the device will impact
the total system time jitter when used in conjunction with the
various oscillators and clock sources, each of which contribute
their own time jitter to the total. In many cases, the time jitter of
the external oscillators and clock sources dominates the system
time jitter.
Rev. PrB | Page 18 of 52
Page 19
Preliminary Technical Data AD9510
TYPICAL PERFORMANCE CHARACTERISTICS
–110
–100
FREQUENCY (dBc/Hz)
FREQUENCY (dBc/Hz)
–120
–130
–140
–150
–160
–170
–120
–130
–140
–150
–160
1010M100M1M100k10k1k100
FREQUENCY (Hz)
Figure 7. Phase Noise—LVPECL: 245.76 MHz
05046-004
FREQUENCY (dBc/Hz)
–110
–120
–130
–140
–150
–160
–170
1010M1M100k10k1k100
FREQUENCY (Hz)
Figure 10. Phase Noise−CMOS: 245.76 MHz
05046-007
FREQUENCY (dBc/Hz)
–170
1010M80M1M100k10k1k100
–100
–110
–120
–130
–140
–150
–160
–170
1010M1M100k10k1k100
FREQUENCY (Hz)
Figure 8. Phase Noise—LVPECL: 622MHz
FREQUENCY (Hz)
Figure 9. Phase Noise—CMOS: 61.44MHz
05046-005
05046-006
Figure 11.
Figure 12.
Rev. PrB | Page 19 of 52
Page 20
AD9510 Preliminary Technical Data
TYPICAL MODES OF OPERATION
PLL WITH EXTERNAL VCXO/VCO FOLLOWED BY
CLOCK DISTRIBUTION
This is the most common operational mode for the AD9510.
An external oscillator (shown as VCO/VCXO) is phase locked
to a reference input frequency applied to REFIN. The loop filter
is usually a passive design. A VCO or a VCXO may be used.
The CLK2 input is connected internally to the feedback divider,
N. The CLK2 input provides the feedback path for the PLL. If
the VCO/VCXO frequency exceeds maximum frequency of the
output(s) being used, an appropriate divide ratio must be set in
the corresponding divider(s) in the Distribution Section.
PLL
REF
CHARGE
PUMP
LVPECL
LVPECL
LVPECL
LVPECL
LVDS/CMOS
LVDS/CMOS
LVDS/CMOS
LVDS/CMOS
LOOP
FILTER
VCXO,
VCO
CLOCK
OUTPUTS
REFERENCE
INPUT
V
REF
AD9510
REFIN
FUNCTION
CLK1CLK2
SERIAL
PORT
DIVIDE
DIVIDE
DIVIDE
DIVIDE
DIVIDE
DIVIDE
DIVIDE
DIVIDE
R
PFD
N
STATUS
∆T
∆T
Figure 13. PLL and Clock Distribution Mode
05046-010
CLOCK
INPUT 1
V
REF
REFIN
FUNCTION
CLK1CLK2
SERIAL
PORT
DIVIDE
DIVIDE
DIVIDE
DIVIDE
DIVIDE
DIVIDE
DIVIDE
DIVIDE
AD9510
R
N
PFD
STATUS
∆
T
∆
T
Figure 14. Clock Distribution Mode
PLL
REF
CHARGE
PUMP
LVPECL
LVPECL
LVPECL
LVPECL
LVDS/CMOS
LVDS/CMOS
LVDS/CMOS
LVDS/CMOS
CLOCK
INPUT 2
CLOCK
OUTPUTS
05046-011
CLOCK DISTRIBUTION ONLY
In this mode, the PLL is not used. A customer can save power
by initiating a PLL power-down and by powering down any
unused clock channels.
In distribution mode, both CLK1 and CLK2 inputs are available
for distribution to outputs via a low jitter multiplexer (MUX).
Rev. PrB | Page 20 of 52
Page 21
Preliminary Technical Data AD9510
PLL WITH EXTERNAL VCO AND BAND-PASS
FILTER FOLLOWED BY CLOCK DISTRIBUTION
An external band-pass filter may be used to possibly improve
the phase noise and spurious characteristics of the PLL output.
This option is most appropriate when the desire is to optimize
cost by choosing a less expensive VCO combined with a
moderately priced filter. Note that the BPF is shown outside of
the VCO-to-N divider path, with the BP filter outputs routed to
CLK1.
REFERENCE
INPUT
PFD
PLL
REF
CHARGE
PUMP
LVPECL
LVPECL
LVPECL
LVPECL
LVDS/CMOS
LVDS/CMOS
LVDS/CMOS
LVDS/CMOS
V
REF
DIVIDE
DIVIDE
DIVIDE
DIVIDE
DIVIDE
DIVIDE
DIVIDE
DIVIDE
AD9510
R
N
STATUS
∆
T
∆
T
REFIN
FUNCTION
CLK1CLK2
SERIAL
PORT
Figure 15. AD9510 with VCO and BPF Filter
LOOP
FILTER
VCO
BPF
CLOCK
OUTPUTS
05046-012
Rev. PrB | Page 21 of 52
Page 22
AD9510 Preliminary Technical Data
GNDVSVCP
RSET
CPRSET
REFIN
REFINB
FUNCTION
CLK1
CLK1B
SCLK
SDIO
SDO
CSB
DISTRIBUTION
REF
R DIVIDER
SYNCB,
RESETB
PDB
SERIAL
CONTROL
PORT
N DIVIDER
PROGRAMMABLE
DIVIDERS AND
PHASE ADJUST
/1, /2, /3... /31, /32
/1, /2, /3... /31, /32
/1, /2, /3... /31, /32
/1, /2, /3... /31, /32
/1, /2, /3... /31, /32
/1, /2, /3... /31, /32
/1, /2, /3... /31, /32
/1, /2, /3... /31, /32
Figure 16. Functional Block Diagram
AD9510
PHASE
FREQUENCY
DETECTOR
∆
∆
T
T
PLL
REF
CHARGE
PUMP
PLL
SETTINGS
LVPECL
LVPECL
LVPECL
LVPECL
LVDS/CMOS
LVDS/CMOS
LVDS/CMOS
LVDS/CMOS
CP
STATUS
CLK2
CLK2B
OUT0
OUT0B
OUT1
OUT1B
OUT2
OUT2B
OUT3
OUT3B
OUT4
OUT4B
OUT5
OUT5B
OUT6
OUT6B
OUT7
OUT7B
05046-013
Rev. PrB | Page 22 of 52
Page 23
Preliminary Technical Data AD9510
FUNCTIONAL DESCRIPTION
V
OVERALL
Figure 16 shows a block diagram of the AD9510. The chip
combines a programmable PLL core with a configurable clock
distribution system. A complete PLL requires the addition of a
suitable external VCO (or VCXO) and loop filter. This PLL can
lock to a reference input signal and produce an output that is
related to the input frequency by the ratio defined by the
programmable R and N dividers. The PLL offers some jitter
clean up of the external reference signal, depending on the loop
bandwidth and the phase noise performance of the VCO
(VCXO).
The output from the VCO (VCXO) can be applied to the clock
distribution section of the chip, where it can be divided by any
integer value from 1 to 32. The duty cycle and relative phase of
the outputs can be selected. There are four LVPECL outputs,
(OUT0, OUT1, OUT2, and OUT3) and four outputs that can be
either LVDS or CMOS level outputs (OUT4, OUT5, OUT6, and
OUT7). Two of these outputs (OUT5 and OUT6) can also make
use of a variable delay block.
S
10kΩ12kΩ
REFIN
REFINB
10kΩ10kΩ
150Ω
150Ω
Figure 17. REFIN Equivalent Circuit
05046-033
VCO/VCXO CLOCK INPUT—CLK2
The CLK2 differential input is used to connect an external VCO
or VCXO to the PLL. Only the CLK2 input port has a
connection to the PLL N divider. This input can receive up to
1.5 GHz. These inputs are internally self-biased and must be accoupled via capacitors.
Alternatively, CLK2 may be used as an input to the Distribution
Section. This is accomplished by setting Register 45h<0> = 0.
The default condition is for CLK1 to feed the Distribution
Section.
Alternatively, the clock distribution section can be driven
directly by an external clock signal, and the PLL can be powered
off. Whenever the clock distribution section is used alone, there
is no clock clean-up. The jitter of the input clock signal is
passed along directly to the distribution section and may
dominate at the clock outputs.
PLL SECTION
The AD9510 is partitioned into two sections: PLL and
distribution. If desired, the PLL section can be used separately
from the Distribution Section.
The AD9510 has a complete PLL core on-chip, requiring only
an external loop filter and VCO/VCXO. This PLL is based on
the ADF4106, a PLL noted for its superb low phase noise
performance. The operation of the AD9510 PLL is nearly
identical to that of the ADF4106, offering an advantage to those
with experience with the ADF series of PLLs. Differences
include the addition of differential inputs at REFIN and CLK2,
a different control register architecture, and the prescaler has
been changed to allow N as low as 1. The AD9510 PLL also
implements the digital lock detect feature somewhat differently
than the ADF4106 does offering improved functionality at
higher PFD rates. See Register Map Description section.
PLL REFERENCE INPUT—REFIN
The REFIN and REFINB pins can drive differentially or singleended. These pins are internally self-biased; therefore, they
should always be ac-coupled via capacitors. This also applies to
the unused side when single-ended input is used. Figure 17
shows the equivalent circuit of REFIN.
See Figure 18 for the equivalent circuit of CLK1 and CLK2.
CLOCK INPUT
V
S
CLK
CLKB
2.5kΩ2.5kΩ
5kΩ
5kΩ
Figure 18 CLK1, CLK2 Equivalent Input Circuit
STAGE
05046-016
PLL REFERENCE DIVIDER—R
The REFIN/REFINB inputs are routed to reference divider, R,
which is a 14-bit counter. R may be programmed to any value
from 0 to 16383 via its control register (OBh<5:0>, OCh<7:0>).
The output of the R divider goes to one of the phase/frequency
detector inputs. The maximum allowable frequency into the
phase, frequency detector (PFD) must not be exceeded. This
means that the REFIN frequency divided by R must be less than
the maximum allowable PFD frequency. See Figure 17.
VCO/VCXO FEEDBACK DIVIDER—N (P, A, B)
The N divider is a combination of a prescaler, P, (3 bits) and two
counters, A (6 bits) and B (13 bits). Although the AD9510’s PLL
is similar to the ADF4106, the AD9510 has a redesigned
prescaler that allows for lower values of N. The prescaler has
both a dual modulus (DM) and a fixed divide (FD) mode. The
AD9510 prescaler modes are shown in Table 14.
Rev. PrB | Page 23 of 52
Page 24
AD9510 Preliminary Technical Data
Table 14. PLL Prescaler Modes
Mode
(FD = Fixed Divide
DM = Dual Modulus) Value in 0Ah<4:2> Divide By
When using the prescaler in FD mode, the A counter is not
used, and the B counter may need to be bypassed. The DM
prescaler modes set some upper limits on the frequency, which
can be applied to CLK2, see Table 15.
Table 15. Frequency Limits of Each Prescaler Mode
Mode (DM = Dual Modulus) CLK2
P = 2 DM (2/3) <500 MHz
P = 4 DM (4/5) <750 MHz
P = 8 DM (8/9) <1.5 GHz
P = 16 DM <1.5 GHz
P = 32 DM <1.5 GHz
A AND B COUNTERS
The AD9510 B counter has a bypass mode (B = 1) that is not
available on the ADF4106. The B counter bypass mode is only
valid when using the prescaler in FD mode. The B counter is
bypassed by writing 1 to the B counter bypass bit in the register
map. Note that the A counter is not used when the prescaler is
in FD mode.
Note also that the A/B counters have their own reset bit that is
primarily intended for testing. The A and B counters can also be
reset using the shared R, A, and B counters reset bit.
DETERMINING VALUES FOR P, A, B, AND R
When operating the AD9510 in a dual-modulus mode, the
input reference frequency, F
frequency, F
F
VCO
VCO.
= (F
/R) × (PB + A) = F
REF
When operating the prescaler in fixed divide mode, the A
counter is not used and the equation simplifies to
= (F
F
VCO
/R) × (PB) = F
REF
By using combinations of dual modulus and fixed divide
modes, the AD9510 can achieve values of N all the way down to
N = 1. Table 16 shows how a 10 MHz reference input may be
locked to any integer multiple of N. Note that the same value of
N may be derived in different ways, as illustrated by the case of
N = 12.
, is related to the VCO output
REF
× N/R
REF
× N/R
REF
Rev. PrB | Page 24 of 52
Page 25
Preliminary Technical Data AD9510
V
Table 16. P, A, B, R—Smallest Values for N
F
R P A B N F
REF
10 1 1 X 1 1 10 FD P = 1, B = 1 (Bypassed)
10 1 2 X 1 2 20 FD P = 2, B = 1 (Bypassed)
10 1 1 X 3 3 30 FD P = 1, B = 3
10 1 1 X 4 4 40 FD P = 1, B = 4
10 1 1 X 5 5 50 FD P = 1, B = 5
10 1 2 X 3 6 60 FD P = 2, B = 3
10 1 2 0 3 6 60 DM P/P + 1 = 2/3, A = 0, B = 3
10 1 2 1 3 7 70 DM P/P + 1 = 2/3, A = 1, B = 3
10 1 2 2 3 8 80 DM P/P + 1 = 2/3, A = 2, B = 3
10 1 2 1 4 9 90 DM P/P + 1 = 2/3, A = 1, B = 4
10 1 2 X 5 10 100 FD P = 2, B = 5
10 1 2 0 5 10 100 DM P/P + 1 = 2/3, A = 0, B = 5
10 1 2 1 5 11 110 DM P/P + 1 = 2/3, A = 1, B = 5
10 1 2 X 6 12 120 FD P = 2, B = 6
10 1 2 0 6 12 120 DM P/P + 1 = 2/3, A = 0, B = 6
10 1 4 0 3 12 120 DM P/P + 1 = 4/5, A = 0, B = 3
10 1 4 1 3 13 130 DM P/P + 1 = 4/5, A = 1, B = 3
Mode Notes
VCO
PHASE FREQUENCY DETECTOR (PFD) AND
CHARGE PUMP
The PFD takes inputs from the R counter and N counter
(N = BP + A) and produces an output proportional to the phase
and frequency difference between them. Figure 19 is a
simplified schematic. The PFD includes a programmable delay
element that controls the width of the antibacklash pulse. This
pulse ensures that there is no dead zone in the PFD transfer
function and minimizes phase noise and reference spurs. Two
bits in Register 0Dh <1:0> control the width of the pulse.
P
CHARGE
GND
PUMP
CP
05046-014
D1 Q1
U1
CLR1
PROGRAMMABLE
ANTIBACKLASH
CLR2
D2 Q2
U2
UP
DELAY
PULSE WIDTH
DOWN
U3
HI
R DIVIDER
HI
N DIVIDER
Figure 19. PFD Simplified Schematic and Timing (In Lock)
ANTIBACKLASH PULSE
The PLL features a programmable antibacklash pulse width that
is set by the value in Register 0Dh<1:0>. The default
antibacklash pulse width is 1.3 ns. The antibacklash pulse
eliminates the dead zone around the phase-locked condition
and thereby reduces the potential for certain spurs that could be
impressed on the VCO signal.
STATUS PIN
The output multiplexer on the AD9510 allows access to various
signals and internal points on the chip at the STATUS pin.
Figure 20 shows a block diagram of the STATUS pin section.
The function of the STATUS pin is controlled by Register
08h<5:2>.
PLL Digital Lock Detect
The STATUS pin can display two types of PLL lock detect:
digital (DLD) and analog (ALD). Whenever digital lock detect
is desired, the STATUS pin provides a CMOS level signal, which
can be active high or active low.
The digital lock detect has one of two time windows, as selected
by Register 0Dh<5>. The default (ODh<5> = 0) requires the
signal edges on the inputs to the PFD to be coincident within
9.5 ns in order to set the DLD true, which then must separate by
at least 15 ns in order to give DLD = false.
The other setting (ODh<5> = 1) makes these coincidence times
3.5 ns for DLD = true and 7 ns for DLD = false.
The DLD may be disabled by writing to Register 0Dh<6> = 1.
Rev. PrB | Page 25 of 52
Page 26
AD9510 Preliminary Technical Data
DIGITAL LOCK DETECT (ACTIVE HIGH)
DIGITAL LOCK DETECT (ACTIVE LOW)
ANALOG LOCK DETECT (N-CHANNEL OPEN DRAIN)
LOSS OF REFERENCE (ACTIVE HIGH)
ANALOG LOCK DETECT (P-CHANNEL OPEN DRAIN)
LOSS OF REFERENCE OR LOCK DETECT (ACTIVE HIGH)
LOSS OF REFERENCE OR LOCK DETECT (ACTIVE LOW)
LOSS OF REFERENCE (ACTIVE LOW)
OFF (LOW) (DEFAULT)
N DIVIDER OUTPUT
R DIVIDER OUTPUT
A COUNTER OUTPUT
PRESCALER OUTPUT (NCLK)
PFD UP PULSE
PFD DOWN PULSE
Figure 20. STATUS Pin Circuit CLK1 Clock Input
PLL Analog Lock Detect
An analog lock detect (ALD) signal may be selected. When
ALD is selected, the signal at the STATUS pin is either an opendrain, p-channel (08h<5:2> = 1100) or an open-drain, nchannel (08h<5:2> = 0101).
The analog lock detect signal is true (relative to the selected
mode) with brief false pulses. These false pulses get shorter as
the inputs to the PFD are nearer to coincidence and longer as
they are further from coincidence.
In order to extract a usable analog lock detect signal, an
external RC network is required in order to provide an analog
filter with the appropriate RC constant to allow for the
discrimination of a lock condition by an external voltage
comparator. A 1 kΩ resistor in parallel with a small capacitance
usually fulfills this requirement. However, some
experimentation may be required to get the desired operation.
The analog lock detect function may introduce some spurious
energy into the clock outputs. It is prudent to limit the use of
the ALD when the best possible jitter/phase noise performance
is required on the clock outputs.
LOSS OF REFERENCE
The AD9510 PLL can warn of a loss-of-reference signal at
REFIN. The loss-of-reference monitor internally sets a flag
called LREF. Externally, this signal can be observed in several
ways on the STATUS pin, depending on the PLL MUX control
settings in Register 08h<5:2>. The LREF alone can be observed
as an active high signal by setting 08h<5:2> = <1010> or as an
active low signal by setting 08h<5:2> = <1111>.
TRI-STATE
PLL MUX CONTROL
08h <5:2>
SYNC
DETECT
SYNC DETECT ENABLE
58h <0>
V
S
STATUS
PIN
LOCK DETECT MODE
CONTROL FOR ANALOG
GND
5046-015
The digital lock detect (DLD) block of the AD9510 requires a
PLL reference signal to be present in order for the digital lock
detect output to be valid. It is possible to have a digital lock
detect indication (DLD = true) that remains true even after a
loss-of-the reference signal. For this reason, the digital lock
detect signal alone cannot be relied upon if the reference has
been lost. There is a way to combine the DLD and the LREF
into a single signal at the STATUS pin. Set 08h<5:2> = <1101>
to get a signal that is the logical OR of the DLD and the LREF
active high. If an active low version of this signal is desired, set
08h<5:2> = <1110>.
The reference monitor is enabled only after the DLD signal has
been high for the number of PFD cycles set by the value in
07h<6:5>. This delay is measured in PFD cycles. The delay
ranges from 3 PFD cycles (default) to 24 PFD cycles. When the
reference goes away, LREF goes true and the charge pump goes
into tri-state.
User intervention is required to take the part out of this state.
First, 07h<2> = 0 must be written in order to disable the loss-ofreference circuit, taking the charge pump out of tri-state and
causing LREF to go false. A second write of 07h<2> = 1 is
required to re-enable the loss-of-reference circuit.
PLL LOOP LOCKS
DLD GOES TRUE
WRITE 07h<2> = 0
LREF SET FALSE
CHARGE PUMP COMES
OUT OF TRI-STATE
WRITE 07h<2> = 1
LOR ENABLED
LREF IS FALSE
n PFD CYCLES WITH
DLD TRUE
(n SET BY 07h<6:5>)
The loss-of-reference circuit is clocked by the signal from the
VCO, which means that there must be a VCO signal present in
order to detect a loss of reference.
Rev. PrB | Page 26 of 52
CHARGE PUMP
GOES INTO TRI-STATE.
LREF SET TRUE.
CHECK FOR PRESENCE
OR REFERENCE.
LREF STAYS FALSE IF
MISSING
REFERENCE
DETECTED
REFERENCE IS DETECTED.
Figure 21. Loss of Reference Sequence of Events
5046-034
Page 27
Preliminary Technical Data AD9510
FUNCTION PIN
The FUNCTION pin (16) has three functions that are selected
by the value in Register 58h<6:5>. There is an internal 30 kΩ
pull-down resistor on this pin.
RESETB: 58h<6:5> = 00b (Default)
In its default mode, the FUNCTION pin acts as RESETB, which
generates an asynchronous reset or hard reset when pulled low.
The resulting reset writes the default values into the serial
control port buffer registers as well as loading them into the
chip control registers. The AD9510 immediately resumes
operation according to the default values. When the pin is taken
high again, an asynchronous sync is issued (see the SYNCB:
58h<6:5> = 01b section).
SYNCB: 58h<6:5> = 01b
The FUNCTION pin may be used to cause a synchronization or
alignment of phase among the various clock outputs. The
synchronization applies only to clock outputs that:
CLK1 CLOCK INPUT
Either CLK1 or CLK2 may be selected as the input to the
distribution section. The CLK1 input can only be connected to
drive the distribution section. CLK1 is selected as the source for
the distribution section by setting Register 45h<0> = 1. This is
the power-up default state. CLK1 works for inputs up to
1500 MHz.
See Figure 18 for the CLK1 and CLK2 equivalent input circuit.
The CLK1 input is fully differential and self-biased. The signal
should be ac-coupled using capacitors. If a single-ended input
must be used, this can be accommodated by ac coupling to one
side of the differential input only. The other side of the input
should be bypassed to a quiet ac ground by a capacitor.
If only the distribution section of the AD9510 is used, the
unselected clock input should be powered down in order to
eliminate any possibility of unwanted crosstalk between the
selected clock input and the unselected clock input.
• are not powered down
• the divider is not masked (no sync = 0)
• are not bypassed (bypass = 0)
SYNCB is level and rising edge sensitive. When SYNCB is low,
the set of affected outputs are held in a predetermined state,
defined by each divider’s start high bit. On a rising edge, the
dividers begin after a predefined number of fast clock cycles
(fast clock is the selected clock input, CLK1 or CLK2) as
determined by the values in the divider’s phase offset bits.
The SYNCB application of the FUNCTION pin is always active,
regardless of whether the pin is also assigned to perform reset
or power-down. When the SYNCB function is selected, the
FUNCTION pin does not act as either RESETB or PDB.
PDB: 58h<6:5> = 11b
The FUNCTION pin may also be programmed to work as an
asynchronous full chip power-down, PDB. In PDB mode, the
FUNCTION pin is active low. The chip remains in a powerdown state until PDB is returned to logic high. The chip returns
to the settings programmed prior to the power-down.
See the Chip Power-Down or Sleep Mode—PDB section for
more details on what occurs during a PDB initiated powerdown.
DISTRIBUTION SECTION
As previously mentioned, the AD9510 is partitioned into two
operational sections: PLL and distribution. The PLL Section
was discussed previously. If desired, the distribution section can
be used separately from the PLL section.
DIVIDERS
Each of the eight clock outputs of the AD9510 has its own
divider. The divider may be bypassed in order to get an output
at the same frequency as the input (1×). When a divider is
bypassed, it is also powered down to save power.
All integer divide ratios from 2 to 32 may be selected.
Each divider may be configured for divide ratio, phase, and
duty cycle. The phase and duty cycle values that can be selected
depend on the divide ratio that is chosen.
Setting the Divide Ratio
The divide ratio is determined by the values written via the SCP
to the registers that control each individual output, OUT0 to
OUT7. These are the even numbered registers beginning at 48h
and going through 56h. Each of these registers are divided into
bits that control the number of clock cycles that the divider
output stays high (high_cycles <3:0>) and the number of clock
cycles that the divider output stays low (low_cycles <7:4>). Each
value is 4 bits and has the range of 0 to 15.
The divide ratio is set by
Divide Ratio = (high_cycles + 1) + (low_cycles + 1)
Example 1:
Set the Divide Ratio = 2
high_cycles = 0
low_cycles = 0
Divide Ratio = (0 + 1) + (0 + 1) = 2
Rev. PrB | Page 27 of 52
Page 28
AD9510 Preliminary Technical Data
Example 2:
Set Divide Ratio = 8
high_cycles = 3
low_cycles = 3
Divide Ratio = (3 + 1) + (3 + 1) = 8
Note that a Divide Ratio of 8 may also be obtained by setting:
Although the second set of settings produce the same divide
ratio, the resulting duty cycle is not the same. See the Setting the
Duty Cycle section for an explanation of how the duty cycle and
divide ratio are related.
Setting the Duty Cycle
Different divide ratios have different duty cycle options. For
example, if Divide Ratio = 2, the only duty cycle possible is 50%.
If the Divide Ratio = 4, the duty cycle may be 25%, 50%, or 75%.
13 69 3 8
13 31 8 3
13 77 2 9
13 23 9 2
13 85 1 A
13 15 A 1
13 92 0 B
13 8 B 0
14 50 6 6
14 57 5 7
14 43 7 5
14 64 4 8
14 36 8 4
14 71 3 9
14 29 9 3
14 79 2 A
14 21 A 2
14 86 1 B
14 14 B 1
14 93 0 C
14 7 C 0
15 53 6 7
15 47 7 6
15 60 5 8
15 40 8 5
15 67 4 9
15 33 9 4
15 73 3 A
15 27 A 3
15 80 2 B
15 20 B 2
15 87 1 C
15 13 C 1
15 93 0 D
15 7 D 0
16 50 7 7
16 56 6 8
16 44 8 6
16 63 5 9
16 38 9 5
16 69 4 A
16 31 A 4
16 75 3 B
16 25 B 3
16 81 2 C
16 19 C 2
16 88 1 D
16 13 D 1
16 94 0 E
16 6 E 0
17 53 7 8
LO <7:4> HI<3:0>
Divide Ratio Duty Cycle (%)
17 47 8 7
17 59 6 9
17 41 9 6
17 65 5 A
17 35 A 5
17 71 4 B
17 29 B 4
17 76 3 C
17 24 C 3
17 82 2 D
17 18 D 2
17 88 1 E
17 12 E 1
17 94 0 F
17 6 F 0
18 50 8 8
18 56 7 9
18 44 9 7
18 61 6 A
18 39 A 6
18 67 5 B
18 33 B 5
18 72 4 C
18 28 C 4
18 78 3 D
18 22 D 3
18 83 2 E
18 17 E 2
18 89 1 F
18 11 F 1
19 53 8 9
19 47 9 8
19 58 7 A
19 42 A 7
19 63 6 B
19 37 B 6
19 68 5 C
19 32 C 5
19 74 4 D
19 26 D 4
19 79 3 E
19 21 E 3
19 84 2 F
19 16 F 2
20 50 9 9
20 55 8 A
20 45 A 8
20 60 7 B
20 40 B 7
20 65 6 C
20 35 C 6
48h to 56h
LO <7:4> HI<3:0>
Rev. PrB | Page 29 of 52
Page 30
AD9510 Preliminary Technical Data
48h to 56h
Divide Ratio Duty Cycle (%)
20 70 5 D
20 30 D 5
20 75 4 E
20 25 E 4
20 80 3 F
20 20 F 3
21 52 9 A
21 48 A 9
21 57 8 B
21 43 B 8
21 62 7 C
21 38 C 7
21 67 6 D
21 33 D 6
21 71 5 E
21 29 E 5
21 76 4 F
21 24 F 4
22 50 A A
22 55 9 B
22 45 B 9
22 59 8 C
22 41 C 8
22 64 7 D
22 36 D 7
22 68 6 E
22 32 E 6
22 73 5 F
22 27 F 5
23 52 A B
23 48 B A
23 57 9 C
23 43 C 9
23 61 8 D
23 39 D 8
23 65 7 E
23 35 E 7
23 70 6 F
23 30 F 6
24 50 B B
24 54 A C
24 46 C A
24 58 9 D
LO <7:4> HI<3:0>
Divide Ratio Duty Cycle (%)
24 42 D 9
24 63 8 E
24 38 E 8
24 67 7 F
24 33 F 7
25 52 B C
25 48 C B
25 56 A D
25 44 D A
25 60 9 E
25 40 E 9
25 64 8 F
25 36 F 8
26 50 C C
26 54 B D
26 46 D B
26 58 A E
26 42 E A
26 62 9 F
26 38 F 9
27 52 C D
27 48 D C
27 56 B E
27 44 E B
27 59 A F
27 41 F A
28 50 D D
28 54 C E
28 46 E C
28 57 B F
28 43 F B
29 52 D E
29 48 E D
29 55 C F
29 45 F C
30 50 E E
30 53 D F
30 47 F D
31 52 E F
31 48 F E
32 50 F F
48h to 56h
LO <7:4> HI<3:0>
Rev. PrB | Page 30 of 52
Page 31
Preliminary Technical Data AD9510
Divider Phase Offset
The phase of each output may also be selected, depending on
the divide ratio chosen. This is selected by writing the
appropriate values to the registers which set the phase and start
high/low bit for each output. These are the odd numbered
registers from 49h to 57h. Each divider has a 4-bit phase offset
<3:0> and a start high or low bit <4>.
Following a sync pulse, the phase offset word determines how
many fast clock (CLK1 or CLK2) cycles to wait before initiating
a clock output edge. The Start H/L bit determines if the divider
output starts low or high. By giving each divider a different
phase offset, output-to-output delays can be set in increments of
the fast clock period, t
CLK
.
Figure 22 shows four dividers, each set for DIV = 4, 50% duty
cycle. By incrementing the phase offset from 0 to 3, each output
is offset from the initial edge by a multiple of t
CLOCK INPUT
U
O
I
V
I
D
E
R
D
D
Y
U
D
T
I
V
=
4
,
A
T
S
P
H
A
A
S
T
P
H
A
A
S
T
P
H
A
A
T
S
P
H
A
Figure 22. Phase Offset—All Dividers Set for DIV = 4, Phase Set from 0 to 3
0123456789101112131415
CLK
T
S
T
P
U
5
0
%
=
0
,
=
R
T
=
0
E
S
0
,
R
T
=
=
1
E
S
=
0
,
R
T
E
=
2
S
0
,
=
R
T
E
=
3
S
t
CLK
t
CLK
2
× t
3
× t
CLK
CLK
CLK
.
For example:
CLK1 = 491.52 MHz,
= 1/491.52 = 2.0345 ns
t
CLK1
For DIV = 4
Phase Offset 0 = 0 ns
Phase Offset 1 = 2.0345 ns
Phase Offset 2 = 4.069 ns
Phase Offset 3 = 6.104 ns
The four outputs may also be described as:
OUT1 = 0°
OUT2 = 90°
OUT3 = 180°
OUT4 = 270°
05046-035
Setting the phase offset to Phase = 4 results in the same relative
phase as the first channel, Phase = 0° or 360°.
In general, by combining the 4-bit phase offset and the Start
H/L bit, there are 32 possible phase offset states (see Table 18).
Phase offsets may be related to degrees by calculating the phase
step for a particular divide ratio:
Phase Step = 360°/(Divide Ratio) = 360°/DIV
Using some of the same examples:
DIV = 4
Phase Step = 360°/4 = 90°
Unique Phase Offsets in Degrees Are Phase = 0°, 90°,
180°, 270°
DIV = 7
Phase Step = 360°/7 = 51.43°
Unique Phase Offsets in Degrees Are Phase = 0°, 51.43°,
102.86°, 154.29°, 205.71°, 257.15°, 308.57°
DELAY BLOCK
OUT5 and OUT6 (LVDS/CMOS) include an analog delay
element that can be programmed (Register 34h to Register 3Ah)
to give variable time delays (Δt) in the clock signal passing
through that output, with respect to the other outputs that are
not delayed.
CLOCK INPUT
÷
N
∅SELECT
OUT5
ONLY
OUT6
FULL-SCALE: 1ns TO 10ns
Figure 23. Analog Delay (OUT5 andOUT6)
∆
T
FINE DELAY ADJUST
(32 STEPS)
The amount of delay that can be used is determined by the
frequency of the clock being delayed. The amount of delay can
approach one-half cycle of the clock period. For example, for a
10 MHz clock, the delay can extend to the full 10 ns maximum
of which the delay element is capable. However, for a 100 MHz
clock, the maximum delay is less than 5 ns (or half of the
period).
MUX
LVDS
CMOS
OUTPUT
DRIVER
05046-036
This path adds some jitter greater than that specified for the
nondelay outputs. This means that the delay function should be
used primarily for clocking digital chips, such as FPGA, ASIC,
DUC, and DDC, rather than data converters. The jitter is higher
for long full scales (~10 ns). This is because the delay block uses
a ramp and trip points to create the variable delay. A longer
ramp means more noise has a chance of being introduced.
Calculating the Delay
The following values and equations are used to calculate the
delay of the delay block.
Value of Ramp Current Control Bits (Register 35h or Register 39h
<2:0>) = Iramp_bits
= 200 µA × (Iramp_bits + 1)
I
RAMP
No. of Caps = No. of 0s + 1 in Ramp Control Capacitor (Register
35h or Register 39h <5:3>) that is, 101 = 1 + 1 = 2; 110 = 2;
The AD9510 offers three different output level choices:
LVPECL, LVDS, and CMOS. OU T 0 t o O U T 3 are LVPECL only.
OUT4 to OUT7 may be selected as either LVDS or CMOS. Each
output may be enabled or turned off as needed, to save power.
The simplified equivalent circuit of the LVPECL outputs is
shown in Figure 24.
3.3V
OUT
OUTB
OUT5 and OUT6 allow for a full-scale delay in the range 1 ns to
10 ns. The full-scale delay is selected by choosing a combination
of ramp current and the number of capacitors by writing the
appropriate values into Register 35h and Register 39h. There are
32 fine delay settings for each full scale, set by Register 36h and
Register 3Ah.
The PDB chip power-down turns off most of the functions and
currents in the AD9510. When the PDB mode is enabled, a chip
power-down is activated by taking the FUNCTION pin to a
logic low level. The chip remains in this power-down state until
PDB is brought back to logic high. When woken up, the
AD9510 returns to the settings programmed into its registers
prior to the power-down.
The PDB power-down mode shuts down the currents on the
chip, except the bias current necessary to maintain the LVPECL
outputs in a safe shutdown mode. This is needed to protect the
LVPECL output circuitry from damage that could be caused by
certain termination and load configurations. Because this is not
a complete power-down, it can be called sleep mode.
When the AD9510 is in a PDB power-down or sleep mode, the
chip is in the following state:
• The PLL is off (asynchronous power-down).
• All clocks and sync circuits are off.
• All dividers are off.
• All LVDS/CMOS outputs are off.
• All LVPECL outputs are in safe off mode.
• The serial control port is active, and the chip responds to
commands.
PLL Power-Down
The PLL section of the AD9510 may be selectively powered
down. There are three PLL power-down modes, set by the
values in Register 0Ah<1:0>, as shown in Table 19.
Table 19. Register 0Ah: PLL Power-Down
<1> <0> Mode
0 0 Normal Operation
0 1 Asynchronous Power-Down
1 0 Normal Operation
1 1 Synchronous Power-Down
In synchronous power-down mode, the PLL power-down is
gated by the charge pump to prevent unwanted frequency
jumps. The device goes into power-down on the occurrence of
the next charge pump event after the registers are updated.
Distribution Power-Down
The distribution section can be powered down by writing to
Register 58h<3> = 1. This turns off the bias to the distribution
section. The power-down mode also shuts off the protection
circuitry that keeps the LVPECL outputs in safe mode. This
mode should be avoided if any of the LVPECL outputs are
terminated in ways that could cause reverse biasing of the
output transistors, resulting in damage to these devices.
When combined with the PLL power-down (above), this mode
results in the lowest possible power-down current for the AD9510.
Individual Clock Output Power-Down
Any of the eight clock distribution outputs may be powered
down individually by writing to the appropriate registers via the
SCP. The register map details the individual power-down
settings for each output. The LVDS/CMOS outputs may be
powered down regardless of their output load configuration.
However, the LVPECL outputs have multiple power-down
modes, which gives the user flexibility in dealing with either
loaded/terminated or unloaded/unterminated outputs.
Individual Circuit Block Power Downs
Many of the AD9510 circuit blocks (CLK1, CLK2, REFIN, and
so on) may be powered down individually. This gives flexibility
in configuring the part for power savings when all chip
functionality is not needed.
RESET MODES
The AD9510 has several ways to force the chip into a reset
condition.
Power-On Reset—Start-Up Conditions when VS Is
Applied
A power-on reset (POR) is issued when the VS power supply is
turned on. This initializes the chip to the power-on conditions
that are determined by the default register settings. These are
indicated in the default value column of Table 23).
Asynchronous Reset via the FUNCTION Pin
As mentioned in the Function Pin section, a hard reset,
RESETB: 58h<6:5> = 00b (Default), restores the chip to the
default settings.
Soft Reset via the Serial Port
The serial control port allows for a soft reset by writing to
Register 00h<5> = 1. When <5> is set to 1, the chip executes a
soft reset. This restores the default values to the internal
registers. Once this is completed, <5> is cleared automatically.
Rev. PrB | Page 33 of 52
Page 34
AD9510 Preliminary Technical Data
SINGLE-CHIP SYNCHRONIZATION
The AD9510 clocks can be synchronized to each other at any
time. The outputs of the clocks are forced into a known state
with respect to each other and then allowed to continue
clocking from that state in synchronicity. Before a
synchronization is done, the Function Pin should be set as the
SYNCB: 58h<6:5> = 01b input. Synchronization is done by
forcing the FUNCTION pin low, creating a SYNCB signal and
then releasing it.
See the SYNCB: 58h<6:5> = 01b section for a more detailed
description of what happens when the SYNCB: 58h<6:5> = 01b
signal is issued.
MULTICHIP SYNCHRONIZATION
The AD9510 provides a means of synchronizing two or more
AD9510s. This is not an active synchronization; it requires user
monitoring and action. The arrangement of two AD9510s to be
synchronized is shown in Figure 26.
Synchronization of two or more AD9510s requires a fast clock
and a slow clock. The fast clock can be up to 1 GHz and may be
the clock driving the master AD9510 CLK1 input or one of the
outputs of the master. The fast clock acts as the input to the
distribution section of the slave AD9510 and is connected to its
CLK1 input. The PLL may be used on the master, but the slave
PLL is not used.
The slow clock is the clock that is synchronized across the two
chips. This clock must be no faster than one-fourth of the fast
clock, and no greater than 250 MHz. The slow clock is taken
from one of the outputs of the master AD9510 and acts as the
REFIN (or CLK2) input to the slave AD9510. One of the
outputs of the slave must provide this same frequency back to
the CLK2 (or REFIN) input of the slave.
Multichip synchronization is enabled by writing Register
58h<0> = 1 on the slave AD9510. When this bit is set, the
STATUS pin becomes the output for the SYNC signal. A low
signal indicates an in-sync condition, and a high indicates an
out-of-sync condition.
Register 58h<1> selects the number of fast clock cycles that are
the maximum separation of the slow clock edges that are
considered synchronized. When 58h<1> = 0 (default), the slow
clock edges must be coincident within 1 to 1.5 high speed clock
cycles. If the coincidence of the slow clock edges is closer than
this amount, the SYNC flag stays low. If the coincidence of the
slow clock edges is greater than this amount, the SYNC flag is
set high. When Register 58h<1> = 1, the amount of coincidence
required is 0.5 fast clock cycles to 1 fast clock cycles.
Whenever the SYNC flag is set (high), indicating an out-of-sync
condition, a SYNCB signal applied simultaneously at the
FUNCTION pins of both AD9510s brings the slow clocks into
synchronization.
AD9510
SYNCB
MASTER
FUNCTION
(SYNCB)
CLK2REFIN
AD9510
SLAVE
FAST CLOCK
<1GHz
CLK1
FUNCTION
(SYNCB)
Figure 26. Multichip Synchronization
FAST CLOCK
SLOW CLOCK
<250MHz
SYNC
DETECT
<1GHz
<250MHz
SLOW
CLOCK
OUTN
OUTM
F
SYNC
OUTY
STATUS
(SYNC)
F
SYNC
05046-039
Rev. PrB | Page 34 of 52
Page 35
Preliminary Technical Data AD9510
S
SERIAL CONTROL PORT
The AD9510 serial control port is a flexible, synchronous, serial
communications port that allows an easy interface with many
industry-standard microcontrollers and microprocessors. The
AD9510 serial control port is compatible with most
synchronous transfer formats, including both the Motorola SPI®
and Intel® SSR protocols. The serial control port allows
read/write access to all registers that configure the AD9510.
Single or multiple byte transfers are supported, as well as MSB
first or LSB first transfer formats. The AD9510 serial control
port can be configured for single pin I/O (SDIO only) or two
unidirectional pins for in/out (SDIO/SDO).
SERIAL CONTROL PORT PIN DESCRIPTIONS
SCLK (serial clock) is the serial shift clock. This pin is an input.
SCLK is used to synchronize serial control port reads and
writes. Write data bits are registered on the rising edge of this
clock, and read data bits are registered on the falling edge. This
pin is internally pulled down by a 30 kΩ resistor to ground.
SDIO (serial data input/output) is a dual-purpose pin and acts
as either an input only in 4-wire mode or as an input/output in
3-wire mode. The AD9510 defaults to 3-wire mode (single pin
I/O—SDIO only). Four-wire mode (two unidirectional pins for
I/O—SDIO/SDO) may be enabled by setting 1 into the SDO
enable register at 00h<7>.
SDO (serial data out) is used only in the 4-wire mode as a
separate output pin for readback data. The AD9510 defaults to
3-wire mode. Four-wire mode may be enabled by setting 1 into
the SDO enable register at 00h<7>.
CSB (chip select bar) is an active low control that gates the read
and write cycles. When CSB is high, SDO and SDIO are in a
high impedance state. This pin is internally pulled down by a
30 kΩ resistor to ground.
SCLK (PIN 18)
SDIO (PIN 19)
SDO (PIN 20)
CSB (PIN 21)
Figure 27. Serial Control Port
AD9510
SERIAL
CONTROL
PORT
05046-017
Write
If the instruction word (Phase 1) is for a write operation
(I15 = 0), then Phase 2 is the transfer of data into the serial
control port buffer of the AD9510. The length of the transfer (1,
2, 3, or 4 data bytes) is indicated by 2 bits (W1:W0) in the
instruction byte. Multibyte data transfer is the preferred
method. Single-byte data transfers are useful to reduce CPU
overhead when only one byte of data needs to be loaded. CSB
can be raised after each sequence of 8 bits (except the last byte)
to stall the bus. The serial transfer resumes when CSB is
lowered. Stalling on nonbyte boundaries resets the serial control
port.
Since data is written into a serial control port buffer area, not
directly into the AD9510’s actual control registers, a Phase 3
operation is needed in order to transfer the serial control port
buffer contents to the actual control registers of the AD9510,
thereby causing them to take effect. Phase 3 consists of writing a
high bit (one) to Address 5Ah, Bit <0>. This update bit is selfclearing (it is not required to write 0 to it in order to clear it).
Since any number of bytes of data may be changed before
issuing an update, the update simultaneously enables all register
changes since any previous update.
Read
If the instruction word (Phase 1) is for a read operation
(I15 = 1), the next N × 8 SCLK cycles clock out the data from
the address specified in the instruction word, where N is 1 to 4
as determined by W1:W0. The readback data is valid on the
falling edge of SCLK.
The default mode of the AD9510 serial control port is 3-wire
mode; therefore, the requested data normally appears on the
SDIO pin. It is possible to set the AD9510 to 4-wire mode by
setting 1 into the SDO enable register at 00h<7>. In 4-wire
mode, the readback data appears on the SDO pin.
A readback request reads the data that is in the serial control
port buffer area not the active data in the AD9510’s actual
control registers.
GENERAL OPERATION OF SERIAL CONTROL PORT
There are three phases to a communication cycle with the
AD9510. Phase 1 is the instruction cycle, which is the writing of
a 16-bit instruction word into the AD9510, coincident with the
first 16 SCLK rising edges. The instruction word provides the
AD9510 serial control port with information regarding the data
transfer cycle (Phase 2) of the communication cycle. The
Phase 1 instruction word defines whether the upcoming data
transfer is read or write, the number of bytes in the data
transfer, and the starting register address for the first byte of the
data transfer.
Rev. PrB | Page 35 of 52
CLK
SDIO
SDO
CSB
SERIAL
CONTROL
PORT
Figure 28. Relationship Between Serial Control Port Register Buffers and
Control Registers of the AD9510
UPDATE
REGISTERS
5Ah <0>
REGISTER BUFFERS
CONTROL REGISTERS
AD9510
CORE
05046-018
Page 36
AD9510 Preliminary Technical Data
The AD9510 uses Addresses 00h to 5Ah. Although the AD9510
serial control port allows for both 8-bit and 16-bit instructions,
the 8-bit instruction mode provides access to five address bits
(A4 to A0) only, which restricts its use to the address space 00h
to 01F. The AD9510 defaults to 16-bit instruction mode on
power-up. The 8-bit instruction mode (although defined for
this serial control port) is not useful for the AD9510; therefore,
it is not discussed in this data sheet.
THE INSTRUCTION WORD (16 BITS)
The MSB of the instruction word is R/W , which indicates
whether the instruction will be a read or a write. The next two
bits, W1:W0, indicate the length of the transfer in bytes. The
final 13 bits are the address (A12:A0) at which to begin the read
or write operation. For a write, the instruction word is followed
by the number of bytes of data indicated by Bits W1:W0, which
is interpreted according to Table 20.
Table 20. Byte Transfer Count
W1 W0 Bytes to Transfer
0 0 1
0 1 2
1 0 3
1 1 4
A12:A0: These 13 bits select the address within the register map
that is written to or read from during the data transfer portion
of the communications cycle. For multibyte transfers, this
address is the starting byte address. In MSB first mode,
subsequent bytes increment the address.
MSB/LSB FIRST TRANSFERS
The AD9510 instruction word and byte data may be MSB first
or LSB first. The default for the AD9510 is MSB first. The LSB
first mode may be set by writing 1 to Address 00h, Bit <6>. This
takes effect immediately (since it only affects the operation of
the serial control port) and does not require that an update be
executed. Immediately after the LSB first bit is set, all serial
control port operations are changed to LSB first order.
When MSB first mode is active, the instruction and data bytes
must be written from MSB to LSB. Multibyte data transfers in
MSB first format start with an instruction byte that includes the
register address of the most significant data byte. Subsequent
data bytes must follow in order from high address to low
address. In MSB first mode, the serial control port internal byte
address generator decrements for each data byte of the
multibyte transfer cycle.
When LSB_First = 1 (LSB first), the instruction and data bytes
must be written from LSB to MSB. Multibyte data transfers in
LSB first format start with an instruction byte that includes the
register address of the least significant data byte followed by
multiple data bytes. The serial control port internal byte address
generator increments for each byte of the multibyte transfer
cycle.
The AD9510 serial control port data address decrements from
the data address written toward 0x00 for multibyte I/O
operations if the MSB first mode is active. The serial control
port address increments from the data address written toward
0x1F for multibyte I/O operations if the LSB first mode is active.
Table 21. Serial Control Port, 16-Bit Instruction Word, MSB First
16-BIT INSTRUCTION HEADERREGISTER (N) DATAREGISTER (N + 1) DATA
Figure 33. Serial Control Port Write—LSB First, 16-Bit Instruction, 2 Bytes Data
Rev. PrB | Page 37 of 52
DON'T CARE
DON'T CARE
05046-023
Page 38
AD9510 Preliminary Technical Data
CSB
SCLK
t
S
t
CLK
t
HI
t
DS
t
DH
t
LO
t
H
SDIO
BI NBI N + 1
Figure 34 Serial Control Port Timing—Write
Table 22. Serial Control Port Timing
Parameter Meaning
tDS Setup time between data and rising edge of SCLK
tDH Hold time between data and rising edge of SCLK
t
Period of the clock
CLK
tS Setup time between CSB and SCLK
tH Hold time between CSB and SCLK
tHI Minimum period that SCLK should be in a logic high state
tLO Minimum period that SCLK should be in a logic low state
05046-040
Rev. PrB | Page 38 of 52
Page 39
Preliminary Technical Data AD9510
REGISTER MAP AND DESCRIPTION
SUMMARY TABLE
Table 23. AD9510 Register Map
Def.
Addr
(Hex)
00
01 Not Used
02 Not Used
03 Not Used
04 A Counter Not Used 6-Bit A Counter <5:0> 00
05 B Counter Not Used 13-Bit B Counter Bits 12:8 (MSB) <4:0> 00
06 B Counter 13-Bit B Counter Bits 7:0 (LSB) <7:0> 00
07 PLL 1
08 PLL 2
09 PLL 3
0A PLL 4
0B R Divider Not Used 14-Bit R Divider Bits 13:8 (MSB) <5:0> 00 R Divider
0C R Divider 14-Bit R Divider Bits 13:8 (MSB) <7:0> 00 R Divider
0D PLL 5
OE33
34 Delay Bypass 5 Not Used Bypass 01
35
36
37 Not Used 04
38 Delay Bypass 6 Not Used Bypass 01
39
3A
3B Not Used 04
Parameter
Name
Serial
Control Port
Configuration
PLL
Not Used
FINE DELAY
ADJUST
Delay FullScale 5
Delay Fine
Adjust 5
Delay FullScale 6
Delay Fine
Adjust 6
Bit 7
(MSB) Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1
SDO
Active
Not
Used
Not
Used
Not
Used
Not
Used B Bypass
Not
Used
LSB
First
LOR Lock_Del
PFD
Polarity
Digital
Lock
Det.
Enable
Not Used Ramp Capacitor <5:3> Ramp Current <2:0> 00
Not Used 5-Bit Fine Delay <5:1> Not Used 00
Not Used Ramp Capacitor <5:3> Ramp Current <2:0> 00
Not Used 5-Bit Fine Delay <5:1> Not Used 00
Soft
Reset
<6:5>
CP Current <6:4> Not Used
Not
Used
Digital
Lock
Det.
Window
Long_Ins Long_Ins
Not Used
PLL Mux Select <5:2> CP Mode <1:0> 00
Prescaler P <4:2> Power-Down <1:0> 01
Not Used
Soft
Reset
LOR
Enable
Reset R
Counter
LSB
First
Reset N
Counter
Antibacklash Pulse-
Bit 0
(LSB)
SDO
Active
Not Used 00
Reset All
Counters
Width <1:0>
Value
(Hex) Notes
10
00
00
<7:4>
Mirror
<3:0>
PLL Starts
in PowerDown
N Divider
(A)
N Divider
(B)
N Divider
(B)
N Divider
(P)
Fine
Delays
Bypassed
Bypass
Delay
Max. Delay
Full-Scale
Min. Delay
Value
Bypass
Delay
Max. Delay
Full-Scale
Min. Delay
Value
Rev. PrB | Page 39 of 52
Page 40
AD9510 Preliminary Technical Data
Def.
Addr
(Hex)
3C LVPECL OUT0 Not Used Output Level <3:2> Power-Down <1:0> 0A OFF
3D LVPECL OUT1 Not Used Output Level <3:2> Power-Down <1:0> 08 ON
3E LVPECL OUT2 Not Used Output Level <3:2> Power-Down <1:0> 08 ON
3F LVPECL OUT3 Not Used Output Level <3:2> Power-Down <1:0> 08 ON
40
41
42
43
44 Not Used
45
46, 47 Not Used
48 Divider 0 Low Cycles <7:4> High Cycles <3:0> 00 Divide by 2
49 Divider 0 Bypass No Sync Force Start H/L Phase Offset <3:0> 00 Phase = 0
4A Divider 1 Low Cycles <7:4> High Cycles <3:0> 00 Divide by 2
4B Divider 1 Bypass No Sync Force Start H/L Phase Offset <3:0> 00 Phase = 0
4C Divider 2 Low Cycles <7:4> High Cycles <3:0> 11 Divide by 4
4D Divider 2 Bypass No Sync Force Start H/L Phase Offset <3:0> 00 Phase = 0
4E Divider 3 Low Cycles <7:4> High Cycles <3:0> 33 Divide by 8
4F Divider 3 Bypass No Sync Force Start H/L Phase Offset <3:0> 00 Phase = 0
50 Divider 4 Low Cycles <7:4> High Cycles <3:0> 00 Divide by 2
51 Divider 4 Bypass No Sync Force Start H/L Phase Offset <3:0> 00 Phase = 0
52 Divider 5 Low Cycles <7:4> High Cycles <3:0> 11 Divide by 4
53 Divider 5 Bypass No Sync Force Start H/L Phase Offset <3:0> 00 Phase = 0
54 Divider 6 Low Cycles <7:4> High Cycles <3:0> 00 Divide by 2
55 Divider 6 Bypass No Sync Force Start H/L Phase Offset <3:0> 00 Phase = 0
56 Divider 7 Low Cycles <7:4> High Cycles <3:0> 00 Divide by 2
57 Divider 7 Bypass No Sync Force Start H/L Phase Offset <3:0> 00 Phase = 0
58
59 Not Used
5A
END
Parameter
Name
OUTPUTS
LVDS_CMOS
OUT 4
LVDS_CMOS
OUT 5
LVDS_CMOS
OUT 6
LVDS_CMOS
OUT 7
CLK1 AND
CLK2
Clocks Select,
Power-Down
(PD) Options
DIVIDERS
FUNCTION
FUNCTION Pin
and Sync
Update
Registers
Bit 7
(MSB) Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1
Not Used
Not
Used
Not Used
Not Used
Not Used
Not Used
CLKs in
PD
Set FUNCTION Pin PD Sync
CMOS
Inverted
Driver On
CMOS
Inverted
Driver On
CMOS
Inverted
Driver On
CMOS
Inverted
Driver On
REFIN PD
Not Used
Logic
Select
Logic
Select
Logic
Select
Logic
Select
CLK to
PLL PD
PD All
Ref.
Output Level <2:1>
Output Level <2:1>
Output Level <2:1>
Output Level <2:1>
CLK2
PD
Sync
Reg.
CLK1
PD
Sync
Select
Bit 0
(LSB)
Output
Power
Output
Power
Output
Power
Output
Power
Select
CLK IN
Sync
Enable
Update
Registers
Value
(Hex) Notes
02 LVDS, ON
02 LVDS, ON
03 LVDS, OFF
03 LVDS, OFF
Input
Receivers
01
00
00
All Clocks
ON, Select
CLK1
FUNCTION
Pin =
RESETB
SelfClearing
Bit
Rev. PrB | Page 40 of 52
Page 41
Preliminary Technical Data AD9510
REGISTER MAP DESCRIPTION
Table 24 lists the AD9510 control registers by hexadecimal address. A specific bit or range of bits within a register is indicated by angle
brackets. For example, <3> refers to Bit 3, while <5:2> refers to the range of bits from Bit 5 through Bit 2. Table 24 describes the
functionality of the control registers on a bit-by-bit basis. For a more concise (but less descriptive) table, see Table 23.
Table 24. AD9510 Register Descriptions
Reg.
Addr.
(Hex) Bit(s) Name Description
00 <0> SDO Active
00 <1> LSB First
00 <2> Soft Reset
00 <3> Long Instruction
00 <4> Long Instruction Same as <3>
00 <5> Soft Reset Same as <2>
00 <6> LSB First Same as <1>
00 <7> SDO Active Same as <0>
Not Used
01 <7:0> Not Used
02 <7:0> Not Used
03 <7:0> Not Used
PLL Settings
04 <5:0> A Counter 6-Bit A Counter <5:0>
04 <7:6> Not Used
05 <4:0> B Counter MSBs 13-Bit B Counter (MSB) <12:8>
05 <7:5> Not Used
06 <7:0> B Counter LSBs 13-Bit B Counter (LSB) <7:0>
07 <1:0> Not Used
07 <2> LOR Enable 1 = Enables the Loss-of-Reference (LOR) Function; (Default = 0)
07 <4:3> Not Used
07 <6:5>
Note: <7:4> mirror <3:0> to ensure that this register can be accessed regardless of the state of
<1> or <6> (the bit that sets LSB first).
When set causes SDO to become active. When clear, the SDO pin remains in tri-state, and all
read data is routed to the SDIO pin. (Default = 0.)
When set causes input and output data to be oriented as LSB first. Additionally, addressing
increments. If this bit is clear, data is oriented as MSB first and addressing decrements (Default
= 0, MSB first).
When 1 is written to this bit, the chip executes a soft reset, restoring default values to the
internal registers.
This bit is self-clearing. A 0 does not have to be written to clear it.
When set, the instruction phase is 16 bits. When clear, the instruction phase is 8 bits.
The default, and only, mode for this part is long instruction (Default = 1).
LOR Initial Lock Detect Delay. Once a lock detect is indicated, this is the number of phase
frequency detector (PFD) cycles that occur prior to turning on the LOR monitor.
<6> <5> LOR Initial Lock Detect Delay
<1> <0> Charge Pump Mode
Rev. PrB | Page 41 of 52
Page 42
AD9510 Preliminary Technical Data
Reg.
Addr.
(Hex) Bit(s) Name Description
08 <5:2> PLL Mux Control
0 0 0 0 Off (Signal Goes Low) (Default)
0 0 0 1 Digital Lock Detect (Active High)
0 0 1 0 N Divider Output
0 0 1 1 Digital Lock Detect (Active Low)
0 1 0 0 R Divider Output
0 1 0 1 Analog Lock Detect (N Channel, Open-Drain)
0 1 1 0 A Counter Output
0 1 1 1 Prescaler Output (NCLK)
1 0 0 0 PFD Up Pulse
1 0 0 1 PFD Down Pulse
1 0 1 0 Loss-of-Reference (Active High)
1 0 1 1 Tri-State
1 1 0 0 Analog Lock Detect (P Channel, Open-Drain)
1 1 0 1 Loss-of-Reference or Lock Detect (Active High)
1 1 1 0 Loss-of-Reference or Lock Detect (Active Low)
1 1 1 1 Loss-of-Reference (Active Low)
MUXOUT is the PLL portion of the STATUS output MUX
08 <6>
Phase-Frequency
Detector (PFD) Polarity
08 <7> Not Used
09 <0> Reset All Counters 0 = Normal (Default), 1 = Reset R, A, and B Counters
09 <1> N-Counter Reset 0 = Normal (Default), 1 = Reset A and B Counters
09 <2> R-Counter Reset 0 = Normal (Default), 1 = Reset R Counter
09 <3> Not Used
09 <6:4>
Charge Pump (CP)
Current Setting
0 0 0 0.62
0 0 1 1.25
0 1 0 1.87
0 1 1 2.50
1 0 0 3.12
1 0 1 3.75
1 1 0 4.37
1 1 1 5.00
Default = 000
These currents assume: CPR
Actual current can be calculated by: CP_lsb = 3.1875/CPR
09 <7> Not Used
0A <1:0> PLL Power-Down 01 = Asynchronous Power-Down (Default)
0 0 Normal Operation
0 1 Asynchronous Power-Down
1 0 Normal Operation
1 1 Synchronous Power-Down
0D <7> Not Used
Unused
0E-33 Not Used
Fine Delay Adjust <0> Delay Control Delay Block Control Bit
34 OUT5 Bypasses Delay Block and Powers It Down (Default = 1)
(38) (OUT6)
34 <7:1> Not Used
(38)
<2:0> Ramp Current
35 OUT5 The slowest ramp (200 µs) sets the longest full scale of approximately 10 ns.
(39) (OUT6)
14-Bit Reference
Counter, MSBs
14-Bit Reference
Counter, R LSBs
Antibacklash PulseWidth
Digital Lock Detect
Window
Digital Lock Detect
Window
<4> <3> <2> Mode Prescaler Mode
Only valid when operating the prescaler in fixed divide (FD) mode. When this bit is set, the B
counter is divide by 1. This allows the prescaler setting to determine the divide for the N
divider.
R Divider (MSB) <13:8>
R Divider (MSB) <7:0>
<1> <0> Antibacklash Pulse Width (ns)
<5> Digital Lock
Detect Window (ns)
If the time difference of the rising edges at the inputs to the PFD are less than the lock detect
window time, the digital lock detect flag is set. The flag remains set until the time difference is
greater than the loss-of-lock threshold.
0 0 RESETB (Default)
0 1 SYNCB
1 0 Test Only; Do Not Use
1 1 PDB
58 <7> Not Used
59 <7:0> Not Used
5A <0> Update Registers
5A <7:1> Not Used
END
Soft SYNC bit works the same as the FUNCTION pin when in SYNCB mode, except that this bit’s
polarity is reversed. That is, a high level forces selected outputs into a known state, and a high
> low transition triggers a sync (default = 0).
<6> <5> Function
1 written to this bit updates all registers and transfers all serial control port register buffer
contents to the control registers on the next rising SCLK edge. This is a self-clearing bit. 0 does
not have to be written in order to clear it.
Rev. PrB | Page 47 of 52
Page 48
AD9510 Preliminary Technical Data
APPLICATIONS
USING THE AD9510 OUTPUTS FOR ADC CLOCK
APPLICATIONS
Any high speed analog-to-digital converter (ADC) is extremely
sensitive to the quality of the sampling clock provided by the
user. An ADC can be thought of as a sampling mixer; and any
noise, distortion, or timing jitter on the clock is combined with
the desired signal at the A/D output. Clock integrity
requirements scale with the analog input frequency and
resolution, with higher analog input frequency applications at
>= 14-bit resolution being the most stringent. The theoretical
SNR of an ADC is limited by the ADC resolution and the jitter
on the sampling clock. Considering an ideal ADC of infinite
resolution where the step size and quantization error can be
ignored, the available SNR can be expressed approximately by
1
⎤
SNR
where f is the highest analog frequency being digitized, and t
the rms jitter on the sampling clock. Figure 35 shows the
required sampling clock jitter as a function of the analog
frequency and effective number of bits (ENOB)
120
100
80
SNR (dB)
60
40
20
131030100
FULL-SCALE SINE WAVE ANALOG INPUT FREQUENCY (MHz)
Figure 35. ENOB and SNR vs. Analog Input Frequency
log20
×=
tj = 0.1ps
⎡
⎢
⎣
2π
tj = 50fs
ft
J
tj = 1ns
⎥
⎦
tj = 1ps
tj = 10ps
tj = 100ps
SNR = 20log
is
j
1
10
2πft
j
18
16
14
12
ENOB
10
8
6
4
05046-024
CMOS CLOCK DISTRIBUTION
The AD9510 provides four clock outputs (OUT4 to OUT7) that
are selectable as either CMOS or LVDS levels. When selected as
CMOS, these outputs provide a way to drive devices requiring
CMOS level logic at their clock inputs. Due to factors inherent
to CMOS logic, the jitter performance of these outputs cannot
equal that of the LVPECL and LVDS outputs. However, for
many clocking needs within a system, CMOS clock levels are
appropriate.
Whenever single-ended CMOS clocking is used, some of the
following general guidelines should be followed.
Point-to-point nets should be designed such that a driver has
one receiver only on the net, if possible. This allows for simple
termination schemes and minimizes ringing due to possible
mismatched impedances on the net. Series termination at the
source is generally required to provide transmission line
matching and/or to reduce current transients at the driver. The
value of the resistor is dependent on the board design and
timing requirements (typically 10 Ω to 100 Ω is used). CMOS
outputs are also limited in terms of the capacitive load or trace
length that they can drive, typically trace lengths less than 3
inches are recommended to preserve signal rise/fall times and
preserve signal integrity. Simulation results for the AD9510
CMOS outputs with a 1-inch and 3-inch trace load are shown in
Figure 37. In this example, the series resistor is 10 Ω and the
trace impedance is 60 Ω. Signal integrity, in this example, has
started to degrade already at a 3-inch trace length.
AD9540
OUT4
Figure 36. Series Termination of CMOS Output
10Ω
60.4Ω
1.0 INCH
MICROSTRIP
50pF
GND
05046-025
See Application Note AN-501 at www.analog.com.
Many high performance ADC’s feature differential clock inputs
to simplify the task of providing the required low jitter clock on
a noisy PCB. (Distributing a single-ended clock on a noisy PCB
can result in coupled noise on the sample clock. Differential
distribution has inherent common-mode rejection that can
provide superior clock performance in a noisy environment.) The
AD9510 features both LVPECL and LVDS outputs that provide
differential clock outputs, which enable clock solutions that
maximize converter SNR performance. The input requirements
of the ADC (differential or single-ended, logic level,
termination) should be considered when selecting the best
clocking/converter solution.
Rev. PrB | Page 48 of 52
Figure 37. CMOS Output Waveforms
Page 49
Preliminary Technical Data AD9510
Termination at the far end of the PCB trace is a second option.
The CMOS outputs of the AD9510 do not supply enough
current to provide a full voltage swing with a low impedance
resistive, far-end termination, as shown in Figure 39. The far
end termination network should match the PCB trace
impedance and provide the desired switching point. The
reduced signal swing may still meet receiver input requirements
in some applications. This can be useful when driving long
trace lengths on less critical nets.
= 3.3V
V
PULLUP
CMOS
10Ω
OUT4, OUT5, OUT6, OUT7
SELECTED AS CMOS
50Ω
Figure 38. CMOS Output with Far-End Termination
100Ω
100Ω
3pF
05046-027
3.3V
3.3V
LVPECL
50
Ω
SINGLE-ENDED
(NOT COUPLED)
50
VT = VCC– 1.3V
Ω
127
83
127
Ω
83
Ω
Ω
Figure 40. LVPECL Far-End Termination
3.3V
LVPECL
200Ω200Ω
0.1nF
0.1nF
DIFFERENTIAL
(COUPLED)
100Ω
Figure 41 LVPECL with Parallel Transmission Line
3.3V
Ω
LVPECL
3.3V
LVPECL
05046-030
05046-031
Figure 39. Far-End Termination of CMOS Output Waveform
Because of the limitations of single-ended CMOS clocking,
consider using differential outputs when driving high speed
signals over long traces. The AD9510 offers both LVPECL and
LVDS outputs that are better suited for driving long traces
where the inherent noise immunity of differential signaling
provides superior performance for clocking converters.
LVPECL CLOCK DISTRIBUTION
The low voltage, positive emitter-coupled, logic (LVPECL)
outputs of the AD9510 provide the lowest jitter clock signals
available from the AD9510. The LVPECL outputs (because they
are open emitter) require a dc termination to bias the output
transistors. A simplified equivalent circuit in Figure 24 shows
the LVPECL output stage.
In most applications, a standard LVPECL far-end termination is
recommended, as shown in Figure 40. The resistor network is
designed to match the transmission line impedance (50 Ω) and
the desired switching threshold (1.3 V). Figure 42 shows a
typical LVPECL clock waveform.
Figure 42. Typical LVPECL Outputs
LVDS CLOCK DISTRIBUTION
Low voltage differential signaling (LVDS) i s a s econd
differential output option for the AD9510. LVDS provides clock
signals with jitter performance nearly as good as that obtainable
from LVPECL and better than CMOS. LVDS uses a current
mode output stage with several user-selectable current levels. A
3.5 mA output current yields 350 mV output swing across a
standard LVDS output termination of 100 Ω, meeting ANSI 644
requirements.
A recommended termination circuit for the LVDS outputs is
shown in Figure 43.
3.3V
LVDS
DIFFERENTIAL (COUPLED)
100Ω
Figure 43. LVDS Output Termination
100Ω
3.3V
LVDS
05046-032
Rev. PrB | Page 49 of 52
Page 50
AD9510 Preliminary Technical Data
A typical LVDS output waveform is shown in Figure 44.
See Application Note AN-586 at www.analog.com for more
information on LVDS.
Figure 44. Typical LVDS Output Waveforms
POWER AND GROUNDING CONSIDERATIONS AND
POWER SUPPLY REJECTION
Many applications seek high speed and performance under less
than ideal operating conditions. In these application circuits,
the implementation and construction of the PCB is as
important as the circuit design. Proper RF techniques must be
used for device selection, placement, and routing, as well as
power supply bypassing and grounding to ensure optimum
performance.
Figure 45. Differential LC Filter for Single 3.3 V Applications
Rev. PrB | Page 50 of 52
Page 51
Preliminary Technical Data AD9510
OUTLINE DIMENSIONS
0.30
0.25
0.18
64
1
PIN 1
INDICATOR
*
4.85
4.70 SQ
4.55
BSC SQ
PIN 1
INDICATOR
9.00
TOP
VIEW
8.75
BSC SQ
0.60 MAX
49
48
0.60 MAX
EXPOSED PAD
(BOTTOM VIEW)
1.00
0.85
0.80
12° MAX
SEATING
PLANE
0.45
0.40
0.35
0.80 MAX
0.65 TYP
0.50 BSC
*
COMPLIANT TO JEDEC STANDARDS MO-220-VMMD
EXCEPT FOR EXPOSED PAD DIMENSION
0.05 MAX
0.02 NOM
0.20 REF
33
32
7.50
REF
16
17
Figure 46. 64-Lead Lead Frame Chip Scale Package [LFCSP_VQ]
9 mm × 9 mm Body, Very Thin Quad (CP-64-1)
Dimensions shown in millimeters
ORDERING GUIDE
Model Temperature Range Package Description Package Options