ADF5610BCCZ [ADI]
Microwave Wideband Synthesizer Integrated VCO;型号: | ADF5610BCCZ |
厂家: | ADI |
描述: | Microwave Wideband Synthesizer Integrated VCO |
文件: | 总51页 (文件大小:1016K) |
中文: | 中文翻译 | 下载: | 下载PDF数据表文档文件 |
Microwave Wideband Synthesizer with
Integrated VCO
Data Sheet
ADF5610
power supplies for the PLL circuitry range from 3.1 V to 3.5 V,
and the VCO supplies are between 4.75 V and 5.25 V. The
charge pump supply voltage can be extended up to 3.6 V for
improved frequency band overlap and extended upper
frequency range.
FEATURES
RF output frequency range: 57 MHz to 14,600 MHz
RFOUT: 7300 MHz to 14,600 MHz
PDIV/NDIV: 57 MHz to 14,600 MHz
Fractional-N synthesizer and Integer N synthesizer modes
24-bit fractional modulus
Exact frequency mode for 0 Hz frequency error
Typical PFD spurious: <−105 dBc
The ADF5610 has an integrated VCO with a fundamental
frequency of 3650 MHz to 7300 MHz. These frequencies are
internally doubled and routed to the RFOUT pin. An additional
differential output allows the doubled VCO frequency to be
divided by 1, 2, 4, 8, 16, 32, 64, or 128, allowing the user to
generate RF output frequencies as low as 57 MHz. A simple
3-wire serial port interface (SPI) provides control of all on-chip
registers. To conserve power, this divider block can be disabled
when not needed through the SPI interface. Likewise, the output
power for both the single-ended output and the differential
output are programmable via the VCO register settings. The
ADF5610 also contains various power-down modes for the
VCO circuitry and PLL circuitry.
Integrated rms jitter: <40 fs
Normalized inband phase noise floor FOM
High current mode: −232 dBc/Hz (integer) and −229 dBc/Hz
(fractional)
Normal mode: −229 dBc/Hz (integer) and −226 dBc/Hz
(fractional)
Maintains frequency lock over −40°C to +85°C (lock and leave)
Low phase noise VCO
−115 dBc/Hz typical at 100 kHz (7.3 GHz)
−114 dBc/Hz typical at 100 kHz (10 GHz)
−109 dBc/Hz typical at 100 kHz (14.6 GHz)
RFOUT power: 5 dBm
The integrated phase detector (PD) and delta-sigma (Δ-Σ)
modulator, capable of operating at up to 100 MHz, permit wide
loop bandwidths and fast frequency tuning with a typical
spurious level of −100 dBc.
Programmable divide by 1, 2, 4, 8, 16, 32, 64, or 128 output
Programmable output power level
Typical power dissipation: 815 mW
With phase noise levels from −115 dBc/Hz at 7.3 GHz to
−109 dBc/Hz at 14.6 GHz, the ADF5610 is equipped to
minimize blocker effects, and to improve receiver sensitivity
and transmitter spectral purity. The low phase noise floor
eliminates any contribution to modulator and mixer noise floor
in transmitter applications.
Programmable low current and power dissipation: <700 mW
Fast frequency hopping (autocalibration enabled): <40 µs
48-terminal, 7 mm × 7 mm LGA package: 49 mm2
APPLICATIONS
Military and defense
Test equipment
Clock generation
Wireless infrastructure
Satellite and very small aperture terminal (VSAT)
Microwave radio
The ADF5610 is a PLL with integrated VCO. The device features
an innovative programmable performance technology that
enables the ADF5610 to tailor current consumption and
corresponding noise performance to individual applications by
selecting either a low current consumption mode or a high
performance mode for improved phase noise performance.
GENERAL DESCRIPTION
The ADF5610 allows implementation of fractional-N or Integer N
phase-locked loop (PLL) frequency synthesizers when used
with an external loop filter and an external reference source.
The wideband microwave voltage controlled oscillator (VCO)
design permits frequency operation from 7300 MHz to
Additional features of the ADF5610 include approximately 3 dB
of RFOUT gain control in 1.5 dB steps and 5 dB of control on the
differential port in approximately 2.5 dB steps. Finally, the Δ-Σ
modulator with exact frequency mode enables users to generate
output frequencies with 0 Hz frequency error.
14600 MHz at a single radio frequency (RF) output. A series of
frequency dividers with a differential frequency output allows
operation from 57 MHz to 14600 MHz. Analog and digital
Rev. 0
Document Feedback
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.
One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A.
Tel: 781.329.4700
Technical Support
©2019 Analog Devices, Inc. All rights reserved.
www.analog.com
ADF5610
Data Sheet
TABLE OF CONTENTS
Features .............................................................................................. 1
Legacy Mode: Serial Port Read Operation.............................. 36
Open Mode: Serial Port Write Operation............................... 37
Open Mode: Serial Port Read Operation................................ 37
PLL Register Map ........................................................................... 39
Applications....................................................................................... 1
General Description......................................................................... 1
Revision History ............................................................................... 2
Functional Block Diagram .............................................................. 3
Specifications..................................................................................... 4
Timing Specifications .................................................................. 9
Absolute Maximum Ratings.......................................................... 10
Thermal Resistance .................................................................... 10
ESD Caution................................................................................ 10
Pin Configuration and Function Descriptions........................... 11
Typical Performance Characteristics ........................................... 13
Theory of Operation ...................................................................... 16
PLL Subsystem Overview.......................................................... 16
VCO Subsystem Overview........................................................ 16
SPI Configuration of PLL and VCO Subsystems................... 17
VCO Subsystem.......................................................................... 18
PLL Subsystem............................................................................ 24
Soft Reset and Power-On Reset................................................ 32
Power-Down Mode .................................................................... 32
General-Purpose Output (GPO).............................................. 32
Chip Identification ..................................................................... 33
Power Supply............................................................................... 33
Programmable Performance Technology................................ 33
Loop Filter and Frequency Changes........................................ 33
Mute Mode .................................................................................. 34
Serial Port Interface........................................................................ 35
SPI Modes of Operation ............................................................ 35
Register 0x00 Modes .................................................................. 35
Serial Port Mode Decision After Power-On Reset................. 35
Serial Port Legacy Mode............................................................ 35
Serial Port Legacy Mode: Single PLL....................................... 35
Serial Port Open Mode.............................................................. 35
Legacy Mode: Serial Port Write Operation............................. 36
ID Register, Reset Strobe Register, and Open Mode Read
Register ........................................................................................ 39
RST Register................................................................................ 39
Reference Divider (RDIV) Register......................................... 39
Frequency Register, Integer Part .............................................. 40
Frequency Register, Fractional Part......................................... 40
VCO SPI Register....................................................................... 40
Δ-Σ Configuration Register ...................................................... 41
Lock Detect Register.................................................................. 41
Analog Enable (EN) Register.................................................... 42
Charge Pump Register............................................................... 43
Autocalibration Register............................................................ 43
Phase Detector (PD) Register................................................... 44
Exact Frequency Mode Register............................................... 44
General-Purpose, SPI, and Reference Divider
(GPO_SPI_RDIV) Register ...................................................... 44
VCO Tune Register .................................................................... 45
Successive Approximation Register ......................................... 46
General-Purpose 2 Register...................................................... 46
Built In Self Test (BIST) Register.............................................. 46
VCO Subsystem Register Map...................................................... 47
VCO Power Control Register ................................................... 47
VCO Differential Output Divider Register............................. 48
Applications Information.............................................................. 50
Evaluation Printed Circuit Board (PCB) ................................ 50
Evaluation Kit Contents ............................................................ 50
Outline Dimensions....................................................................... 51
Ordering Guide .......................................................................... 51
REVISION HISTORY
2/2019—Revision 0: Initial Version
Rev. 0 | Page 2 of 51
Data Sheet
ADF5610
FUNCTIONAL BLOCK DIAGRAM
VTUNE
XREFP
CEN SEN SCK SDI
SPI CONTROL
R
CAL
VCO
DIVIDER
PHASE
FREQUENCY
DETECTOR
CHARGE
PUMP
CP
×1
×2
AMP
RFOUT
PDIV_OUT
NDIV_OUT
÷1/÷2/÷4/
÷8/÷16/÷32/
÷64/÷128
N
÷1/÷2
DIVIDER
MODULATOR
ADF5610
Figure 1.
Rev. 0 | Page 3 of 51
ADF5610
Data Sheet
SPECIFICATIONS
Analog supply (AVDD) = digital power supply (DVDD) = 3.3 V 5%, power supply for charge pump digital section (VDDLS), power
supply for charge pump (VPPCP), reference path supply (RVDD), phase detector supply (VCCPD), prescaler supply (VCCPS), power
supply for PLL RF section (VCCHF) = 3.3 V 5%, ꢀias for VCO digital logic, SPI ꢀuffer, and input ꢀuffer to PLL (VDD1), reference
voltage supply (VDD2), differential output divider supply (VDD3) = 3.3 V 5%, VCO supply (VCOVCC) = 5.0 V 5%, ground (GND)
= 0 V, minimum and maximum specifications across the temperature range of −40°C to +85°C.
Table 1.
Parameter
Symbol Test Conditions/Comments
Min
Typ
Max
Unit
RF OUTPUT
CHARACTERISTICS
RFOUT Frequency
VCO Frequency at PLL
Input
7300
3650
14,600
7300
MHz
MHz
Frequency Range for
PDIV_OUT and
NDIV_OUT
57
14,600
MHz
OUTPUT POWER
RFOUT Power
Across all frequencies, high performance
0
5
10
dBm
mode (VCO Register 0x01, Bits[4:2] = 0x3),
maximum gain setting (VCO Register 0x01,
Bits[1:0] = 0x3)
RFOUT Power Control
Range
RFOUT Power Variation vs.
Temperature
RFOUT Power Variation vs.
Frequency
PDIV_OUT and NDIV_OUT
Power
1 dB steps
4
1
2
0
dB
dBm
dBm
dBm
Maximum gain setting (VCO Register 0x01,
Bits[6:5] = 0x3), single-ended
−6
6
Divide by 2 to divide by 128
Three settings, bypass mode (divide by 1)
2
6
dBm
dB
PDIV_OUT and NDIV_OUT
Control Range
HARMONICS (RFOUT)
½ Harmonic
1.5 Harmonic
Second Harmonic
2.5 Harmonic
Third Harmonic
3650 MHz to 7300 MHz
−20
−30
−30
−35
−30
dBc
dBc
dBc
dBc
dBc
HARMONICS (PDIV_OUT and
NDIV_OUT)
Single-Ended
½ Harmonic
Fundamental feedthrough
N = divide by 1
−20
−60
dBc
dBc
N = divide by 2
Second Harmonic
Third Harmonic
Push/push feedthrough
N = divide by 1
N = divide by 2
N = divide by 1
N = divide by 2
−26
−24
−24
−17
dBc
dBc
dBc
dBc
Differential
½ Harmonic
Fundamental feedthrough (N = 1)
Push/push feedthrough (N = 2)
N = divide by 1
−20
−52
−31
−27
dBc
dBc
dBc
dBc
Second Harmonic
N = divide by 2
Rev. 0 | Page 4 of 51
Data Sheet
ADF5610
Parameter
Symbol Test Conditions/Comments
N = divide by 1
Min
Typ
−30
−16
67
Max
Unit
dBc
dBc
Third Harmonic
N = divide by 2
RFOUT (VT = 1.65 V)
VCO Supply Pushing
(VDD1)1
MHz/V
VCO Supply Pushing
RFOUT (VT = 1.65 V)
1.0
1.0
MHz/V
(VCOVCC)1
VCO Load Pulling (2.0:1)2
VCO Frequency Drift
RFOUT (VT = 1.65 V)
RFOUT (VT = 1.65 V)
Operating at 7.3 GHz
Operating at 14.6 GHz
MHz p-p
0.5
1.02
MHz/°C
MHz/°C
PLL RF INPUT PRESCALER
Input Frequency Range
Approximate drive level = −10 dBm
Maximum = 219 – 1
57
4000
MHz
PLL RF DIVIDER
CHARACTERISTICS
19-Bit N Divider Range
(Integer N)
19-Bit N Divider Range
(Fractional-N)
16
20
524,287
524,283
Fractional nominal divide ratio varies ( 4)
dynamically maximum
REFERENCE INPUT
CHARACTERISTICS
Maximum XREFP Input
Frequency
XREFP Input Level
XREFP Input Capacitance
14-Bit R Divider Range
PHASE DETECTOR (PD)4
DC
−6
50
350
MHz
AC-coupled3
+12
5
16,383
dBm
pF
1
PD Frequency Fractional
Mode5
PD Frequency Integer
Mode
Fractional Mode B
DC
DC
100
100
MHz
MHz
CHARGE PUMP
Output Current
Charge Pump Gain Step
Size
0.02
40
2.54
60
mA
µA
20
50
LOGIC INPUTS
Logic Switching Threshold
Input high voltage (VIH) and input low
voltage (VIL) within 50 mV of VSW
% of
DVDD
(VSW
)
SDO LOGIC OUTPUT
Output High Voltage
Complementary metal-oxide semiconductor −0.2
(CMOS) 3.3 V mode (Register 0x0F, Bits[9:8] =
00b)
V
Open-drain mode (Register 0x0F, Bits[9:8] =
01b)6
CMOS mode (Register 0x0F, Bits[9:8] = 00b)
3.3
0.1
V
V
Output Low Voltage (VOL
)
Open-drain mode (Register 0x0F, Bits[9:8] =
0.4
6
01b)7
Serial Clock (SCK)
Frequency Rate
CMOS mode (Register 0x0F, Bits[9:8] = 00b)8
50
MHz
Open-drain mode (Register 0x0F[9:8] = 01b)9
CMOS mode (Register 0x0F, Bits[9:8] = 00b)
5
10
10
20
10
MHz
pF
pF
Capacitive Load
Load Current
Open-drain mode (Register 0x0F, Bits[9:8] =
01b)10
CMOS mode (Register 0x0F, Bits[9:8] = 00b)11
3.6
mA
Rev. 0 | Page 5 of 51
ADF5610
Data Sheet
Parameter
Symbol Test Conditions/Comments
Min
Typ
Max
Unit
Open-drain mode (Register 0x0F, Bits[9:8] =
7.2
mA
01b)12
Output Resistance When
Open-drain mode (Register 0x0F, Bits[9:8] =
01b)
Open-drain mode (Register 0x0F, Bits[9:8] =
01b)
100
200
Ω
Ω
Driver Is Low (RON
)
Pull-Up Resistor (RUP
)
500
1000
Rise Time
Fall Time
SCK to SDO Turnaround
Time
RF Divider Input Below
4 GHz, Integer Mode
RF Divider Input Above
4 GHz, Integer Mode
RF Divider Input Below
4 GHz, Fractional Mode
RF Divider Input Above
4 GHz, Fractional Mode
CMOS mode (Register 0x0F, Bits[9:8] = 00b)13
CMOS mode (Register 0x0F, Bits[9:8] = 00b)13
CMOS mode (Register 0x0F, Bits[9:8] = 00b)13
0.5 + 0.3 × CLOAD
1.5 + 0.2 × CLOAD
0.9 + 0.1 × CLOAD
7
10
12
ns
ns
ns
19-bit, all values14
16
32
20
40
524,287
19-bit, even values only
19-bit, all values
1,048,574
524,283
19-bit, even values only
1,048,566
POWER SUPPLY VOLTAGES
3.3 V Supplies
AVDD, VCCHF, VCCPS, VCCPD, RVDD, DVDD,
VPPCP, VDDLS, VDD1, VDD2, VDD3
VCOVCC
3.1
3.3
5.0
3.5
V
V
5.0 V Supplies
4.75
5.25
POWER DISSIPATION
Typical Power Dissipation
Programmable Low
Current and Power
Dissipation
815
<700
mW
mW
POWER SUPPLY CURRENTS
AVDD
RVDD
3.3 V
3.3 V
3.3 V
3.3 V
3.3 V
3.3 V
3.3 V
3.3 V
1.6
6
4
29.7
1.0
11.9
5.6
0.6
3
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
VCCHF
VCCPS
VCCPD
DVDD
VPPCP15
VDDLS15
VDD1
3.3 V
3.3 V
VDD2
VDD3
1
62
Divide by 1
Divide by 2
Divide by 4
Divide by 8
Divide by 16
Divide by 32
Divide by 64
Divide by 128
5.0 V
73.7
78.8
82.4
85.9
88.1
90
92
110
VCOVCC
Power-Down
PLL16
Power down via SPI Register 0x01, Bit 0 = 0
and Register 0x01, Bit 1 = 0
Power down using VCO Register 0x02,
Bit 3 = 1
Power down using VCO Register 0x01,
Bit 8 = 0
100
13
µA
VCOVCC17
Divider17
mA
mA
1.5
Rev. 0 | Page 6 of 51
Data Sheet
ADF5610
Parameter
Symbol Test Conditions/Comments
Min
Typ
Max
Unit
POWER-ON RESET
Typical Reset Voltage on
DVDD
Minimum DVDD Voltage
for No Reset
700
mV
V
1.5
Power-On Reset Delay
250
µs
VCO OPEN-LOOP PHASE
NOISE
RFOUT at 7.3 GHz
10 kHz Offset
100 kHz Offset
1 MHz Offset
10 MHz Offset
100 MHz Offset
RFOUT at 10 GHz
10 kHz Offset
100 kHz Offset
1 MHz Offset
10 MHz Offset
100 MHz Offset
RFOUT at 11 GHz
10 kHz Offset
100 kHz Offset
1 MHz Offset
10 MHz Offset
100 MHz Offset
RFOUT at 14.6 GHz
10 kHz Offset
100 kHz Offset
1 MHz Offset
−92
dBc/Hz
dBc/Hz
dBc/Hz
dBc/Hz
dBc/Hz
−115
−135
−155
−168
−92
dBc/Hz
dBc/Hz
dBc/Hz
dBc/Hz
dBc/Hz
−114
−135
−155
−164
−89
dBc/Hz
dBc/Hz
dBc/Hz
dBc/Hz
dBc/Hz
−112
−132
−152
−161
−87
−109
−130
−150
−163
dBc/Hz
dBc/Hz
dBc/Hz
dBc/Hz
10 MHz Offset
100 MHz Offset
PLL
Phase Noise at 20 kHz
Offset, 50 MHz Phase
Frequency Detector
(PFD) Rate
Over process with 3.3 V power supply at
25°C, measured with >200 kHz loop
bandwidth
Fast Frequency Hopping
(Autocalibration
Enabled)
<40
µs
µs
Hz
Lock Time
Depends on loop filter bandwidth, PFD rate,
definition of lock (to within Hz or degrees
of settling), and Register 0x0A configuration
Depends on PFD rate and VCO output
divider setting
3.5 GHz to 7.5 GHz output, at typical phase
detector frequency (fPD) of 50 MHz, typical
resolution = 3 Hz
100
Frequency Resolution
Fundamental Mode
fPD/224
Divider Mode
<3.5 GHz output, resolution depends on
VCO output divider setting
fPD/(224 × output
divider)
Hz
Integer Boundary Spurs
Phase Detector Spurs
Reference Spurs
−45
−105
−100
dBc
dBc
dBc
Normalized Inband Phase
Noise Floor Figure of
Merit (FOM)
Normalized to 1 Hz
Rev. 0 | Page 7 of 51
ADF5610
Data Sheet
Parameter
Symbol Test Conditions/Comments
Min
Typ
Max
Unit
Floor Integer Mode
Floor High Current Mode18
Integer Mode
−229
−232
dBc/Hz
dBc/Hz
Floor Fractional Mode
Floor High Current Mode
Fractional Mode
−226
−229
dBc/Hz
dBc/Hz
Flicker Noise (Integer
Mode and Fractional
Mode)
−268
<40
dBc/Hz
fs
Integrated RMS Jitter
VCO CHARACTERISTICS
VCO Tuning Sensitivity at
RFOUT
Measured on RFOUT (VT = 1.65 V)
7300 MHz
9000 MHz
11000 MHz
13000 MHz
14600 MHz
101
80
128
109
96
175
154
135
10
MHz/V
MHz/V
MHz/V
MHz/V
MHz/V
pF
pF
pF
ns
Tune Port Capacitance
VT = 0.5 V dc
VT = 1.65 V dc
VT = 2.8 V dc
VCO Register 0x00, Bit 0 switch time (VT to
VPRST)
Autocalibration Delay
(Enable or Disable)
VCO OUTPUT DIVIDER
VCO RF Divider Range
1, 2, 4, 6, 8, … 128
1
128
VCO Supply Pushing
(VCOVCC)1
VCO Supply Pushing
(VDD1)17
PDIV_OUT and NDIV_OUT, divide by 1, VT =
1.65 V
PDIV_OUT and NDIV_OUT, divide by 1, VT =
1.65 V
0.7
67
MHz/V
MHz/V
RETURN LOSS
RFOUT
PDIV_OUT (N > 1)
5
14
dB
dB
NDIV_OUT (N > 1)
14
dB
1 Pushing refers to a change in VCO frequency due to a change in the power supply voltage. To derive pushing at other divide ratios, divide the nominal pushing value
by the divide ratio.
2 Pulling refers to a change in VCO frequency due to a change in the load impedance.
3 Measured with 100 Ω external termination. See the Reference Input Stage section for more details.
4 Slew rate of ≥0.5 ns/V is recommended. See the Reference Input Stage section for more details. Frequency is guaranteed across process voltage and temperature from
−40°C to +85°C.
5 This maximum PD frequency can only be achieved if the minimum N value is respected. For example, in the case of fractional mode, the maximum PD frequency =
fVCO/20 or 100 MHz, whichever is less.
6 External 1 kΩ pull-up resistor to 1.8 V.
7 Limited by the 1 kΩ pull-up resistor and NMOS RON
8 10 pF load capacitor.
.
9 10 pF load capacitor, 1 kΩ pull-up resistor. In general, open-drain mode can support higher frequencies at the expense of maximum VOL. The maximum frequency for a
given pull-up resistor and load capacitor is approximately 1/(10 × RPULL-UP × CLOAD). For example, a 10 pF load capacitor and 1 kΩ pull-up resistor can support up to
10 MHz, where VOL maximum = VDD × RON/(1 kΩ + RON) ≈ 164 mV. With a 500 Ω pull-up resistance and a 10 pF load, a 20 MHz maximum frequency is possible, and the
maximum VOL increases to 300 mV.
10 1 kΩ pull-up resistor.
11 The minimum resistive load to ground in CMOS mode is 1 kΩ.
12 The SDO pin does not have short-circuit protection. Do not exceed the maximum current of 7.2 mA under any condition.
13
C
is the capacitive load in pF. CLOAD maximum = 20 pF. For example, with a 5 pF capacitive load, the rise time is 0.5 + 0.3 × 5 = 2 ns.
LOAD
14 Values shown are the product of the prescaler divide ratio × N counter value.
15 VCCPD and VDDLS can be operated at 3.6 V maximum to increase band and VCO core overlap and extend the upper end of the frequency range. The typical current is
expected to be less than 1.5 mA. Both must be equal. Therefore, if one changes, so must the other. Exceeding 3.6 V results in ESD diodes drawing current.
16 Reference disconnected.
17 Some circuits remain on.
18 High current mode is a register setting that is enabled and works in conjunction with the functional mode in which the device is operating (integer mode or fractional
mode).
Rev. 0 | Page 8 of 51
Data Sheet
ADF5610
TIMING SPECIFICATIONS
SPI Open Mode: Write Timing Characteristics
AVDD = DVDD = 3.3 V, exposed pad (EPAD) = 0 V. See Figure 30.
Table 2.
Parameter
Test Conditions/Comments
Min
3
3
10
10
10
20
Typ
Max
Unit
ns
ns
ns
ns
ns
ns
MHz
t1
t2
t3
t4
t5
t6
fSCK
Serial data input (SDI) setup time to SCK rising edge
Serial clock (SCK) rising edge to SDI hold time
Serial enable (SEN) low duration
SEN high duration
SCK 32nd rising edge to SEN rising edge
Recovery time (not shown in Figure 30)
Maximum SPI clock speed
50
SPI Open Mode: Read Timing Characteristics
AVDD = DVDD = 3.3 V, E PAD = 0 V. See Figure 31 and Figure 32.
Table 3.
Parameter
Test Conditions/Comments
SDI setup time to SCK rising edge.
SCK rising edge to SDI hold time.
SEN low duration.
SEN high duration.
SCK rising edge to SDO time. Add 0.2 ns for each pF of CLOAD
Min
Typ
Max
Unit
ns
ns
ns
ns
t1
t2
t3
t4
3
3
10
10
1
t5
.
8.2
ns
t6
t7
Recovery time.
SCK 32nd rising edge to SEN rising edge.
10
10
ns
ns
1 An extra 0.2 ns delay is required for every 1 pF load on SDO.
SPI Legacy Mode: Write Timing Characteristics
AVDD = DVDD = 3.3 V 10%, analog ground (AGND) = digital ground (DGND) = 0 V. See Figure 28.
Table 4.
Parameter
Test Conditions/Comments
SEN to SCK setup time
SDI to SCK setup time
SCK to SDI hold time
SEN low duration
Min
8
3
3
20
Typ
Max
Units
t1
t2
t3
t4
ns
ns
ns
ns
Maximum SPI Clock Frequency
50
MHz
SPI Legacy Mode: Read Timing Characteristics
AVDD = DVDD = 3.3 V, 10%, AGND = DGND = 0 V. See Figure 29.
Table 5.
Parameter
Test Conditions/Comments
SEN to SCK setup time
SDI to SCK setup time
SCK to SDI hold time
SEN low duration
Min
8
3
3
20
Typ
Max
Units
ns
ns
ns
ns
t1
t2
t3
t4
t5
SCK to SDO delay. Add 0.2 ns for each pF of CLOAD
8.2
ns
Rev. 0 | Page 9 of 51
ADF5610
Data Sheet
ABSOLUTE MAXIMUM RATINGS
THERMAL RESISTANCE
Table 6.
Thermal performance is directly linked to printed circuit board
(PCB) design and operating environment. Careful attention to
PCB thermal design is required.
Parameter
Rating
AVDD to GND, DVDD to GND
AVDD to DVDD
VDDLS, VPPCP to GND
VDDLS to VPPCP1
RVDD, VCCPD, VCCPS, VCCHF to GND
XREFP
Minimum Digital Load
−0.3 V to +3.6 V
−0.5 V to +0.5 V
−0.3 V to +3.8 V
−0.5 V to +0.5 V
−0.3 V to +3.6 V
18 dBm, 5.6 V peak
1.0 kΩ
θJA is the natural convection junction to ambient thermal
resistance measured in a one cubic foot sealed enclosure. θJC is
the junction to case thermal resistance.
Table 7. Thermal Resistance
Package Type1
θJA
θJC
Unit
Digital Input 1.4 V to 1.7 V Minimum Rise 20 ns
Time
CC-48-1
31.2
6.9
°C/W
Digital Input Voltage Range
−0.25 V to DVDD +
0.5 V
1 Thermal impedance measured values are based on use of a 2S2P board with
the thermal impedance pad (EPAD) soldered to GND (GND = 0 V). A cold plate
is set to 20°C and attached to the bottom side of the PCB using 100 μm
thermal interface material (3.56 W/mK).
VDD1, VDD2, VDD3
VCOVCC
Operating Temperature Range
Storage Temperature Range
Maximum Junction Temperature
Reflow Soldering
−0.3 V to +3.6 V
−0.3 V to +5.5 V
−40°C to +85°C
−65°C to +120°C
150°C
ESD CAUTION
Peak Temperature
260°C
40 sec
Time at Peak Temperature
Electrostatic Discharge (ESD)
Charged Device Model
Human Body Model
500 V
1000 V
1 VDDLS must be equal to VPPCP.
Stresses at or above those listed under Absolute Maximum
Ratings may cause permanent damage to the product. This is a
stress rating only; functional operation of the product at these
or any other conditions above those indicated in the operational
section of this specification is not implied. Operation beyond
the maximum operating conditions for extended periods may
affect product reliability.
Rev. 0 | Page 10 of 51
Data Sheet
ADF5610
PIN CONFIGURATION AND FUNCTION DESCRIPTIONS
48 47 46 45 44 43 42 41 40 39 38 37
36
35
34
33
32
31
30
29
28
27
26
25
1
2
3
GND
RVDD
GND
VDDLS
GND
CP
GND
VPPCP
GND
GND
VPRST
VDD2
VCOVCC
GND
RFOUT
GND
VDD3
PDIV_OUT
NDIV_OUT
GND
4
5
6
ADF5610
TOP VIEW
(Not to Scale)
7
8
9
10
11
12
AVDD
GND
GND
GND
13 14 15 16 17 18 19 20 21 22 23 24
NOTES
1. EXPOSED PAD. THE EXPOSED PAD OR GROUND PADDLE ON THE BACKSIDE OF THE PACKAGE
MUST BE TIED TO DC GROUND FOR ELECTRICAL, MECHANICAL, AND THERMAL REASONS.
NOTE THAT RF GROUND AND DC GROUND ARE THE SAME IN THIS CASE.
Figure 2. Pin Configuration
Table 8. Pin Function Descriptions
Pin No. Mnemonic Description
1, 3, 5, 7, 9, 11, 12, 14, 16, GND
18, 20, 22, 24 to 26,
30, 32, 36, 38, 40, 45,
47
Ground.
2
4
RVDD
VDDLS
Reference Path Supply. 3.3 V 5%.
Power Supply for Charge Pump Digital Section. 3.3 V 5% typical, 3.6 V maximum. VDDLS must
be equal to VPPCP.
6
8
CP
Charge Pump Output. Place first pole of low-pass loop filter close to this pin if loop filter path is
electrically long.
Power Supply for Charge Pump. 3.3 V 5% typical, 3.6 V maximum. VPPCP must be equal to
VDDLS.
VPPCP
10
13
AVDD
BIAS
Analog Supply. 3.3 V 5%. AVDD must be equal to DVDD.
External Bypass Decoupling for Precision Bias Circuits. Note that the 1.920 V 20 mV reference
voltage is generated internally and cannot drive an external load. The reference voltage must be
measured with a 10 GΩ meter, such as the Agilent 34410A. A normal 10 MΩ digital voltmeter
reads erroneously.
15
17
19
21
VCCPD
VCCPS
SDO
Phase Detector Supply. 3.3 V 5%.
Prescaler Supply. 3.3 V 5%.
Serial Data Output. Lock detect and various other functions are available via an internal mux.
Power Supply for PLL RF Section. 3.3 V 5%. Place a decoupling capacitor as close as possible to
this pin.
VCCHF
23
27
28
VDD1
Bias for VCO Digital Logic, VG Pin, SPI Buffer, and Input Buffer to PLL. Nominally 3.3 V 5%. Retain
bias to hold the VCO register contents when the VCO is powered down.
NDIV_OUT Complementary Output of Differential Frequency Divider. N = 1, 2, 4, 8, 16, 32, 64, or 128. DC block
required. A broadband 100 nF capacitor is recommended.
PDIV_OUT
Primary Output of Differential Frequency Divider. N = 1, 2, 4, 8, 16, 32, 64, or 128. DC block
required. A broadband 100 nF capacitor is recommended.
29
31
VDD3
RFOUT
Differential Output Divider Supply. 3.3 V 5%.
RF Output (7300 MHz to 14,600 MHz). Internal dc block. DC block using a broadband 100 nF
capacitor.
33
34
35
37
VCOVCC
VDD2
VPRST
VT
VCO Power Supply. 5.0 V 5%.
Reference Voltage Supply. 3.3 V 5%.
Temperature Dependent, Calibration Preset. Decouple with a 470 nF capacitor.
VCO Tuning Port. 0 V to 3.6 V, 5%. Place last pole of low-pass filtered charge pump output close
to this pin if loop filter path is electrically long.
Rev. 0 | Page 11 of 51
ADF5610
Data Sheet
Pin No.
Mnemonic Description
39
VG
Gate Voltage Bypassing. Decouple to GND with a low effective series resistance (ESR), 10 µF
capacitor.
41
42
43
44
46
SCK
SDI
SEN
CEN
DVDD
Serial Port Interface Clock Input.
Serial Port Data Input.
Serial Port Interface Enable Input. Active high.
Hardware Chip Enable Input. Active high. Logic low powers down the PLL section.
Digital Power Supply. 3.3 V 5%. Retain bias to hold the PLL register contents when powering
down the PLL.
48
XREFP
EPAD
External Reference Input. For 50 Ω match, ac couple to the XREFP pin using a low reactance
capacitor value and add a 100 Ω resistor to ground.
Exposed Pad. The exposed pad or ground paddle on the backside of the package must be tied to
dc ground for electrical, mechanical, and thermal reasons. Note that RF ground and dc ground are
the same in this case.
Rev. 0 | Page 12 of 51
Data Sheet
ADF5610
TYPICAL PERFORMANCE CHARACTERISTICS
For Figure 3 through Figure 7, Agilent E5052B is the test equipment used to measure the phase noise.
–20
–30
–30
–40
1: 100Hz –92.4277dBc/Hz
1: 1kHz
–63.6667dBc/Hz
–92.2744dBc/Hz
–114.9648dBc/Hz
–135.5251dBc/hz
–155.2561dBc/Hz
2: 1kHz
–102.7103dBc/Hz
2: 10kHz
3: 100kHz
4: 1MHz
5: 10MHz
3: 10kHz –110.4199dBc/Hz
4: 100kHz –112.4860dBc/hz
–40
–50
5: 1MHz
–131.1319dBc/Hz
–50
–60
6: 10MHz –154.3151dBc/Hz
7: 40MHz –162.6025dBc/Hz
8: 95MHz –167.0970dBc/Hz
9: 100MHz–145.2447dBc/Hz
x: START:100Hz
6: 91.203029MHz –168.2089dBc/Hz
–60
–70
1
–70
–80
–80
–90
STOP: 100MHz
CENTER: 50.00005MHz
SPAN: 99.9999MHz
2
–90
–100
–110
–120
–130
–140
–150
–160
–170
–180
–190
1
–100
–110
–120
–130
–140
–150
–160
–170
–180
2
3
3
4
4
5
9
ANALYSIS RANGE X: BAND MARKER
ANALYSIS RANGE Y: BAND MARKER
INTG NOISE: –56.8411dBc/100MHz
RMS NOISE: 2.03452mrad
5
6
116.569mdeg
6
RMS JITTER: 40.475fsec
RESIDUAL FM: 8.51901kHz
7
8
1k
10k
100k
1M
10M
100M
100
1k
10k
100k
1M
10M
100M
FREQUENCY (Hz)
FREQUENCY (Hz)
Figure 3. Open-Loop VCO Phase Noise at 7300 MHz, RFOUT, High
Performance Mode (VCO Register 0x01, Bits[4:2] = 7 (Decimal)
Figure 6. Closed-Loop Phase Noise Illustrating Low Jitter Capability at
8.0 GHz, Wenzel 100 MHz Reference, 100 MHz PFD, RFOUT, High
Performance Mode (VCO Register 0x01, Bits[4:2] = 7 (Decimal)), 25°C
–20
–30
1: 100Hz –87.2769dBc/Hz
1: 1kHz
–63.5734dBc/Hz
–90.9588dBc/Hz
–113.2091dBc/Hz
–134.1621dBc/hz
–154.6000dBc/Hz
–30
–40
–40
–50
2: 1kHz
–97.5534dBc/Hz
2: 10kHz
3: 100kHz
4: 1MHz
5: 10MHz
3: 10kHz –107.4160dBc/Hz
4: 100kHz –128.5294dBc/hz
5: 1MHz
–131.1319dBc/Hz
–50
–60
6: 10MHz –149.9961dBc/Hz
7: 40MHz –159.6728dBc/Hz
8: 95MHz –163.7876dBc/Hz
9: 100MHz –164.0052dBc/Hz
x: START:100Hz
6: 91.203029MHz –164.5740dBc/Hz
–60
–70
1
–70
–80
–80
–90
STOP: 100MHz
CENTER: 50.00005MHz
SPAN: 99.9999MHz
–90
2
–100
–110
–120
–130
–140
–150
–160
–170
–180
–190
1
–100
–110
–120
–130
–140
–150
–160
–170
–180
2
3
3
4
4
5
ANALYSIS RANGE X: BAND MARKER
ANALYSIS RANGE Y: BAND MARKER
INTG NOISE: –52.4262dBc/100MHz
RMS NOISE: 3.38226mrad
5
9
6
193.789mdeg
6
7
RMS JITTER: 37.382fsec
RESIDUAL FM: 6.46437kHz
8
100
1k
10k
100k
1M
10M
100M
1k
10k
100k
1M
10M
100M
FREQUENCY (Hz)
FREQUENCY (Hz)
Figure 4. Open-Loop VCO Phase Noise at 10000 MHz, RFOUT, High
Performance Mode (VCO Register 0x01, Bits[4:2] = 7 (Decimal))
Figure 7. Closed-Loop Phase Noise Illustrating Low Jitter Capability at 14.4 GHz,
Wenzel 100 MHz Reference, 100 MHz PFD, RFOUT, High Performance Mode
(VCO Register 0x01, Bits[4:2] = 7 (Decimal)), 25°C
–30
–40°C
+25°C
+85°C
1: 1kHz
–60.6675dBc/Hz
–87.3627dBc/Hz
–109.1086dBc/Hz
–129.9533dBc/hz
–150.5663dBc/Hz
–40
–50
2: 10kHz
3: 100kHz
4: 1MHz
5: 10MHz
–60
6: 91.203029MHz –163.5237dBc/Hz
–70
1
–80
–90
2
–100
–110
–120
–130
–140
–150
–160
–170
–180
–190
3
4
5
6
95
55
15
255
215
175
135
VCO BAND
1k
10k
100k
1M
10M
100M
FREQUENCY (Hz)
Figure 5. Open-Loop VCO Phase Noise 14.6 GHz, RFOUT, High Performance
Mode (VCO Register 0x01, Bits[4:2] = 7 (Decimal))
Figure 8. Carrier Frequency vs. VCO Band, Autocalibration Enabled,
VDDLS = 3.3 V, VPPCP = 3.3 V
Rev. 0 | Page 13 of 51
ADF5610
Data Sheet
–40°C
+25°C
+85°C
–40°C
+25°C
+85°C
6
7
8
9
10
11
12
13
14
15
16
7
8
9
10
11
12
13
14
15
16
CARRIER FREQUENCY (GHz)
CARRIER FREQUENCY (GHz)
Figure 12. Second Harmonic vs. Carrier Frequency at RFOUT, Maximum
Output Gain (VCO Register 0x01, Bits[1:0] = 3 (Decimal)), High Performance
Mode (VCO Register 0x01, Bits[4:2] = 7 (Decimal))
Figure 9. RFOUT Power vs. Carrier Frequency, Maximum Output Gain =
(VCO Register 0x01, Bits[1:0] = 3 (Decimal)), High Performance Mode
(VCO Register 0x01, Bits[4:2] = 7 (Decimal))
0
200
–40°C
–40°C
+25°C
+85°C
+25°C
–5
+85°C
180
160
140
120
100
–10
–15
–20
–25
–30
–35
–40
–45
–50
80
60
40
7
8
9
10
11
12
13
14
15
16
6
7
8
9
10
11
12
13
14
15
16
CARRIER FREQUENCY (GHz)
CARRIER FREQUENCY (GHz)
Figure 10. Tuning Sensitivity vs. Carrier Frequency, VT = 1.65 V, Maximum Output
Gain (VCO Register 0x01, Bits[1:0] = 3 (Decimal)), High Performance Mode
(VCO Register 0x01, Bits[4:2] = 7 (Decimal))
Figure 13. Third Harmonic vs. Carrier Frequency, Maximum Output Gain
(VCO Register 0x01, Bits[1:0] = 3 (Decimal)), High Performance Mode
(VCO Register 0x01, Bits[4:2] = 7 (Decimal))
0
0
–10
–20
–30
–40
–50
–60
–40°C
+25°C
–5
–10
–15
–20
–25
–30
–35
–40
–45
+85°C
7
8
9
10
11
12
13
14
15
16
7
8
9
10
11
12
13
14
15
CARRIER FREQUENCY (GHz)
CARRIER FREQUENCY (GHz)
Figure 11. 0.5 Harmonic vs. Carrier Frequency (Fundamental Feedthrough at
RFOUT), Maximum Output Gain = (VCO Register 0x01, Bits[1:0] = 3 (Decimal)),
High Performance Mode (VCO Register 0x01, Bits[4:2] = 7 (Decimal))
Figure 14. Integer Boundary Spurs (IBS) vs. Carrier Frequency, 10 kHz Spur,
100 kHz Loop, 50 MHz PFD, Maximum Output Gain (VCO Register 0x01, Bits[1:0]
= 3 (Decimal)), High Performance Mode (VCO Register 0x01, Bits[4:2] =
7 (Decimal)), Low Current Mode (VCO Register 0x01, Bits[4:2] = 0 (Decimal))
Rev. 0 | Page 14 of 51
Data Sheet
ADF5610
0
–10
–20
–30
–40
–50
–60
–70
–60
–70
HIGH PERFORMANCE MODE, 25°C
HIGH PERFORMANCE MODE,10kHz
LOW CURRENT MODE,10kHz
–80
–90
–100
–110
–120
6.5
7.5
8.5
9.5
10.5 11.5 12.5 13.5 14.5 15.5
6.5
7.5
8.5
9.5
10.5 11.5 12.5 13.5 14.5 15.5
VCO BAND
CARRIER FREQUENCY (GHz)
Figure 17. Reference Spurs vs. Carrier Frequency, 100 MHz PFD Frequency,
Maximum Output Gain (VCO Register 0x01, Bits[1:0] = 3 (Decimal)), High
Performance Mode (VCO Register 0x01, Bits[4:2] = 7 (Decimal))
Figure 15. Integer Boundary Spur vs. VCO Band, 20 kHz Spur, 100 kHz Loop,
50 MHz PFD, Maximum Output Gain (VCO Register 0x01, Bits[1:0] = 3 (Decimal)),
High Performance Mode (VCO Register 0x01, Bits[4:2] = 7 (Decimal)), Low Current
Mode (VCO Register 0x01, Bits[4:2] = 0 (Decimal)
–60
HIGH PERFORMANCE MODE, 25°C
–70
–80
–90
–100
–110
–120
6.5
7.5
8.5
9.5
10.5 11.5 12.5 13.5 14.5 15.5
CARRIER FREQUENCY (GHz)
Figure 16. Phase Frequency Detector Spurs vs. Carrier Frequency, 100 MHz PFD
Frequency, Maximum Output Gain (VCO Register 0x01, Bits[1:0] = 3 (Decimal)),
High Performance Mode (VCO Register 0x01, Bits[4:2] = 7 (Decimal))
Rev. 0 | Page 15 of 51
ADF5610
Data Sheet
THEORY OF OPERATION
SEN
SDI
4
DIVIDER
BUFFER
CONTROL
DIVIDER
SCK
DIV PWR CNTL
PDIV_OUT
NDIV_OUT
DIVIDER EN
÷1, ÷2, ÷4, ÷8, ÷16,
÷32, ÷64, ÷128
SDO
3
MODULATOR
CAL
VSPI
CHARGE
PUMP
CP
VSPI
VT SELECT
VCO EN
RFOUT
BIAS CNTL
VCO EN
RFOUT PWR CNTL
VCO BUFER
N
PHASE
FREQUENCY
DETECTOP
DIVIDER
R
XREFP
CEN
DIVIDER
VT
PLL BUFF
PLL ONLY
Figure 18. PLL and VCO Subsystems
The ADF5610 PLL with integrated VCO is composed of two
subsystems, a PLL subsystem and a VCO subsystem, as shown
in Figure 18.
VCO SUBSYSTEM OVERVIEW
The VCO subsystem consists of a switched capacitor, step tuned
VCO, and an output stage. The topology of the VCO allows both
fundamental and doubled frequency outputs. This arrangement
allows the lower, fundamental frequency of the oscillator to be
routed to the input of the PLL, which reduces the N counter of
the PLL by a factor of 2. This reduction of the N counter results
in not only a 3 dB improvement in the close in-phase noise, but
also prevents residual phase noise degradation at offsets beyond
10 MHz.
PLL SUBSYSTEM OVERVIEW
The PLL subsystem divides down the VCO output to the desired
comparison frequency via the N divider (integer value set in
Register 0x03, fractional value set in Register 0x04), compares
the divided VCO signal to the divided reference signal (reference
divider set in Register 0x02) in the phase detector (PD), and
drives the VCO tuning voltage via the charge pump (CP)
(configured in Register 0x09) to the VCO subsystem. Some of
the additional PLL subsystem functions include
The VCO tuning is a two-step process consisting of a coarse then
fine tune. During normal operation, autocalibration is enabled
(Register 0x0A, Bit 11), which allows the PLL finite state
machine (FSM) to perform a binary search of the VCO bands
(coarse tune). After the proper band is located, the charge pump
output from the PLL takes control of the tuning port (VT) on
the VCO and adjusts to the proper fundamental frequency
(3650 MHz to 7300 MHz), and is phase locked by the PLL.
Δ-Σ configuration (Register 0x06).
Exact frequency mode (configured in Register 0x0C,
Register 0x03, and Register 0x04).
LD configuration (use Register 0x07 to configure LD and
Register 0x0F to configure the SDO output pin).
External CEN pin used for the hardware PLL enable pin.
The CEN pin does not affect the VCO subsystem.
The VCO subsystem controls the output stages of the ADF5610,
enabling configuration of the following:
Only writes to the divider registers of the PLL subsystem are
required for ADF5610 output frequency changes.
User defined performance settings (see the Programmable
Performance Technology section) that are configured via
VCO Register 0x01, Bits[4:2].
The divider registers of the PLL subsystem (Register 0x03 and
Register 0x04) set the fundamental frequency (3650 MHz to
7300 MHz) of the VCO subsystem. Output frequencies ranging
from 57 MHz to 14,600 MHz are generated by tuning to the
appropriate fundamental VCO frequency (3650 MHz to
7300 MHz), by programming the N divider (Register 0x03 and
Register 0x04), and programming the output divider (divide by
1 to 128, in VCO Register 0x02) in the VCO subsystem.
Depending on the step size, high performance applications may
require one or more additional register writes to optimize spurs,
loop bandwidth and settling time, or to set the lock detect
window at the new frequency.
VCO output divider settings that are configured in VCO
Register 0x02, Bits[2:0] (divide by 2 to 128 to generate
frequencies from 3650 MHz to 57 MHz, respectively) or
divide by 1 to generate frequencies between 7300 MHz and
14,600 MHz.
RFOUT gain settings (VCO Register 0x01, Bits[1:0]).
PDIV_OUT and NDIV_OUT gain settings (VCO
Register 0x01, Bits[6:5]).
Power down VCO (VCO Register 0x02, Bit 3).
Power down divider (VCO Register 0x01, Bit 8).
For detailed frequency tuning information and an example, see
the Frequency Tuning section.
Rev. 0 | Page 16 of 51
Data Sheet
ADF5610
When concatenated, the result is Register 0x05 = 0x0310.
SPI CONFIGURATION OF PLL AND VCO
SUBSYSTEMS
During autocalibration, the autocalibration controller writes
into the VCO register address specified by the VCO_ID and
VCO_REGADDR, as stored in Register 0x05, Bits[2:0] and
Register 0x05, Bits[6:3], respectively. Autocalibration requires
that the values for VCO_ID be zero (Register 0x05, Bits[2:0] =
0). When the VCO_ID values are not zero (Register 0x05,
Bits[2:0] ≠ 0), autocalibration does not function. Because VCO
Register 0x00 is the register that manages the VCO bands and
frequencies, Register 0x05, Bits[6:0] = 0.
The two subsystems (PLL and VCO) each have their own
register maps, as shown in the Serial Port Interface section and
the VCO Subsystem Register Map section. Typically, writes to
both register maps are required for initialization and frequency
tuning operations.
As shown in Figure 18, the PLL subsystem is connected directly
to the SPI interface of the ADF5610, whereas VCO SPI commands
are passed through the PLL subsystem to the ADF5610 VCO. As
a result, writes to the PLL register map are written directly and
immediately, whereas the writes to the VCO subsystem register
map are written to PLL Register 0x05 and forwarded via the
internal VCO SPI (VSPI) to the VCO subsystem, which is a
form of indirect addressing.
To ensure that frequency changes can occur after configuring the
VCO subsystem settings in Register 0x05, it is critical to write
Register 0x05, Bits[6:0] = 0 after the last VCO subsystem write,
but prior to the output frequency change. The frequency change
is triggered by the final PLL register write, which is Register 0x03
for integer values or Register 0x04 for fractional frequencies.
Register 0x03 must be the last write in integer mode because it
triggers autocalibration to implement a frequency change. When
operating in fractional mode, Register 0x04 is written last to trig-
ger autocalibration to make a frequency change. If the last write
to Register 0x05, Bits[6:0] ≠ 0), autocalibration does not function.
The VCO registers are write only and cannot be read. However,
the VCO tuning band that is currently enabled can be read back
via PLL Register 0x0A. For more information, see the VCO
Serial Port Interface (VSPI) section.
VCO Serial Port Interface (VSPI)
The ADF5610 communicates with the internal VCO subsystem
via an internal 16-bit VSPI. The internal serial port controls the
step tuned VCO and other VCO subsystem functions.
Note that it is impossible to write only Register 0x05, Bits[6:0] =
0 (VCO_ID and VCO_REGADDR) without also writing to
VCO_DATA (Register 0x05, Bits[15:7]). Therefore, if it is
desired to remain in the existing VCO band and only change
the divide ratio (see previous example), the user must ensure
that VCO_DATA (Register 0x05, Bits[15:7]) is not changed. To
accomplish this, it is required to read the band settings provided
in Register 0x10, Bits[7:0], and then rewrite them to Register 0x05,
Bits[15:7] as follows:
The internal VSPI runs at the rate of the autocalibration FSM
clock (tFSM) (see the VCO Autocalibration section), where the
FSM clock frequency cannot be greater than 50 MHz. The VSPI
clock rate is set by PLL Register 0x0A, Bits[14:13].
Writes to the control registers of the VCO are handled indi-
rectly via writes to Register 0x05 of the ADF5610. A write to
Register 0x05 causes the internal PLL subsystem to forward the
packet, MSB first, across its internal serial link to the VCO
subsystem, where the packet is interpreted.
1. Read Register 0x10.
2. Write the following to Register 0x05:
a. Register 0x05, Bits[15:14] = Register 0x10, Bits[7:6].
b. Register 0x05, Bit 13 = 1 (reserved bit).
c. Register 0x05, Bits[12:8] = Register 0x10, Bits[4:0].
d. Register 0x05, Bits[7:0] = 0.
VSPI Use of Register 0x05
The packet data written into Register 0x05 is subparsed by logic
at the VCO subsystem into the following three fields:
Changing the VCO subsystem configuration (see the VCO
Subsystem Register Map section) without following the preceding
procedure results in a failure to lock to the desired frequency.
•
•
•
Field 1, Bits[2:0]: 3-bit VCO_ID, target subsystem address =
000b.
Field 2, Bits[6:3]: 4-bit VCO_REGADDR, the internal
register address inside the VCO subsystem.
Field 3, Bits[15:7]: 9-bit VCO_DATA, the data field to
write to the VCO register.
For applications not using the read functionality of the ADF5610
SPI, in which Register 0x10 cannot be read, it is possible to write
Register 0x05 = 0 to set Register 0x05, Bits[6:0] = 0, which also
sets the VCO subband setting equal to zero (Register 0x05,
Bits[15:7] = 0), effectively programming an incorrect VCO
subband settings and causing the ADF5610 to lose lock. This
procedure is then immediately followed by a write to the following:
For example, to write 0 0000 0110 into VCO Register 0x02 of
the VCO subsystem (VCO_ID = 000b) and set the VCO output
divider to divide by 2, the following must be written to
Register 0x05.
•
•
Register 0x03, if in integer mode
Register 0x04, if in fractional mode
•
•
•
•
•
0b (VCO_DATA, Bit 15)
0000b (VCO_DATA, Bits[14:11])
0110b (VCO_DATA, Bits[10:7])
0010b (VCO_REGADDR)
000b (VCO_ID)
This write effectively retriggers the FSM, forcing the ADF5610
to relock whether in integer mode or fractional mode.
Rev. 0 | Page 17 of 51
ADF5610
Data Sheet
Lock times within 10° of phase settling as low as 30 μs can be
achieved, but are dependent on both the loop filter design (loop
filter bandwidth and loop filter phase margin) and the
VCO SUBSYSTEM
See Figure 19 and Figure 20.
configuration of the autocalibration register (Register 0x0A).
For applications where lock time is critical, it is recommended
that lock time be verified at the operating temperature extremes.
SPI
(SCK, SEN, SDI)
SDO
VCO REGISTER
0x01[4:2]
VCO SUBSYSTEM
PERFORMANCE
TUNING
VSPI
VCO
CONTROL
PDIV_OUT
PLL
REGISTER
0x0A[11]
VCO REGISTER
0x01[8]
÷1, ÷2, ÷4, ÷8, ... ÷128
EN
NDIV_OUT
VCO REGISTER 0x02[2:0]
VCO REGISTER 0x01[6:5]
CONTROL CAL
MODULATOR
VCO REGISTER 0x01[8:7]
RFOUT
VT
VCO
VCO REGISTER
0x00[0]
VCO CAL
VOLTAGE
VCO REGISTER 0x01[1:0]
VCO REGISTER 0x02[3], EN
N
DIVIDER
VPRST
VG
VCO REGISTER 0x00[8:1]
PHASE
FREQUENCY
DETECTOR
LOOP
FILTER
CP
CHARGE
PUMP
XREFP
R
DIVIDER
Figure 19. Detail of PLL and VCO Subsystems
VSCK
SCK
SDI
DIGITAL TUNING
VCO
SUBBAND
VSDO
VSLE
VCO
VSPI
SELECT
HOST
SEN
SYNTHESIZER
VT
CP
LOOP
FILTER
VCO INPUT
RF
OUT
VCO
Figure 20. Simplified Step Tuned VCO
Rev. 0 | Page 18 of 51
Data Sheet
ADF5610
The ADF5610 contains a VCO subsystem that can be
configured to provide the following output frequencies:
lock results in a phase hit during this transition period. Phase
hits are unacceptable for many applications. If a different band is
selected upon relock, the tuning sensitivity changes slightly,
resulting in a minor shift to the 3 dB corner for the loop filter. A
different band results in a minor change to phase settling time
and spurious suppression. If the loss of lock was initially due to
selection of a band where the frequency is near the band edge,
the accompanying tuning voltage may be close to the charge
pump rail. As the charge pump approaches its limits of
operation (rails), its performance becomes nonlinear, resulting
in increased spurious and PLL FOM and consequently degrades
phase noise performance at frequency offsets inside the loop
filter bandwidth.
•
Fundamental frequency (fO) mode (3650 MHz to
7300 MHz) is available at the differential outputs by setting
the output frequency divider to N = 2.
•
The push/push (doubled) frequency is always available at
RFOUT. The push/push frequency is also available at the
differential outputs by setting N = 1 to bypass the divider.
Divide by N mode, where N = 1, 2, 4, 6, 8 … 32, 64, 128
(57 MHz to 1750 MHz).
All modes are VCO register programmable, as shown in
Figure 19. One loop filter design can be used for the entire
frequency of operation of the ADF5610.
•
•
The ADF5610 simultaneously solves these potential issues by
integrating a temperature compensation circuit within the
autocalibration routine. This circuit produces a very low noise,
temperature dependent voltage, and depending on the state of
the autocalibration bit (VCO Register 0x00, Bit 7), routes the
voltage to the tuning port of the VCO during the autocalibration
process. When the operating temperature of the ADF5610 is
low (−40°C), a band where the desired frequency occurs at a
low tune voltage is required so that as the temperature increases,
higher tune voltages are available. Likewise, when the ADF5610
is operating at high temperatures (85°C), a band where the
desired frequency falls at a higher tune voltage is needed to
maintain lock as the temperature drops. The reference voltage
used for band selection ranges linearly from about 0.85 V dc at
−40°C to 1.75 V dc at +85°C. The voltage range, which is
centered around 1.3 V dc at room temperature, is offset with
respect to the available tune voltage range of 0 V dc to 3.3 V dc.
This offset allows additional voltage at the upper end of each
VCO core to compensate for the reduction in tuning sensitivity.
By limiting the voltage range from approximately 0.85 V dc to
1.70 V dc, bands are selected that allow operation away from
the charge pump rails over the full operating temperature range.
Band selection where tuning voltage falls from 0.85 V dc to
1.70 V dc allows phase lock using a single subband over
temperature extremes (lock and leave) to be met as long as
sufficient charge pump voltage exists to compensate for the
frequency drift. Some applications may require increasing the
charge pump supply (VPPCP) and the charge pump digital
supply level shifter (VDDLS) to 3.6 V dc to meet lock and leave
requirements. Note that the voltage for the VPPCP and VDDLS
must be equal. Phase lock is maintained within a single VCO
core and VCO subband over the full temperature range (−40°C
to +85°C) regardless of the operating temperature when
autocalibration initially selected the subband.
VCO Calibration
VCO Autocalibration
The ADF5610 incorporates a step tuned VCO. A simplified step
tuned VCO is shown in Figure 20. Step tuned VCOs are unique
because they incorporate not only varactor diodes for frequency
tuning, but also a digitally selectable, binary bank of capacitors.
This bank of capacitors, along with the varactor diodes, defines
the frequency range of each subband within a given VCO core.
The ADF5610 contains two VCO cores, each with their own
bank of capacitors. Multiple VCO cores allow not only broader
frequency coverage, but also more consistent tuning sensitivity
across the band.
As capacitors are selected or deselected within the bank, the
nominal center frequency of the VCO tank circuit is stepped up
or stepped down. Multiple capacitors can be changed at once or
one at a time if required. Essentially, the capacitor banks
provide coarse frequency tuning in the form of 128 subbands
within each VCO core while the varactor diodes manage fine
tuning within each of these subbands. To guarantee continuous
frequency coverage over process, voltage, and temperature (PVT),
the ends of each frequency subband and each VCO core overlap.
The frequency overlap implies that certain frequencies exist in
more than one subband. It is known that the frequency of a VCO
varies (drifts) with temperature. Selection of a subband where
the desired frequency is too close to the edge may lead to loss of
lock if the operating temperature change is large enough and in
a direction that allows the frequency to drift beyond the edge of
the subband. During normal operation, the ADF5610 is designed
to remain locked and in the same band as long as the temperature
remains within the specified operating range. During manual
calibration (autocalibration disabled, refer to the Manual VCO
Calibration for Fast Frequency Hopping section), it is possible
to select a band where the desired frequency exists, but is too
close to the transition point where multiple bands overlap.
Having the frequency too close to the transition point may
result in the selection of a different band as the operating
temperature changes. If automatic relock is enabled (Register 0x07,
Bit 13 = 1), the autocalibration routine runs one additional time
in an attempt to relock the synthesizer. This temporary loss of
Rev. 0 | Page 19 of 51
ADF5610
Data Sheet
3.3
3.0
2.7
2.4
2.1
1.8
1.5
1.2
0.9
0.6
0.3
0
When the ADF5610 PLL receives a request to change the
frequency such that a new VCO band must be selected, the
process is as follows. Note that the PLL portion of the
autocalibration circuitry, the FSM, is assumed to already be
enabled (Register 0x0A, Bit 11 = 0).
–40°C
+85°C
1. The PLL registers that must be updated to facilitate lock at
the new frequency as well as any changes to the VCO
registers (gain settings, divide ratio) are sent to the
ADF5610 via the SPI. There is no need for users to change
the VCO autocalibration bit when changing frequencies.
The VCO autocalibration bit state is handled automatically
when PLL Register 0x0A, Bit 11 = 0.
2. The FSM takes control of the VCO autocalibration bit
(VCO Register 0x00, Bit 7), changing the bit from 0 to 1,
temporarily opening the loop filter path and switching the
source of the VT pin from the charge pump to the
VPRST pin.
3. With VT set to a value that provides optimal performance,
the PLL FSM begins a binary search for the VCO core and
subband with a frequency that is closest to the desired
frequency. Note that the setting for the VT resolution
(Register 0x0A, Bits[2:0]) can impact band selection as well
as the ability to retain lock over the full operating
temperature range. When in doubt, use the highest setting
for the best resolution.
4. After the proper band has been determined, the VCO
autocalibration bit (VCO Register 0x00, Bit 7) changes
from 1 back to 0, closing the loop filter path and returning
the source of the VT pin to the charge pump.
6.9 7.6 8.3 9.1 9.8 10.6 11.4 12.1 12.9 13.6 14.4 15.2
CARRIER FREQUENCY (GHz)
Figure 21. Tuning Voltage vs. Carrier Frequency After Calibration, Maximum
Tuning Resolution, Register 0x0A, Bits[2:0] = 7
Although the digital VCO bands are automatically controlled by
the ADF5610 using the autocalibration feature, the VCO bands
can also be controlled directly via Register 0x05 for testing or
for special purpose operations. Other control bits specific to the
VCO are also sent via Register 0x05. Refer to the VCO
Subsystem Register Map section for further details.
As mentioned in the VCO Subsystem Overview section, during
autocalibration, coarse tuning is provided by the FSM as it
selects the most appropriate band for operation over the full
operating temperature range. Fine tuning is achieved by
varactor diodes after the VCO autocalibration bit (VCO
Register 0x00, Bit 7) is reset by the FSM. The reset gives control
of the tuning voltage back to the phase detector and charge pump.
When the band is selected, a narrow voltage range on the varactor
is all that is needed for phase lock. Note that the tuning voltage
stays in a narrow range over a wide range of output frequencies.
5. The new VCO frequency is then routed to the input of the
PLL, divided down, and compared to the reference
frequency. The charge pump adjusts the VCO tuning
voltage as necessary to phase lock the divided down VCO
frequency to the reference frequency.
The calibration is normally run automatically, once for every
change of frequency. Autocalibration ensures optimum
selection of VCO band settings vs. time and temperature. The
user does not need to be concerned about which band setting is
used for a given frequency because this selection is handled by
the autocalibration routine.
The temperature compensated voltage that is used during
autocalibration is transferred to the VPRST pin (Pin 35) of the
ADF5610. This voltage correlates to the temperature of the
silicon and can be used for a real-time estimate of the junction
temperature of the die. The voltage at VPRST is a function of
the ambient temperature as well as the PCB design and
assembly quality. Take caution if it is desired to continuously
monitor this voltage. The VPRST pin is tied to a low noise
circuit, and although the internal switch has sufficient isolation,
it may still allow external noise to be coupled onto the precision
reference circuit, which results in unwanted spurs, sidebands,
and degraded phase noise.
The accuracy required in the calibration affects the amount of
time required to tune the VCO. The calibration routine searches
for the step setting where the desired frequency falls within the
frequency range (determined by the FSM resolution setting)
that is centered around the temperature compensated VPRST
value being used during the search. When located, the loop
filter path is closed and the VCO is locked at the programmed
frequency. If the proper resolution is used, the VCO remains
locked and operates away from the charge pump rails over its
full temperature range without additional calibration, regardless
of the temperature at which the VCO was calibrated. Note that
when autocalibration is enabled, preselection of the VCO core
that autocalibration starts in is not possible.
Rev. 0 | Page 20 of 51
Data Sheet
ADF5610
Autocalibration can be disabled, thereby allowing manual band
selection and even faster VCO tuning. Refer to the Manual
VCO Calibration for Fast Frequency Hopping section for more
information about manual tuning.
where:
N
N
INT is the integer set in Register 0x03.
FRAC is the fractional part set in Register 0x04.
The autocalibration FSM and the data transfers to the internal
VCO VSPI run at the rate of the FSM clock (tFSM) where the
FSM clock frequency must not be greater than 50 MHz.
Autocalibration Using Register 0x05
Autocalibration transfers band control data to the VCO
subsystem via Register 0x05. The address of the VCO subsystem
in Register 0x05 is not altered by the autocalibration routine.
The address and ID of the VCO subsystem in Register 0x05
must be set to the correct value before autocalibration is
executed. For more information, see the VCO Serial Port
Interface (VSPI) section.
t
FSM = tXTAL × 2m
(3)
where m is 0, 2, 4, or 5 as determined by Register 0x0A, Bits[14:13].
The VSPI and FSM clock rate are determined by the value of m.
It is recommended to set Register 0x0A, Bits[14:13] = 2.
The expected number of VCO counts is given by
VCO Counts = floor(N × 2n)
(4)
(5)
(6)
Automatic Relock on Lock Detect Failure
The nominal VCO frequency measured (fVCOM) is given by
It is possible, by setting Register 0x07, Bit 13, to have the VCO
subsystem automatically rerun the calibration routine and
relock itself if the lock detect bit indicates an unlocked condition
for any reason. With this option, the system attempts to relock
only once.
f
VCOM = VCO Counts × fXTAL/(2n × R)
where the worst case measurement error (fERR) is
ERR ≈ fPD/2n + 1
For example, the autocalibration time (tCAL) for an 8-bit step
f
VCO Autocalibration on Frequency Change
tuned VCO (where the total number of bits includes those needed
to switch VCO cores) is as follows. First, a 20 × (k + 1) bit wait
state occurs. Register 0x0A, Bits[9:3] set the value of the wait
state (k) for wideband devices to k = 1. Wideband VCOs like
the ADF5610 require additional wait time or one additional PD
cycle (k = 1), whereas narrowband VCOs operate with no
additional wait time (k = 0). The wait period is followed by the
Assuming Register 0x0A, Bit 11 = 0, the VCO calibration starts
automatically whenever a frequency change is requested. To
rerun the autocalibration routine for any reason at the same
frequency, rewrite the frequency change with the same value,
and the autocalibration routine executes again without
changing the final frequency.
t
MMT for calibration, then the VSPI data transfers of 20 clock
VCO Autocalibration Time and Accuracy
cycles each. This process repeats eight times because there is
one bit to select the VCO core and seven bits to select the VCO
band for a total of eight bits. Total calibration time, worst case,
is given by the following equation:
The VCO frequency is counted for the period of a single
autocalibration measurement cycle (tMMT).
t
MMT = tXTAL × R × 2n
where:
XTAL is the period of the external reference (crystal) oscillator.
(1)
t
CAL = tXTAL (8R × 2n + (160 + (128 × (k + 1)) × 2m))
(8)
t
For guaranteed lock retention across temperature extremes
using a single subband (lock and leave), the frequency error
must be less than 1/8 of the frequency step required to induce a
R is the reference path division ratio currently in use, set in
Register 0x02.
n is set by Register 0x0A, Bits[2:0] and results in measurement
periods that are multiples of the PD period, tXTAL × R. The
frequency resolution used by the FSM during autocalibration is
dependent on the value of n and the PD frequency. Higher values
of n result in reduced frequency error (higher frequency
resolution) during autocalibration. The highest value is
currently recommended.
VCO subband change.
VCO Autocalibration Example
The VCO subsystem must satisfy the maximum fPD limited by
the two following conditions:
N ≥ 16 (fINT), N ≥ 20.0 (fFRAC
where:
N = fVCO/fPD. fPD ≤ 100 MHz.
)
The VCO autocalibration counter, on average, expects to
register N counts, rounded down (floor) to the nearest integer,
for every PD cycle.
f
INT is integer mode.
f
FRAC is fractional-N mode. The minimum N values change
N is the ratio of the target VCO frequency (fVCO) to the
frequency of the PD (fPD) where N can be any rational number
supported by the N divider.
depending on the operating mode.
For example, if the VCO subsystem output frequency is to
operate at 7.01 GHz and the crystal frequency is fXTAL = 50 MHz,
R = 1, and m = 2 (see Figure 22), then tFSM = 80 ns (12.5 MHz).
N is set by the integer register and fractional register contents
using Equation 2.
When using autocalibration, do not allow the autocalibration
FSM clock to exceed 50 MHz (see Register 0x0A, Bits[14:13]).
The FSM clock does not affect the accuracy of the measurement.
N = NINT + NFRAC/224
(2)
Rev. 0 | Page 21 of 51
ADF5610
Data Sheet
The FSM clock only affects the time to produce the result. This
same time standard clocks the 16-bit VSPI.
Using an input crystal of 50 MHz (R = 1 and fPD = 50 MHz), the
times and accuracies for calibration using Equation 6 and
Equation 8 are listed in Table 9, where minimal tuning time is
1/8 of the VCO band spacing.
If the time to change frequencies is not a concern, use the longest
calibration time, because this provides the minimum frequency
error during band selection. The longest calibration time results
in the selection of bands where the desired frequency is as close
as possible to the recommended tuning voltage value as
determined by the temperature compensation circuit.
tPD
CALIBRATION WINDOW
REG 0x02
÷ R
n
tMMT
= tXTAL × R × 2
XREFP
n
÷ 2
START
STOP
REG 0x0A[14:13]
m = [0, 2, 4, 5]
REG 0x0A[2:0]
n = [0, 1, 2, 3, 5, 6, 7, 8]
m
÷ 2
50MHz MAX FOR
FSM + VSPI CLOCKS
V
CTR
VCO
FSM
Figure 22. VCO Calibration
Table 9. Autocalibration Example with fXTAL = 50 MHz, R = 1, m = 2
VCO Autocalibration Configuration Register 0x0A, Bits[2:0]
n
0
1
2
3
5
6
7
8
2n
1
2
4
tMMT (µs)
0.02
0.04
0.08
0.16
0.64
1.28
2.56
5.12
tCAL (µs)
33.44
33.60
33.92
34.56
38.40
43.52
53.76
74.24
fERR Maximum
25 MHz
12.5 MHz
6.25 MHz
3.125 MHz
781 kHz
390 kHz
95 kHz
98 kHz
0
1
2
3
4
5
6
7
8
32
64
128
256
Rev. 0 | Page 22 of 51
Data Sheet
ADF5610
For example, it is determined that across all VCO subbands, a
measurement resolution of at least 800 kHz results in the
selection of the correct frequency band. Setting m = 2 and n = 5
provides 781 kHz of resolution and adds 38.4 μs of autocalibration
time to a normal frequency hop. After autocalibration selects
the band (38.4 μs after the frequency change command in the
preceding example), the fractional register is loaded, and the
loop locks with the normal transient predicted by the loop
dynamics. This means that, as shown in this example,
the fractional frequency change is loaded immediately into
the modulator when the register is written with no
adjustment to the VCO. Write this register last for fractional
frequencies because it triggers the frequency change.
The VCO SPI register, Register 0x05. This write is required
only for manual control of VCO if Register 0x0A, Bit 11 = 1,
autocalibration is disabled, or to enable or change the VCO
output divider value, VCO bias, or the divider power level
or RFOUT power level. See Figure 19 for more information.
•
autocalibration adds 38.4 μs to the normal time to achieve
frequency lock. Autocalibration is recommended for all but the
most extreme frequency hopping requirements.
Small steps in frequency in fractional mode, with autocalibration
enabled (Register 0x0A, Bit 11 = 0), usually require only a single
write to the fractional register. In a worst case scenario, three
main serial port transfers to the ADF5610 may be required to
change frequencies in fractional mode. If the frequency step is
small such that the integer part of the frequency does not change,
the integer register is not changed. In all cases, in fractional mode,
it is necessary to write to the fractional register, Register 0x04,
for frequency changes.
Manual VCO Calibration for Fast Frequency Hopping
When switching frequencies quickly is needed, it is possible to
eliminate the autocalibration time by calibrating the VCO in
advance and storing the step number vs. frequency information
in the host, which is accomplished by initially locking the
ADF5610 on each desired frequency using autocalibration, then
reading and storing the selected VCO band settings. The VCO
band settings are available in Register 0x10, Bits[7:0] after every
autocalibration operation. The host must then program the
VCO band settings directly when changing frequencies.
Registers Required for Frequency Changes in Integer
Mode
In integer mode (Register 0x06, Bit 11 = 0), a change of frequency
requires main serial port writes to the following registers:
Manual writes to the VCO bands are executed immediately as are
writes to the integer and fractional registers when autocalibration
is disabled. Therefore, frequency changes with manual control
and autocalibration disabled requires a minimum of two serial
port transfers to the PLL, once to select the VCO band and once
to set the PLL frequency.
•
VCO SPI register, Register 0x05. This write is required only
for manual control of the VCO when Register 0x0A, Bit 11
= 1 (autocalibration disabled) or when the VCO configura-
tion or the output frequency divider value must change
(VCO Register 0x02).
•
Integer register, Register 0x03. In integer mode, an integer
register write triggers autocalibration when Register 0x0A,
Bit 11 = 0 and the integer frequency change is loaded into
the prescaler automatically after autocalibration runs. If
autocalibration is disabled (Register 0x0A, Bit 11 = 1), the
integer frequency change is loaded into the prescaler
immediately when written with no adjustment to the VCO.
Normally, changes to the integer register cause large steps
in the VCO frequency. Therefore, the VCO band settings
must be adjusted. Autocalibration enabled is the recom-
mended method for integer mode frequency changes. If
autocalibration is disabled (Register 0x0A, Bit 11 = 1), a
prior knowledge of the correct VCO band setting and the
corresponding adjustment to the VCO is required before
executing the integer frequency change. Failure to select
the proper band may result in loss of lock at temperature
extremes or degraded phase noise performance.
When autocalibration is disabled (Register 0x0A, Bit 11 = 1), the
VCO updates its registers immediately with the value written
via Register 0x05. The VCO internal transfer requires 16 VCO
SPI clock (VSCK) cycles after the completion of a write to
Register 0x05. VSCK and the autocalibration controller clock
are equal to the input reference divided by 0, 4, 16, or 32, as
controlled by Register 0x0A, Bits[14:13].
For settling time requirements faster than 50 µs, contact Analog
Devices, Inc., applications support. Settling times under 50 µs
are possible, but certain limitations on performance may exist.
Registers Required for Frequency Changes in Fractional
Mode
In fractional mode (Register 0x06, Bit 11 = 1), a large change of
frequency may require main serial port writes to one of the
three following registers:
•
The integer register, Register 0x03. This write is required
only if the integer part changes. For integer frequencies,
write this register last because it triggers the frequency
change.
The fractional register, Register 0x04. The fractional
register write triggers autocalibration when Register 0x0A,
Bit 11 = 0, and the fractional frequency change is loaded
into the modulator automatically after the autocalibration
runs. If autocalibration is disabled (Register 0x0A, Bit 11 = 1),
VCO Built In Self Test (BIST) with Autocalibration
The frequency limits of the VCO can be measured using the BIST
features of the autocalibration FSM by setting Register 0x0A,
Bit 10 = 1, which freezes the VCO bands in one position. VCO
bands can then be written manually with the varactor biased at
the nominal midrail voltage used for autocalibration. For example,
to measure the VCO maximum frequency, use Subband 0 (see
•
Rev. 0 | Page 23 of 51
ADF5610
Data Sheet
Table 35), written to the VCO subsystem via Register 0x05 =
000000001 0000 VCO_ID, where VCO_ID = 000b.
Charge Pump Gain
Charge pump up and charge pump down gains are set by
Register 0x09, Bits[13:7] and Register 0x09, Bits[6:0], respectively.
The current gain of the pump in amps per radian is equal to the
gain setting of this register (Register 0x09) divided by 2π.
When autocalibration is enabled (Register 0x0A, Bit 11 = 0),
and a new frequency is written, autocalibration runs. The VCO
frequency error relative to the command frequency is measured
and the results are written to Register 0x11, Bits[19:0], where
Register 0x11, Bit 19 is the sign bit. The result is written in
terms of VCO count error (see Equation 4 through Equation 6).
The typical CP gain setting is set from 2 mA to 2.5 mA.
However, lower values can also be used. Note that values less
than 1 mA may result in degraded phase noise performance.
Using generic values as an example, if the expected frequency at
the RFOUT pin is 7.0 GHz, the PFD frequency is 50 MHz, and
Register 0x0A, Bits[2:0] = 8 (n = 8), expect to measure
For example, if both Register 0x09, Bits[13:7] and Register 0x09,
Bits[6:0] are set to 50 decimal, the output current of each pump
is 1 mA, and the phase frequency detector gain is kP = 1 mA/2π
radians, or 159 μA/rad. See the Charge Pump (CP) and Phase
Detector (PD) section for more information. Setting both the
charge pump up (Register 0x09, Bits[13:7] = 0) and charge pump
down (Register 0x09, Bits[6:0] = 0) tristates the charge pump.
3500/(50/28) = 17,920 counts. Frequencies less than 8.0 GHz
route the fundamental frequency directly to the N counter.
Therefore, in this case, 3500 MHz is routed to the N counter.
For RFOUT frequencies greater than or equal to 8.0 GHz, the
internal prescaler of the PLL (Register 0x08, Bit 19) must be
enabled. If a difference of −5 counts is measured in Register 0x11,
that means 17,915 counts were actually measured. With a 7 GHz
VCO, 50 MHz reference, and n = 8, one count is 195.3125 kHz.
Therefore, the actual frequency of the VCO is low in frequency by
5 × 195.3 kHz for a fundamental frequency of 3499.0234375 MHz
or a push/push frequency of 6998.04687 MHz at RFOUT.
Tristating can also be achieved using Register 0x0B. Refer to the
Phase Detector Functions section for more information.
Charge Pump Phase Offset
In integer mode, the phase detector operates with zero offset.
The divided reference signal and the divided VCO signal arrive
at the phase detector inputs at the same time. Integer mode does
not require any CP offset current. When operating in integer
mode, disable the CP offset in both directions (up and down)
by writing Register 0x09, Bits[22:21] = 00b, and set the CP offset
magnitude to zero by writing Register 0x09, Bits[20:14] = 0 decimal.
PLL SUBSYSTEM
Charge Pump (CP) and Phase Detector (PD)
The PD has two inputs, one from the reference path divider and
one from the RF path divider. When in lock, these two inputs
are at the same average frequency and are fixed at a constant
average phase offset with respect to each other. The frequency
of operation of the phase detector is fPD. Most formulas related
to step size, Δ-Σ modulation, and timers, are functions of the
operating frequency of the phase detector (fPD). The fPD is also
referred to as the comparison frequency of the phase detector.
In fractional mode, CP linearity is of paramount importance.
Any nonlinearity degrades phase noise and spurious performance.
These nonlinearities are eliminated by operating the PD with an
average phase offset, either positive or negative (either the
reference or the VCO edge always leads or arrives at the PD).
A programmable CP offset current source adds dc current to
the loop filter and creates the desired phase offset. Positive
current causes the VCO to lead, whereas negative current
causes the reference to lead. Applications requiring an active
loop filter use the up offset, or positive offset. The ADF5610
uses a passive loop filter. Therefore, a down offset, or negative
offset, is recommended. Note that there can be conditions
where the offset that is not typically recommended may provide
improved spurious suppression, but these are often over very
narrow bandwidths.
The phase detector compares the phase of the RF path signal
with that of the reference path signal and controls the charge
pump output current as a linear function of the phase difference
between the two signals. The output current varies linearly over
a full 2π radians ( 360°) of input phase difference.
Charge Pump
A simplified diagram of the charge pump is shown in Figure 24.
The charge pump consists of four programmable current sources.
There are two sources controlling the charge pump gain (up gain,
Register 0x09, Bits[13:7], and down gain, Register 0x09, Bits[6:0])
and two sources controlling the charge pump offset, where the
magnitude of the offset is set by Register 0x09, Bits[20:14] and
the direction is selected by Register 0x09, Bit 21 = 1 for up offset
and Register 0x09, Bit 22 = 1 for down offset.
The CP offset is controlled via Register 0x09. Increasing the
offset current causes the phase offset to scale from 0° to 360°.
The specific level of the charge pump offset current (Register 0x09,
Bits[20:14]) is calculated using Equation 9 and shown in Figure 23.
Required CP Offset = min ((4.3 × 10−9 × fPD × ICP), 0.25 × ICP) (9)
Charge pump gain is used at all times, whereas charge pump
offset is recommended for fractional mode only. Typically, the
charge pump up and charge pump down gain settings are set to
the same value (Register 0x09, Bits[13:7] = Register 0x09, Bits[6:0])
to minimize spurs.
where:
f
PD is the comparison frequency of the phase detector (Hz).
I
CP is the full-scale current setting (A) of the switching charge
pump (set in Register 0x09, Bits[6:0] and Register 0x09,
Bits[13:7]).
Rev. 0 | Page 24 of 51
Data Sheet
ADF5610
700
600
500
400
300
200
100
0
Do not allow the required CP offset current to exceed 25% of
the programmed CP current. It is recommended to enable the
up offset and disable the down offset by writing Register 0x09,
Bits[22:21] = 01b.
CP CURRENT = 2.5mA
CP CURRENT = 2mA
Operation with CP offset influences the required configuration
of the lock detect function. See the description of the lock
detect function in the Lock Detect section.
CP CURRENT = 1mA
0
20
40
60
80
100
PHASE DETECTOR FREQUENCY (MHz)
Figure 23. Recommended CP Offset Current vs. Phase Detector Frequency for
Typical CP Gain Currents, Calculated Using Equation 9
UP OFFSET REG 0x09[21]
0µA TO 635µA
UP GAIN
REG 0x09[13:7]
5µA STEP
REG 0x09[20:14]
0mA TO 2.54mA
20µA STEP
UP
DN
PD
REF PATH
VCO PATH
LOOP
FILTER
DOWN OFFSET REG 0x09[22]
0µA TO 635µA
DOWN GAIN
REG 0x09[6:0]
5µA STEP
REG 0x09[20:14]
0mA TO 2.54mA
20µA STEP
Figure 24. Charge Pump Gain and Offset Control
Rev. 0 | Page 25 of 51
ADF5610
Data Sheet
these regions, the device usually still operates, but with degraded
reference path phase noise performance.
Phase Detector Functions
Register 0x0B, the phase detector register, provides access to
special phase detector features.
When operating at 50 MHz, the input referred phase noise of
the PLL is between −148 dBc/Hz and −150 dBc/Hz at a 10 kHz
offset, depending on the mode of operation. To avoid degra-
dation of the PLL noise contribution, the input reference signal
must be 10 dB lower (−158 dBc/Hz to −160 dBc/Hz), or lower
at a 10 kHz offset, than this floor. Such low levels are only
necessary if the PLL is the dominant noise contributor and
these levels are required for the system goals.
Setting Register 0x0B, Bit 5 = 0 masks the PD up output, which
prevents the charge pump from pumping up.
Setting Register 0x0B, Bit 6 = 0 masks the PD down output,
which prevents the charge pump from pumping down.
Clearing both Register 0x0B, Bit 5 and Register 0x0B, Bit 6
tristates the charge pump while leaving all other functions
operating internally.
Reference Path R Divider
The CP_FORCE_UP (Register 0x0B, Bit 9 = 1) and CP_
FORCE_DOWN (Register 0x0B, Bit 10 = 1) bits allow the
charge pump to be forced up or down, respectively. These bits
force the VCO to the ends of the tuning range, which is useful
in testing the VCO or the continuity of the loop filter path.
The reference path R divider is based on a 14-bit counter and
can divide input signals by values from 1 to 16,383 and is
controlled via Register 0x02.
RF Path, N Divider
The main RF path divider is capable of average divide ratios
between 219 − 5 (524,283) and 20 in fractional mode, and between
219 − 1 (524,287) and 16 in integer mode. Normally, the VCO
frequency range divided by the minimum N divider value can
result in practical restrictions on the maximum usable PD
frequency. For example, a VCO operating at 1.5 GHz in fractional
mode with a minimum N divider value of 20 has a maximum
PD frequency of 75 MHz, which is not the case with the ADF5610.
Regardless of the mode of operation or the PD frequency, the
minimum N value never occurs as long as the specified
maximum phase detector frequency is respected. For RFOUT
frequencies less than 8 GHz, the minimum frequency at the
input to the PLL is 3650 MHz. Therefore, N is always greater
than the minimum regardless of the mode. When RFOUT is set
to 8.0 GHz, the input to the PLL is 4.0 GHz, and the internal
prescaler must be enabled (Register 0x08, Bit 19 = 1). Enabling
the internal prescaler results in the 4.0 GHz fundamental
frequency being divided down so that 2.0 GHz are supplied to
the PD input. Operating in integer mode allows PD rates up to
100 MHz. This results in an N value of 20, which is not a
problem. Operation in fractional mode limits the PD rate to
100 MHz, which results in the minimum value for N when
RFOUT equals 8.0 GHz, but no less than this minimum value.
Reference Input Stage
RVDD
AC COUPLE
XREFP
20Ω
80Ω
100Ω
V
BIAS
Figure 25. Reference Path Input Stage
The reference buffer provides the path from an external reference
source (generally crystal based) to the R divider and eventually
to the phase detector. The buffer has two modes of operation
controlled by Register 0x08, Bit 21. High gain (Register 0x08,
Bit 21 = 0) is recommended below 200 MHz, and high frequency
(Register 0x08, Bit 21 = 1) for 200 MHz to 350 MHz operation.
The buffer is internally dc biased with 100 Ω internal termination.
For a 50 Ω match, add an external 100 Ω resistor to ground,
followed by an ac coupling capacitor (impedance less than 1 Ω).
At low frequencies, a relatively square reference is recommended to
maintain a high input slew rate. At higher frequencies, use a
square or sinusoid. Table 10 shows the recommended operating
regions for different reference frequencies. If operating outside
Table 10. Reference Sensitivity1
Square Input, Slew > 0.5 V/ns
Sinusoidal Input
Recommended Swing (V p-p)
Recommended Power Range (dBm)
Reference Input
Frequency (MHz)
Recommended
Minimum
0.6
Maximum
2.5
Recommended
Minimum
Maximum
<10
10
25
50
100
150
200
200 to 350
Yes
Yes
Yes
Yes
No
No
No
No
8
No
No
15
15
15
12
8
0.6
2.5
0.6
2.5
Okay
Yes
Yes
Yes
Yes
Yes
0.6
2.5
6
5
4
3
Yes
0.6
2.5
Okay
Okay
No
0.9
2.5
1.2
2.5
No
No
5
10
1 Okay means the setting works. For example, a 150 MHz input square wave is sufficient but 100 MHz may provide improved performance.
Rev. 0 | Page 26 of 51
Data Sheet
ADF5610
Lock Detect
1
LDWindow (sec) =
(11)
2× fPD
The lock detect (LD) function verifies that the ADF5610 is
generating the desired frequency. LD is enabled by writing
Register 0x07, Bit 3 = 1. The ADF5610 provides an LD indicator
in the following two ways:
where:
f
I
PD is the comparison frequency of the phase detector.
CP_OFFSET is the charge pump offset current (Register 0x09,
Bits[20:14]).
CP is the full-scale current setting of the switching charge pump
•
As an output available on the SDO pin of the ADF5610
(configuration is required to use the SDO pin for LD
purposes). For more information, see the Configuring the
SDO Pin for LD Output section.
I
(Register 0x09, Bits[6:0] or Register 0x09, Bits[13:7]).
If the result provided by Equation 10 is equal to 10 ns, analog
LD can be used (Register 0x07, Bit 6 = 0). Otherwise, digital LD
is necessary (Register 0x07, Bit 6 = 1). Table 11 lists the required
Register 0x07 settings to appropriately program the digital LD
window size. From Table 11, select the closest value in the digital
LD window size columns to the ones calculated in Equation 10
and Equation 11, and program Register 0x07, Bits[11:10] and
Register 0x07, Bits[9:7] accordingly.
•
Reading from Register 0x12, Bit 1, where Bit 1 = 1 indicates
a locked condition and Bit 1 = 0 indicates an unlocked
condition.
The LD circuit expects the divided VCO edge and the divided
reference edge to appear at the PD within a user specified time
period (window), repeatedly. Either signal may arrive first. Only
the difference in arrival times is significant. The arrival of the
two edges within the designated window increments an internal
counter. When the count reaches and exceeds a user specified
value (Register 0x07, Bits[2:0]), the ADF5610 declares lock.
Digital Window Configuration Example
For this example, assume the device is in fractional mode, with
a 50 MHz PD and the following conditions:
Failure in registering the two edges in any one window resets
the counter and immediately declares an unlocked condition.
Lock is deemed to be reestablished when the counter reaches
the user specified value (Register 0x07, Bits[2:0]) again.
•
Charge pump gain of 2 mA (Register 0x09, Bits[13:7] =
0x64, Register 0x09, Bits[6:0] = 0x64)
Up offset (Register 0x09, Bits[22:21] = 01b)
Offset current magnitude of 400 μA (Register 0x09,
Bits[20:14] = 0x50)
•
•
The ADF5610 supports the following two lock detect modes:
•
Analog LD supports a fixed window size of 10 ns. Analog
LD mode is selected by writing Register 0x07, Bit 6 = 0.
Digital LD supports a user configurable window size,
programmed in Register 0x07, Bits[11:7]. Digital LD is
selected by writing Register 0x07, Bit 6 = 1.
Apply Equation 10 to calculate the required LD window size.
LD Window (sec) =
•
0.4×10−3(A)
1
+ 2.66×10−9(sec) +
50×106(Hz)×2×10−3(A)
50×106(Hz)
2
Lock Detect Configuration
=13.33 ns
Optimal spectral performance in fractional mode requires CP
current and CP offset current configuration, described in detail
in the Charge Pump (CP) and Phase Detector (PD) section.
Locate the Table 11 value that is closest to this result, which is, in
this case, 13.3 ≈ 13.33. To set the digital LD window size, program
Register 0x07, Bits[11:10] = 10b and Register 0x07, Bits[9:7] =
010b, according to Table 11. For a given operating point, there is
typically a correct solution for the lock detect window. However,
one solution does not fit all operating points. As observed from
Equation 10 and Equation 11, if the charge pump offset or PD
frequency is changed significantly, the lock detect window may
need to be adjusted.
The PD frequency and settings in the CP (Register 0x09) impact
the required LD window size in fractional mode of operation.
To function properly, the required lock detect window size is
provided by Equation 10 in fractional mode and Equation 11 in
integer mode.
LD Window sec =
(
)
I
CP _OFFSET (A)
1
+ 2.66×10−9(sec) +
(10)
fPD(Hz)× ICP (A)
f
PD (Hz)
2
Table 11. Typical Digital Lock Detect Window for Various LD Timer Divide Settings, Register 0x07, Bits[9:7]
Digital Lock Detect Window Size Nominal Value (ns)
LD Timer Speed, Register 0x07, Bits[11:10]
000
6.5
7
7.1
7.6
001
8
8.9
9.2
10.2
010
11
12.8
13.3
15.4
011
17
21
22
26
100
29
36
38
47
101
53
68
72
88
110
100
130
138
172
111
195
255
272
338
00 (Fastest)
01
10
11 (Slowest)
Rev. 0 | Page 27 of 51
ADF5610
Data Sheet
The fundamental band of operation (3650 MHz to 7300 MHz)
is input to the internal PLL and is not directly available to the
user. For more information, see the VCO Autocalibration section.
Configuring the SDO Pin for LD Output
Setting Register 0x0F, Bit 7 = 1 and Register 0x0F, Bits[4:0] = 1
displays the lock detect flag on the SDO pin of the ADF5610.
When locked, SDO is high. As the name suggests, the SDO pin
is multiplexed between the LD and the serial data output (SDO)
signals. Therefore, LD is available on the SDO pin at all times
except when a serial port read is requested, in which case the
pin reverts temporarily to the serial data output pin and returns
to the lock detect flag after the read is completed.
Integer Mode
The ADF5610 is capable of operating in integer mode. For
integer mode, set the following registers:
•
•
Disable the fractional modulator, Register 0x06, Bit 11 = 0
Bypass the modulator circuit, Register 0x06, Bit 7 = 1
In integer mode, the VCO step size is fixed to that of the PD
frequency. Integer mode typically has a 3 dB lower phase noise
than fractional mode for a given PD operating frequency. Integer
mode, however, often requires a lower PD frequency to meet
step size requirements. The fractional mode advantage is that
higher PD frequencies can be used. Therefore, lower phase
noise can often be realized in fractional mode. Disable the
charge pump offset when in integer mode.
LD can be made available on the SDO pin at all times by writing
Register 0x0F, Bit 6 = 1. In that case, the ADF5610 does not
provide any readback functionality because the SDO signal is
not available.
Cycle Slip Prevention (CSP)
When changing the VCO frequency and the VCO is not yet
locked to the reference, the instantaneous frequencies of the two
PD inputs are different, and the phase difference of the two
inputs at the PD varies rapidly over a range much greater than
2π radians. Because the gain of the PD varies linearly with
phase up to 2π, the gain of a conventional PD cycles from high
gain, when the phase difference approaches a multiple of 2π, to
low gain, when the phase difference is slightly larger than a
multiple of 0 radians. The output current from the charge pump
cycles from maximum to minimum, even though the VCO has
not yet reached its final frequency.
Integer Frequency Tuning
In integer mode, the digital Σ-Δ modulator is shut off and the
N divider (Register 0x03) can be programmed to any integer
value in the range of 16 to 219 − 1. To run in integer mode,
configure Register 0x06 (as described in the Configuring the
SDO Pin for LD Output section), then program the integer
portion of the frequency as explained by Equation 12, ignoring
the fractional part.
The charge on the loop filter small capacitor may actually
discharge slightly during the low gain portion of the cycle. This
discharge can make the VCO frequency reverse temporarily
during locking. This phenomenon is known as cycle slipping.
Cycle slipping causes the pull in rate during the locking phase
to vary cyclically. Cycle slipping increases the time to lock to a
value greater than predicted by normal small signal Laplace
transform analysis.
The following sequence illustrates how to reconfigure the PLL
registers to switch from fractional mode to integer mode and
enable the output divider at the PDIV_OUT and NDIV_OUT
pins.
1. Disable the fractional modulator, Register 0x06, Bit 11 = 0.
2. Bypass the Δ-Σ modulator, Register 0x06, Bit 7 = 1.
3. Because the fundamental frequency (3650 MHz to
7300 MHz) range is supplied to the PLL, the fundamental
frequency is the basis for all programming. Determine the
fundamental frequency being used during operation.
4. If the fundamental frequency is <4000 MHz (RFOUT <
8000 MHz), set the internal prescaler to divide by 1
(Register 0x08, Bit 19 = 0). Otherwise, set the internal
prescaler to 1.
The ADF5610 PD features an ability to reduce cycle slipping
during acquisition. The CSP feature increases the PD gain
during large phase errors. The specific phase error that triggers
the momentary increase in PD gain is set via Register 0x0B,
Bits[8:7].
Frequency Tuning
5. To tune to frequencies (<7300 MHz), select the appropriate
output divider value (VCO Register 0x02, Bits[2:0]), then
enable the differential output divider (VCO Register 0x01,
Bit 8 = 1).
The ADF5610 VCO subsystem always operates in the doubled
frequency range of operation (7300 MHz to 14600 MHz), which
is always available at the RFOUT port or at the differential
divider ports, PDIV_OUT and NDIV_OUT, when the divider is
set to divide by 1 in VCO Register 0x02, Bits[2:0].
Writing to the VCO subsystem registers (VCO Register 0x02,
Bits[2:0] and VCO Register 0x01, Bit 8 in this case) is
accomplished indirectly through PLL Register 0x05. For more
information on communicating with the VCO subsystem
through PLL Register 0x05, see the VCO Serial Port Interface
(VSPI) section.
Frequencies at or below the VCO fundamental frequency of
operation (57 MHz to 7300 MHz) are realized by tuning to the
appropriate doubled frequency and selecting the appropriate
output divider setting (divide by 2 to 128) in VCO Register 0x02,
Bits[2:0].
Rev. 0 | Page 28 of 51
Data Sheet
ADF5610
In this example, the differential divider output frequency of
Fractional Mode
1752.5 MHz is achieved by programming the 19-bit binary value
of 70 decimal = 0x46 into the INTG_REG bit in Register 0x03,
and the 24-bit binary value of 1677722 decimal = 0x19999A
into the FRAC bit in Register 0x04. Eliminate the 0.596 Hz
quantization error using the exact frequency mode, if required.
In this example, the RFOUT value is divided by 4. Specific control
of the output divider is required. See the VCO Subsystem
Register Map section for details.
Set the following registers to place the ADF5610 in fractional
mode:
•
•
Enable the fractional modulator, Register 0x06, Bit 11 = 1.
Connect the Δ-Σ modulator in circuit, Register 0x06, Bit 7 = 0.
Fractional Frequency Tuning
This sections provides a generic example with the goal of
explaining how to program the output frequency. Actual
variables are dependent on the reference in use.
Exact Frequency Tuning
Due to quantization effects, the absolute frequency precision of
a fractional PLL is normally limited by the number of bits in the
fractional modulator. For example, the frequency resolution of a
24-bit fractional modulator is set by the PD comparison rate
divided by 224. The 224 value in the denominator is sometimes
referred to as the modulus. Analog Devices PLLs use a fixed
modulus, which is a binary number. In some types of fractional
PLLs, the modulus is variable, allowing exact frequency steps to
be achieved with decimal step sizes. Unfortunately, small steps
using small modulus values result in large spurious outputs at
multiples of the modulus period (channel step size). For this
reason, Analog Devices PLLs use a large fixed modulus. Normally,
the step size is set by the size of the fixed modulus. In the case
of a 50 MHz PD rate, a modulus of 224 results in a 2.98 Hz step
resolution, or 0.0596 ppm. In some applications, it is necessary
to have exact frequency steps, and even an error of 3 Hz cannot
be tolerated.
The ADF5610 in fractional mode achieves frequencies at
fractional multiples of the reference. The frequency of the
ADF5610 is given by the following equation:
f
RF = 2 × fVCO
where:
f
fVCO
=
XTAL (NINT + NFRAC ) = fINT + fFRAC
(12)
(13)
R
and
f
DIV = fRF/k
where:
f
f
RF is the doubled frequency (RFOUT).
VCO is the fundamental frequency (input to PLL) after any
necessary correction. The fVCO value must always be less than or
equal to 4000 MHz. When fVCO is greater than or equal to 4000
MHz, the internal divide by 2 prescaler (Register 0x08, Bit 19)
must be enabled.
Fractional PLLs are able to generate exact frequencies (with zero
frequency error) if N can be exactly represented in binary (for
example, N = 50.0, 50.5, 50.25, 50.75). Some common frequencies
cannot be exactly represented. For example, NFRAC = 0.1 = 1/10
must be approximated as round((0.1 × 224)/224) ≈ 0.100000024.
At fPD = 50 MHz, the NFRAC value translates to a 1.2 Hz error.
The exact frequency mode of the ADF5610 addresses this issue
and can eliminate quantization error by programming the
channel step size to fPD/10 in Register 0x0C to 10 (in this example).
More generally, this feature can be used whenever the desired
frequency (fVCO) can be exactly represented on a step plan where
there is an integer number of steps (<214) across Integer N
boundaries. Mathematically, this situation is satisfied if
f
XTAL is the frequency at the reference oscillator input.
R is the reference path division ratio (set in Register 0x02).
INT is the integer division ratio (set in Register 0x03), an
integer number between 20 and 524,284.
FRAC is the fractional part, from 0.0 to 0.99999…, NFRAC
Register 0x04/224.
DIV is the output frequency after any potential dividers.
k is 1 for fundamental, or k = 2 to 128 depending on the
N
N
=
f
selected output divider value (Register 0x05, Bits[9:7] indirectly
address VCO Register 0x02, Bits[2:0]).
For example, fDIV = 1752.5 MHz, k = 4, fVCO = 3505 MHz, fRF
7010 MHz, fXTAL = 50 MHz, R = 1, fPD = 50 MHz, NINT = 70, and
=
f
VCOk mod(fGCD) = 0
where:
VCOk is the channel step frequency, 0 < k < 224 − 1, as shown in
(17)
NFRAC = 0.1. The fPD value is the PD operating frequency, fXTAL/R.
Register 0x04 = round(0.1 × 224) = round(1,677,721.6) =
1,677,722.
f
Figure 26.
GCD is the greatest common divisor.
If fVCO ≤ 4000 MHz, then Register 0x08, Bit 19 = 0 (divide by 1).
If fVCO > 4000 MHz, then Register 0x08, Bit 19 = 1.
f
PD
214
fGCD = GCD( fVCO1, fPD ) and fGCD
≥
f
VCO (50 × 106/1) (70 + 1677722/224) = 3505 MHz +
1.192 Hz error
(14)
(15)
where fPD is the frequency of the phase detector.
f
RF = 2 × fVCO = 7010 MHz + 2.384 Hz error
Some fractional PLLs are able to achieve these exact frequencies
by shortening the length of the phase accumulator (the
denominator or the modulus of the Δ-Σ modulator) so that the
fDIV = fRF/k = (7010 MHz + 2.384 Hz error)/4 =
1752.5 MHz + 0.596 Hz error
(16)
Δ-Σ modulator phase accumulator repeats at an exact period
Rev. 0 | Page 29 of 51
ADF5610
Data Sheet
related to the interval frequency (fVCOk − fVCO(k − 1)) in Figure 26.
Consequently, the shortened accumulator results in more
frequent repeating patterns and, as a result, often leads to spurious
emissions at multiples of the repeating pattern period, or at
harmonic frequencies of fVCOk − fVCO(k − 1). For example, in some
applications, these intervals may represent the spacing between
radio channels, with the spurious occurring at multiples of the
channel spacing.
For example, to configure the ADF5610 for exact frequency
mode at fVCO = 2800.2 MHz, where the PD rate (fPD) =
61.44 MHz, proceed by checking Equation 17 to confirm that
the exact frequency mode for this fVCO is possible.
f
PD
fGCD GCD( fVCO , fPD ) and fGCD
214
f
GCD = GCD(2800.2 × 106, 61.44 × 106) =
61.44 106
In comparison, the Analog Devices method is able to generate
exact frequencies between adjacent Integer N boundaries while
still using the full 24-bit phase accumulator modulus. This
method achieves exact frequency steps with a high phase
detector comparison rate, which allows Analog Devices PLLs to
maintain excellent phase noise and spurious performance in the
exact frequency mode.
120 × 103 >
= 3750
214
Because Equation 17 is satisfied, the ADF5610 can be configured
for exact frequency mode at fVCO = 2800.2 MHz by continuing
with the following steps:
1. Calculate NINT
.
N
INT = Register 0x03 =
Using Exact Frequency Mode
6
fVCO1
fPD
2800.2 10
If the constraint in Equation 17 is satisfied, the ADF5610 is able
to generate signals with zero frequency error at the desired
VCO frequency. Exact frequency mode can be reconfigured for
each target frequency or be set up for a fixed fGCD that applies to
all channels.
floor
floor
61.44 106
45 decimal 0x2D
2. Calculate the value for Register 0x0C.
Register 0x0C =
Configuring Exact Frequency Mode for a Particular
Frequency
fPD
GCD(( fVCOk1 fVCOk ), fPD
61.44106
)
The following sequence shows how to properly derive the
various PLL register values for exact frequency mode operation.
An example is provided as well.
GCD(100103, 61.44106)
61.44106
3072 decimal 0xC00
1. Calculate and program the integer register setting.
20000
Register 0x03 = NINT = floor(fVCO/fPD)
3. To program Register 0x04, the closest integer N boundary
frequency (fN) that is less than the desired VCO frequency
(fVCO) must be calculated by fN = fPD × NINT. Using the
current example,
where the floor function is the rounding down to the
nearest integer.
2. Then calculate the low side integer boundary frequency.
fN = NINT × fPD
fN = fPD × NINT = 45 × 61.44 × 106 = 2764.8 MHz
3. Calculate and program the exact frequency register value.
then,
Register 0x04 =
Register 0x0C = fPD/fGCD
24
6
2 ( fVCO fN )
where fGCD = GCD(fVCO, fPD).
4. Calculate and program the fractional register setting.
ceil
ceil
fPD
24
6
24
2 (2800.210 2764.810 )
61.44106
2 ( fVCOk fN )
Register 0x04 ceil
fPD
9666560 decimal 0x938000
where ceil is the ceiling function, meaning round up to the
nearest integer.
Rev. 0 | Page 30 of 51
Data Sheet
ADF5610
fVCO = fVCO2
INTEGER
BOUNDARY
INTEGER
BOUNDARY
14
fN + 1
14 – 2
14 – 1
fN
fVCO1
fVCO2
fVCO3
fVCO4
fVCO
fVCO
fVCO
fN + 1
– fN = fPD
Figure 26. Exact Frequency Tuning
Exact Frequency Channel Mode
When multiple, equally spaced, exact frequency channels are
needed that fall within the same interval (that is, fN ≤ fVCOk
N + 1) where fVCOk is shown in Figure 26 and 1 ≤ k ≤ 214, it is
f
f
GCD1 GCD( fVCO1, fPD ) and
f
PD
GCD1
and fGCD2 GCD( fVCO2 , fPD )
214
<
(18)
f
and
possible to maintain the same Integer N register (Register 0x03)
and exact frequency register (Register 0x0C) settings and only
update the fractional register (Register 0x04) setting. The exact
frequency channel mode is possible when Equation 17 is
satisfied for at least two equally spaced adjacent frequency
channels, that is, the channel step size.
f
PD
f
GCD2
214
f
GCD1 GCD(2800.2 106 , 61.44 106 )
61.44 106
120 103
3750
214
To configure the ADF5610 for exact frequency channel mode,
initially and only at the beginning, the Integer N register
(Register 0x03) and exact frequency register (Register 0x0C)
must be programmed for the smallest fVCO frequency (fVCO1 in
Figure 26). Calculate and program the integer register setting by
fGCD2 GCD(2800.3 106 , 61.44 106 )
61.44 106
20 103
3750
214
2. If Equation 17 is satisfied for at least two of the equally
spaced interval (channel) frequencies (fVCO1, fVCO2, fVCO3, ...
VCON) as it is in Equation 18, the ADF5610 exact frequency
Register 0x03 = NINT = floor(fVCO1/fPD)
f
where fVCO1 is shown in Figure 26 and corresponds to the
minimum channel VCO frequency. Then, the lower integer
boundary frequency is given by fN = NINT × fPD.
channel mode is possible for all desired channel frequencies,
and can be configured as follows:
Register 0x03 =
Calculate and program the exact frequency register value by
Register 0x0C = fPD/fGCD, where fGCD = GCD((fVCOk + 1 − fVCOk), fPD),
greatest common divisor of the desired equidistant channel
spacing and the PD frequency ((fVCOk + 1 − fVCOk) and fPD). To
switch between various equally spaced intervals (channels) only,
the fractional register (Register 0x04) must be programmed to
the desired VCO channel frequency (fVCOk) in the following manner:
6
fVCO1
fPD
2800.2 10
61.44 106
floor
floor
45 decimal
0x2D
Register 0x0C =
fPD
GCD(( fVCOk1 fVCOk ), fPD
61.44106
)
2
24 ( fVCOk fN
)
61.44106
20000
Register 0x04 =
NFRAC ceil
GCD(100103,61.44106)
3072 decimal 0xC00
fPD
where:
fN = floor(fVCO1/fPD).
VCO1, as shown in Figure 26, represents the smallest channel
where (fVCOk+1 − fVCOk) is the desired channel spacing
(100 kHz in this example).
f
VCO frequency that is greater than fN.
3. To program Register 0x04, the closest Integer N boundary
frequency (fN) that is less than the smallest channel VCO
frequency (fVCO1) must be calculated (fN = floor(fVCO1/fPD))
using the following example:
For example, to configure the ADF5610 for the exact frequency
mode for equally spaced intervals of 100 kHz, where the first
channel (Channel 1) = fVCO1 = 2800.200 MHz and the PD rate
(fPD) = 61.44 MHz, proceed as follows:
6
2800.210
1. Check that the exact frequency mode for fVCO1
=
fN fPD floor
61.44106
2800.2 MHz (Channel 1) and fVCO2 = 2800.2 MHz +
100 kHz = 2800.3 MHz (Channel 2) is possible.
4561.44106 2764.8 MHz
Rev. 0 | Page 31 of 51
ADF5610
Data Sheet
Then, for Channel 1,
Register 0x04 = ceil
2. Power down the PLL by pulling the CEN pin (Pin 44) low
(assuming there are no SPI overrides (Register 0x01, Bit 0 =
1)). Pulling the CEN pin low disables all analog functions
and internal clocks. Current consumption typically drops
to 100 μA in the power-down state if the reference input is
disconnected. The serial port still responds to normal
communication in power-down mode.
2
24 ( fVCO1 fN )
fPD
where fVCO1 = 2800.2 MHz.
The following equation provides the value for Register 0x04
for operating in Channel 1 (fVCO1 = 2800.2 MHz).
It is possible to ignore the CEN pin by setting Register 0x01,
Bit 0 = 0. Control of the power-down mode then comes from
the serial port register, Register 0x01, Bit 1.
24
6
6
2
(2800.2 10 2764.8 10 )
61.44 106
ceil
9666560 decimal
0x938000
It is also possible to leave various blocks turned on when in
power-down (see Register 0x01), as listed in Table 12.
4. To change from Channel 1 (fVCO1 = 2800.2 MHz) to
Channel 2 (fVCO2 = 2800.3 MHz), only Register 0x04 must
be programmed, as long as all of the desired exact
frequencies (fVCOk, see Figure 26) fall between the same
Integer N boundaries (fN < fVCOk < fN + 1). In that case,
Table 12. Bit and Block Assignments for Register 0x01
Bit Assignment
Block Assignment
Internal bias reference sources
PD block
CP block
Reference path buffer
VCO path buffer
Bit 2
Bit 3
Bit 4
Bit 5
Bit 6
Bit 7
Register 0x04 =
24
6
6
2 (2800.310 2764.8 10 )
61.44 106
ceil
9693867 decimal 0x93EAAB
Digital I/O test pads (GPO driver)
Seed Register
The start phase of the fractional modulator digital phase
accumulator (DPA) can be set to one of four possible default
values via the SEED_SELECT bits, Register 0x06, Bits[1:0]. The
ADF5610 automatically reloads the start phase (seed value) into
the DPA every time a new fractional frequency is selected.
Certain zero seed values or binary seed values may cause
spurious energy correlation at specific frequencies. For most
cases, a random (not zero and not binary) start seed is
recommended (Register 0x06, Bits[1:0] = 2).
GENERAL-PURPOSE OUTPUT (GPO)
The PLL shares the SDO (lock detect/serial data output) pin to
perform various functions. Although the pin is most commonly
used to read back registers from the chip via the SPI, the pin is
also capable of exporting a variety of signals and real-time test
waveforms (including lock detect). The SDO pin is driven by a
tristate CMOS driver with ~200 Ω ROUT. The SDO pin has logic
associated with it to dynamically select whether the driver is
enabled, and to determine which data to export from the chip.
SOFT RESET AND POWER-ON RESET
In the default configuration of the output driver, after POR, the
output driver is disabled, and drives only during appropriately
addressed SPI reads. This configuration allows the ADF5610 to
share its output with other devices on the same bus.
The ADF5610 features a hardware power-on reset (POR). All
chip registers are reset to default states approximately 250 μs
after power-up.
The PLL subsystem SPI registers can also be soft reset by an SPI
write to Register 0x00. Note that the soft reset does not clear the
SPI mode of operation referred to in the SPI Modes of Operation
section. The VCO subsystem is not affected by the PLL soft reset.
The VCO subsystem registers can only be reset by removing the
power supply.
The pin driver is enabled if the chip is addressed. That is, the
last three bits of the SPI cycle = 000b before the rising edge of
SEN. If SEN rises before SCK has clocked in an invalid (nonzero)
chip address, the ADF5610 starts to drive the bus.
To monitor any of the GPO signals, including lock detect, set
Register 0x0F, Bit 7 = 1 to keep the SDO driver always on. This
setting stops the LDO driver from tristating and means that the
SDO line cannot be shared with other devices.
If external power supplies or regulators have rise times slower than
250 μs, write to the SPI RST_SWRST bit (Register 0x00, Bit 5 = 1)
immediately after power-up, before any other SPI activity. This
write procedure ensures starting from a known state.
The ADF5610 naturally switches from the GPO data and exports
the SDO signal during an SPI read. To prevent this automatic data
selection and to always select the GPO signal, set Register 0x0F,
Bit 6 = 1 to prevent automuxing of the SDO pin. The phase
noise performance at this output is poor and uncharacterized.
Also, do not toggle the GPO during normal operation because
toggling may degrade the spectral performance.
POWER-DOWN MODE
The VCO subsystem is not affected by the CEN pin or PLL soft
reset. Therefore, device power-down consists of the following
two-step process:
1. Power down the VCO by writing 1 to VCO Register 0x02,
Bit 3 via PLL Register 0x05.
Rev. 0 | Page 32 of 51
Data Sheet
ADF5610
Additional controls are available that may be helpful when
sharing the bus with other devices, and are as follows:
isolation is also achievable with the HMC1060LP3E quad low
noise regulator, which conserves board space.
•
To disable the driver completely, set Register 0x08, Bit 5 = 0
(this bit takes precedence over all other SDO driver bit
settings).
To disable either the pull-up or pull-down sections of the
driver, set Register 0x0F, Bit 8 = 1 or Register 0x0F, Bit 9 =
1, respectively.
PROGRAMMABLE PERFORMANCE TECHNOLOGY
For low power applications that do not require maximum noise
floor performance, the ADF5610 features the ability to reduce
current consumption by 18 mA (power consumption by 90 mW)
•
High performance is enabled by writing VCO Register 0x01,
Bits[4:2] = 7 decimal, and high performance is disabled (low
current consumption mode enabled) by writing VCO Register
0x01, Bits[4:2] = 0 decimal. High performance mode improves
phase noise and output power at the cost of increased current
consumption.
Example scenarios are listed in Table 13. The signals that are
available on the GPO are selected by changing the GPO_
SELECT bits (Register 0x0F, Bits[4:0]).
The ADF5610 SPI supports both 1.8 V and 3.3 V input voltage
levels. Input pins including SDI, SCK, and SEN support both
voltage levels without the need for any configuration.
LOOP FILTER AND FREQUENCY CHANGES
R3
R4
CP
VT
C4
C1
C3
R2
C2
The SPI output (SDO pin) only supports 3.3 V levels when driven
and 1.8 V and 3.3 V in open-drain mode. The CMOS or open-
drain configuration is register programmable via Register 0x0F,
Bits[9:8]. Open-drain mode requires an external pull-up resistor.
See Table 1 for more information.
Figure 27. Loop Filter Design
All PLLs with integrated VCOs exhibit integer boundary spurs
at harmonics of the reference frequency. Figure 15 shows the
worst case spurious scenario where the harmonic of the reference
frequency (50 MHz) is within the loop filter bandwidth of the
fundamental frequency of the ADF5610.
Table 13. Driver Scenarios
Scenario
Action
Drive SDO During Reads,
Tristate Otherwise (Allows
Bus Sharing)
No action required.
A tunable reference can be used to vary the reference frequency
and to distance the harmonic of the reference frequency (spurious
emissions) from the fundamental output frequency of the
ADF5610 so that the harmonic of the reference frequency is
filtered by the loop filter. The internal ADF5610 setup and divide
ratios can be changed in the opposite direction accordingly so that
the ADF5610 generates an identical output frequency without
the spurious emissions inside the loop bandwidth using these
same procedures.
Drive SDO During Reads, Lock
Detect Otherwise
Set GPO select, Register 0x0F,
Bits[4:0] = 00001b (default)
Set Register 0x0F, Bit 6 = 0,
enable automux of SDO
Set Register 0x0F, Bit 7 = 1,
prevent GPO driver disable
Always Drive Lock Detect, 3.3 V Set Register 0x0F, Bit 6 = 1,
Output
prevent automux of SDO
Set the GPO select,
Register 0x0F, Bits[4:0] = 00001
(default)
Set Register 0x0F, Bit 7 = 1,
prevent GPO driver disable
The ADF5610 features an internal autocalibration process that
seamlessly calibrates the ADF5610 when a frequency change
is executed. When the ADF5610 is calibrated at any temperature,
the calibration setting is guaranteed to hold across the entire
operating range of the ADF5610 (−40°C to +85°C). The tuning
voltage is maintained within a narrow operating range for worst
case scenarios where calibration is executed at one temperature
extreme and the device is operating at the other temperature
extreme. For optimum temperature and spurious performance,
use EIA character code C0G capacitors of the NP0 or C0G type.
The use of capacitors with a temperature coefficient worse than
X7R is not recommended.
CHIP IDENTIFICATION
Identifies the PLL subsystem version information by reading the
content of Register 0x00. It is not possible to read the VCO
subsystem version.
POWER SUPPLY
The ADF5610 is a high performance, low noise device. Phase
noise and spurious performance may be degraded by noisy
power supplies. To achieve maximum performance and ensure
that power supply noise does not degrade the performance of
the ADF5610, it is recommended to use Analog Devices low
noise, high power supply rejection ratio (PSRR) regulators.
Preferred regulators include the LT3042, LT3045, ADM7150,
and the ADM7151. Optimal spectral performance and supply
For applications that require fast frequency changes, the
ADF5610 supports manual calibration that enables faster
settling times. Manual calibration must be characterized once
for each individual ADF5610 device, at any temperature, and is
valid across all temperature operating ranges of the ADF5610.
Manual calibration is achieved by using autocalibration to
initially determine the proper band for each frequency. The
frequency vs. band information is stored in a look up table
Rev. 0 | Page 33 of 51
ADF5610
Data Sheet
(LUT) for future access. For more information about manual
calibration, see the Manual VCO Calibration for Fast Frequency
Hopping section. Any size frequency hop has a similar settling
time with autocalibration disabled (Register 0x0A, Bit 11 = 1).
MUTE MODE
The ADF5610 has some support for muting the outputs. The
differential divider outputs, PDIV_OUT and NDIV_OUT, can
be disabled be setting VCO Register 0x01, Bit 8 = 0 to power
down the divider when not in use. The RFOUT port can only
be muted by powering down the VCO from VCO Register 0x02,
Bit 3. However, powering down the VCO also effectively mutes
the differential divider outputs simultaneously because no
signal is present at the input of the buffer to the divider.
Rev. 0 | Page 34 of 51
Data Sheet
ADF5610
SERIAL PORT INTERFACE
The ADF5610 SPI supports both 1.8 V and 3.3 V input voltage
levels. Input pins including SDI, SCK, and SEN support both
voltage levels without the need for any configuration.
SERIAL PORT LEGACY MODE
Typical serial port legacy mode operation can be run with SCK
at speeds up to 50 MHz. This mode is used with the legacy
evaluation software for Analog Devices PLL products with an
HMC prefix. The ADF5610 incorporates the HMC704 PLL.
The ADI HMC PLL VCO evaluation software can be used to
communicate with the ADF5610.
The SPI output (SDO pin) only supports 3.3 V levels when driven
and 1.8 V and 3.3 V in open-drain mode. See Table 1 and the
General-Purpose Output (GPO) section for more information.
The SPI protocol has the following general features:
SERIAL PORT LEGACY MODE: SINGLE PLL
•
3-bit chip address. Can address up to eight devices
connected to the serial bus.
Wide compatibility with multiple protocols from multiple
vendors.
Legacy mode SPI operation can only address and communicate
with a single PLL. The legacy mode protocol for the SPI is
designed for a 4-wire interface with a fixed protocol featuring
•
•
•
•
•
Simultaneous read/write during the SPI cycle.
5-bit address space.
3-wire for write only capability, 4-wire for read/write capability.
Typical serial port interface operation can be run with SCK
at speeds up to 50 MHz.
•
Compatibility with the Analog Devices line of multichip,
microwave, VCO PLL solutions, including the HMC764,
HMC765, HMC783, and HMC807.
Primarily utilized for the ADI HMC PLL VCO evaluation
software and the HMC700, HMC701, and HMC702
microwave products.
One read/write bit
Six address bits
24 data bits
•
SPI MODES OF OPERATION
•
•
•
The PLL VCO serial port interface can operate in two different
modes of operation, as follows:
•
•
Legacy mode: single slave per SPI bus.
Open mode: up to eight slaves per SPI bus.
SERIAL PORT OPEN MODE
The serial port open mode features:
Both protocols support five bits of register address space. Legacy
mode can support up to six bits of register address, but is restricted
to five bits when compatibility with open mode is offered.
•
Compatibility with general serial port protocols that use a
shift and strobe approach to communication.
Up to eight supported devices of various types can be
addressed from a single serial port bus.
•
REGISTER 0x00 MODES
Register 0x00 has a dedicated function in each mode. Open
mode allows wider compatibility with SPI protocols of other
manufacturers.
The open mode protocol has the following general features:
•
3-bit chip address. Can address up to eight devices
connected to the serial bus.
SERIAL PORT MODE DECISION AFTER POWER-ON
RESET
•
Wide compatibility with multiple protocols from multiple
vendors.
On power-up, both legacy mode and open mode are active and
listening. All digital inputs and outputs must be low at power-up.
•
•
•
Simultaneous read/write during the SPI cycle.
5-bit register address space.
3-wire for write only capability, 4-wire for read/write
capability.
Selection of the desired serial port mode (protocol) is determined
on the first occurrence of SEN or SCK, after which the serial
port mode is fixed and only changeable by a power-down. If a
rising edge on SEN is detected first, legacy mode is selected. If a
rising edge on SCK is detected first, open mode is selected.
The HMC767, HMC769, and HMC778 RF PLLs with
integrated VCOs also support open mode. The HMC700,
HMC701, HMC702, and some generations of microwave PLLs
with integrated VCOs do not support open mode. Typical open
mode serial port operation can be run with SCK at speeds up to
50 MHz.
Table 14. Register 0 Comparison: Legacy Mode vs. Open Mode
Type
Read
Write
Legacy Mode (Single User)
Open Mode (Single or Multiple User)
Chip ID, 24 bits
Chip ID, 24 bits
Soft reset, general strobes
Read address, Bits[4:0]
Soft reset, Bit 5
General strobes, Bits[24:6]
Rev. 0 | Page 35 of 51
ADF5610
Data Sheet
The following list is the sequence of operations for a read cycle
while communicating in legacy mode:
LEGACY MODE: SERIAL PORT WRITE OPERATION
A typical legacy mode write cycle is shown in Figure 28.
1. The master (host) asserts both SEN and SDI to indicate a
read cycle, followed by a rising edge on SCK. Note that the
lock detect (LD) function is usually multiplexed onto the
SDO pin. It is suggested that LD only be considered valid
when SEN is low. In fact, LD does not toggle until the first
active data bit toggles on SDO and is restored immediately
after the trailing edge of the LSB of the serial data output,
as shown in Figure 29.
The following list is the sequence of operations for a write cycle
while communicating in legacy mode.
1. The master (host) both asserts serial port enable (SEN) and
clears SDI to indicate a write cycle, followed by a rising
edge of SCK.
2. The slave (PLL) reads SDI on the first rising edge of SCK
WR
after SEN. SDI low indicates a write cycle (
).
3. The host places the six address bits on the next six falling
edges of SCK, with the MSB first.
4. The slave registers the address bits in the next six rising
edges of SCK (two to seven clock cycles).
5. The host places the 24 data bits on the next 24 falling edges
of SCK, with the MSB first.
6. The slave registers the data bits on the next 24 rising edges
2. The slave (PLL) reads SDI on the first rising edge of SCK
after SEN. SDI high initiates the read cycle.
3. The host places the six address bits on the next six falling
edges of SCK, with the MSB first.
4. The slave registers the address bits on the next six rising
edges of SCK (two to seven clock cycles).
5. The slave switches from lock detect and places the
requested 24 data bits on SDO on the next 24 rising edges
of SCK (eight to 31 clock cycles) with the MSB first.
6. The host registers the data bits on the next 24 falling edges
of SCK (8 to 31 clock cycles).
of SCK (8 to 31 clock cycles).
7. SEN is cleared on the 32nd falling edge of SCK.
8. The 32nd falling edge of SCK completes the cycle.
LEGACY MODE: SERIAL PORT READ OPERATION
7. The slave restores lock detect on the 32nd rising edge of SCK.
8. SEN is deasserted on the 32nd falling edge of SCK.
9. The 32nd falling edge of SCK completes the read cycle.
A typical legacy mode read cycle is shown in Figure 29.
1
2
3
4
5
6
7
8
28
29
30
31
32
SCK
SDI
t2
t3
X
WR
A4
A3
A2
A1
A0
D23
D22
D3
D2
D1
D0
X
t1
SEN
t4
Figure 28. Legacy Mode: Serial Port Timing Diagram, Write Cycle
1
2
3
4
5
6
7
8
28
29
30
31
32
SCK
t2
A3
t3
X
RD
A5
A4
A2
A1
A0
X
SDI
t1
SEN
SDO
t5
t4
LD
LOCK DETECT (LD)
D23
D22
D3
D2
D1
D0
Figure 29. Legacy Mode: Serial Port Timing Diagram, Read Cycle
Rev. 0 | Page 36 of 51
Data Sheet
ADF5610
as shown in Figure 31 and Figure 32. Set Bits[D23:D5] to
zero. Bits[D4:D0] is the address of the register to be read
on the next cycle.
OPEN MODE: SERIAL PORT WRITE OPERATION
A typical write cycle is shown in Figure 30.
The following list is the sequence of operations for write cycle
while communicating in open mode:
2. The slave (PLL) shifts in data on SDI on the first 24 rising
edges of SCK.
3. The master places a 5-bit register address, Bits[R4:R0] (the
read address register), with the MSB first, on the next five
falling edges of SCK (23 to 29 clock cycles). Bits[R4:R0] =
00000.
1. The master (host) places 24 bit data, Bits[D23:D0], with
the MSB first, on SDI on the first 24 falling edges of SCK.
2. The slave (PLL) shifts in data on SDI on the first 24 rising
edges of SCK.
4. The slave shifts the register bits on the next five rising
edges of SCK (23 to 29 clock cycles).
5. The master places a 3-bit chip address, Bits[A2:A0], with
the MSB first, on the next three falling edges of SCK (30 to
32 clock cycles). Chip address is always 000 for RF PLL
VCOs.
3. The master places a 5-bit register address to be written to,
Bits[R4:R0], with the MSB first, on the next five falling
edges of SCK (25 to 29 clock cycles).
4. The slave shifts the register bits on the next five rising
edges of SCK (25 to 29 clock cycles).
5. The master places a 3-bit chip address, Bits[A2:A0], with
the MSB first, on the next three falling edges of SCK (30 to
32 clock cycles). Analog Devices reserves chip address
Bits[A2:A0] = 000 for all RF PLL VCOs.
6. The slave shifts the chip address bits on the next three
rising edges of SCK (30 to 32 clock cycles).
7. The master asserts SEN after the 32nd rising edge of SCK.
8. The slave registers the SDI data on the rising edge of SEN.
9. The master clears SEN to complete the address transfer of
the two-part read cycle.
6. The slave shifts the chip address bits on the next three
rising edges of SCK (30 to 32 clock cycles).
7. The master asserts SEN after the 32nd rising edge of SCK.
8. The slave registers the SDI data on the rising edge of SEN.
9. The master clears SEN to complete the write cycle.
10. If the user does not wish to write data to the chip at the
same time as the second cycle, it is recommended to simply
rewrite the same contents on SDI to Register 0x00 on the
readback part of the cycle.
11. The master places the same SDI data as the previous cycle
on the next 32 falling edges of SCK.
OPEN MODE: SERIAL PORT READ OPERATION
A typical read cycle is shown in Figure 31 and Figure 32.
In general, in open mode, the SDO line is always active during
the write cycle. During any open mode SPI cycle, SDO contains
the data from the current address written in Register 0x00,
Bits[4:0]. If Register 0x00, Bits[4:0] are not changed, the same
data is always present on SDO when an open mode cycle is in
progress. If it is desired to read from a specific address, it is
necessary in the first SPI cycle to write the desired address to
Register 0x00, Bits[4:0]. Then, in the next SPI cycle, the desired
data is available on SDO.
12. The slave (PLL) shifts the SDI data on the next 32 rising
edges of SCK.
13. The slave places the desired data (for example, data from
the address in Register 0x00, Bits[4:0]) on SDO on the next
32 rising edges of SCK. Lock detect is disabled.
14. The master asserts SEN after the 32nd rising edge of SCK to
complete the cycle and revert back to lock detect on SDO.
Note that if the chip address bits are unrecognized (Bits[A2:A0]),
the slave tristates the SDO output to prevent a possible
contention issue.
An example of the open mode two-cycle procedure to read
from any random address is as follows:
1. The master (host), on the first 24 falling edges of SCK
places 24 bit data, Bits[D23:D0], with the MSB first, on SDI,
t1
t2
t5
2
3
22
23
24
25
26
31
32
SCK
SDI
X
D22
D2
D1
D0
R4
R3
R0
A2
A1
A0
X
SEN
t3
t4
Figure 30. Open Mode: Serial Port Timing Diagram, Write Cycle
Rev. 0 | Page 37 of 51
ADF5610
Data Sheet
FIRST CYCLE
25 26
t1
t2
20
t5
2
19
21
24
29
30
31
32
SCK
X
D5
D4
D0
R4
R3
R0
A2
A1
A0
X
SDI
READ ADDRESS
REGISTER ADDRESS = 00000
CHIP ADDRESS = 000
t5
SDO
LD
X
X
X
X
X
X
X
X
X
X
X
LD
SEN
t6
t4
Figure 31. Open Mode: Serial Port Timing Diagram, Read Operation, First Cycle
SECOND CYCLE
2
19
20
21
24
25
26
30
31
32
SCK
SDI
X
D23
D5
D4
D0
R4
R0
A2
A1
A0
X
1
LD
SDO
SEN
LD
D31
D30
D10
D9
D8
D7
D6
D3
D2
D1
D0
t3
1
READBACK ON SDO CAN FUNCTION WITHOUT SEN; HOWEVER, SEN
RISING EDGE IS REQUIRED TO RETURN THE SDO TO THE LOCK DETECT STATE.
Figure 32. Open Mode: Serial Port Timing Diagram, Read Operation, Second Cycle
Rev. 0 | Page 38 of 51
Data Sheet
ADF5610
PLL REGISTER MAP
ID REGISTER, RESET STROBE REGISTER, AND
OPEN MODE READ REGISTER
The ID register is read only and returns the chip ID when read.
The reset strobe register only contains one functional bit field
for resetting all digital content and registers to the default values.
The reset strobe register is write only.
Table 15. Register 0x00, ID Register (Read Only)
Bits
Type
Name
Default
Description
23:0
Read
CHIP_ID
A7975
ADF5610 chip ID
Table 16. Register 0x00, Reset Strobe Register (Write Only)
Bits Type Name
Default
Description
4:0
Write Reserved
Not
Reserved. Set all bits to 0 for proper operation.
Applicable
5
Write RST_SWRST Not
Applicable
Not
Applicable
Strobe generates soft reset. Resets all registers to the default state. Can be used in both
legacy mode and open mode.
23:6 Write Reserved
Reserved. Set all bits to 0 for proper operation.
Table 17. Register 0x00, Open Mode Read Address Register (Write Only)
Bits Type Name
Default
Description
4:0
5
Write OPEN_MODE_READ_ADDRESS
Write RST_SWRST
Not applicable
Not Applicable
Not applicable
Read address for next cycle when in open mode, two-cycle read.
Strobe generates soft reset. Resets all registers to the default state.
Reserved. Set all bits to 0 for proper operation.
23:6 Write Reserved
RST REGISTER
Table 18. Register 0x01, RST Register (Default 0x000002)
Bits Type Name Default Description
0
Read/Write RST_CHIPEN_PIN_SELECT
0
1 = power down the PLL via the CEN pin (see the Power-Down Mode section).
0 = power down the PLL via the SPI (RST_CHIPEN_FROM_SPI), Register 0x01,
Bit 1.
1
2
3
4
5
6
7
8
9
Read/Write RST_CHIPEN_FROM_SPI
Read/Write KEEP_BIAS_ON
Read/Write KEEP_PFD_ON
Read/Write KEEP_CP_ON
1
0
0
0
0
0
0
0
0
PLL enable bit of the SPI.
Keeps internal bias generators on, ignores chip enable control.
Keeps PFD circuit on, ignores chip enable control.
Keeps charge pump on, ignores chip enable control.
Keeps reference buffer block on, ignores chip enable control.
Keeps VCO divider buffer block on, ignores chip enable control.
Keeps GPO output driver on, ignores chip enable control.
Reserved. Program 0.
Read/Write KEEP_REF_BUF_ON
Read/Write KEEP_VCO_BUF_ON
Read/Write KEEP_GPO_DRIVER_ON
Read/Write Reserved
Read/Write Reserved
Reserved. Program 0.
REFERENCE DIVIDER (RDIV) REGISTER
Table 19. Register 0x02, RDIV Register (Default 0x000001)
Bits Type
Name Default Description
13:0 Read/Write RDIV
1
Reference divider R value (see Equation 12). Using the divider requires the reference path buffer to
be enabled (Register 0x08, Bit 3 = 1). 1 decimal ≤ RDIV ≤ 16,383 decimal.
Rev. 0 | Page 39 of 51
ADF5610
Data Sheet
FREQUENCY REGISTER, INTEGER PART
Table 20. Register 0x03, Frequency Register, Integer Part (Default 0x000019)
Bits Type
18:0 Read/Write INTG_REG 25
decimal
Name
Default
Description
Integer divider register. These bits are the VCO divider integer part, used in all modes (see
Equation 12).
Fractional mode:
Minimum = 0x14 = 20 decimal.
Maximum = 219 − 4 = 0x7FFFC = 524,284 decimal.
Integer mode:
Minimum = 0x10 = 16 decimal.
Maximum = 219− 1 = 0x7FFFF = 524,287 decimal.
FREQUENCY REGISTER, FRACTIONAL PART
Table 21. Register 0x04, Frequency Register, Fractional Part (Default 0x000000)
Bits Type
Name Default Description
23:0 Read/Write FRAC
0
VCO divider fractional part (24-bit unsigned). See the Fractional Frequency Tuning section.
These bits are used in fractional mode only (NFRAC = Register 0x04/224). Minimum = 0 decimal.
Maximum = 224 − 1.
divider ratio to be configured as well as powering down the
VCO. Initialize the output divider register only if the functions
are needed for the application.
VCO SPI REGISTER
Register 0x05 is a special register used for indirect addressing of
the VCO subsystem. Writes to Register 0x05 are automatically
forwarded to the VCO subsystem by the VCO SPI FSM controller.
For autocalibration, Register 0x05, Bits[6:0] must be set to 0.
This write is required whenever autocalibration is used and is
always the final register write. This write allows Register 0x03,
when programming an integer frequency, or Register 0x04, for
fractional frequencies, to trigger autocalibration. Setting
Register 0x05, Bits[6:0] to zero is not required when manually
selecting a VCO band.
The auxiliary SPI outputs the contents of Register 0x05 upon
receipt of a frequency change command. The auxiliary SPI data
is output at the auxiliary SPI clock rate, which is based on the
f
PD rate (Register 0x05, Bit 6). A single auxiliary SPI transfer
requires 16 auxiliary SPI cycles plus four overhead cycles.
Register 0x05 is a read/write register. However, Register 0x05
holds only the contents of the last transfer to the VCO subsystem.
Therefore, it is not possible to read the full contents of the VCO
subsystem. Only the content of the last transfer to the VCO sub-
system can be read. The ADF5610 VCO has three registers, each
containing various functionality. All VCO register bits default to
zero at power-up. As long as autocalibration is enabled, the FSM
controls VCO Register 0x00, which is the tuning register. This
register provides access to the VCO cores and frequency bands.
When autocalibration is enabled, there is no need to write to VCO
Register 0x00. Writes are only required when manually selecting
a VCO core or frequency band. VCO Register 0x01 is the power
control register. Unless the default settings are required, this
register must be initialized to configure the VCO bias, output
power levels, and to enable the frequency divider. Lastly, VCO
Register 0x02 is the output divider register, which allows the
VCO register writes are typically performed in descending
order. However, other sequences may work as well. Depending
on the application, there is a total of one to four Register 0x05
writes. For additional information, refer to the VCO Subsystem
Register Map section.
Rev. 0 | Page 40 of 51
Data Sheet
ADF5610
Table 22. Register 0x05, VCO SPI Register (Default 0x000000)
Bits Type
Name
Default Description
2:0
6:3
Read/Write VCO_ID
0
0
Internal VCO subsystem ID.
Read/Write VCO_REGADDR
VCO subsystem register address. These bits are for interfacing with the VCO. See the VCO
Serial Port Interface (VSPI) section.
15:7 Read/Write VCO_DATA
0
VCO subsystem data. These bits are used to write the data to the VCO subsystem.
Δ-Σ CONFIGURATION REGISTER
Table 23. Register 0x06, Δ-Σ Configuration Register (Default 0x200B4A)
Bit
Type
Name
Default
Description
1:0
Read/Write SEED_SELECT
2
Selects the seed in fractional mode. Writes to this register are stored in
the ADF5610 and are loaded into the modulator only when a frequency
change is executed and when Register 0x06, Bit 8 = 1.
0: 0 seed.
1: LSB seed.
2: 0xB29D08 seed.
3: 0x50F1CD seed.
Reserved.
6:2
7
Read/Write Reserved
18
decimal
(0x12)
Read/Write FRAC_BYPASS
0
Bypass fractional mode. When bypassing the fractional modulator, the
output is ignored, but the fractional modulator continues to be clocked
when SD enable = 1. Use this bit to test the isolation of the digital
fractional modulator from the VCO output in integer mode.
0: use modulator, required for fractional mode.
1: bypass modulator, required for integer mode.
10:8
11
Read/Write Reserved
3 decimal Program to 7 decimal. Default initialization
Read/Write SD_ENABLE
This bit controls whether autocalibration starts on an integer write or a
fractional write.
1
0: disables fractional core, use for integer mode or integer mode with
CSP.
1: enables fractional core (required for fractional mode), or integer
isolation testing.
20:12 Read/Write Reserved
0
1
0
Reserved.
21
22
Read/Write AUTO_CLK_CONFIGURATION
Read/Write Reserved
Program to 0.
Reserved.
LOCK DETECT REGISTER
Table 24. Register 0x07, Lock Detect Register (Default 0x00014D)
Bit
Type
Name
Default
Description
2:0
Read/Write LKD_WINCNT_MAX
5
The lock detect window sets the number of consecutive counts of
the divided VCO that must be within the lock detect window to
declare lock.
decimal
0: 5.
1: 32.
2: 96.
3: 256.
4: 512.
5: 2048.
6: 8192.
7: 65,535.
3
Read/Write LKD_ENABLE
Read/Write Reserved
1
0
Enable internal lock detect. See the Serial Port Interface section.
Reserved.
5:4
Rev. 0 | Page 41 of 51
ADF5610
Data Sheet
Bit
Type
Name
Default
Description
6
Read/Write LKD_ONESHOT_SEL
1
Lock detection window timer selection.
1: digital programmable timer.
0: analog one shot, nominal 10 ns window.
Lock detection, digital window duration.
0: half cycle.
9:7
Read/Write LKD_ONESHOT_DURATION
2
1: one cycle.
2: two cycles.
3: four cycles.
4: eight cycles.
5: 16 cycles.
6: 32 cycles.
7: 64 cycles.
11:10 Read/Write LKD_CONFIG
0
Lock detect digital timer frequency control (see the Lock Detect
section for more information).
00: fastest.
11: slowest.
Reserved.
12
13
Read/Write Reserved
0
0
Read/Write LKD_AUTOCAL_WHEN_UNLOCKED
1: autorelock attempts to relock one time, if lock detect fails for any
reason.
ANALOG ENABLE (EN) REGISTER
Table 25. Register 0x08, Analog EN Register (Default 0xC1BEFF)
Bit
Type
Name
Default
Description
0
Read/Write BIAS_EN
Read/Write CP_EN
1
1
1
1
1
1
Enables main chip bias reference.
Charge pump enable.
1
2
Read/Write PD_EN
PD enable.
3
Read/Write REFBUF_EN
Read/Write VCOBUF_EN
Read/Write GPO_PAD_EN
Reference path buffer enable.
VCO path RF buffer enable.
0: disables the SDO pin.
4
5
1: enables the GPO port or allows a shared SPI.
When Bit 5 = 1 and Register 0xF, Bit 7 = 1, the SDO pin is always driven, which is
required for use of the GPO port.
When Bit 5 = 1 and Register 0xF, Bit 7 = 0, SDO is off when an unmatched chip
address is seen on the SPI. Allows shared SPI with other compatible devices.
9:6
10
Read/Write Reserved
11 decimal
(0x0B)
Reserved.
Read/Write VCO_SIDE_BIAS
1
VCO buffer and prescaler bias enable.
Reserved. Program to 55 decimal (0x37).
18:11 Read/Write Reserved
55 decimal
(0x37)
19
Read/Write DIV_SEL
0
8 GHz divide by 2 enable. Allows PLL operation up to 8 GHz.
When input to PLL is less than 4000 MHz, set to 0 decimal to bypass internal
divider.
When input to PLL is greater than or equal to 4.0 GHz, set to 1 decimal to enable
internal fixed divide by 2.
20
21
Read/Write Reserved
0
0
Reserved. Program 0.
Read/Write HI_FREQ_RF_SEL
High frequency reference select. Program to 1 for reference frequencies from
200 MHz to 350 MHz. Program to 0 for reference frequencies less than 200 MHz.
23:22 Read/Write Reserved
3 decimal
Reserved.
Rev. 0 | Page 42 of 51
Data Sheet
ADF5610
CHARGE PUMP REGISTER
Table 26. Register 0x09, Charge Pump Register (Default 0x403264)
Bit
Type
Name
Default
Description
6:0
Read/Write CP_DOWN_MAG 100 decimal,
0x64
Charge pump down gain control, 20 µA per step. Affects fractional phase noise
and lock detect settings.
0 = 0 µA.
1 = 20 µA.
2 = 40 µA.
…
127 = 2.54 mA.
13:7
Read/Write CP_UP_MAG
100 decimal,
0x64
Charge pump up gain control, 20 µA per step. Affects fractional phase noise and
lock detect settings.
0 = 0 µA.
1 = 20 µA.
2 = 40 µA.
…
127 = 2.54 mA.
20:14 Read/Write CP_LEAK_MAG
0
Charge pump offset control, 5 µA per step. Affects fractional phase noise and
lock detect settings.
0 = 0 µA.
1 = 5 µA.
2 = 10 µA.
…
127 = 635 µA.
21
22
23
Read/Write CP_LEAK_UP_EN
Read/Write CP_LEAK_DN_EN
Read/Write CP_HIK_EN
0
1
0
Charge pump up offset. Recommended setting in fractional mode using a
passive loop = 0 decimal. Set to 0 decimal in integer mode.
Charge pump down offset. Recommended setting in fractional mode using a
passive loop equals 1 decimal. Set to 0 decimal in integer mode.
High current charge pump. Low noise, narrow compliance range. Requires an
op amp.
AUTOCALIBRATION REGISTER
Table 27. Register 0x0A, VCO Autocalibration Configuration Register (Default 0x002205), Program to 0x002047
Bit
Type
Name
Default Description
2:0
Read/Write VTUN_RESOLUTION
5
R divider cycles.
0: 1 cycle.
1: 2 cycles.
2: 4 cycles.
…
7: 256 cycles.
9:3
10
11
12
Read/Write Reserved
64d
0
Reserved. Program 8 decimal.
Read/Write VTUN_FORCE_CAPS
Read/Write VTUN_DISABLE
Read/Write VCO_SPI_DISABLE
Force capacitors programmed in Register 0x05. Program 0 for normal operation.
Disable autocalibration. Program 0 for normal operation using VCO autocalibration.
0: normal operation.
0
0
1: this bit disables the serial transfers to the VCO subsystem (via Register 0x05).
These bits set the autocalibration FSM and VSPI clock (100 MHz maximum).
0: input crystal reference.
1: input crystal reference divide by 4.
2: input crystal reference divide by 16.
14:13 Read/Write AUX_CLK_DIV_SEL
1
0
3: input crystal reference divide by 32.
16:15 Read/Write Reserved
Reserved.
Rev. 0 | Page 43 of 51
ADF5610
Data Sheet
PHASE DETECTOR (PD) REGISTER
Table 28. Register 0x0B, PD Register (Default 0x0F8061)
Bit
2:0
4:3
5
Type
Name
Default
Description
Read/Write PFD_DEL_SEL
Read/Write Reserved
1
0
1
1
0
Sets the PFD reset path delay (recommended setting is 001).
Reserved.
Read/Write PFD_UP_EN
Read/Write PFD_DN_EN
Read/Write FIX_CP_MODE
Enables the PFD up output.
6
Enables the PFD down output.
8:7
Cycle slip prevention mode. This delay varies by 10% with temperature
and 12% with process. Extra current (~8 mA) is driven into the loop filter
when the phase error is larger than the following:
0: Disabled.
1: 5.4 ns.
2: 14.4 ns.
3: 24.1 ns.
Only use CSP with comparison frequencies ≤ 50 MHz, otherwise disable.
Always confirm loop stability when using CSP.
9
Read/Write CP_FORCE_UP
Read/Write CP_FORCE_DN
Read/Write CP_FORCE_MID
Read/Write Reserved
0
0
0
Forces CP up output to turn on. Use for test only.
Forces CP down output to turn on. Use for test only.
Forces CP to midrail. Use for test only.
Reserved.
10
11
19:12
248
decimal
(0xF8)
[21:20] Read/Write DIV_PULSEWIDTH_SEL
0
Divider pulse width.
0: shortest.
3: longest.
[23:22] Read/Write Reserved
0
Reserved.
EXACT FREQUENCY MODE REGISTER
Table 29. Register 0x0C, Exact Frequency Mode Register (Default 0x000000)
Bit Type Name Default Description
13:0 Read/Write CHANNELSPERBOUND
Number of channels per fPD. The comparison frequency divided by the correction
0
rate must be an integer. Frequencies at exactly the correction rate have zero
frequency error.
0: disabled.
1: disabled.
2: 16383 decimal (0x3FFF).
GENERAL-PURPOSE, SPI, AND REFERENCE DIVIDER (GPO_SPI_RDIV) REGISTER
Table 30. Register 0x0F, GPO_SPI_RDIV Register (Default 0x000001)
Bit Type
Name
Default Description
4:0 Read/Write GPO_SELECT
1d
The signal selected by this bit is an output to the SDO pin when the SDO pin is
enabled via Register 0x08, Bit 5.
0: data from Register 0x0F, Bit 5.
1: lock detect output.
2: lock detect trigger.
3: lock detect window output.
4: ring oscillator test.
5: pull-up resistor is ~230 Ω from CSP.
6: pull-down resistor is ~230 Ω from CSP.
7: reserved.
8: reference buffer output.
9: reference divider output.
Rev. 0 | Page 44 of 51
Data Sheet
ADF5610
Bit Type
Name
Default Description
10: VCO divider output.
11: modulator clock from VCO divider.
12: auxiliary clock.
13: auxiliary SPI clock.
14: auxiliary SPI enable.
15: auxiliary SPI data output.
16: PD down.
17: PD up.
18: internal clock path third-order, Δ-Σ modulator (DSM3) clock delay.
19: DSM3 core clock.
20: autostrobe integer write.
21: autostrobe fractional write
22: autostrobe auxiliary SPI.
23: SPI latch enable.
24: VCO divider sync reset.
25: seed load strobe.
26 to 29: do not use.
30: SPI output buffer enable.
31: soft reset.
5
6
Read/Write GPO_STATIC_TEST
Read/Write GPO_ALWAYS
0
0
1: GPO test data.
Prevent automux on SDO.
0: automuxes between SDO and GPO data.
1: outputs GPO data only.
7
Read/Write GPO_OUTBUFF_HIGH
0
LDO driver always on.
0: SDO pin driver only on during SPI read cycle.
1: SDO pin driver always on.
Prevents SPI from disabling SDO. Set to 1 if using SPI mode.
Disable P channel MOSFET transistor (PFET).
0: enable SDO pin high drive.
1: disable SDO pin high drive.
Disable N channel MOSFET transistor (NFET).
0: enable SDO pin low drive.
8
9
Read/Write GPO_PULLUP_DISABLE
Read/Write GPO_PULLDN_DISABLE
0
0
1: disable SDO pin low drive.
VCO TUNE REGISTER
The VCO tune register is a read only register.
Table 31. Register 0x10, VCO Tune Register (Default 0x000020)
Bit Type Name Default Description
7:0 Read SAR_DATA_IN 32
Indicates the VCO band setting selected by the autocalibration FSM to yield the nearest free
running VCO frequency to the desired operating frequency. Not valid when Register 0x10, Bit 8 =
1, autocalibration busy. When a manual change is made to the VCO band settings, this register
does not indicate the current VCO band position. VCO subsystems may not use all the MSBs, in
which case the unused bits are don’t care bits.
0: highest frequency.
1: second highest frequency.
…
255: lowest frequency.
8
Read VTUNE_BUSY
0
Busy when the autocalibration FSM is searching for the nearest band setting to the requested
frequency.
Rev. 0 | Page 45 of 51
ADF5610
Data Sheet
SUCCESSIVE APPROXIMATION REGISTER
The successive approximation register (SAR) is a read only register.
Table 32. Register 0x11, Successive Approximation Register (Default 0x07FFFF)
Bit
18:0
19
Type
Read
Read
Name
Default
219 to 1
0
Description
SAR_ERR_ABS_MIN
SAR_ERR_SIGN
SAR error magnitude counts
SAR error sign
0: error is positive
1: error is negative
GENERAL-PURPOSE 2 REGISTER
The GPO2 register is a read only register.
Table 33. Register 0x12, GPO2 Register (Default 0x000000)
Bit
Type
Read
Read
Name
Default
Description
GPO state
0
GPO
0
0
1
LOCK_DETECT
Lock detect status
1: locked
0: unlocked
BUILT IN SELF TEST (BIST) REGISTER
The BIST register is a read only register.
Table 34. Register 0x13, BIST Register (Default 0x001259)
Bit
Type
Name
Default
Description
16:0
Read
Reserved
4697 decimal
Reserved
Rev. 0 | Page 46 of 51
Data Sheet
ADF5610
VCO SUBSYSTEM REGISTER MAP
The VCO subsystem uses indirect addressing via Register 0x05,
unlike the PLL registers. Because the VCO contains three registers
and all are accessed via PLL Register 0x05, it is necessary to add
the VCO Chip ID (Register 0x000) and VCO Register Address
(Register 0x00xx) prior to sending the VCO data bits shown in
Table 35, Table 36, and Table 37. For more detailed information
on how to write to the VCO subsystem, see the VCO Serial Port
Interface (VSPI) section.
Note that VCO Register 0x00, Bit 0 controls the switch that
selects either the internal, temperature compensated tuning
voltage or VT.
The VCO tuning register is write only.
Table 35. VCO Register 0x00, Tuning Register
Bit Type
Name
Default
Description
0
Write
VT_SEL
0
0: VCO tune voltage provided by an internal, temperature compensated calibration voltage
1: VCO tune voltage provided by VT pin
8:1 Write
CAPS
16
VCO subband selection
0000 0000 = Subband 0 (contains maximum frequency)
1111 1111 = Subband 255 (contains minimum frequency)
VCO POWER CONTROL REGISTER
The VCO power control register is a write only register.
Table 36. VCO Register 0x01, Power Control Register
Bit
Type
Name
Default
Description
1:0
Write
RFOUT_PWR_CNTL
0
RFOUT power control
0: 2 dBm
1: 3 dBm
2: 4 dBm
3: 5 dBm
4:2
6:5
Write
Write
VCO_BIAS_CNTL
DIV_PWR_CNTL
0
0
VCO bias control
0: low current mode
7: high performance mode
Divider output power
0: −4 dBm
1: 0 dBm
2: 0 dBm
3: 3 dBm (nominal single-ended power level of PDIV_OUT and NDIV_OUT)
Reserved. Program to 0 decimal
7
8
Write
Write
0
0
DIV_EN
Divider enable
0: power down differential output divider and buffer amplifier
1: enable differential output divider and buffer amplifier
Rev. 0 | Page 47 of 51
ADF5610
Data Sheet
Example: Optimizing Power and Phase Noise
Performance at RFOUT Using the VCO Subsystem
•
•
•
Register 0x05, Bits[8:7] = 11b. RFOUT maximum output
power, 5 dBm nominally.
Register 0x05, Bits[11:9] = 111b. VCO bias, high
performance mode.
Register 0x05, Bits[13:12] = 11b. Differential divider output
power set to maximum. Note that disabling the differential
output divider, VCO Register 0x01, Bit 8, also disables the
differential buffer amplifier.
For optimum phase noise on all ports and maximum power at
the RFOUT port, configure VCO Register 0x01, Bits[4:0] data
bits by accessing the VCO subsystem of the ADF5610. The VCO
subsystem register is accessed via a write to the PLL subsystem
using Register 0x05 = 0011111110001000b = 0x3F88.
The following settings illustrate how settings for VCO
Register 0x01 translate to Register 0x05.
•
•
Register 0x05, Bit 14 = 0b. Reserved.
Register 0x05, Bit 15 = 0b. Disable differential output
divider (improves RFOUT power level by ~1 dBm).
•
Register 0x05, Bits[2:0] = 000b. 3-bit VCO_ID, target
subsystem address = 000b.
VCO DIFFERENTIAL OUTPUT DIVIDER REGISTER
•
Register 0x05, Bits[6:3] = 0001b. VCO subsystem register
address.
This is a write only register. To make use of the divider ratio
settings, the divider must be enabled via VCO Register 0x01,
Bit 8 = 1.
The remaining bits, Register 0x05, Bits[15:7], are data bits for
VCO Register 0x01, Bits[8:0].
Table 37. VCO Register 0x02, Output Divider Register
Bit
Type
Name
Default
Description
2:0
Write
DIFF_DIV_RATIO
0
RF divider ratio.
0: fO/128
1: fO/64.
2: fO/32.
3: fO/16.
4: fO/8.
5: fO/4.
6: fO/2.
7: fO/1 (bypass).
3
Write
Write
PD_VCO
Reserved
0
0
0: enable VCO and buffer amplifier.
1: power down VCO and buffer amplifier.
Reserved. Program 0 decimal.
8:4
Rev. 0 | Page 48 of 51
Data Sheet
ADF5610
Example: Setting the Differential Output Divider to
Divide by 64 via the VCO Subsystem
Example: Changing Frequency after Configuring the
VCO
The VCO registers are written in reverse order (highest to
lowest). First, set the divide ratio then enable the divider. Finally,
select the band.
If writing to the synthesizer for the first time or if the user is
changing frequencies, a write to Register 0x05 = 0 to allow a
subsequent frequency change via autocalibration is required.
The last write in the PLL register sequence triggers the
frequency update. If an integer value results at the phase
detector input, the last write in the register sequence is PLL
Register 0x03. If a fractional value results at the phase detector
input, the last write in the register sequence is PLL Register 0x04.
To write 0 0000 0001 into VCO Register 0x02 (VCO_ID = 000b)
and set the VCO output divider to divide by 64, the following
must be written to Register 0x05 = 0 0000 0001 0010 000 =
0x0090:
•
Register 0x05, Bits[2:0] = 000. 3-bit VCO_ID, target
subsystem address = 000b.
If the band for the desired RFOUT frequency (prior to division)
is already known, the VCO band can be written to directly via
VCO Register 0x00 by disabling the temperature controlled VT
calibration voltage and manually selecting the band via VCO
Register 0x00.
•
•
Register 0x05, Bits[6:3] = 0010. VCO Register 0x02 address.
Register 0x05, Bits[16:7] = 0 0000 0001. Divide by 64,
maximum output RF gain.
Next, use the same settings for Bits[14:0] in the Example:
Optimizing Power and Phase Noise Performance at RFOUT
Using the VCO Subsystem section for VCO Register 0x01. To
enable the differential output divider, simply change VCO
Register 0x01, Bit 8 = 1. The full VCO subsystem write via PLL
Register 0x05 = 0011111110001000 = 0xBF88. The following
settings illustrate how settings for VCO Register 0x01 translate
to Register 0x05.
For example, assume that the desired RFOUT frequency prior
to division is 8 GHz and 8 GHz falls in Subband 205 (see Table 35)
when autocalibration is used at room temperature. First, set
VCO Register 0x02 = 0x0090 for divide by 64 and maximum
output power at PDIV_OUT and NDIV_OUT. Next, set VCO
Register 0x01 = 0xBF88 for maximum output power on the
RFOUT port and the optimal phase noise performance on all
ports. Lastly, configure the VCO Register 0x00 subsystem write
via PLL Register 0x05 = 1100110100000000 = 0xCD00. The
following are individual bit field settings used in Register 0x05
to achieve the write to VCO Register 0x00.
•
•
•
•
Register 0x05, Bits[2:0] = 000b. 3-bit VCO_ID, target
subsystem address = 000b.
Register 0x05, Bits[6:3] = 0001b. VCO subsystem register
address.
Register 0x05, Bits[8:7] = 11b. RFOUT maximum output
power, 5 dBm nominally.
Register 0x05, Bits[11:9] = 111b. VCO bias, high
performance mode.
•
Register 0x05, Bits[2:0] = 000b. 3-bit VCO_ID, target
subsystem address = 000b.
•
Register 0x05, Bits[6:3] = 0000b. VCO subsystem register
address.
•
•
Register 0x05, Bit 7 = 0b. Disable autocalibration.
Register 0x05, Bits[15:8] = 1100101 (205 decimal).
Subband 205 = Subband 77 of the lower VCO core (see
Table 35).
•
•
•
Register 0x05, Bits[13:12] = 11b. Divider output power, 3 dBm.
Register 0x05, Bit 14 = 0b. Reserved.
Register 0x05, Bit 15 = 1b. Enable differential output
divider (note that RFOUT power level drops by ~1 dBm).
If no VCO band change is required, register writes are complete.
Rev. 0 | Page 49 of 51
ADF5610
Data Sheet
APPLICATIONS INFORMATION
The ADF5610 contains two multiband VCOs that cover an
octave frequency range. To ensure optimum performance, it is
essential to provide biasing to all supply pins using low noise
voltage regulators, such as the LT3042 and LT3045. The use of
three regulators is ideal because it maximizes isolation by
allowing one regulator for VCO biasing at 5.0 V, one regulator
for the 3.3 V digital supply pins (VDDLS, VPPCP, VCCPD, and
DVDD), and one regulator for the 3.3 V analog supply pins
(VDD1, VDD2, VDD3, RVDD, BIAS, VCCPS, and VCCHF).
Alternatively, a single LT3045 can be used to combine the 3.3 V
analog and digital supply pins, but combining the pins may lead
to some degradation in spurious performance due to the
reduction in isolation between these circuits.
EVALUATION PRINTED CIRCUIT BOARD (PCB)
The PCB used in the application uses RF circuit design
techniques. Signal lines have 50 Ω impedance, whereas the
package ground leads and exposed pad are connected directly
to the ground plane. Use a sufficient number of thermal vias to
connect the top and bottom ground planes. Use of double
shielded cables with the proper adapters is strongly recommended
for the dc connections. High quality RF cables that can support
microwave frequencies are required for the RF outputs. The
upper end of the frequency range for the RF cable must ideally
include any harmonics that are important to consider at the
highest frequency of operation. The evaluation board details are
available on www.analog.com/EVAL-ADF5610.
EVALUATION KIT CONTENTS
The evaluation kit contains one EV-ADF5610SD1Z evaluation
board, a CD ROM that contains the user manual, evaluation
PCB schematic, evaluation software, and Analog Devices PLL
design software.
Rev. 0 | Page 50 of 51
Data Sheet
ADF5610
OUTLINE DIMENSIONS
0.30
0.25
0.20
7.10
7.00
6.90
0.15
BSC
0.43
0.38
0.33
PIN 1
PIN 1
INDICATOR
CORNER AREA
48
37
36
1
2.56 BSC
SQ
5.50 REF
SQ
25
24
12
13
0.50
BSC
0.48
0.43
0.38
TOP VIEW
SIDE VIEW
BOTTOM VIEW
0.10
BSC
0.53 REF
0.85
MAX
FOR PROPER CONNECTION OF
THE EXPOSED PADS, REFER TO
THE PIN CONFIGURATION AND
FUNCTION DESCRIPTIONS
0.19
0.17
0.15
SEATING
PLANE
SECTION OF THIS DATA SHEET.
Figure 33. 48-Terminal Land Grid Array [LGA]
(CC-48-1)
Dimensions shown in millimeters
ORDERING GUIDE
Lead
Moisture Sensitivity
Level (MSL) Rating4
Package
Option
Model1, 2
Finish3
Temperature Range
−40°C to +85°C
−40°C to +85°C
Package Description
ADF5610BCCZ
ENEPIG
MSL3
MSL3
48-Terminal Land Grid Array [LGA]
48-Terminal Land Grid Array [LGA]
Evaluation Board
CC-48-1
CC-48-1
ADF5610BCCZ-RL7 ENEPIG
EV-ADF5610SD1Z
1 Z = RoHS Compliant Part.
2 USB interface board, EVAL-SDP-CS1Z, must be ordered separately.
3 ENEPIG is electroless nickel electroless palladium immersion gold.
4 See the Absolute Maximum Ratings section for more information.
©2019 Analog Devices, Inc. All rights reserved. Trademarks and
registered trademarks are the property of their respective owners.
D16799-0-2/19(0)
Rev. 0 | Page 51 of 51
相关型号:
©2020 ICPDF网 联系我们和版权申明