AD ADF4193BCPZ-RL7

Low Phase Noise, Fast Settling PLL
Frequency Synthesizer
ADF4193
FEATURES
GENERAL DESCRIPTION
New, fast settling, fractional-N PLL architecture
Single PLL replaces ping-pong synthesizers
Frequency hop across GSM band in 5 μs with phase settled
by 20 μs
0.5° rms phase error at 2 GHz RF output
Digitally programmable output phase
RF input range up to 3.5 GHz
3-wire serial interface
On-chip, low noise differential amplifier
Phase noise figure of merit: −216 dBc/Hz
Loop filter design possible using ADI SimPLL
The ADF4193 frequency synthesizer can be used to implement
local oscillators in the upconversion and downconversion
sections of wireless receivers and transmitters. Its architecture
is specifically designed to meet the GSM/EDGE lock time
requirements for base stations. It consists of a low noise, digital
phase frequency detector (PFD), and a precision differential
charge pump. There is also a differential amplifier to convert
the differential charge pump output to a single-ended voltage
for the external voltage-controlled oscillator (VCO).
The Σ-Δ based fractional interpolator, working with the N
divider, allows programmable modulus fractional-N division.
Additionally, the 4-bit reference (R) counter and on-chip
frequency doubler allow selectable reference signal (REFIN)
frequencies at the PFD input. A complete phase-locked loop
(PLL) can be implemented if the synthesizer is used with an
external loop filter and a VCO. The switching architecture
ensures that the PLL settles inside the GSM time slot guard
period, removing the need for a second PLL and associated
isolation switches. This decreases cost, complexity, PCB area,
shielding, and characterization on previous ping-pong GSM
PLL architectures.
APPLICATIONS
GSM/EDGE base stations
PHS base stations
Instrumentation and test equipment
FUNCTIONAL BLOCK DIAGRAM
SDVDD
DVDD1
DVDD2
DVDD3
AVDD1
VP 1
VP2
VP3
RSET
REFERENCE
/2
DIVIDER
+
PHASE
FREQUENCY
DETECTOR
–
SW1
CPOUT+
CHARGE +
PUMP –
CPOUT–
SW2
VDD
HIGH Z
MUXOUT
4-BIT R
COUNTER
×2
DOUBLER
REFIN
DGND
OUTPUT
MUX
CMR
LOCK DETECT
DIFFERENTIAL
AMPLIFIER
–
AIN–
+
AIN+
RDIV
NDIV
AOUT
N COUNTER
SW3
FRACTIONAL
INTERPOLATOR
CLK
DATA
LE
24-BIT
DATA
REGISTER
RFIN+
RFIN–
FRACTION
REG
MODULUS
REG
INTEGER
REG
AGND1
AGND2
DGND1
DGND2
DGND3
SDGND
SWGND
05328-001
ADF4193
Figure 1.
Rev. B
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
www.analog.com
Fax: 781.461.3113
©2006 Analog Devices, Inc. All rights reserved.
ADF4193
TABLE OF CONTENTS
Features .............................................................................................. 1
Function Register (R3) .............................................................. 18
Applications....................................................................................... 1
Charge Pump Register (R4) ...................................................... 19
General Description ......................................................................... 1
Power-Down Register (R5) ....................................................... 20
Functional Block Diagram .............................................................. 1
Mux Register (R6) ...................................................................... 21
Revision History ............................................................................... 2
Programming .................................................................................. 22
Specifications..................................................................................... 3
Worked Example ........................................................................ 22
Timing Characteristics ................................................................ 4
Spur Mechanisms ....................................................................... 22
Absolute Maximum Ratings............................................................ 5
Power-Up Initialization ............................................................. 23
ESD Caution.................................................................................. 5
Changing the Frequency of the PLL and the Phase Look-Up
Table ............................................................................................. 23
Pin Configuration and Function Descriptions............................. 6
Typical Performance Characteristics ............................................. 8
Theory of Operation ...................................................................... 11
Reference Input Section............................................................. 11
RF Input Stage............................................................................. 11
Register Map.................................................................................... 14
FRAC/INT Register (R0)........................................................... 15
Applications..................................................................................... 25
Local Oscillator for A GSM Base Station ................................ 25
Interfacing ................................................................................... 27
PCB Design Guidelines for Chip Scale Package .................... 27
Outline Dimensions ....................................................................... 28
Ordering Guide .......................................................................... 28
MOD/R Register (R1) ................................................................ 16
Phase Register (R2) .................................................................... 17
REVISION HISTORY
6/06—Rev A. to Rev. B
Changes to Table 1............................................................................ 3
Changes to Figure 32...................................................................... 18
Changes to Power-Up Initialization Section............................... 23
Changes to Timer Values for Tx Section and Timer Values for
Rx Section ........................................................................................ 25
11/05—Rev 0. to Rev. A
Updated Format..................................................................Universal
Changes to Features Section............................................................ 1
Changes to Table 1............................................................................ 3
Changes to Reference Input Section ............................................ 11
Changes to RF N Divider Section ................................................ 11
Changes to the Lock Detect Section ............................................ 13
Changes to Figure 29...................................................................... 15
Changes to the 8-Bit INT Value Section ..................................... 15
Changes to Figure 33...................................................................... 19
Replaced Figure 35 ......................................................................... 21
Changes to the Σ-Δ and Lock Detect Modes Section................ 21
Changes to the Power-Up Initialization Section ........................ 23
Changes to Table 8.......................................................................... 23
Changes to the Local Oscillator for a GSM
Base Station Section ....................................................................... 25
Changes to the Timer Values for Rx Section .............................. 25
Changes to Figure 36...................................................................... 26
Updates to the Outline Dimensions ............................................ 28
Changes to the Ordering Guide ................................................... 28
4/05—Revision 0: Initial Version
Rev. B | Page 2 of 28
ADF4193
SPECIFICATIONS
AVDD = DVDD = SDVDD = 3 V ± 10%, VP1, VP2 = 5 V ± 10%, VP3 = 5.35 V ± 5%, AGND = DGND = GND = 0 V, RSET = 2.4 kΩ, dBm
referred to 50 Ω, TA = TMIN to TMAX, unless otherwise noted.
Table 1.
Parameter
RF CHARACTERISTICS
RF Input Frequency (RFIN)
RF Input Sensitivity
Maximum Allowable Prescaler Output Frequency 2
REFIN CHARACTERISTICS
REFIN Input Frequency
REFIN Edge Slew Rate
REFIN Input Sensitivity
REFIN Input Capacitance
REFIN Input Current
PHASE DETECTOR
Phase Detector Frequency
CHARGE PUMP
ICP Up/Down
High Value
Low Value
Absolute Accuracy
RSET Range
ICP Three-State Leakage
ICP Up vs. Down Matching
ICP vs. VCP
ICP vs. Temperature
DIFFERENTIAL AMPLIFIER
Input Current
Output Voltage Range
VCO Tuning Range
Output Noise
LOGIC INPUTS
VIH, Input High Voltage
VIL, Input Low Voltage
IINH, IINL, Input Current
CIN, Input Capacitance
LOGIC OUTPUTS
VOH, Output High Voltage
VOL, Output Low Voltage
POWER SUPPLIES
AVDD
DVDD
VP1, VP2
VP3
IDD (AVDD + DVDD + SDVDD)
IDD (VP1 + VP2)
IDD (VP3)
IDD Power-Down
B Version 1
Unit
Test Conditions/Comments
0.4/3.5
–10/0
470
GHz min/max
dBm min/max
MHz max
See Figure 21 for input circuit
300
300
0.7/VDD
0 to VDD
10
±100
MHz max
V/μs min
V p-p min/max
V max
pF max
μA max
For f > 120 MHz, set REF/2 bit = 1
26
MHz max
6.6
104
5
1/4
1
0.1
1
1
mA typ
μA typ
% typ
kΩ min/max
nA typ
% typ
% typ
% typ
1
1.4/(VP3 − 0.3)
1.8/(VP3 − 0.8)
7
nA typ
V min/max
V min/max
nV/√Hz typ
1.4
0.7
±1
10
V min
V max
μA max
pF max
VDD – 0.4
0.4
V min
V max
2.7/3.3
AVDD
4.5/5.5
5.0/5.65
27
27
30
10
V min/V max
V min/V max
V min/V max
mA max
mA max
mA max
μA typ
Rev. B | Page 3 of 28
AC-coupled
CMOS-compatible
With RSET = 2.4 kΩ
With RSET = 2.4 kΩ
Nominally RSET = 2.4 kΩ
0.75 V ≤ VCP ≤ VP – 1.5 V
0.75 V ≤ VCP ≤ VP – 1.5 V
0.75 V ≤ VCP ≤ VP – 1.5 V
@ 20 kHz offset
IOH = 500 μA
IOL = 500 μA
AVDD ≤ VP1, VP2 ≤ 5.5 V
VP1, VP2 ≤ VP3 ≤ 5.65 V
22 mA typ
22 mA typ
24 mA typ
ADF4193
Parameter
SW1, SW2, and SW3
RON (SW1 and SW2)
RON SW3
NOISE CHARACTERISTICS
900 MHz Output 3
1800 MHz Output 4
Phase Noise Figure of Merit 5
B Version 1
Unit
65
75
Ω typ
Ω typ
–108
–102
–216
dBc/Hz typ
dBc/Hz typ
dBc/Hz typ
Test Conditions/Comments
@ 5 kHz offset and 26 MHz PFD frequency
@ 5 kHz offset and 13 MHz PFD frequency
@ VCO output with dither off
1
Operating temperature range is from –40°C to +85°C.
The prescaler value is chosen to ensure that the RF input is divided down to a frequency that is less than this value.
3
fREFIN = 26 MHz; fSTEP = 200 kHz; fRF = 900 MHz; loop BW = 40 kHz.
4
fREFIN = 13 MHz; fSTEP = 200 kHz; fRF = 1850 MHz; loop BW = 60 kHz.
5
Calculated from the phase noise measured at 5 kHz with a 60 kHz loop BW. Increased noise contribution from the differential amplifier if the loop BW is reduced.
2
TIMING CHARACTERISTICS
AVDD = DVDD = 3 V ± 10%, VP1, VP2 = 5 V ± 10%, VP3 = 5.35 V ± 5%, AGND = DGND = GND = 0 V, RSET = 2.4 kΩ, dBm referred to
50 Ω, TA = TMIN to TMAX, unless otherwise noted.
Table 2.
Limit (B Version) 1
10
10
10
15
15
10
15
Parameter
t1
t2
t3
t4
t5
t6
t7
Test Conditions/Comments
LE setup time
DATA to CLOCK setup time
DATA to CLOCK hold time
CLOCK high duration
CLOCK low duration
CLOCK to LE setup time
LE pulse width
Operating temperature is from −40°C to +85°C.
t4
t5
CLK
t2
DATA
DB23
(MSB)
t3
DB22
DB2
DB1
(CONTROL BIT C2)
DB0 (LSB)
(CONTROL BIT C1)
t7
LE
t1
t6
LE
05238-002
1
Unit
ns min
ns min
ns min
ns min
ns min
ns min
ns min
Figure 2. Timing Diagram
Rev. B | Page 4 of 28
ADF4193
ABSOLUTE MAXIMUM RATINGS
TA = 25°C, unless otherwise noted.
Table 3.
Parameter
AVDD to GND
AVDD to DVDD, SDVDD
VP to GND
VP to AVDD
Digital I/O Voltage to GND
Analog I/O Voltage to GND
REFIN, RFIN+, RFIN− to GND
Operating Temperature Range
Industrial (B Version)
Storage Temperature Range
Maximum Junction Temperature
LFCSP θJA Thermal Impedance
(Paddle Soldered)
Reflow Soldering
Peak Temperature
Time at Peak Temperature
Rating
–0.3 V to +3.6 V
–0.3 V to +0.3 V
–0.3 V to +5.8 V
–0.3 V to +5.8 V
–0.3 V to VDD + 0.3 V
–0.3 V to VP + 0.3 V
–0.3 V to VDD + 0.3 V
–40°C to +85°C
–65°C to +125°C
150°C
27.3°C/W
Stresses above those listed under Absolute Maximum Ratings
may cause permanent damage to the device. This is a stress
rating only; functional operation of the device at these or any
other conditions above those indicated in the operational
section of this specification is not implied. Exposure to absolute
maximum rating conditions for extended periods may affect
device reliability.
This device is a high performance RF integrated circuit with an
ESD rating of <2 kV, and it is ESD sensitive. Proper precautions
need to be taken for handling and assembly.
Transistor Count
75,800 (MOS), 545 (BJT)
260°C
40 sec
ESD CAUTION
ESD (electrostatic discharge) sensitive device. Electrostatic charges as high as 4000 V readily accumulate on
the human body and test equipment and can discharge without detection. Although this product features
proprietary ESD protection circuitry, permanent damage may occur on devices subjected to high energy
electrostatic discharges. Therefore, proper ESD precautions are recommended to avoid performance
degradation or loss of functionality.
Rev. B | Page 5 of 28
ADF4193
32 VP3
31 AIN+
30 CPOUT+
29 SW1
28 SWGND
27 SW2
26 CPOUT–
25 AIN–
PIN CONFIGURATION AND FUNCTION DESCRIPTIONS
1
2
3
4
5
6
PIN 1
INDICATOR
ADF4193
TOP VIEW
24 VP2
23 RSET
22 AGND2
21 DGND3
20 VP1
19 LE
18 DATA
8
17 CLK
05328-003
7
DGND1 9
DVDD2 10
REFIN 11
DGND2 12
DVDD3 13
SDGND 14
SDVDD 15
MUXOUT 16
CMR
AOUT
SW3
AGND1
RFIN–
RFIN+
AVDD1
DVDD1
Figure 3. Pin Configuration
Table 4. Pin Function Descriptions
Pin No.
1
Mnemonic
CMR
2
3
4
5
AOUT
SW3
AGND1
RFIN−
6
7
RFIN+
AVDD1
8
DVDD1
9
10
DGND1
DVDD2
11
REFIN
12
13
14
15
DGND2
DVDD3
SDGND
SDVDD
16
MUXOUT
17
CLK
18
DATA
19
LE
20
VP1
21
22
DGND3
AGND2
Description
Common-Mode Reference Voltage for the Differential Amplifier’s Output Voltage Swing. Internally biased to
three-fifths of VP3. Requires a 0.1 μF capacitor to ground.
Differential Amplifier Output to Tune the External VCO.
Fast-Lock Switch 3. Closed while SW3 timeout counter is active.
Analog Ground. This is the ground return pin for the differential amplifier and the RF section.
Complementary Input to the RF Prescaler. This point must be decoupled to the ground plane with a small bypass
capacitor, typically 100 pF.
Input to the RF Prescaler. This small signal input is ac-coupled to the external VCO.
Power Supply Pin for the RF Section. Nominally 3 V. A 100 pF decoupling capacitor to the ground plane should be
placed as close as possible to this pin.
Power Supply Pin for the N Divider. Should be the same voltage as AVDD1. A 0.1 μF decoupling capacitor to ground
should be placed as close as possible to this pin.
Ground Return Pin for DVDD1.
Power Supply Pin for the REFIN Buffer and R Divider. Nominally 3 V. A 0.1 μF decoupling capacitor to ground
should be placed as close as possible to this pin.
Reference Input. This is a CMOS input with a nominal threshold of VDD/2 and a dc equivalent input resistance of
100 kΩ (see Figure 15). This input can be driven from a TTL or CMOS crystal oscillator or it can be ac-coupled.
Ground Return Pin for DVDD2 and DVDD3.
Power Supply Pin for the Serial Interface Logic. Nominally 3 V.
Ground Return Pin for the Σ-Δ Modulator.
Power Supply Pin for the Digital Σ-Δ Modulator. Nominally 3 V. A 0.1 μF decoupling capacitor to the ground plane
should be placed as close as possible to this pin.
Multiplexer Output. This multiplexer output allows either the lock detect, the scaled RF, or the scaled reference
frequency to be accessed externally (see Figure 35).
Serial Clock Input. Data is clocked into the 24-bit shift register on the CLK rising edge. This input is a high
impedance CMOS input.
Serial Data Input. The serial data is loaded MSB first with the three LSBs as the control bits. This input is a high
impedance CMOS input.
Load Enable, CMOS Input. When LE goes high, the data stored in the shift register is loaded into the register that is
selected by the three LSBs.
Power Supply Pin for the Phase Frequency Detector (PFD). Nominally 5 V, should be at the same voltage at VP2.
A 0.1 μF decoupling capacitor to ground should be placed as close as possible to this pin.
Ground Return Pin for VP1.
Ground Return Pin for VP2.
Rev. B | Page 6 of 28
ADF4193
Pin No.
23
Mnemonic
RSET
Description
Connecting a resistor between this pin and GND sets the charge pump output current. The nominal voltage bias at
the RSET pin is 0.55 V. The relationship between ICP and RSET is
ICP = 0.25/RSET
24
VP2
25
26
27
28
29
30
31
32
AIN−
CPOUT−
SW2
SWGND
SW1
CPOUT+
AIN+
VP3
So, with RSET = 2.4 kΩ, ICP = 104 μA.
Power Supply Pin for the Charge Pump. Nominally 5 V, should be at the same voltage at VP1. A 0.1 μF decoupling
capacitor to ground should be placed as close as possible to this pin.
Differential Amplifier’s Negative Input Pin.
Differential Charge Pump’s Negative Output Pin. Should be connected to AIN− and the loop filter.
Fast Lock Switch 2. This switch is closed to SWGND while the SW1/2 timeout counter is active.
Common for SW1 and SW2 Switches. Should be connected to the ground plane.
Fast Lock Switch 1. This switch is closed to SWGND while the SW1/2 timeout counter is active.
Differential Charge Pump’s Positive Output Pin. Should be connected to AIN+ and the loop filter.
Differential Amplifier’s Positive Input Pin.
Power Supply Pin for the Differential Amplifier. This can range from 5.0 V to 5.5 V. A 0.1 μF decoupling capacitor to
ground should be placed as close as possible to this pin. Also requires a 10 μF decoupling capacitor to ground.
Rev. B | Page 7 of 28
ADF4193
TYPICAL PERFORMANCE CHARACTERISTICS
0
FREQ. UNIT GHz KEYWORD R
PARAM TYPE S
IMPEDANCE 50
DATA FORMAT
MA
MAGS11
0.67107
0.66556
0.6564
0.6333
0.61406
0.5977
0.5655
0.5428
0.51733
0.49909
0.47309
0.45694
0.44698
0.43589
0.42472
0.41175
0.41055
0.40983
–5
ANGS11
–75.8206
–77.6851
–80.3101
–82.5082
–85.5623
–87.3513
–89.7605
–93.0239
–95.9754
–99.1291
–102.208
–106.794
–111.659
–117.986
–125.62
–133.291
–140.585
–147.97
4/5 PRESCALER
–10
8/9 PRESCALER
–15
–20
–25
–30
–35
05328-005
FREQ.
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3.0
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
4.0
0
–30
–30
GSM900 Rx SETUP, 40kHz LOOP BW, DITHER OFF
RF = 1092.8MHz, FREF = 26MHz, MOD = 130
N = 42 4/130
INTEGER BOUNDARY SPUR: –103dBc @ 800kHz
–40
–50
–60
PHASE NOISE (dBc/Hz)
–70
–80
–90
–100
–110
–120
–130
–90
–100
–110
–120
–130
–140
–150
–160
10k
100k
1M
10M
–160
–170
1k
100M
FREQUENCY (Hz)
–60
DCS1800 Tx SETUP WITH DITHER OFF,
60kHz LOOP BW, 13MHz PFD.
MEASURED ON EVAL-ADF4193-EB1 BOARD
1M
10M
100M
DCS1800 Tx SETUP WITH DITHER OFF,
60kHz LOOP BW, 13MHz PFD.
MEASURED ON EVAL-ADF4193-EB1 BOARD
–70
400kHz SPURS @ 25°C
–80
–90
–100
–80
600kHz SPURS @ 25°C
–90
–100
–110
400kHz SPURS @ 85°C
1859
FREQUENCY (MHz)
05328-010
–110
–120
1846
100k
Figure 8. SSB Phase Noise Plot at 1842.6 MHz (DCS1800 Tx Setup)
SPUR LEVEL (dBc)
SPUR LEVEL (dBc)
–70
10k
FREQUENCY (Hz)
Figure 5. SSB Phase Noise Plot at 1092.8 MHz (GSM900 Rx Setup) vs.
Free Running VCO Noise
–60
DCS1800 Tx SETUP, 60kHz LOOP BW, DITHER OFF
RF = 1842.6MHz, FREF = 13MHz, MOD = 65
DSB INTEGRATED PHASE ERROR = 0.46° RMS
SIRENZA 1843T VCO
–80
–150
–170
1k
5000
–70
–140
05328-006
PHASE NOISE (dBc/Hz)
–60
4000
–120
1846
1872
Figure 6. 400 kHz Fractional Spur Levels Across All DCS1800 Tx Channels over
Two-Integer Multiples of the PFD Reference
600kHz SPURS @ 85°C
1859
FREQUENCY (MHz)
05328-011
–50
2000
3000
RFIN FREQUENCY (MHz)
Figure 7. RF Input Sensitivity
Figure 4. S Parameter Data for the RF Input
–40
1000
05328-007
ANGS11
–16.6691
–19.9279
–23.561
–26.9578
–30.8201
–34.9499
–39.0436
–42.3623
–46.322
–50.3484
–54.3545
–57.3785
–60.695
–63.9152
–66.4365
–68.4453
–70.7986
–73.7038
RFIN LEVEL (dBm)
MAGS11
0.8897
0.87693
0.85834
0.85044
0.83494
0.81718
0.80229
0.78917
0.77598
0.75578
0.74437
0.73821
0.7253
0.71365
0.70699
0.7038
0.69284
0.67717
05328-038
FREQ.
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2.0
2.1
2.2
1872
Figure 9. 600 kHz Fractional Spur Levels Across All DCS1800 Tx Channels over
Two-Integer Multiples of the PFD Reference
Rev. B | Page 8 of 28
ADF4193
5
5
4
DCS1800 Tx SETUP, 60kHz LOOP BW.
MEASURED ON EVAL-ADF4193-EB1
EVALUATION BOARD.
TIMERS: ICP = 28, SW1/SW2, SW3 = 35.
FREQUENCY LOCK IN WIDE BW MODE @ 5μs.
4
VTUNE
3
3
CPOUT–
(V)
(V)
CPOUT+
2
2
VTUNE
1
CPOUT+
0
1
2
3
4
5
TIME (μs)
6
7
8
05328-040
0
–1
0
–1
9
Figure 10. VTUNE Settling Transient for a 75 MHz Jump from 1818 MHz to
1893 MHz with Sirenza 1843T VCO
50
30
10
0
–40°C
–20
+85°C
3
4
5
TIME (μs)
6
7
8
9
20
10
0
–10
–40°C
–20
+85°C
–40
–50
–5
0
5
10
15
20
25
TIME (μs)
30
35
40
05328-009
–30
05328-008
–30
–40
–50
–5
45
Figure 11. Phase Settling Transient for a 75 MHz Jump from 1818 MHz to
1893 MHz (VTUNE 1.8 V to 3.7 V with Sirenza 1843T VCO)
0
5
10
15
20
25
TIME (μs)
30
35
40
45
Figure 14. Phase Settling Transient for a 75 MHz Jump from 1893 MHz to
1818 MHz (VTUNE = 3.7 V to 1.8 V with Sirenza 1843T VCO)
2.0
8
ICPOUT + P, ICPOUT – P
IUP = | ICPOUT + P | + | ICPOUT – N |
IDOWN = | ICPOUT – P | + | ICPOUT + N |
1.0
CHARGE PUMP MISMATCH (%)
0
0
–2
–0.5
NORMAL OPERATING RANGE
–4
–1.0
–6
–1.5
(V)
0.5
2
AOUT (= VTUNE )
4
MISMATCH (%)
4
VP1 = VP2 = 5V
VP3 = 5.5V
VCMR = 3.3V
5
1.5
6
ICP (mA)
2
DCS1800 Tx SETUP, 60kHz LOOP BW.
MEASURED ON EVAL-ADF4193-EB1
EVALUATION BOARD WITH AD8302
PHASE DETECTOR.
TIMERS: ICP = 28, SW1/SW2, SW3 = 35.
+25°C PEAK PHASE ERROR < 5° @ 19.2μs
40
20
–10
1
50
PHASE ERROR (Degrees)
PHASE ERROR (Degrees)
30
0
Figure 13. VTUNE Settling Transient for a 75 MHz Jump Down from 1893 MHz to
1818 MHz, the Bottom of the Allowed Tuning Range with the Sirenza 1843T VCO
DCS1800 Tx SETUP, 60kHz LOOP BW.
MEASURED ON EVAL-ADF4193-EB1
EVALUATION BOARD WITH AD8302
PHASE DETECTOR.
TIMERS: ICP = 28, SW1/SW2, SW3 = 35.
+25°C PEAK PHASE ERROR < 5° @ 17.8μs
40
05328-041
CPOUT–
DCS1800 Tx SETUP, 60kHz LOOP BW.
MEASURED ON EVAL-ADF4193-EB1
EVALUATION BOARD.
TIMERS: ICP = 28, SW1/SW2, SW3 = 35.
FREQUENCY LOCK IN WIDE BW MODE @ 4μs.
1
3
CPOUT+ (= AIN+)
2
ICPOUT + N, ICPOUT – N
0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
–2.0
5.0
CPOUT– (= AIN–)
0
1780
CPOUT + / CPOUT – VOLTAGE (V)
Figure 12. Differential Charge Pump Output Compliance Range and
Charge Pump Mismatch with VP1 = VP2 = 5 V
1800
1820
1840
1860
1880
1900
1920
05328-013
–8
05328-012
1
1940
FREQUENCY (MHz)
Figure 15. Tuning Range with a Sirenza 1843T VCO and a 5.5 V Differential
Amplifier Power Supply Voltage
Rev. B | Page 9 of 28
ADF4193
100
7nV/ Hz @ 20kHz
10
10k
100k
FREQUENCY (Hz)
1M
1.2
0.9
0.6
0.3
0
10M
Figure 16. Voltage Noise Density Measured at the Differential Amplifier Output
1.5
0
13
26
+85°C
SW1/
SW2
+85°C
+25°C
70
REFIN
40
AGILENT
HP8663A
SIG. GEN.
TUNING VOLTAGE RANGE
130
1880MHz
20
3
4
TEKTRONIX
TDS714L
OSCILLOSCOPE
AD8302
EVB
1880MHz
R&S
SMT03
SIG. GEN.
05328-014
10
2
VPHS
INPB
10MHz
EXT REF
30
1
117
INPA
–40°C
0
104
RFOUT
1805
+25°C
50
0
91
ADF4193
EVAL BOARD
–40°C
60
RON (Ω)
104MHz
5dBm
SW3
80
52
65
78
PHASE CODE
Figure 18. Detected RF Output Phase for Phase Code Sweep from 0 to MOD
100
90
39
5
INTERVAL BETWEEN R0 WRITES SHOULD BE A MULTIPLE OF MOD
REFERENCE CYCLES (5µs) FOR COHERENT PHASE MEASUREMENTS
DRAIN VOLTAGE (V)
Figure 19. Test Setup for Phase Lock Time Measurement
Figure 17. On Resistance of Loop Filter Switches SW1/SW2 and SW3
Rev. B | Page 10 of 28
05328-045
1
1k
MEASURED USING AD8302 PHASE DETECTOR
Y-AXIS SCALE: 10mV/DEGREE
RF = 1880MHz, PFD = 26MHz, MOD = 130
X-AXIS SCALE: 2.77°/STEP
05328-044
PHASE DETECTOR OUTPUT (V)
1.8
05328-042
NOISE (nV/ Hz)
1000
ADF4193
THEORY OF OPERATION
The ADF4193 is targeted at GSM base station requirements,
specifically to eliminate the need for ping-pong solutions. It
works based on fast lock, using a wide loop bandwidth during a
frequency change and narrowing the loop bandwidth once
frequency lock is achieved. Widening the loop bandwidth is
achieved by increasing the charge pump current. Switches are
included to change the loop filter component values to maintain
stability with the changing charge pump current. The narrow
loop bandwidth ensures that phase noise and spur specifications
are met. A differential charge pump and loop filter topology are
used to ensure that the fast lock time benefit from widening the
loop bandwidth is maintained when the loop is restored to
narrow bandwidth mode for normal operation.
RF INPUT STAGE
The RF input stage is shown in Figure 21. It is followed by a
2-stage limiting amplifier to generate the CML clock levels
needed for the prescaler. Two prescaler options are selectable: a
4/5 and an 8/9. The 8/9 prescaler is selected for N divider values
greater than 80.
BIAS
GENERATOR
500Ω
100kΩ
TO R COUNTER
S3
NO
05328-016
BUFFER
S1
Figure 21. RF Input Stage
RF N Divider
The RF N divider allows a fractional division ratio in the PLL
feedback path. The integer and fractional parts of the division
are programmed using separate registers, as shown in Figure 22
and described in the INT, FRAC, and MOD Relationship
section. Integer division ratios from 26 to 255 are allowed and a
third-order, Σ-Δ modulator interpolates the fractional value
between the integer steps.
POWER-DOWN
CONTROL
S2
AGND
05328-017
RFIN–
The reference input stage is shown in Figure 20. Switches S1 and
S2 are normally closed, and S3 is normally open. During powerdown, S3 is closed, and S1 and S2 are opened to ensure that
there is no loading of the REFIN pin. The falling edge of REFIN is
the active edge at the positive edge triggered PFD.
REFIN NC
500Ω
RFIN+
REFERENCE INPUT SECTION
NC
AVDD
1.6V
Figure 20. Reference Input Stage
RF N DIVIDER
FROM RF
INPUT STAGE
The 4-bit R counter allows the input reference frequency to be
divided down to produce the reference clock to the phase
frequency detector (PFD). A toggle flip-flop can be optionally
inserted after the R counter to give a further divide-by-2. Using
this option has the additional advantage of ensuring that the
PFD reference clock has a 50/50 mark-space ratio. This ratio
gives the maximum separation between the fast lock timer
clock, which is generated off the falling edge of the PFD
reference, and the rising edge, which is the active edge in the
PFD. It is recommended that this toggle flip-flop be enabled for
all even R divide values greater than 2. It must be enabled if
dividing down a REFIN frequency that is greater than 120 MHz.
An optional doubler before the 4-bit R counter can be used for
low REFIN frequencies, up to 20 MHz. With these programmable
options, reference division ratios from 0.5 to 30 between REFIN
and the PFD are possible.
TO PFD
N COUNTER
THIRD-ORDER
FRACTIONAL
INTERPOLATOR
INT
REG
MOD
REG
FRAC
VALUE
05328-018
R Counter and Doubler
N = INT + FRAC/MOD
Figure 22. Fractional-N Divider
INT, FRAC, and MOD Relationship
The INT, FRAC, and MOD values, programmed through the
serial interface, make it possible to generate RF output frequencies
that are spaced by fractions of the PFD reference frequency.
The N divider value, shown inside the brackets of the following
equation for the RF VCO frequency (RFOUT), is made up of an
integer part (INT) and a fractional part (FRAC/MOD):
RFOUT = FPFD × [INT + (FRAC/MOD)]
where:
RFOUT is the output frequency of the external VCO.
FPFD is the PFD reference frequency.
Rev. B | Page 11 of 28
ADF4193
PFD and Charge Pump
The PFD takes inputs from the R divider and N divider and
produces up and down outputs with a pulse width difference
proportional to the phase difference between the inputs. The
charge pump outputs a net up or down current pulse of a width
equal to this difference, to pump up or pump down the voltage
that is integrated onto the loop filter, which in turn increases or
decreases the VCO output frequency. If the N divider phase lags
the R divider phase, a net up current pulse is produced that
increases the VCO frequency (and thus the phase). If the
N divider phase leads the R divider edge, then a net down pulse
is produced to reduce the VCO frequency and phase. Figure 23
is a simplified schematic of the PFD and charge pump. The
charge pump is made up of an array of 64 identical cells, each of
which is fully differential. All 64 cells are active during fast lock,
but only one is active during normal operation. Because a
single-ended control voltage is required to tune the VCO, an
on-chip, differential-to-single-ended amplifier is provided for
this purpose. In addition, because the phase-lock loop only
controls the differential voltage generated across the charge
pump outputs, an internal common-mode feedback (CMFB)
loop biases the charge pump outputs at a common-mode
voltage of approximately 2 V.
D
R DIVIDER
CLR
CHARGE
PUMP
ARRAY
[64:1]
CMFB
CLR
CPOUT–
Q
N DIVIDER
EN[64:1]
05328-019
D
P
VBIAS P
P
UP
DOWN
CPOUT+
CPOUT–
DOWN
UP
N
VBIAS N
N
Figure 24. Differential Charge Pump Cell with External Loop Filter Components
Fast Lock Timeout Counters
CPOUT+
Q
To pump up, the up switches are on and PMOS current is
sourced out through CPOUT+; this increases the voltage on the
external loop filter capacitors connected to CPOUT+. Similarly,
the NMOS current sink on CPOUT− decreases the voltage on the
external loop filter capacitors connected to CPOUT−. Therefore,
the differential voltage between CPOUT+ and CPOUT− increases.
To pump down, PMOS current sources out through CPOUT− and
NMOS current sinks in through CPOUT+, which decreases the
(CPOUT+, CPOUT−) differential voltage. The charge pump up/
down matching is improved by an order of magnitude over the
conventional single-ended charge pump that depended on the
matching of two different device types. The up/down matching
in this structure depends on how a PMOS matches a PMOS and
an NMOS matches an NMOS.
05328-035
The value of MOD is chosen to give the desired channel step
with the available reference frequency. Thereafter, program the
INT and FRAC words for the desired RF output frequency. See
the Worked Example section for more information.
Figure 23. PFD and Differential Charge Pump Simplified Schematic
Differential Charge Pump
The charge pump cell (see Figure 24) has a fully differential
design for best up-to-down current matching. Good matching
is essential to minimize the phase offset created when switching
the charge pump current from its high value (in fast lock mode)
to its nominal value (in normal mode).
Timeout counters, clocked at one quarter the PFD reference
frequency, are provided to precisely control the fast locking
operation (see Figure 25). Whenever a new frequency is
programmed, the fast lock timers start and the PLL locks into
wide BW mode with the 64 identical 100 μA charge pump cells
active (6.4 mA total). When the ICP counter times out, the
charge pump current is reduced to 1× by deselecting cells in
binary steps over the next six timer clock cycles, until just one
100 μA cell is active. The charge pump current switching from
6.4 mA to 100 μA equates to an 8-to-1 change in loop bandwidth. The loop filter must be changed to ensure stability when
this happens. That is the job of the SW1, SW2, and SW3 switches.
The application circuit (shown in Figure 36) shows how they
can be used to reconfigure the loop filter time constants. The
application circuits close to short out external loop filter resistors
during fast lock and open when their counters time out to restore
the filter time constants to their normal values for the 100 μA
charge pump current. Because it takes six timer clock cycles to
reduce the charge pump current to 1×, it is recommended that
both switch timers be programmed to the value of the ICP
timer + 7.
Rev. B | Page 12 of 28
ADF4193
START
MUXOUT and Lock Detect
ICP
TIMEOUT
COUNTER
FPFD
SW1/SW2
TIMEOUT
COUNTER
The output multiplexer on the ADF4193 allows the user to
access various internal points on the chip. The state of MUXOUT
is controlled by M4 to M1 in the MUX register. Figure 35 shows
the full truth table. Figure 27 shows the MUXOUT section in
block diagram form.
SW3
TIMEOUT
COUNTER
÷4
SW3
CHARGE PUMP
ENABLE LOGIC
AOUT
SW1
SW2
EN[64:1]
SWGND
DVDD
05328-036
WRITE
TO R0
LOGIC LOW
SERIAL DATA OUTPUT
R DIVIDER OUTPUT
N DIVIDER OUTPUT
Figure 25. Fast Lock Timeout Counters
MUX
MUXOUT
CONTROL
THREE-STATE OUTPUT
Differential Amplifier
TIMER OUTPUTS
VAOUT = (VAIN+ − VAIN−) + VCMR
The CMR offset voltage is internally biased to three-fifths of
VP3, the differential amplifier power supply voltage, as shown in
Figure 26. Connect a 0.1 μF capacitor to ground to the CMR pin
to roll off the thermal noise of the biasing resistors.
As can be seen in Figure 15, the differential amplifier output
voltage behaves according to the previous equation over a 4 V
range from approximately 1.2 V minimum up to VP3 − 0.3 V.
However, fast settling is guaranteed only over a tuning voltage
range from 1.8 V up to VP3 − 0.8 V. This is to allow sufficient
room for overshoot in the PLL frequency settling transient.
Noise from the differential amplifier is suppressed inside the
PLL bandwidth. For loop bandwidths >20 kHz, the 1/f noise has
a negligible effect on the PLL output phase noise. Outside the
loop bandwidth, the differential amplifier’s noise FM modulates
the VCO. The passive filter network following the differential
amplifier, shown in Figure 36, suppresses this noise contribution to
below the VCO noise from offsets of 400 kHz and above. This
network has a negligible effect on lock time because it is
bypassed when SW3 is closed while the loop is locking.
AIN–
500Ω
500Ω
AOUT
VP3
AIN+
20kΩ
CMR
500Ω
30kΩ
Figure 26. Differential Amplifier Block Diagram
C EXT =
0.1µF
05328-020
500Ω
DIGITAL LOCK DETECT
LOGIC HIGH
DGND
NOTE:
NOT ALL MUXOUT MODES SHOWN REFER TO MUX REGISTER
05328-021
The internal, low noise, differential-to-single-ended amplifier is
used to convert the differential charge pump output to a singleended control voltage for the tuning port of the VCO. Figure 26
shows a simplified schematic of the differential amplifier. The
output voltage is equal to the differential voltage, offset by the
voltage on the CMR pin, according to
Figure 27. MUXOUT Circuit
Lock Detect
MUXOUT can be programmed to provide a digital lock detect
signal. Digital lock detect is active high. Its output goes high if
there are 40 successive PFD cycles with an input error of less
than 3 ns. For reliable lock detect operation with RF frequencies
<2 GHz, it is recommended that this threshold be increased to
10 ns by programming Register R6. The digital lock detect goes
low again when a new channel is programmed or when the
error at the PFD input exceeds 30 ns for one or more cycles.
Input Shift Register
The ADF4193 serial interface section includes a 24-bit input
shift register. Data is clocked in MSB first on each rising edge
of CLK. Data from the shift register is latched into one of eight
control registers, R0 to R7, on the rising edge of latch enable
(LE). The destination register is determined by the state of
the three control bits (Control Bit C3, Control Bit C2, and
Control Bit C1) in the shift register. The three LSBs are Bit DB2,
Bit DB1, and Bit DB0, as shown in the timing diagram of Figure 2.
The truth table for these bits is shown in Table 5. Figure 28
shows a summary of how the registers are programmed.
Table 5. C3, C2, and C1 Truth Table
C3
0
0
0
0
1
1
1
1
Rev. B | Page 13 of 28
Control Bits
C2
C1
0
0
0
1
1
0
1
1
0
0
0
1
1
0
1
1
Name
FRAC/INT
MOD/R
Phase
Function
Charge Pump
Power-Down
Mux
Test Mode
Register
R0
R1
R2
R3
R4
R5
R6
R7
ADF4193
REGISTER MAP
RESERVED
FRAC/INT REGISTER (R0)
8-BIT RF INT VALUE
CONTROL
BITS
12-BIT RF FRAC VALUE
DB23
DB22
DB21
DB20
DB19
DB18
DB17
DB16
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
0
N8
N7
N6
N5
N4
N3
N2
N1
F12
F11
F10
F9
F8
F7
F6
F5
F4
F3
F2
F1
DB2
DB1
DB0
C3 (0) C2 (0) C1 (0)
MOD/R REGISTER (R1)
DOUBLER
ENABLE
DBB
PRESCALER
DBB
RESERVED
DBB
REF/2
DBB
CP ADJ
DBB
DB23
DB22
DB21
DB20
DB19
DB18
DB17
DB16
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
F5
F4
0
F2
F1
R4
R3
R2
R1
M12
M11
M10
M9
M8
M7
M6
M5
M4
M3
M2
M1
4-BIT RF R
COUNTER
CONTROL
BITS
12-BIT MODULUS
DB2
DB1
DB0
C3 (0) C2 (0) C1 (1)
PHASE REGISTER (R2)
RESERVED
DBB
CONTROL
BITS
12-BIT PHASE
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
0
P12
P11
P10
P9
P8
P7
P6
P5
P4
P3
P2
P1
DB2
DB1
DB0
C3 (0) C2 (1) C1 (0)
CPO GND
RESERVED
PFD
POLARITY
FUNCTION REGISTER (R3)
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
0
0
0
0
0
0
0
0
0
1
F3
1
F1
RESERVED
CONTROL
BITS
DB2
DB1
DB0
C3 (0) C2 (1) C1 (1)
CHARGE PUMP REGISTER (R4)
RESERVED
TIMER
SELECT
9-BIT TIMEOUT COUNTER
DB23
DB22
DB21
DB20
DB19
DB18
DB17
DB16
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
0
0
0
0
0
0
0
0
0
1
C9
C8
C7
C6
C5
C4
C3
C2
C1
F2
F1
CONTROL
BITS
DB2
DB1
DB0
C3 (1) C2 (0) C1 (0)
CP
3-STATE
COUNTER
RESET
DB7
DB6
DB5
DB4
DB3
F5
F4
F3
F2
F1
PD
DIFF AMP
PD
CHARGE
PUMP
POWER-DOWN REGISTER (R5)
CONTROL
BITS
DB2
DB1
DB0
C3 (1) C2 (0) C1 (1)
MUX REGISTER (R6)
SIGMA-DELTA
AND
LOCK DETECT MODES
RESERVED
CONTROL
BITS
MUXOUT
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
M13
M12
M11
M10
0
0
0
0
0
M4
M3
M2
M1
DB2
DB1
DB0
C3 (1) C2 (1) C1 (0)
TEST MODE REGISTER (R7)
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
0
0
0
0
0
0
0
0
0
0
0
0
0
DBB = DOUBLE BUFFERED BIT(S)
Figure 28. Register Map
Rev. B | Page 14 of 28
DB2
DB1
DB0
C3 (1) C2 (1) C1 (1)
05328-022
CONTROL
BITS
RESERVED
ADF4193
RESERVED
FRAC/INT REGISTER (R0)
8-BIT RF INT VALUE
CONTROL
BITS
12-BIT RF FRAC VALUE
DB23
DB22
DB21
DB20
DB19
DB18
DB17
DB16
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
0
N8
N7
N6
N5
N4
N3
N2
N1
F12
F11
F10
F9
F8
F7
F6
F5
F4
F3
F2
F1
F12
0
0
0
0
.
.
.
1
1
1
1
F11
0
0
0
0
.
.
.
1
1
1
1
F10
0
0
0
0
.
.
.
1
1
1
1
F3
0
0
0
0
.
.
.
1
1
1
1
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
F2
0
0
1
1
.
.
.
0
0
1
1
F1
0
1
0
1
.
.
.
0
1
0
1
DB2
DB1
DB0
C3 (0) C2 (0) C1 (0)
FRACTIONAL VALUE (FRAC)
0
1
2
3
.
.
.
4092
4093
4094
4095
N8
N7
N6
N5
N4
N3
N2
N1
INTEGER VALUE (INT)
0
.
.
.
1
0
.
.
.
1
0
.
.
.
1
1
.
.
.
1
1
.
.
.
1
0
.
.
.
1
1
.
.
.
1
0
.
.
.
1
26
.
.
.
255
05328-023
0 = < FRAC < MOD
Figure 29. FRAC/INT Register (R0)
R0, the FRAC/INT register, is used to program the synthesizer
output frequency. On the next PFD cycle following a write to
R0, the N divider section is updated with the new INT and
FRAC values. At the same time, the PLL automatically enters
fast lock mode and the charge pump current is increased to its
maximum value and stays at this value until the ICP timeout
counter times out, and switches SW1, SW2, and SW3 closed
and remains closed until the SW1, SW2, and SW3 timeout
counters time out.
Once all registers are programmed during the initialization
sequence (see Table 8), all that is required thereafter to program
a new channel is a write to R0. However, as described in the
Programming section, it can also be desirable to program R1
and R2 register settings on a channel-by-channel basis. These
settings are double buffered by the write to R0. This means that
while the data is loaded through the serial interface on the
respective R1 and R2 write cycles, the synthesizer is not updated
with their data until the next write to Register R0.
Control Bits
The three LSBs, Control Bit C3, Control Bit C2, and Control Bit C1,
should be set to 0, 0, 0, respectively, to select R0, the FRAC/INT
register.
Reserved Bit
Bit DB23 is reserved and must be set to 0.
8-Bit INT Value
These eight bits set the INT value, which determines the integer
part of the feedback division factor. All integer values from 26
to 255 are allowed. See the Worked Example section.
12-Bit FRAC Value
The 12 FRAC bits set the numerator of the fraction that is input
to the Σ-Δ modulator. This, along with INT, specifies the new
frequency channel that the synthesizer locks to, as shown in the
Worked Example section. FRAC values from 0 to MOD − 1
cover channels over a frequency range equal to the PFD
reference frequency.
Rev. B | Page 15 of 28
ADF4193
CP
ADJ
REF/2
RESERVED
PRESCALER
DOUBLER
ENABLE
MOD/R REGISTER (R1)
DB23
DB22
DB21
DB20
DB19
DB18
DB17
DB16
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
F5
F4
0
F2
F1
R4
R3
R2
R1
M12
M11
M10
M9
M8
M7
M6
M5
M4
M3
M2
M1
4-BIT RF
R COUNTER
CONTROL
BITS
12-BIT MODULUS
DB2
DB1
DB0
C3 (0) C2 (0) C1 (1)
F4 REF/2
DISABLE
ENABLE
F2 PRESCALER
F1 DOUBLER ENABLE
0
1
0
1
4/5
8/9
DOUBLER DISABLED
DOUBLER ENABLED
F5 CP ADJ
0
1
NOMINAL
ADJUSTED
R4
0
0
0
0
.
.
.
1
1
1
1
R3
0
0
0
1
.
.
.
1
1
1
1
M12
0
0
0
.
.
.
1
1
1
1
M11
0
0
0
.
.
.
1
1
1
1
R2
0
1
1
0
.
.
.
0
0
1
1
R1
1
0
1
0
.
.
.
0
1
0
1
M10
0
0
0
.
.
.
1
1
1
1
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
M3
1
1
1
.
.
.
1
1
1
1
M2
0
1
1
.
.
.
0
0
1
1
M1
1
0
1
.
.
.
0
1
0
1
INTERPOLATOR MODULUS VALUE (MOD)
13
14
15
.
.
.
4092
4093
4094
4095
RF R COUNTER DIVIDE RATIO
1
2
3
4
.
.
.
12
13
14
15
05328-024
0
1
Figure 30. MOD/R Register (R1)
This register is used to set the PFD reference frequency and the
channel step size, which is determined by the PFD frequency
divided by the fractional modulus. Note that the MOD, R
counter, REF/2, CP ADJ, and doubler enable bits are double
buffered. They do not take effect until the next write to R0
(FRAC/INT register) is complete.
Control Bits
Reserved Bit
Reserved Bit DB21 must be set to 0.
Doubler Enable
Setting this bit to 1 inserts a frequency doubler between REFIN
and the 4-bit R counter. Setting this bit to 0 bypasses the
doubler.
4-Bit RF R Counter
With C3, C2, and C1 set to 0, 0, 1, respectively, the MOD/R
register (R1) is programmed.
It allows the REFIN frequency to be divided down to produce the
reference clock to the PFD. All integer values from 1 to 15 are
allowed. See the Worked Example section.
CP ADJ
When this bit is set to 1, the charge pump current is scaled up
25% from its nominal value on the next write to R0. When this
bit is set to 0, the charge pump current stays at its nominal value
on the next write to R0. See the Programming section for more
information on how this feature can be used.
REF/2
12-Bit Interpolator Modulus
For a given PFD reference frequency, the fractional denominator or modulus sets the channel step resolution at the RF
output. All integer values from 13 to 4095 are allowed. See the
Programming section for additional information and guidelines
for selecting the value of MOD.
Setting this bit to 1 inserts a divide-by-2, toggle flip-flop
between the R counter and PFD, which extends the maximum
REFIN input rate.
Rev. B | Page 16 of 28
ADF4193
CONTROL
BITS
12-BIT PHASE
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
0
P12
P11
P10
P9
P8
P7
P6
P5
P4
P3
P2
P1
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
P3
0
0
0
.
.
.
1
1
1
1
P12
0
0
0
.
.
.
1
1
1
1
P11
0
0
0
.
.
.
1
1
1
1
P10
0
0
0
.
.
.
1
1
1
1
P2
0
0
1
.
.
.
0
0
1
1
P1
0
1
0
.
.
.
0
1
0
1
10
DB2
DB1
DB0
C3 (0) C2 (1) C1 (0)
PHASE VALUE1
0
1
2
.
.
.
4092
4093
4094
4095
= < PHASE VALUE < MOD
05328-025
RESERVED
PHASE REGISTER (R2)
Figure 31. Phase Register (R2)
12-Bit Phase
The phase word sets the seed value of the Σ-Δ modulator. It can
be programmed to any integer value from 0 to MOD. As the
phase word is swept from 0 to MOD, the phase of the VCO
output sweeps over a 360° range in steps of 360°/MOD.
Note that the phase bits are double buffered. They do not take
effect until the LE of the next write to R0 (FRAC/INT register).
Therefore, if it is desired to change the phase of the VCO output
frequency, it is necessary to rewrite the INT and FRAC values to
R0, following the write to R2.
The output of a fractional-N PLL can settle to any one of the
MOD possible phase offsets with respect to the reference, where
MOD is the fractional modulus.
If it is desired to keep the output at the same phase offset with
respect to the reference, each time that particular output
frequency is programmed, then the interval between writes to
R0 must be an integer multiple of MOD reference cycles.
If it is desired to keep the outputs of two ADF4193-based
synthesizers phase coherent with each other, but not necessarily
with their common reference, then it is only required to ensure
that the write to R0 on both chips is performed during the same
reference cycle. The interval between R0 writes in this case does
not have to be an integer multiple of the MOD cycles.
Reserved Bit
The reserved bit, Bit DB15, should be set to 0.
Rev. B | Page 17 of 28
ADF4193
CPO GND
RESERVED
PFD
POLARITY
FUNCTION REGISTER (R3)
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
0
0
0
0
0
0
0
0
0
1
F3
1
F1
C3 (0) C2 (1) C1 (1)
F1
PFD POLARITY
0
1
NEGATIVE
POSITIVE
F3
CPO GND
0
1
CPO/CPO GND
NORMAL
CONTROL
BITS
DB2
DB1
DB0
05328-026
RESERVED
Figure 32. Function Register (R3)
R3, the function register (C3, C2, C1 set to 0, 1, 1, respectively),
only needs to be programmed during the initialization sequence
(see Table 8).
PFD Polarity
CPO GND
Reserved Bits
When the CPO GND bit is low, the charge pump outputs are
internally pulled to ground. This is invoked during the
initialization sequence to discharge the loop filter capacitors.
For normal operation, this bit should be high.
The Bit DB15 to Bit DB6 are reserved bits and should be
programmed to hex code 001, and Reserved Bit DB4 should be
set to 1.
This bit should be set to 1 for positive polarity and set to 0 for
negative polarity.
Rev. B | Page 18 of 28
ADF4193
CHARGE PUMP REGISTER (R4)
TIMER
SELECT
9-BIT TIMEOUT COUNTER
CONTROL
BITS
DB23
DB22
DB21
DB20
DB19
DB18
DB17
DB16
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
0
0
0
0
0
0
0
0
0
1
C9
C8
C7
C6
C5
C4
C3
C2
C1
F2
F1
C9
C8
C7
0
0
0
0
.
.
.
1
1
1
1
0
0
0
0
.
.
.
1
1
1
1
0
0
0
0
.
.
.
1
1
1
1
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
..........
F2
F1 TIMER SELECT
0
0
1
1
0
1
0
1
DB2
DB1
DB0
C3 (1) C2 (0) C1 (0)
SW1/SW2
SW3
ICP
NOT USED
C3
C2
C1
TIMEOUT COUNTER
xPFD CYCLES
0
0
0
0
.
.
.
1
1
1
1
0
0
1
1
.
.
.
0
0
1
1
0
1
0
1
.
.
.
0
1
0
1
0
1
2
3
.
.
.
508
509
510
511
0
4
8
12
.
.
.
2032
2036
2040
2044
1DELAY
DELAY µs1
0
0.15
0.30
0.46
.
.
.
78.15
78.30
78.46
78.61
WITH 26MHz PFD
05328-027
RESERVED
Figure 33. Charge Pump Register (R4)
Table 6. Recommended Values for a GSM Tx LO
Reserved Bits
Bit DB23 to Bit DB14 are reserved and should be set to hex
code 001 for normal operation.
9-Bit Timeout Counter
These bits are used to program the fast lock timeout counters.
The counters are clocked at one-quarter the PFD reference
frequency, therefore, their time delay scales with the PFD
frequency according to
Delay(s) = (Timeout Counter Value × 4)/(PFD Frequency)
For example, if 35 were loaded with timer select (00) with a
13 MHz PFD, then SW1/SW2 would be switched after
(35 × 4)/13 MHz = 10.8 μs
Timer Select
10
01
00
Timeout Counter
ICP
SW1/2
SW3
Value
28
35
35
Time (μs) with
PFD = 13 MHz
8.6
10.8
10.8
On each write to R0, the timeout counters start. Switch SW3
closes until the SW3 counter times out. Similarly, switches
SW1/SW2 close until the SW1/SW2 counter times out. When
the ICP counter times out, the charge pump current is ramped
down from 64× to 1× in six binary steps. It is recommended
that the SW1, SW2, and SW3 timeout counter values are set
equal to the ICP timeout counter value plus 7, as in the example
of Table 6.
Timer Select
These two address bits select the timeout counter to be
programmed. Note that to set up the ADF4193 correctly
requires setup of these three timeout counters; therefore, three
writes to this register are required in the initialization sequence.
Table 6 shows example values for a GSM Tx synthesizer with a
60 kHz final loop BW. See the Applications section for more
information.
Rev. B | Page 19 of 28
ADF4193
PD
CHARGE
PUMP
CP
3-STATE
COUNTER
RESET
POWER-DOWN REGISTER (R5)
DB7
DB6
DB5
DB4
DB3
F5
F4
F3
F2
F1
PD
DIFF AMP
NORMAL OPERATION
COUNTER RESET
F2
CHARGE PUMP
3-STATE
0
1
NORMAL OPERATION
3-STATE ENABLED
DISABLED
ENABLED
05328-028
DISABLED
ENABLED
DB0
COUNTER RESET
0
1
0 0
1 1
DB1
0
1
CHARGE PUMP
POWER-DOWN
DIFF AMP
POWER-DOWN
DB2
C3 (1) C2 (0) C1 (1)
F1
F3
F5 F4
CONTROL
BITS
Figure 34. Power-Down Register (R5)
R5, the power-down register (C3, C2, C1 set to 1, 0, 1,
respectively) can be used to software power down the PLL and
differential amplifier sections. After power is initially applied,
there must be writes to R5 to clear the power-down bits and to
R2, R1, and R0 before the ADF4193 comes out of power-down.
Power-Down Differential Amplifier
When Bit DB6 and Bit DB7 are set high, the differential
amplifier is put into power-down. When Bit DB6 and Bit DB7
are set low, normal operation is resumed.
Power-Down Charge Pump
For normal operation, Bit DB5 should be set to 0, followed by a
write to R0.
CP Three-State
When this bit is set high, the charge pump outputs are put into
three-state. With the bit set low, the charge pump outputs are
enabled.
Counter Reset
When this bit is set to 1, the counters are held in reset. For
normal operation, this bit should be 0, followed by a write to R0.
Setting Bit DB5 high activates a charge pump power-down and
the following events occur:
• All active dc current paths are removed, except for the
differential amplifier.
• The R and N divider counters are forced to their load state
conditions.
• The charge pump is powered down with its outputs in threestate mode.
• The digital lock detect circuitry is reset.
• The RFIN input is debiased.
• The reference input buffer circuitry is disabled.
• The serial interface remains active and capable of loading and
latching data.
Rev. B | Page 20 of 28
ADF4193
MUX REGISTER (R6)
RESERVED
CONTROL
BITS
MUXOUT
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
M13
M12
M11
M10
0
0
0
0
0
M4
M3
M2
M1
DB2
DB1
DB0
C3 (1) C2 (1) C1 (0)
M13
M12
M11
M10 SIGMA-DELTA MODES
M4
M3
M2
M1
MUXOUT
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
3-STATE
DIGITAL LOCK DETECT
N DIVIDER OUTPUT
LOGIC HIGH
R COUNTER
RESERVED
SERIAL DATA OUT
LOGIC LOW
R DIVIDER/2 OUTPUT
N DIVIDER/2 OUTPUT
RESERVED
RESERVED
ICP TIMEOUT SIGNAL
SW1/2 TIMEOUT SIGNAL
SW3 TIMEOUT SIGNAL
RESERVED
0
1
0
1
1
0
0
1
ALL OTHER STATES
INIT STATE, DITHER OFF,
3ns LOCK DETECT THRESHOLD
DITHER ON
10ns LOCK DETECT THRESHOLD
RESERVED
05328-029
SIGMA-DELTA
AND
LOCK DETECT MODES
Figure 35. MUX Register (R6)
MUXOUT Modes
With C3, C2, and C1 set to 1, 1, 0, respectively, the MUX
register is programmed.
Σ-Δ and Lock Detect Modes
Bit DB15 to Bit DB12 are used to reconfigure certain PLL
operating modes. In the initialization sequence after power is
applied to the chip, the four bits must first be programmed to
all zeros. This initializes the PLL to a known state with dither
off in the Σ-Δ modulator and a 3 ns PFD error threshold in the
lock detect circuit.
To turn on dither in the Σ-Δ modulator, an additional write
should be made to Register R6 to program bits [DB15:DB12] =
[0011]. However, for lowest noise operation, it is best to leave
dither off.
These bits control the on-chip multiplexer. See Figure 35 for the
truth table. This pin is useful for diagnosis because it allows the
user to look at various internal points of the chip, such as the
R divider and INT divider outputs.
In addition, it is possible to monitor the programmed timeout
counter intervals on MUXOUT. For example, if the ICP timeout
counter was programmed to 65 (with a 26 MHz PFD), then
following the next write to R0, a pulse width of 10 μs would be
observed on the MUXOUT pin.
Digital lock detect is available via the MUXOUT pin.
To change the lock detect threshold from 3 ns to 10 ns, a
separate write to R6 should be performed to program bits
[DB15:DB12] = [1001]. This should be done for reliable lock
detect operation when the RF frequency is <2 GHz.
A write to R6 that programs bits [DB15:DB12] = [0000] returns
operation to the default state with both dither off and a 3 ns
lock detect threshold.
Reserved Bits
The reserved bits must all be set to 0 for normal operation.
Rev. B | Page 21 of 28
ADF4193
PROGRAMMING
The ADF4193 can synthesize output frequencies with a channel
step or resolution that is a fraction of the input reference frequency.
For a given input reference frequency and a desired output
frequency step, the first choice to make is the PFD reference
frequency and the MOD. Once these are chosen, the desired
output frequency channels are set by programming the INT
and FRAC values.
WORKED EXAMPLE
In this example of a GSM900 RX system, it is required to
generate RF output frequencies with channel steps of 200 kHz.
A 104 MHz reference frequency input (REFIN) is available. The
R divider setting that set the PFD reference is shown in
Equation 1.
FPFD = REFIN × [(1 + D)/(R × (1 + T))]
(1)
where:
REFIN is the input reference frequency.
D is the doubler enable bit (0 or 1).
SPUR MECHANISMS
The Fractional Spurs, Integer Boundary Spurs, and Reference
Spurs sections describe the three different spur mechanisms
that arise with a fractional-N synthesizer and how the ADF4193
can be programmed to minimize them.
Fractional Spurs
The fractional interpolator in the ADF4193 is a third-order, Σ-Δ
modulator (SDM) with a modulus (MOD) that is programmable to
any integer value from 13 to 4095. If dither is enabled, then the
minimum allowed value of MOD is 50. The SDM is clocked at
the PFD reference rate (fPFD) that allows PLL output frequencies
to be synthesized at a channel step resolution of fPFD/MOD.
With dither turned off, the quantization noise from the Σ-Δ
modulator appears as fractional spurs. The interval between
spurs is fPFD/L, where L is the repeat length of the code sequence
in the digital Σ-Δ modulator. For the third-order modulator
used in the ADF4193, the repeat length depends on the value of
MOD, as shown in Table 7.
R is the 4-bit R counter code (0…15).
Table 7. Fractional Spurs with Dither Off
T is the REF/2 bit (0 or 1).
The maximum PFD reference frequency of 26 MHz is chosen
and the following settings are programmed to give an R divider
value of 4:
Doubler enable = 0
R=2
Condition (Dither Off)
If MOD is divisible by 2,
but not 3
If MOD is divisible by 3,
but not 2
If MOD is divisible by 6
Otherwise
Repeat Length
2 × MOD
Spur Interval
Channel step/2
3 × MOD
Channel step/3
6 × MOD
MOD
Channel step/6
Channel step
REF/2 = 1
RFOUT is the desired RF output frequency.
With dither enabled, the repeat length is extended to 221 cycles,
regardless of the value of MOD, which makes the quantization
error spectrum look like broadband noise. This can degrade the
in-band phase noise at the PLL output by as much as 10 dB.
Therefore, for the lowest noise, dither off is a better choice,
particularly when the final loop BW is low enough to attenuate
even the lowest frequency fractional spur. The wide loop
bandwidth range available with the ADF4193 makes this
possible in most applications.
INT is the integer part of the division.
Integer Boundary Spurs
FRAC is the numerator part of the fractional division.
Another mechanism for fractional spur creation involves
interactions between the RF VCO frequency and the reference
frequency. When these frequencies are not integer related, spur
sidebands appear on the VCO output spectrum at an offset
frequency that corresponds to the beat note or difference
frequency between an integer multiple of the reference and the
VCO frequency.
Next, the modulus is chosen to allow fractional steps of 200 kHz.
MOD = 26 MHz/200 kHz = 130
(2)
Once the channel step is defined, the following equation shows
how output frequency channels are programmed:
RFOUT = [INT + (FRAC/MOD] × [FPFD]
(3)
where:
MOD is the modulus or denominator part of the fractional
division.
For example, the frequency channel at 962.4 MHz is synthesized
by programming the following values:
INT = 37
FRAC = 2
Rev. B | Page 22 of 28
ADF4193
These spurs are attenuated by the loop filter and are more
noticeable on channels close to integer multiples of the reference where the difference frequency can be inside the loop
bandwidth, thus the name integer boundary spurs.
The 8:1 loop bandwidth switching ratio of the ADF4193 makes
it possible to attenuate all spurs to sufficiently low levels for
most applications. The final loop BW can be chosen to ensure
that all spurs are far enough out of band while meeting the lock
time requirements with the 8× bandwidth boost.
The ADF4193’s programmable modulus and R divider can also
be used to avoid integer boundary channels. This option is
described in the Avoiding Integer Boundary Channels section.
Reference Spurs
Reference spurs are generally not a problem in fractional-N
synthesizers as the reference offset is far outside the loop
bandwidth. However, any reference feedthrough mechanism
that bypasses the loop can cause a problem. One such
mechanism is feedthrough of low levels of on-chip reference
switching noise out through the RFIN pin back to the VCO,
resulting in reference spur levels as high as –90 dBc. These
spurs can be suppressed below –110 dBc by inserting sufficient
reverse isolation, for example, through an RF buffer between
the VCO and RFIN pin. In addition, care should be taken in the
PCB layout to ensure that the VCO is well separated from the
input reference to avoid a possible feedthrough path on the board.
POWER-UP INITIALIZATION
Table 8. Power-Up Initialization Sequence
Step
1
2
Register
Bits
R5 [7:0]
R3 [15:0]
Hex
Codes
FD
005B
Wait
10 ms
3
4
R7 [15:0]
R6 [15:0]
0007
000E
5
R6 [15:0]
900E
6
7
8
9
10
R4 [23:0]
R4 [23:0]
R4 [23:0]
R2 [15:0]
R1 [23:0]
004464
00446C
004394
00D2
520209
11
R0 [23:0]
480140
12
R3 [15:0]
007B
13
14
R5 [7:0]
R0 [23:0]
05
480140
Description
Set all power-down bits.
PD polarity = 1, ground CPOUT+/
CPOUT–.
Allow time for loop filter
capacitors to discharge.
Clear test modes.
Initialize PLL modes, digital lock
detect on MUXOUT.
10 ns lock detect threshold,
digital lock detect on MUXOUT.
SW1/SW2 timer = 10.8 μs.
SW3 timer = 10.8 μs.
ICP timer = 8.6 μs.
Phase = 26.
8/9 prescaler, doubler disabled,
R = 4, toggle FF on, MOD = 65.
INT = 144, FRAC = 40 for
1880 MHz output frequency.
PD polarity = 1, release CPOUT+/
CPOUT–.
Clear all power-down bits.
INT = 144, FRAC = 40 for
1880 MHz output frequency.
The ADF4193 powers up after Step 13. It locks to the
programmed channel frequency after Step 14.
CHANGING THE FREQUENCY OF THE PLL AND THE
PHASE LOOK-UP TABLE
After applying power to the ADF4193, a 14-step sequence is
recommended, as described in Table 8.
The divider and timer setting used in the example in Table 8 is
for a DCS1800 Tx synthesizer with a 104 MHz REFIN frequency.
Once the ADF4193 is initialized, a write to Register R0 is all
that is required to program a new output frequency. The N
divider is updated with the values of INT and FRAC on the next
PFD cycle following the LE edge that latches in the R0 word.
However, the settling time and spurious performance of the
synthesizer can be further optimized by modifying R1 and R2
register settings on a channel-by-channel basis. These settings
are double buffered by the write to R0. This means that while
the data is loaded in through the serial interface on the
respective R1 and R2 write cycles, the synthesizer is not
updated with their data until the next write to Register R0.
The R2 register can be used to digitally adjust the phase of the
VCO output relative to the reference edge. The phase can be
adjusted over the full 360° range at RF with a resolution of
360°/MOD. In most frequency synthesizer applications, the
actual phase offset of the VCO output with respect to the
reference is unknown and does not matter. In such applications,
the phase adjustment capability of the R2 register can instead be
used to optimize the settling time performance, as described in
the Phase Look-Up Table section.
Rev. B | Page 23 of 28
ADF4193
Phase Look-Up Table
Avoiding Integer Boundary Channels
The ADF4193’s fast lock sequence is initiated following the
write to Register R0. The fast lock timers are programmed so
that after the PLL has settled in wide BW mode, the charge
pump current is reduced and loop filter resistor switches are
opened to reduce the loop BW. The reference cycle on which
these events occur is determined by the values preprogrammed
into the timeout counters.
A further option when programming a new frequency involves
a write to Register R1 to avoid integer boundary spurs. If it is
found that the integer boundary spur level is too high, an
option is to move the integer boundary away from the desired
channel by reprogramming the R divider to select a different
PFD frequency. For example, if REFIN = 104 MHz and R = 4 for
a 26 MHz PFD reference and MOD = 130 for 200 kHz steps, the
frequency channel at 910.2 MHz has a 200 kHz integer
boundary spur because it is 200 kHz offset from 35 × 26 MHz.
An alternative way to synthesize this channel is to set R = 5 for a
20.8 MHz PFD reference and MOD = 104 for 200 kHz steps.
The 910.2 MHz channel is now 5 MHz offset from the nearest
integer multiple of 20.8 MHz and the 5 MHz beat note spurs are
well attenuated by the loop. Setting double buffered Bit R1 [23] = 1
(CP ADJ bit) increases the charge pump current by 25%, which
compensates for the 25% increase in N with the change to the
20.8 MHz PFD frequency. This maintains constant loop
dynamics and settling time performance for jumps between the
two PFD frequencies. The CP ADJ bit should be cleared again
when jumping back to 26 MHz-based channels.
Figure 10 and Figure 13 show that the lock time to final phase is
dominated by the phase swing that occurs when the BW is
reduced. Once the PLL has settled to final frequency and phase,
in wide BW mode, this phase swing is the same, regardless of
the size of the synthesizer’s frequency jump. The amplitude of
the phase swing is related to the current flowing through the
loop filter zero resistors on the PFD reference cycle that the
SW1/SW2 switches are opened. In an integer-N PLL, this
current is zero once the PLL has settled. In a fractional-N PLL,
the current is zero on average but varies from one reference
cycle to the next, depending on the quantization error sequence
output from the digital Σ-Δ modulator. Because the Σ-Δ
modulator is all digital logic, clocked at the PFD reference rate,
for a given value of MOD, the actual quantization error on any
given reference cycle is determined by the value of FRAC and
the PHASE word that the modulator is seeded with, following
the write to R0. By choosing an appropriate value of PHASE,
corresponding to the value of FRAC, that is programmed on the
next write to R0, the size of the error current on the PFD
reference cycle the SW1/SW2 switches opened, and thus the
phase swing that occurs when the BW is reduced can be
minimized.
With dither off, the fractional spur pattern due to the SDM’s
quantization noise also depends on the phase word the modulator
is seeded with. Tables of optimized FRAC and phase values for
popular SW1/SW2 and ICP timer settings can be down-loaded
from the ADF4193 product page. If making use of a phase table,
first write phase to double buffered Register R2, then write the
INT and FRAC to R0.
The Register R1 settings necessary for integer boundary spur
avoidance are all double buffered and do not become active on
the chip until the next write to Register R0. Register R0 should
always be the last register written to when programming a new
frequency.
Serial Interface Activity
The serial interface activity when programming the R2 or R1
registers causes no noticeable disturbance to the synthesizers
settled phase or degradation in its frequency spectrum.
Therefore, in a GSM application, it can be performed during
the active part of the data burst. Because it takes just 10.2 μs to
program the three registers, R2, R1, and R0, with the 6.5 MHz
serial interface clock rate typically used, this programming can
also be performed during the previous guard period with the
LE edge to latch in the R0 data delayed until it’s time to switch
frequency.
Rev. B | Page 24 of 28
ADF4193
APPLICATIONS
LOCAL OSCILLATOR FOR A GSM BASE STATION
Timer Values for Tx
Figure 36 shows the ADF4193 being used with a VCO to
produce the LO for a GSM1800 base station. For GSM, the
REFIN signal can be any integer multiple of 13 MHz, but the
main requirement is that the slew rate is at least 300 V/μs.
The 5 dBm, 104 MHz input sine wave shown satisfies this
requirement.
To comply with the GSM spectrum due to switching requirements, the Tx synthesizer should not switch frequency until the
PA output power has ramped down by at least 50 dB. If it takes
10 μs to ramp down to this level, then only the last 20 μs of the
30 μs guard period is available for the Tx synthesizer to lock to
final frequency and phase.
Recommended parameters for the various GSM/PCS/DCS
synthesizers are given in Table 9.
In fast lock mode, the Tx loop BW is widened by a factor-of-8
to 480 kHz, and therefore, the PLL achieves frequency lock for a
jump across the entire band in <6 μs. After this, the PA power
can start to ramp up again, and the loop BW can be restored to
the final value. With the ICP timer = 28, the charge pump
current reduction begins at ~8.6 μs. When SW1, SW2, and SW3
timers = 35, the current reaches its final value before the loop
filter switches open at ~10.8 μs.
Table 9. Recommended Setup Parameters
Parameter
Loop BW
PFD (MHz)
MOD
Dither
Prescaler
ICP Timer
SW1, SW2,
SW3 Timers
VCO KV
Tx
60 kHz
13
65
Off
4/5
28
35
GSM900
Rx
40 kHz
26
130
Off
4/5
78
85
18 MHz/V
18 MHz/V
DCS1800/PCS1900
Tx
Rx
60 kHz
40 kHz
13
13
65
65
Off
Off
8/9
8/9
28
38
35
45
38 MHz/V
38 MHz/V
With these timer values, the phase disturbance created when
the bandwidth is reduced settles back to its final value by 20 μs,
in time for the start of the active part of the GSM burst. If faster
phase settling is desired with the 60 kHz BW setting, then the timer
values can be reduced further but should not be brought less than
the 6 μs it takes to achieve frequency lock in wide BW mode.
Loop BW and PFD Frequency
Timer Values for Rx
A 60 kHz loop BW is narrow enough to attenuate the PLL phase
noise and spurs to the required level for a Tx low. A 40 kHz BW
is necessary to meet the GSM900 Rx synthesizer’s particularly
tough phase noise and spur requirements at ±800 kHz offsets.
To get the lowest spur levels at ±800 kHz offsets for Rx, the Σ-Δ
modulator should be run at the highest oversampling rate
possible. Therefore, for GSM900 Rx, a 26 MHz PFD frequency
is chosen and MOD = 130 is required for 200 kHz steps.
Because this value of MOD is divisible by two, certain FRAC
channels have a 100 kHz fractional spur. This is attenuated by
the 40 kHz loop filter and therefore is not a concern. However,
the 60 kHz loop filter recommended for Tx has a closed-loop
response that peaks close to 100 kHz. Therefore, a 13 MHz PFD
with MOD = 65, which avoids the 100 kHz spur, is the best
choice for a Tx synthesizer.
The 40 kHz Rx loop BW is increased by a factor-of-8 to
approximately 320 kHz during fast lock. With the Rx timer
values shown, the BW is reduced after ~12 μs, which allows
sufficient time for the phase disturbance to settle back before
the start of the active part of the Rx time slot at 30 μs. As in the
Tx case, faster Rx settling can be achieved by reducing these
timer values, their lower limit being determined by the time it
takes to achieve frequency lock in wide BW mode. In addition,
the PCS and DCS Rx synthesizers have relaxed 800 kHz blocker
specifications and thus can tolerate a wider loop BW, which
allows correspondingly faster settling.
Dither
Dither off should be selected for the lowest rms phase error.
Prescaler
The 8/9 prescaler should be selected for the PCS and DCS
bands. The 4/5 prescaler allows an N divider range low enough
to cover the GSM900 Tx and Rx bands with either a 13 MHz or
26 MHz PFD frequency.
VCO KV
In general, the VCO gain, KV, should be set as low as possible to
minimize the reference and integer boundary spur levels that arise
due to feedthrough mechanisms. When deciding on the optimum
VCO KV, a good choice is to allow 2 V to tune across the desired
band, centered on the available tuning range. With VP3 regulated
to 5.5 V ± 100 mV, the tuning range available is 2.8 V.
Loop Filter Components
It is important for good settling performance that capacitors
with low dielectric absorption are used in the loop filter.
Ceramic NPO COG capacitors are a good choice for this
application. A 2% tolerance is recommended for loop filter
capacitors and 1% for resistors. A 10% tolerance is adequate
for the inductor, L1.
Rev. B | Page 25 of 28
ADF4193
ADI SimPLL SUPPORT
Also available is a technical note (ADF4193-TN-001) that
outlines a loop filter design procedure that takes full advantage
of the new degree of freedom in the filter design that the
differential amplifier and loop filter switches provide.
The ADF4193 loop filter design is supported on ADI SimPLL
v2.7 or later. Example files for popular applications are available
for download from the applications section of the ADF4193
product page.
10pF 18Ω
18Ω 100pF
RF OUT
5V
10µF
18Ω
+
3V
5.5V
+
100nF
100nF
100nF
15
8, 10, 13
SVD VDD
100pF
DVDD
RFIN+
SW1
RFIN–
100pF
REFERENCE
104MHz, +5dBm
1nF
51Ω
11
19
RSET
2.40kΩ
29
REFIN
SW2
27
18
DATA
17
CLK
CPOUT–
26
23
RSET
MUXOUT
16
14
AGND
4, 22
C2A
1.20nF
R1A2
820Ω
28
LE
SDGND
100nF
7
VP2
32
AVDD
VP3
10µF
R1A2
6.20kΩ
ADF4193
SWGND
1nF
24
30
51Ω
5
100pF
VP 1
CPOUT+
6
100nF
20
C1A
120pF
31
INTEGRATED
DIFFERENTIAL
AMPLIFIER
AIN+
R3
62Ω
2
R2
1.80kΩ
SW3
AIN–
R1B2
6.20kΩ
3
AOUT
25
+
100nF
L1
2.2mH
C3
470pF
Ct
30pF
CMR
R1B1
820Ω
C2B
1.20nF
1
100nF
SIRENZA VCO190-1843T
38MHz/V
C1B
120pF
DGND
9, 12, 21
LOCK DETECT OUT
Figure 36. Local Oscillator for DCS1800 Tx Using the ADF4193
Rev. B | Page 26 of 28
05328-037
10µF
ADF4193
INTERFACING
ADSP-21xx Interface
The ADF4193 has a simple SPI®-compatible serial interface for
writing to the device. CLK, DATA, and LE control the data
transfer. When LE goes high, the 24 bits that have been clocked
into the input register on each rising edge of CLK are latched
into the appropriate register. See Figure 2 for the timing
diagram and Table 5 for the register address table.
Figure 38 shows the interface between the ADF4193 and the
ADSP-21xx digital signal processor. The ADF4193 needs a
24-bit serial word for some writes. The easiest way to accomplish this using the ADSP-21xx family is to use the autobuffered
transmit mode of operation with alternate framing. This
provides a means for transmitting an entire block of serial data
before an interrupt is generated. Set up the word length for
eight bits and use three memory locations for each 24-bit word.
To program each 24-bit word, store the three 8-bit bytes, enable
the autobuffered mode, and then write to the transmit register
of the DSP. This last operation initiates the autobuffer transfer.
ADuC812 Interface
Figure 37 shows the interface between the ADF4193 and the
ADuC812 MicroConverter®. Because the ADuC812 is based on
an 8051 core, this interface can be used with any 8051-based
microcontroller. The MicroConverter is set up for SPI master
mode with CPHA = 0. To initiate the operation, the I/O port
driving LE is brought low. Some registers of the ADF4193
require a 24-bit programming word. This is accomplished by
writing three 8-bit bytes from the MicroConverter to the device.
When the third byte is written, the LE input should be brought
high to complete the transfer.
An I/O port line on the ADuC812 can also be used to detect
lock (MUXOUT configured as lock detect and polled by the
port input).
ADuC812
SCLOCK
MOSI
ADF4193
DATA
LE
05328-033
I/O PORTS
Figure 37. ADuC812 to ADF4193 Interface
ADF4193
SCLK
DT
TFS
I/O FLAGS
CLK
DATA
LE
MUXOUT
(LOCK DETECT)
Figure 38. ADSP-21xx to ADF4193 Interface
PCB DESIGN GUIDELINES FOR CHIP SCALE
PACKAGE
The lands on the chip scale package (CP-32-3) are rectangular.
The printed circuit board (PCB) pad for these should be
0.1 mm longer than the package land length and 0.05 mm wider
than the package land width. The land should be centered on
the pad. This ensures that the solder joint size is maximized.
The bottom of the chip scale package has a central thermal pad.
CLK
MUXOUT
(LOCK DETECT)
ADSP-21xx
05328-034
The maximum allowable serial clock rate is 33 MHz.
The thermal pad on the PCB should be at least as large as the
exposed pad. On the PCB, there should be a clearance of at least
0.25 mm between the thermal pad and the inner edges of the
pad pattern. This ensures that shorting is avoided.
Thermal vias can be used on the PCB thermal pad to improve
the thermal performance of the package. If vias are used, they
should be incorporated in the thermal pad at 1.2 mm pitch grid.
The via diameter should be between 0.3 mm and 0.33 mm, and
the via barrel should be plated with one ounce copper to plug
the via.
The user should connect the PCB thermal pad to AGND.
Rev. B | Page 27 of 28
ADF4193
OUTLINE DIMENSIONS
0.60 MAX
5.00
BSC SQ
0.60 MAX
PIN 1
INDICATOR
TOP
VIEW
0.50
BSC
4.75
BSC SQ
0.50
0.40
0.30
32
1
EXPOSED
PAD
(BOTTOM VIEW)
17
16
3.45
3.30 SQ
3.15
9
8
0.25 MIN
3.50 REF
0.80 MAX
0.65 TYP
12° MAX
1.00
0.85
0.80
PIN 1
INDICATOR
25
24
0.05 MAX
0.02 NOM
SEATING
PLANE
0.30
0.23
0.18
0.20 REF
COPLANARITY
0.08
COMPLIANT TO JEDEC STANDARDS MO-220-VHHD-2
Figure 39. 32-Lead Lead Frame Chip Scale Package [LFCSP_VQ]
5 mm × 5 mm Body, Very Thin Quad
(CP-32-3)
Dimensions shown in millimeters
ORDERING GUIDE
Model
ADF4193BCPZ 1
ADF4193BCPZ-RL1
ADF4193BCPZ-RL71
EVAL-ADF4193EB1
EVAL-ADF4193EB2
1
Temperature Range
–40°C to +85°C
–40°C to +85°C
–40°C to +85°C
Package Description
32-Lead Lead Frame Chip Scale Package [LFCSP_VQ]
32-Lead Lead Frame Chip Scale Package [LFCSP_VQ]
32-Lead Lead Frame Chip Scale Package [LFCSP_VQ]
Evaluation Board (GSM 1800)
Evaluation Board (No VCO or Loop Filter)
Package Option
CP-32-3
CP-32-3
CP-32-3
Z = Pb-free part.
©2006 Analog Devices, Inc. All rights reserved. Trademarks and
registered trademarks are the property of their respective owners.
D05328-0-6/06(B)
T
T
Rev. B | Page 28 of 28