AD EVAL-ADF7021

High Performance
Narrow-Band Transceiver IC
ADF7021-N
FEATURES
On-chip VCO and fractional-N PLL
On-chip, 7-bit ADC and temperature sensor
Fully automatic frequency control loop (AFC)
Digital received signal strength indication (RSSI)
Integrated Tx/Rx switch
0.1 μA leakage current in power-down mode
Low power, narrow-band transceiver
Frequency bands using dual VCO
80 MHz to 650 MHz
842 MHz to 916 MHz
Programmable IF filter bandwidths of
9 kHz, 13.5 kHz, and 18.5 kHz
Modulation schemes: 2FSK, 3FSK, 4FSK, MSK
Spectral shaping: Gaussian and raised cosine filtering
Data rates supported: 0.05 kbps to 24 kbps
2.3 V to 3.6 V power supply
Programmable output power
−16 dBm to +13 dBm in 63 steps
Automatic power amplifier (PA) ramp control
Receiver sensitivity
−130 dBm at 100 bps, 2FSK
−122 dBm at 1 kbps, 2FSK
Patent pending, on-chip image rejection calibration
APPLICATIONS
Narrow-band, short range device (SRD) standards
ARIB STD-T67, ETSI EN 300 220, Korean SRD standard,
FCC Part 15, FCC Part 90, FCC Part 95
Low cost, wireless data transfer
Remote control/security systems
Wireless metering
Wireless medical telemetry service (WMTS)
Home automation
Process and building control
Pagers
FUNCTIONAL BLOCK DIAGRAM
CE
RSET
TEMP
SENSOR
RLNA
MUX
7-BIT ADC
2FSK
3FSK
4FSK
LNA
RFIN
RSSI/
LOG AMP
IF FILTER
RFINB
CREG(1:4)
MUXOUT
LDO(1:4)
TEST MUX
CLOCK
AND DATA
RECOVERY
DEMODULATOR
TxRxCLK
Tx/Rx
CONTROL
TxRxDATA
SWD
GAIN
AGC
CONTROL
SLE
SERIAL
PORT
AFC
CONTROL
PA RAMP
DIV P
÷1/÷2
VCO1
VCO2
VCOIN
SCLK
GAUSSIAN/
RAISED COSINE
FILTER
3FSK
ENCODING
CP
PFD
DIV R
L2
2FSK
3FSK
4FSK
MOD CONTROL
Σ-Δ
MODULATOR
÷2
MUX
L1
N/N + 1
SREAD
CPOUT
OSC
OSC1
OSC2
CLK
DIV
CLKOUT
07246-001
RFOUT
SDATA
Figure 1.
Rev. 0
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
©2008 Analog Devices, Inc. All rights reserved.
ADF7021-N
TABLE OF CONTENTS
Features .............................................................................................. 1
Demodulation, Detection, and CDR ....................................... 32
Applications....................................................................................... 1
Receiver Setup............................................................................. 34
Functional Block Diagram .............................................................. 1
Demodulator Considerations ................................................... 36
Revision History ............................................................................... 2
AFC Operation ........................................................................... 36
General Description ......................................................................... 3
Automatic Sync Word Detection (SWD)................................ 37
Specifications..................................................................................... 4
Applications Information .............................................................. 38
RF and PLL Specifications........................................................... 4
IF Filter Bandwidth Calibration ............................................... 38
Transmission Specifications........................................................ 5
LNA/PA Matching...................................................................... 39
Receiver Specifications ................................................................ 6
Image Rejection Calibration ..................................................... 40
Digital Specifications ................................................................... 9
Packet Structure and Coding.................................................... 42
General Specifications ............................................................... 10
Programming After Initial Power-Up ..................................... 42
Timing Characteristics .............................................................. 11
Applications Circuit ................................................................... 45
Timing Diagrams........................................................................ 12
Serial Interface ................................................................................ 46
Absolute Maximum Ratings.......................................................... 15
Readback Format........................................................................ 46
ESD Caution................................................................................ 15
Interfacing to a Microcontroller/DSP ..................................... 48
Pin Configuration and Function Descriptions........................... 16
Register 0—N Register............................................................... 49
Typical Performance Characteristics ........................................... 18
Register 1—VCO/Oscillator Register ...................................... 50
Frequency Synthesizer ................................................................... 22
Register 2—Transmit Modulation Register ............................ 51
Reference Input........................................................................... 22
Register 3—Transmit/Receive Clock Register........................ 52
MUXOUT.................................................................................... 23
Register 4—Demodulator Setup Register ............................... 53
Voltage Controlled Oscillator (VCO) ...................................... 24
Register 5—IF Filter Setup Register......................................... 54
Choosing Channels for Best System Performance................. 25
Register 6—IF Fine Cal Setup Register ................................... 55
Transmitter ...................................................................................... 26
Register 7—Readback Setup Register...................................... 56
RF Output Stage.......................................................................... 26
Register 8—Power-Down Test Register .................................. 57
Modulation Schemes.................................................................. 26
Register 9—AGC Register......................................................... 58
Spectral Shaping ......................................................................... 28
Register 10—AFC Register ....................................................... 59
Modulation and Filtering Options ........................................... 29
Register 11—Sync Word Detect Register................................ 60
Transmit Latency ........................................................................ 29
Register 12—SWD/Threshold Setup Register........................ 60
Test Pattern Generator............................................................... 29
Register 13—3FSK/4FSK Demod Register ............................. 61
Receiver Section.............................................................................. 30
Register 14—Test DAC Register............................................... 62
RF Front End............................................................................... 30
Register 15—Test Mode Register ............................................. 63
IF Filter......................................................................................... 30
Outline Dimensions ....................................................................... 64
RSSI/AGC.................................................................................... 30
Ordering Guide .......................................................................... 64
REVISION HISTORY
2/08—Revision 0: Initial Version
Rev. 0 | Page 2 of 64
ADF7021-N
GENERAL DESCRIPTION
The ADF7021-N is a high performance, low power, narrowband transceiver based on the ADF7021. The ADF7021-N has
IF filter bandwidths of 9 kHz, 13.5 kHz, and 18.5 kHz, making
it ideally suited to worldwide narrowband standards and
particularly those that stipulate 12.5 kHz channel separation.
It is designed to operate in the narrow-band, license-free ISM
bands and in the licensed bands with frequency ranges of 80
MHz to 650 MHz and 842 MHz to 916 MHz. The part has both
Gaussian and raised cosine transmit data filtering options to
improve spectral efficiency for narrow-band applications. It is
suitable for circuit applications targeted at the Japanese ARIB
STD-T67, the European ETSI EN 300 220, the Korean short
range device regulations, the Chinese short range device
regulations, and the North American FCC Part 15, Part 90, and
Part 95 regulatory standards. A complete transceiver can be
built using a small number of external discrete components,
making the ADF7021-N very suitable for price-sensitive and
area-sensitive applications.
The range of on-chip FSK modulation and data filtering options
allows users greater flexibility in their choice of modulation
schemes while meeting the tight spectral efficiency requirements.
The ADF7021-N also supports protocols that dynamically
switch among 2FSK, 3FSK, and 4FSK to maximize communication range and data throughput.
The transmit section contains two voltage controlled oscillators
(VCOs) and a low noise fractional-N PLL with an output
resolution of <1 ppm. The ADF7021-N has a VCO using an
internal LC tank (421 MHz to 458 MHz, 842 MHz to 916 MHz)
and a VCO using an external inductor as part of its tank circuit
(80 MHz to 650 MHz). The dual VCO design allows dual-band
operation where the user can transmit and/or receive at any
frequency supported by the internal inductor VCO and can also
transmit and/or receive at a particular frequency band
supported by the external inductor VCO.
The frequency-agile PLL allows the ADF7021-N to be used in
frequency-hopping, spread spectrum (FHSS) systems. Both
VCOs operate at twice the fundamental frequency to reduce
spurious emissions and frequency pulling problems.
The transmitter output power is programmable in 63 steps from
−16 dBm to +13 dBm and has an automatic power ramp control
to prevent spectral splatter and help meet regulatory standards.
The transceiver RF frequency, channel spacing, and modulation
are programmable using a simple 3-wire interface. The device
operates with a power supply range of 2.3 V to 3.6 V and can be
powered down when not in use.
A low IF architecture is used in the receiver (100 kHz), which
minimizes power consumption and the external component
count yet avoids dc offset and flicker noise at low frequencies.
The IF filter has programmable bandwidths of 9 kHz, 13.5 kHz,
and 18.5 kHz. The ADF7021-N supports a wide variety of programmable features including Rx linearity, sensitivity, and IF
bandwidth, allowing the user to trade off receiver sensitivity
and selectivity against current consumption, depending on the
application. The receiver also features a patent-pending automatic
frequency control (AFC) loop with programmable pull-in range
that allows the PLL to track out the frequency error in the
incoming signal.
The receiver achieves an image rejection performance of 56 dB
using a patent-pending IR calibration scheme that does not
require the use of an external RF source.
An on-chip ADC provides readback of the integrated temperature sensor, external analog input, battery voltage, and RSSI
signal, which provides savings on an ADC in some applications.
The temperature sensor is accurate to ±10°C over the full operating temperature range of −40°C to +85°C. This accuracy can
be improved by performing a 1-point calibration at room
temperature and storing the result in memory.
Rev. 0 | Page 3 of 64
ADF7021-N
SPECIFICATIONS
VDD = 2.3 V to 3.6 V, GND = 0 V, TA = TMIN to TMAX, unless otherwise noted. Typical specifications are at VDD = 3 V, TA = 25°C.
All measurements are performed with the EVAL-ADF7021-NDBxx using the PN9 data sequence, unless otherwise noted.
RF AND PLL SPECIFICATIONS
Table 1.
Parameter
RF CHARACTERISTICS
Frequency Ranges (Direct Output)
Frequency Ranges (RF Divide-by-2 Mode)
Phase Frequency Detector (PFD) Frequency 1
PHASE-LOCKED LOOP (PLL)
VCO Gain 2
868 MHz, Internal Inductor VCO
426 MHz, Internal Inductor VCO
426 MHz, External Inductor VCO
160 MHz, External Inductor VCO
Phase Noise (In-Band)
868 MHz, Internal Inductor VCO
Min
Typ
160
842
80
421
RF/256
Max
Unit
650
916
325
458
24
MHz
MHz
MHz
MHz
MHz
Test Conditions/Comments
See Table 9 for required VCO_BIAS and
VCO_ADJUST settings
External inductor VCO
Internal inductor VCO
External inductor VCO, RF divide-by-2 enabled
Internal inductor VCO, RF divide-by-2 enabled
67
45
27
6
MHz/V
MHz/V
MHz/V
MHz/V
VCO_ADJUST = 0, VCO_BIAS = 8
VCO_ADJUST = 0, VCO_BIAS = 8
VCO_ADJUST = 0, VCO_BIAS = 3
VCO_ADJUST = 0, VCO_BIAS = 2
−97
dBc/Hz
433 MHz, Internal Inductor VCO
−103
dBc/Hz
426 MHz, External Inductor VCO
−95
dBc/Hz
Phase Noise (Out-of-Band)
−124
dBc/Hz
10 kHz offset, PA = 10 dBm, VDD = 3.0 V,
PFD = 19.68 MHz, VCO_BIAS = 8
10 kHz offset, PA = 10 dBm, VDD = 3.0 V,
PFD = 19.68 MHz, VCO_BIAS = 8
10 kHz offset, PA = 10 dBm, VDD = 3.0 V,
PFD = 9.84 MHz, VCO_BIAS = 3
1 MHz offset, fRF = 433 MHz, PA = 10 dBm,
VDD = 3.0 V, PFD = 19.68 MHz, VCO_BIAS = 8
Normalized In-Band Phase Noise Floor 3
PLL Settling
−203
40
dBc/Hz
μs
REFERENCE INPUT
Crystal Reference 4
External Oscillator4, 5
Crystal Start-Up Time 6
XTAL Bias = 20 μA
XTAL Bias = 35 μA
Input Level for External Oscillator 7
OSC1
OSC2
ADC PARAMETERS
INL
DNL
3.625
3.625
24
24
Measured for a 10 MHz frequency step to within
5 ppm accuracy, PFD = 19.68 MHz, loop bandwidth
(LBW) = 100 kHz
MHz
MHz
0.930
0.438
ms
ms
10 MHz XTAL, 33 pF load capacitors, VDD = 3.0 V
10 MHz XTAL, 33 pF load capacitors, VDD = 3.0 V
0.8
CMOS levels
V p-p
V
Clipped sine wave
±0.4
±0.4
LSB
LSB
VDD = 2.3 V to 3.6 V, TA = 25°C
VDD = 2.3 V to 3.6 V, TA = 25°C
1
The maximum usable PFD at a particular RF frequency is limited by the minimum N divide value.
VCO gain measured at a VCO tuning voltage of 0.7 V. The VCO gain varies across the tuning range of the VCO. The software package ADIsimPLL™ can be used to model this
variation.
3
This value can be used to calculate the in-band phase noise for any operating frequency. Use the following equation to calculate the in-band phase noise performance
as seen at the power amplifier (PA) output: −203 + 10 log(fPFD) + 20 logN.
4
Guaranteed by design. Sample tested to ensure compliance.
5
A TCXO, VCXO, or OCXO can be used as an external oscillator.
6
Crystal start-up time is the time from chip enable (CE) being asserted to correct clock frequency on the CLKOUT pin.
7
Refer to the Reference Input section for details on using an external oscillator.
2
Rev. 0 | Page 4 of 64
ADF7021-N
TRANSMISSION SPECIFICATIONS
Table 2.
Parameter
DATA RATE
2FSK, 3FSK
4FSK
MODULATION
Frequency Deviation (fDEV) 2
Deviation Frequency Resolution
Gaussian Filter BT
Raised Cosine Filter Alpha
TRANSMIT POWER
Maximum Transmit Power 3
Transmit Power Variation vs.
Temperature
Transmit Power Variation vs. VDD
Transmit Power Flatness
Programmable Step Size
ADJACENT CHANNEL POWER (ACP)
426 MHz, External Inductor VCO
12.5 kHz Channel Spacing
25 kHz Channel Spacing
868 MHz, Internal Inductor VCO
12.5 kHz Channel Spacing
25 kHz Channel Spacing
433 MHz, Internal Inductor VCO
12.5 kHz Channel Spacing
25 kHz Channel Spacing
Min
Typ
Max
Unit
Test Conditions/Comments
0.05
0.05
18.5 1
24
kbps
kbps
IF_FILTER_BW = 18.5 kHz
IF_FILTER_BW = 18.5 kHz
0.056
0.306
56
28.26
156
kHz
kHz
Hz
PFD = 3.625 MHz
PFD = 20 MHz
PFD = 3.625 MHz
0.5
0.5/0.7
Programmable
+13
±1
dBm
dB
VDD = 3.0 V, TA = 25°C
−40°C to +85°C
±1
±1
0.3125
dB
dB
dB
2.3 V to 3.6 V at 915 MHz, TA = 25°C
902 MHz to 928 MHz, 3 V, TA = 25°C
−16 dBm to +13 dBm
−50
dBc
−50
dBc
−46
dBm
−43
dBm
−50
dBm
−47
dBm
3.9
9.9
kHz
kHz
4.4
10.2
kHz
kHz
3.9
9.5
kHz
kHz
13.2
kHz
OCCUPIED BANDWIDTH
2FSK Gaussian Data Filtering
12.5 kHz Channel Spacing
25 kHz Channel Spacing
2FSK Raised Cosine Data Filtering
12.5 kHz Channel Spacing
25 kHz Channel Spacing
3FSK Raised Cosine Filtering
12.5 kHz Channel Spacing
25 kHz Channel Spacing
4FSK Raised Cosine Filtering
25 kHz Channel Spacing
PFD = 9.84 MHz
Gaussian 2FSK modulation, measured in a ±4.25 kHz bandwidth
at ±12.5 kHz offset, 2.4 kbps PN9 data, 1.2 kHz frequency deviation,
compliant with ARIB STD-T67
Gaussian 2FSK modulation, measured in a ±8 kHz bandwidth at
±25 kHz offset, 9.6 kbps PN9 data, 2.4 kHz frequency deviation,
compliant with ARIB STD-T67
PFD = 19.68 MHz
Gaussian 2FSK modulation, 10 dBm output power, measured in
a ±6.25 kHz bandwidth at ±12.5 kHz offset, 2.4 kbps PN9 data,
1.2 kHz frequency deviation, compliant with ETSI EN 300 220
Gaussian 2FSK modulation, 10 dBm output power, measured in
a ±12.5 kHz bandwidth at ±25 kHz offset, 9.6 kbps PN9 data,
2.4 kHz frequency deviation, compliant with ETSI EN 300 220
PFD = 19.68 MHz
Gaussian 2FSK modulation, 10 dBm output power, measured in
a ±6.25 kHz bandwidth at ±12.5 kHz offset, 2.4 kbps PN9 data,
1.2 kHz frequency deviation, compliant with ETSI EN 300 220
Gaussian 2FSK modulation, 10 dBm output power, measured in
a ±12.5 kHz bandwidth at ±25 kHz offset, 9.6 kbps PN9 data,
2.4 kHz frequency deviation, compliant with ETSI EN 300 220
99.0% of total mean power; 12.5 kHz channel spacing (2.4 kbps
PN9 data, 1.2 kHz frequency deviation); 25 kHz channel spacing
(9.6 kbps PN9 data, 2.4 kHz frequency deviation)
19.2 kbps PN9 data, 1.2 kHz frequency deviation
Rev. 0 | Page 5 of 64
ADF7021-N
Parameter
SPURIOUS EMISSIONS
Reference Spurs
HARMONICS 4
Second Harmonic
Third Harmonic
All Other Harmonics
OPTIMUM PA LOAD IMPEDANCE 5
fRF = 915 MHz
fRF = 868 MHz
fRF = 450 MHz
fRF = 426 MHz
fRF = 315 MHz
fRF = 175 MHz
Min
Typ
Max
Unit
Test Conditions/Comments
−65
dBc
100 kHz loop bandwidth
13 dBm output power, unfiltered conductive/filtered conductive
−35/−52
−43/−60
−36/−65
dBc
dBc
dBc
39 + j61
48 + j54
98 + j65
100 + j65
129 + j63
173 + j49
Ω
Ω
Ω
Ω
Ω
Ω
1
Using Gaussian or raised cosine filtering. The frequency deviation should be chosen to ensure that the transmit-occupied signal bandwidth is within the receiver
IF filter bandwidth.
2
For the definition of frequency deviation, refer to the Register 2—Transmit Modulation Register section.
3
Measured as maximum unmodulated power.
4
Conductive filtered harmonic emissions measured on the EVAL-ADF7021-NDBxx, which includes a T-stage harmonic filter (two inductors and one capacitor).
5
For matching details, refer to the LNA/PA Matching section.
RECEIVER SPECIFICATIONS
Table 3.
Parameter
SENSITIVITY
Unit
Test Conditions/Comments
Bit error rate (BER) = 10−3, low noise amplifier (LNA)
and power amplifier (PA) matched separately
−130
dBm
Sensitivity at 0.25 kbps
−127
dBm
Sensitivity at 1 kbps
−122
dBm
Sensitivity at 9.6 kbps
−115
dBm
fDEV = 1 kHz, high sensitivity mode, IF_FILTER_BW =
13.5 kHz
fDEV = 1 kHz, high sensitivity mode, IF_FILTER_BW =
13.5 kHz
fDEV = 1 kHz, high sensitivity mode, IF_FILTER_BW =
13.5 kHz
fDEV = 4 kHz, high sensitivity mode, IF_FILTER_BW =
18.5 kHz
Gaussian 2FSK
Sensitivity at 0.1 kbps
−129
dBm
Sensitivity at 0.25 kbps
−127
dBm
Sensitivity at 1 kbps
−121
dBm
Sensitivity at 9.6 kbps
−114
dBm
GMSK
Sensitivity at 9.6 kbps
−113
dBm
fDEV = 2.4 kHz, high sensitivity mode, IF_FILTER_BW =
18.5 kHz
Raised Cosine 2FSK
Sensitivity at 0.25 kbps
−127
dBm
Sensitivity at 1 kbps
−121
dBm
Sensitivity at 9.6 kbps
−114
dBm
fDEV = 1 kHz, high sensitivity mode, IF_FILTER_BW =
13.5 kHz
fDEV = 1 kHz, high sensitivity mode, IF_FILTER_BW =
13.5 kHz
fDEV = 4 kHz, high sensitivity mode, IF_FILTER_BW =
18.5 kHz
2FSK
Sensitivity at 0.1 kbps
Min
Typ
Max
Rev. 0 | Page 6 of 64
fDEV = 1 kHz, high sensitivity mode, IF_FILTER_BW =
13.5 kHz
fDEV = 1 kHz, high sensitivity mode, IF_FILTER_BW =
13.5 kHz
fDEV = 1 kHz, high sensitivity mode, IF_FILTER_BW =
13.5 kHz
fDEV = 4 kHz, high sensitivity mode, IF_FILTER_BW =
18.5 kHz
ADF7021-N
Parameter
3FSK
Sensitivity at 9.6 kbps
Min
Typ
Max
Unit
Test Conditions/Comments
−110
dBm
fDEV = 2.4 kHz, high sensitivity mode, IF_FILTER_BW =
18.5 kHz, Viterbi detection on
Raised Cosine 3FSK
Sensitivity at 9.6 kbps
−110
dBm
fDEV = 2.4 kHz, high sensitivity mode, IF_FILTER_BW =
13.5 kHz, alpha = 0.5, Viterbi detection on
4FSK
Sensitivity at 9.6 kbps
−112
dBm
fDEV (inner) = 1.2 kHz, high sensitivity mode,
IF_FILTER_BW = 13.5 kHz
Raised Cosine 4FSK
Sensitivity at 9.6 kbps
−109
dBm
−3
dBm
fDEV (inner) = 1.2 kHz, high sensitivity mode,
IF_FILTER_BW = 13.5 kHz, alpha = 0.5
Two-tone test, fLO = 860 MHz, F1 = fLO + 100 kHz,
F2 = fLO − 800 kHz
LNA_GAIN = 3, MIXER_LINEARITY = 1
−13.5
−24
dBm
dBm
LNA_GAIN = 10, MIXER_LINEARITY = 0
LNA_GAIN = 30, MIXER_LINEARITY = 0
INPUT IP3
Low Gain Enhanced Linearity
Mode
Medium Gain Mode
High Sensitivity Mode
ADJACENT CHANNEL REJECTION
868 MHz
12.5 kHz Channel Spacing
25 kHz Channel Spacing
426 MHz
40
39
dB
dB
12.5 kHz Channel Spacing
25 kHz Channel Spacing
CO-CHANNEL REJECTION
40
39
dB
dB
868 MHz
IMAGE CHANNEL REJECTION
−5
dB
868 MHz
450 MHz, Internal Inductor
VCO
BLOCKING
26/39
29/50
±1 MHz
±2 MHz
±5 MHz
±10 MHz
SATURATION
(MAXIMUM INPUT LEVEL)
69
75
78
78.5
12
dB
dB
Wanted signal is 3 dB above the sensitivity point
(BER = 10−3); unmodulated interferer is at the center
of the adjacent channel; rejection measured as the
difference between the interferer level and the
wanted signal level in dB
9 kHz IF_FILTER_BW
18.5 kHz IF_FILTER_BW
Wanted signal is 3 dB above the reference sensitivity
point (BER = 10−2); modulated interferer (same
modulation as wanted signal) at the center of the
adjacent channel; rejection measured as the
difference between the interferer level and reference
sensitivity level in dB
9 kHz IF_FILTER_BW, compliant with ARIB STD-T67
18.5 kHz IF_FILTER_BW, compliant with ARIB STD-T67
Wanted signal (2FSK, 9.6 kbps, ±4 kHz deviation) is
3 dB above the sensitivity point (BER = 10−3), modulated interferer
Wanted signal (2FSK, 9.6 kbps, ±4 kHz deviation) is
10 dB above the sensitivity point (BER = 10−3); modulated interferer (2FSK, 9.6 kbps, ±4 kHz deviation) is
placed at the image frequency of fRF − 200 kHz; the
interferer level is increased until BER = 10−3
Uncalibrated/calibrated 1 , VDD = 3.0 V, TA = 25°C
Uncalibrated/calibrated1, VDD = 3.0 V, TA = 25°C
Wanted signal is 10 dB above the input sensitivity
level; CW interferer level is increased until BER = 10−3
dB
dB
dB
dB
dBm
Rev. 0 | Page 7 of 64
2FSK mode, BER = 10−3
ADF7021-N
Parameter
RSSI
Range at Input 2
Linearity
Absolute Accuracy
Response Time
AFC
Pull-In Range
Response Time
Accuracy
Rx SPURIOUS EMISSIONS 3
Internal Inductor VCO
External Inductor VCO
LNA INPUT IMPEDANCE
fRF = 915 MHz
fRF = 868 MHz
fRF = 450 MHz
fRF = 426 MHz
fRF = 315 MHz
fRF = 175 MHz
Min
Typ
Max
−120 to −47
±2
±3
390
0.5
1.5 × IF_
FILTER_BW
Unit
Test Conditions/Comments
dBm
dB
dB
μs
Input power range = −100 dBm to −47 dBm
Input power range = −100 dBm to −47 dBm
See the RSSI/AGC section
kHz
64
0.5
Bits
kHz
−91/−91
dBm
−52/−70
dBm
−62/−72
dBm
−64/−85
dBm
24 − j60
26 − j63
63 − j129
68 − j134
96 − j160
178 − j190
Ω
Ω
Ω
Ω
Ω
Ω
1
The range is programmable in Register 10
(R10_DB[24:31])
Input power range = −100 dBm to +12 dBm
<1 GHz at antenna input, unfiltered conductive/filtered
conductive
>1 GHz at antenna input, unfiltered conductive/filtered
conductive
<1 GHz at antenna input, unfiltered conductive/filtered
conductive
>1 GHz at antenna input, unfiltered conductive/filtered
conductive
RFIN to RFGND
Calibration of the image rejection used an external RF source.
For received signal levels < −100 dBm, it is recommended to average the RSSI readback value over a number of samples to improve the RSSI accuracy at low input powers.
3
Filtered conductive receive spurious emissions are measured on the EVAL-ADF7021-NDBxx, which includes a T-stage harmonic filter (two inductors and one
capacitor).
2
Rev. 0 | Page 8 of 64
ADF7021-N
DIGITAL SPECIFICATIONS
Table 4.
Parameter
TIMING INFORMATION
Chip Enabled to Regulator Ready
Chip Enabled to Tx Mode
TCXO Reference
XTAL
Chip Enabled to Rx Mode
Min
Max
Unit
Test Conditions/Comments
10
μs
CREG (1:4) = 100 nF
32-bit register write time = 50 μs
1
2
ms
ms
32-bit register write time = 50 μs, IF filter coarse
calibration only
TCXO Reference
XTAL
Tx-to-Rx Turnaround Time
LOGIC INPUTS
Input High Voltage, VINH
Input Low Voltage, VINL
Input Current, IINH/IINL
Input Capacitance, CIN
Control Clock Input
LOGIC OUTPUTS
Output High Voltage, VOH
Output Low Voltage, VOL
CLKOUT Rise/Fall
CLKOUT Load
Typ
1.2
2.2
390 μs + (5 × tBIT)
ms
ms
Time to synchronized data out, includes AGC
settling (three AGC levels)and CDR synchronization;
see the AGC Information and Timing section for
more details; tBIT = data bit period
0.7 × VDD
0.2 × VDD
±1
10
50
V
V
μA
pF
MHz
0.4
5
10
V
V
ns
pF
DVDD − 0.4
Rev. 0 | Page 9 of 64
IOH = 500 μA
IOL = 500 μA
ADF7021-N
GENERAL SPECIFICATIONS
Table 5.
Parameter
TEMPERATURE RANGE (TA)
POWER SUPPLIES
Voltage Supply, VDD
TRANSMIT CURRENT CONSUMPTION 1
868 MHz
0 dBm
5 dBm
10 dBm
450 MHz, Internal Inductor VCO
0 dBm
5 dBm
10 dBm
426 MHz, External Inductor VCO
0 dBm
5 dBm
10 dBm
RECEIVE CURRENT CONSUMPTION
868 MHz
Low Current Mode
High Sensitivity Mode
433MHz, Internal Inductor VCO
Low Current Mode
High Sensitivity Mode
426 MHz, External Inductor VCO
Low Current Mode
High Sensitivity Mode
POWER-DOWN CURRENT CONSUMPTION
Low Power Sleep Mode
1
Min
−40
Typ
2.3
Max
+85
Unit
°C
Test Conditions/Comments
3.6
V
All VDD pins must be tied together
VDD = 3.0 V, PA is matched into 50 Ω
VCO_BIAS = 8
20.2
24.7
32.3
mA
mA
mA
19.9
23.2
29.2
mA
mA
mA
13.5
17
23.3
mA
mA
mA
VCO_BIAS = 8
VCO_BIAS = 2
VDD = 3.0 V
VCO_BIAS = 8
22.7
24.6
mA
mA
24.5
26.4
mA
mA
17.5
19.5
mA
mA
VCO_BIAS = 8
VCO_BIAS = 2
0.1
1
μA
CE low
The transmit current consumption tests used the same combined PA and LNA matching network as that used on the EVAL-ADF7021-NDBxx evaluation boards.
Improved PA efficiency is achieved by using a separate PA matching network.
Rev. 0 | Page 10 of 64
ADF7021-N
TIMING CHARACTERISTICS
VDD = 3 V ± 10%, DGND = AGND = 0 V, TA = 25°C, unless otherwise noted. Guaranteed by design but not production tested.
Table 6.
Parameter
t1
t2
t3
t4
t5
t6
t8
t9
t10
t11
t12
t13
t14
t15
Limit at TMIN to TMAX
>10
>10
>25
>25
>10
>20
<25
<25
>10
5 < t11 < (¼ × tBIT)
>5
>5
>¼ × tBIT
>¼ × tBIT
Unit
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
μs
μs
Test Conditions/Comments
SDATA to SCLK setup time
SDATA to SCLK hold time
SCLK high duration
SCLK low duration
SCLK to SLE setup time
SLE pulse width
SCLK to SREAD data valid, readback
SREAD hold time after SCLK, readback
SCLK to SLE disable time, readback
TxRxCLK negative edge to SLE
TxRxDATA to TxRxCLK setup time (Tx mode)
TxRxCLK to TxRxDATA hold time (Tx mode)
TxRxCLK negative edge to SLE
SLE positive edge to positive edge of TxRxCLK
Rev. 0 | Page 11 of 64
ADF7021-N
TIMING DIAGRAMS
Serial Interface
t3
t4
SCLK
t1
SDATA
DB31 (MSB)
t2
DB30
DB1
(CONTROL BIT C2)
DB2
DB0 (LSB)
(CONTROL BIT C1)
t6
07246-002
SLE
t5
Figure 2. Serial Interface Timing Diagram
t1
t2
SCLK
SDATA
REG7 DB0
(CONTROL BIT C1)
SLE
t3
t10
RV16
RV2
RV15
RV1
X
07246-003
X
SREAD
t9
t8
Figure 3. Serial Interface Readback Timing Diagram
2FSK/3FSK Timing
±1 × DATA RATE/32
1/DATA RATE
TxRxCLK
TxRxDATA
07246-004
DATA
Figure 4. TxRxDATA/TxRxCLK Timing Diagram in Receive Mode
1/DATA RATE
TxRxCLK
TxRxDATA
FETCH
07246-005
DATA
SAMPLE
Figure 5. TxRxDATA/TxRxCLK Timing Diagram in Transmit Mode
Rev. 0 | Page 12 of 64
ADF7021-N
4FSK Timing
In 4FSK receive mode, MSB/LSB synchronization should be guaranteed by SWD in the receive bit stream.
REGISTER 0 WRITE
SWITCH FROM Rx TO Tx
tSYMBOL
t13
t12
t11
tBIT
SLE
TxRxCLK
Rx SYMBOL
MSB
Rx SYMBOL
LSB
Rx SYMBOL
MSB
Rx SYMBOL
LSB
Tx SYMBOL
MSB
Tx SYMBOL
LSB
Rx MODE
Tx/Rx MODE
Tx SYMBOL
MSB
Tx MODE
07246-074
TxRxDATA
Figure 6. Receive-to-Transmit Timing Diagram in 4FSK Mode
REGISTER 0 WRITE
SWITCH FROM Tx TO Rx
t15
tSYMBOL
t14
tBIT
SLE
TxRxCLK
Tx/Rx MODE
Tx SYMBOL
MSB
Tx SYMBOL
LSB
Tx SYMBOL
MSB
Tx SYMBOL
LSB
Rx SYMBOL
MSB
Tx MODE
Figure 7. Transmit-to-Receive Timing Diagram in 4FSK Mode
Rev. 0 | Page 13 of 64
Rx SYMBOL
LSB
Rx MODE
07246-075
TxRxDATA
ADF7021-N
UART/SPI Mode
UART mode is enabled by setting R0_DB28 to 1. SPI mode is enabled by setting R0_DB28 to 1 and setting R15_DB[17:19] to 0x7.
The transmit/receive data clock is available on the CLKOUT pin.
tBIT
CLKOUT
(TRANSMIT/RECEIVE DATA
CLOCK IN SPI MODE.
NOT USED IN UART MODE.)
Tx BIT
SAMPLE
Tx BIT
TxRxDATA
(RECEIVE DATA OUTPUT
IN UART/SPI MODE.)
Tx BIT
Tx BIT
Tx BIT
HIGH-Z
Tx/Rx MODE
07246-082
TxRxCLK
(TRANSMIT DATA INPUT
IN UART/SPI MODE.)
FETCH
Tx MODE
Figure 8. Transmit Timing Diagram in UART/SPI Mode
tBIT
CLKOUT
(TRANSMIT/RECEIVE DATA
CLOCK IN SPI MODE.
NOT USED IN UART MODE.)
FETCH SAMPLE
TxRxCLK
(TRANSMIT DATA INPUT
IN UART/SPI MODE.)
Tx/Rx MODE
Rx BIT
Rx BIT
Rx BIT
Rx BIT
Rx MODE
Figure 9. Receive Timing Diagram in UART/SPI Mode
Rev. 0 | Page 14 of 64
Rx BIT
07246-078
TxRxDATA
(RECEIVE DATA OUTPUT
IN UART/SPI MODE.)
HIGH-Z
ADF7021-N
ABSOLUTE MAXIMUM RATINGS
TA = 25°C, unless otherwise noted.
Table 7.
Parameter
VDD to GND1
Analog I/O Voltage to GND
Digital I/O Voltage to GND
Operating Temperature Range
Industrial (B Version)
Storage Temperature Range
Maximum Junction Temperature
MLF θJA Thermal Impedance
Reflow Soldering
Peak Temperature
Time at Peak Temperature
1
Rating
−0.3 V to +5 V
−0.3 V to AVDD + 0.3 V
−0.3 V to DVDD + 0.3 V
−40°C to +85°C
−65°C to +125°C
150°C
26°C/W
260°C
40 sec
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
should be taken for handling and assembly.
ESD CAUTION
GND = CPGND = RFGND = DGND = AGND = 0.
Rev. 0 | Page 15 of 64
ADF7021-N
CVCO
GND1
L1
GND
L2
VDD
CPOUT
CREG3
VDD3
OSC1
OSC2
MUXOUT
48
47
46
45
44
43
42
41
40
39
38
37
PIN CONFIGURATION AND FUNCTION DESCRIPTIONS
PIN 1
INDICATOR
1
36
CLKOUT
2
35
TxRxCLK
VDD1
3
34
TxRxDATA
RFOUT
4
33
SWD
RFGND
5
32
VDD2
RFIN
6
31
CREG2
RFINB
7
30
ADCIN
RLNA
8
29
GND2
VDD4
9
28
SCLK
RSET 10
27
SREAD
CREG4 11
26
SDATA
GND4 12
25
SLE
ADF7021-N
CE 24
TEST_A 23
GND4 22
FILT_Q 21
FILT_Q 20
GND4 19
FILT_I 18
FILT_I 17
MIX_Q 16
MIX_Q 15
MIX_I 14
MIX_I 13
TOP VIEW
(Not to Scale)
07246-006
VCOIN
CREG1
Figure 10. Pin Configuration
Table 8. Pin Function Descriptions
Pin No.
1
Mnemonic
VCOIN
2
CREG1
3
VDD1
4
RFOUT
5
6
RFGND
RFIN
7
8
9
10
11
RFINB
RLNA
VDD4
RSET
CREG4
12, 19, 22
13 to 18
24
GND4
MIX_I, MIX_I,
MIX_Q, MIX_Q,
FILT_I, FILT_I
FILT_Q, FILT_Q,
TEST_A
CE
25
SLE
26
SDATA
27
SREAD
28
SCLK
20, 21, 23
Description
The tuning voltage on this pin determines the output frequency of the voltage controlled oscillator (VCO).
The higher the tuning voltage, the higher the output frequency.
Regulator Voltage for PA Block. Place a series 3.9 Ω resistor and a 100 nF capacitor between this pin and
ground for regulator stability and noise rejection.
Voltage Supply for PA Block. Place decoupling capacitors of 0.1 μF and 100 pF as close as possible to this pin.
Tie all VDD pins together.
The modulated signal is available at this pin. Output power levels are from −16 dBm to +13 dBm. The output
should be impedance matched to the desired load using suitable components (see the Transmitter section).
Ground for Output Stage of Transmitter. All GND pins should be tied together.
LNA Input for Receiver Section. Input matching is required between the antenna and the differential LNA
input to ensure maximum power transfer (see the LNA/PA Matching section).
Complementary LNA Input. (See the LNA/PA Matching section.)
External Bias Resistor for LNA. Optimum resistor is 1.1 kΩ with 5% tolerance.
Voltage Supply for LNA/MIXER Block. This pin should be decoupled to ground with a 10 nF capacitor.
External Resistor. Sets charge pump current and some internal bias currents. Use a 3.6 kΩ resistor with 5% tolerance.
Regulator Voltage for LNA/MIXER Block. Place a 100 nF capacitor between this pin and GND for regulator
stability and noise rejection.
Ground for LNA/MIXER Block.
Signal Chain Test Pins. These pins are high impedance under normal conditions and should be left unconnected.
Signal Chain Test Pins. These pins are high impedance under normal conditions and should be left unconnected.
Chip Enable. Bringing CE low puts the ADF7021-N into complete power-down. Register values are lost when
CE is low, and the part must be reprogrammed after CE is brought high.
Load Enable, CMOS Input. When SLE goes high, the data stored in the shift registers is loaded into one of the
four latches. A latch is selected using the control bits.
Serial Data Input. The serial data is loaded MSB first with the four LSBs as the control bits. This pin is a high
impedance CMOS input.
Serial Data Output. This pin is used to feed readback data from the ADF7021-N to the microcontroller. The
SCLK input is used to clock each readback bit (for example, AFC or ADC) from the SREAD pin.
Serial Clock Input. This serial clock is used to clock in the serial data to the registers. The data is latched into
the 32-bit shift register on the CLK rising edge. This pin is a digital CMOS input.
Rev. 0 | Page 16 of 64
ADF7021-N
Pin No.
29
30
Mnemonic
GND2
ADCIN
31
CREG2
32
33
VDD2
SWD
34
TxRxDATA
35
TxRxCLK
36
CLKOUT
37
MUXOUT
38
OSC2
39
OSC1
40
41
VDD3
CREG3
42
CPOUT
43
44, 46
VDD
L2, L1
45, 47
48
GND, GND1
CVCO
Description
Ground for Digital Section.
Analog-to-Digital Converter Input. The internal 7-bit ADC can be accessed through this pin. Full scale is 0 V to
1.9 V. Readback is made using the SREAD pin.
Regulator Voltage for Digital Block. Place a 100 nF capacitor between this pin and ground for regulator
stability and noise rejection.
Voltage Supply for Digital Block. Place a decoupling capacitor of 10 nF as close as possible to this pin.
Sync Word Detect. The ADF7021-N asserts this pin when it has found a match for the sync word sequence
(see the Register 11—Sync Word Detect Register section). This provides an interrupt for an external
microcontroller indicating that valid data is being received.
Transmit Data Input/Received Data Output. This is a digital pin, and normal CMOS levels apply. In UART/SPI
mode, this pin provides an output for the received data in receive mode. In transmit UART/SPI mode, this pin
is high impedance (see the Interfacing to a Microcontroller/DSP section).
Outputs the data clock in both receive and transmit modes. This is a digital pin, and normal CMOS levels
apply. The positive clock edge is matched to the center of the received data. In transmit mode, this pin
outputs an accurate clock to latch the data from the microcontroller into the transmit section at the exact
required data rate. In UART/SPI mode, this pin is used to input the transmit data in transmit mode. In receive
UART/SPI mode, this pin is high impedance (see the Interfacing to a Microcontroller/DSP section).
A divided-down version of the crystal reference with output driver. The digital clock output can be used to drive
several other CMOS inputs such as a microcontroller clock. The output has a 50:50 mark-space ratio and is inverted
with respect to the reference. Place a series 1 kΩ resistor as close as possible to the pin in applications where the
CLKOUT feature is being used.
Provides the DIGITAL_LOCK_DETECT signal. This signal is used to determine if the PLL is locked to the correct
frequency. It also provides other signals such as REGULATOR_READY, which is an indicator of the status of the
serial interface regulator (see the MUXOUT section for more information).
Connect the reference crystal between this pin and OSC1. A TCXO reference can be used by driving this pin
with CMOS levels and disabling the internal crystal oscillator.
Connect the reference crystal between this pin and OSC2. A TCXO reference can be used by driving this pin
with ac-coupled 0.8 V p-p levels and by enabling the internal crystal oscillator.
Voltage Supply for the Charge Pump and PLL Dividers. Decouple this pin to ground with a 10 nF capacitor.
Regulator Voltage for Charge Pump and PLL Dividers. Place a 100 nF capacitor between this pin and ground
for regulator stability and noise rejection.
Charge Pump Output. This output generates current pulses that are integrated in the loop filter. The
integrated current changes the control voltage on the input to the VCO.
Voltage Supply for VCO Tank Circuit. Decouple this pin to ground with a 10 nF capacitor.
External VCO Inductor Pins. If using an external VCO inductor, connect a chip inductor across these pins to set
the VCO operating frequency. If using the internal VCO inductor, these pins can be left floating. See the
Voltage Controlled Oscillator (VCO) section for more information.
Grounds for VCO Block.
Place a 22 nF capacitor between this pin and CREG1 to reduce VCO noise.
Rev. 0 | Page 17 of 64
ADF7021-N
TYPICAL PERFORMANCE CHARACTERISTICS
–70
PHASE NOISE (dBc/Hz)
ICP = 0.8mA
–90
DR = 9.6kbps
DATA = PRBS9
fDEV = 2.4kHz
RF FREQ = 869.5MHz
RF FREQ = 900MHz
VDD = 2.3V
TEMPERATURE = 25°C
VCO_BIAS = 8
VCO_ADJUST = 3
–80
ICP = 1.4mA
–100
2FSK
GFSK
–110
ICP = 2.2mA
–120
–130
1
10
100
1000
10000
FREQUENCY OFFSET (kHz)
Figure 11. Phase Noise Response at 900 MHz, VDD = 2.3 V
16
VBW 300Hz
8
SPAN 50kHz
SWEEP 2.118s (601pts)
Figure 14. Output Spectrum in 2FSK and GFSK Modes
DR = 9.6kbps
DATA = PRBS9
fDEV = 2.4kHz
RF FREQ = 869.5MHz
PA_BIAS = 11µA
12
PA_BIAS = 9µA
4
0
PA_BIAS = 5µA
–4
2FSK
PA_BIAS = 7µA
–8
–12
–16
–20
–24
RC2FSK
–28
–32
0
4
8
12 16 20 24 28 32 36 40 44 48 52 56 60
PA SETTING
Figure 12. RF Output Power vs. PA Setting
1R
CENTER 869.5 25MHz
RES BW 300Hz
VBW 300Hz
SPAN 50kHz
SWEEP 2.118s (601pts)
07246-048
–40
07246-051
–36
Figure 15. Output Spectrum in 2FSK and Raised Cosine 2FSK Modes
RF FREQ = 440MHz
OUTPUT POWER = 10dBm
FILTER = T-STAGE LC FILTER
MARKER Δ = 52.2dB
SR = 4.8ksym/s
DATA = PRBS9
fDEV = 2.4kHz
RF FREQ = 869.5MHz
4FSK
1
START 300MHz
RES BW 100Hz
STOP 3.5GHz
VBW 100Hz SWEEP 385.8ms (601pts)
CENTER 869.493 8MHz
RES BW 300Hz
Figure 13. PA Output Harmonic Response with T-Stage LC Filter
VBW 300Hz
SPAN 100kHz
SWEEP 4.237s (601pts)
07246-049
RC4FSK
07246-050
RF OUTPUT POWER (dBm)
CENTER 869.5 25MHz
RES BW 300Hz
07246-047
–150
07246-060
–140
Figure 16. Output Spectrum in 4FSK and Raised Cosine 4FSK Modes
Rev. 0 | Page 18 of 64
ADF7021-N
REF 15dBm
SAMP LOG 10dB/
0
ATTEN 25dB
DR = 9.6kbps
DATA = PRS9
fDEV = 2.4kHz
RF FREQ = 869.5MHz
DATA RATE = 1kbps
fDEV = 1kHz
RF FREQ = 135MHz
IF BW = 12.5kHz
–1
–2
VAVG 100
V1 V2
S3 FC
LOG BER
–3
3FSK
3.0V, +25°C
2.3V, +85°C
–4
–5
RC3FSK
3.6V, –40°C
–6
CENTER 869.5MHz
RES BW 300Hz
VBW 300Hz
SPAN 50kHz
SWEEP2.226s (401pts)
07246-070
–8
–130 –128 –126 –124 –122 –120 –118 –116 –114 –112 –110 –108
RF INPUT POWER (dBm)
Figure 20. 2FSK Sensitivity vs. VDD and Temperature, fRF = 135 MHz
Figure 17. Output Spectrum in 3FSK and Raised Cosine 3FSK Modes
RAMP RATE:
CW ONLY
256 CODES/BIT
128 CODES/BIT
64 CODES/BIT
32 CODES/BIT
10
0
TRACE = MAX HOLD
PA ON/OFF RATE = 3Hz
PA ON/OFF CYCLES = 10,000
VDD = 3.0V
–2
–10
–30
–5
–40
–6
–50
–7
–50
0
50
100
FREQUENCY OFFSET (kHz)
–8
–120
0
–2
3.0V, +25°C
2.3V, +85°C
3.6V, –40°C
–6
–6
–7
–7
–116
–114
–112
–110
–108
–106
–104
RF INPUT POWER (dBm)
07246-052
–5
–118
–95
–4
–5
–120
–100
–3
LOG BER
–4
–105
DATA RATE = 19.6kbps
SYMBOL RATE = 9.8ksym/s
fDEV (inner) = 2.4kHz
MOD INDEX = 0.5
RF FREQ = 420MHz
IF BW = 12.5kHz
–1
–2
–8
–122
–110
Figure 21. 3FSK Sensitivity vs. VDD and Temperature, fRF = 440 MHz
DATA RATE = 9.6kbps
fDEV = 4kHz
RF FREQ = 868MHz
IF BW = 25kHz
–1
–115
RF INPUT POWER (dBm)
Figure 18. Output Spectrum in Maximum Hold
for Various PA Ramp Rate Options
0
2.3V +25°C
3.0V +25°C
3.6V +25°C
2.3V –40°C
3.0V –40°C
3.6V –40°C
2.3V +85°C
3.0V +85°C
3.6V +85°C
Figure 19. 2FSK Sensitivity vs. VDD and Temperature, fRF = 868 MHz
–8
–120
2.3V +25°C
3.0V +25°C
3.6V +25°C
2.3V –40°C
3.0V –40°C
3.6V –40°C
2.3V +85°C
3.0V +85°C
3.6V +85°C
–115
–110
–105
–100
–95
RF INPUT POWER (dBm)
Figure 22. 4FSK Sensitivity vs. VDD and Temperature, fRF = 420 MHz
Rev. 0 | Page 19 of 64
07246-066
–60
–100
LOG BER
–4
07246-065
LOG BER
–3
–20
–3
3FSK MODULATION
DATA RATE = 9.6kbps
fDEV = 2.4kHz
MOD INDEX = 0.5
RF FREQ = 440 MHz
–1
07246-068
OUTPUT POWER (dBm)
0
07246-053
–7
90
2.5
80
0
–2.5
–5.0
–7.5
60
–10.0
–12.5
ATTENUATION (dB)
70
50
RF FREQ = 868MHz
WANTED SIGNAL
(10dB ABOVE SENSITIVITY
POINT) = 2FSK,
fDEV = 4kHz,
DATA RATE = 9.8kbps
BLOCKER = 2FSK,
fDEV = 4kHz,
DATA RATE = 9.8kbps
VDD = 3.0V
TEMPERATURE = 25°C
30
20
10
0
–10
–22
–18
–14
–10
–6
–2 0 2
6
10
14
18
22
FREQUENCY OFFSET (MHz)
–15.0
–17.5
–40°C
–20.0
–22.5
–25.0
–27.5
–30.0
–32.5
–35.0
Figure 23. Wideband Interference Rejection
–37.5
90
92
94
96
98
100
–100
106
108
110
RF FREQ = 860MHz
2FSK MODULATION
DATA RATE = 9.6kbps
IF BW = 25kHz
VDD = 3.0V
TEMPERATURE = 25°C
SENSITIVITY POINT (dBm)
–102
–60
–80
–100
104
Figure 26. Variation of IF Filter Response with Temperature
(IF_FILTER_BW = 9 kHz, Temperature Range is −40°C to +90°C in 10° Steps)
RSSI
READBACK LEVEL
–40
102
IF FREQUENCY (kHz)
–20
RSSI LEVEL (dBm)
+90°C
007246-091
40
07246-059
BLOCKING (dB)
ADF7021-N
ACTUAL RF INPUT LEVEL
–104
–106
–108
DISCRIMINATOR BANDWIDTH =
2× FSK FREQUENCY DEVIATION
–110
–112
–114
–120
–116
–52.5
–42.5
CALIBRATED
60
0.4
0.6
0.8
1.0
1.2
0
RF FREQ = 430MHz
EXTERNAL VCO INDUCTOR
DATA RATE = 9.6kbps
TEMPERATURE = 25°C, VDD = 3.0V
–1
THRESHOLD DETECTION
–2
UNCALIBRATED
20
–3
–4
–5
10
–6
0
–10
429.80 429.85 429.90 429.95 430.00 430.05 430.10 430.15 430.20
RF FREQUENCY (MHz)
VITERBI DETECTION
3FSK MODULATION
VDD = 3.0V, TEMP = 25°C
DATA RATE = 9.6kbps
fDEV = 2.4kHz
RF FREQ = 868MHz
IF BW = 18.75kHz
–7
–120 –118 –116 –114 –112 –110 –108 –106 –104 –102 –100
INPUT POWER (dBm)
Figure 28. 3FSK Receiver Sensitivity Using Viterbi Detection and
Threshold Detection
Figure 25. Image Rejection, Uncalibrated vs. Calibrated
Rev. 0 | Page 20 of 64
07246-062
LOG BER
40
07246-054
BLOCKING (dB)
0.2
Figure 27. 2FSK Sensitivity vs. Modulation Index vs. Correlator
Discriminator Bandwidth
50
30
0
MODULATION INDEX
Figure 24. Digital RSSI Readback Linearity
70
DISCRIMINATOR BANDWIDTH =
1× FSK FREQUENCY DEVIATION
07246-058
–62.5
–118
07246-055
–140
–122.5 –112.5 –102.5 –92.5 –82.5 –72.5
RF INPUT (dBm)
ADF7021-N
–70
+3
HIGH MIXER
LINEARITY
SENSITIVITY (dBm)
+1
0
–1
–90
22452 ACQS
IF BW = 25kHz
POST DEMOD BW = 12.4kHz
M 50µs
IP3 = –9dBm
IP3 = –3dBm
IP3 = –20dBm
–110
–120
DEFAULT
MIXER
LINEARITY
IP3 = –13.5dBm
IP3 = –24dBm
–130
3, 72
10, 72
30, 72
(LOW GAIN MODE)
(MEDIUM GAIN MODE)
(HIGH GAIN MODE)
LNA GAIN, FILTER GAIN
Figure 31. Receive Sensitivity vs. LNA/IF Filter Gain and Mixer Linearity Settings
(The input IP3 at each setting is also shown)
+1
0
–1
4
RF I/P LEVEL = –70dBm
DATA RATE = 10kbps
fDEV = 2.5kHz
20834 ACQS
IF BW = 12.5kHz
POST DEMOD BW = 12.4kHz
M 20µs
C13
1.7V
07246-063
RECEIVER SYMBOL LEVEL
Figure 29. 4FSK Receiver Eye Diagram Measured Using the Test DAC Output
07246-069
RF I/P LEVEL = –70dBm
DATA RATE = 9.7kbps
fDEV (inner) = 1.2kHz
IP3= –5dBm
–100
–3
07246-064
RECEIVER SYMBOL LEVEL
–80
MODULATION = 2FSK
DATA RATE = 9.6kbps
fDEV = 4kHz
IF BW = 12.5kHz
DEMOD = CORRELATOR
SENSITIVITY @ 1E-3 BER
Figure 30. 3FSK Receiver Eye Diagram Measured Using the Test DAC Output
Rev. 0 | Page 21 of 64
ADF7021-N
FREQUENCY SYNTHESIZER
REFERENCE INPUT
CLKOUT Divider and Buffer
The on-board crystal oscillator circuitry (see Figure 32) can use
a quartz crystal as the PLL reference. Using a quartz crystal with
a frequency tolerance of ≤10 ppm for narrow-band applications
is recommended. It is possible to use a quartz crystal with >10 ppm
tolerance, but to comply with the absolute frequency error
specifications of narrow-band regulations (for example, ARIB
STD-T67 and ETSI EN 300 220), compensation for the
frequency error of the crystal is necessary.
The CLKOUT circuit takes the reference clock signal from the
oscillator section, shown in Figure 32, and supplies a divideddown, 50:50 mark-space signal to the CLKOUT pin. The CLKOUT
signal is inverted with respect to the reference clock. An even
divide from 2 to 30 is available. This divide number is set in
R1_DB[7:10]. On power-up, the CLKOUT defaults to divide-by-8.
DVDD
CLKOUT
ENABLE BIT
OSC1
07246-083
CP1
DIVIDER
1 TO 15
÷2
CLKOUT
Figure 33. CLKOUT Stage
To disable CLKOUT, set the divide number to 0. The output
buffer can drive up to a 20 pF load with a 10% rise time at
4.8 MHz. Faster edges can result in some spurious feedthrough
to the output. A series resistor (1 kΩ) can be used to slow the
clock edges to reduce these spurs at the CLKOUT frequency.
OSC2
CP2
OSC1
07246-008
The oscillator circuit is enabled by setting R1_DB12 high. It is
enabled by default on power-up and is disabled by bringing CE
low. Errors in the crystal can be corrected by using the automatic
frequency control feature or by adjusting the fractional-N value
(see the N Counter section).
R Counter
Figure 32. Oscillator Circuit on the ADF7021-N
Two parallel resonant capacitors are required for oscillation at
the correct frequency. Their values are dependent on the crystal
specification. They should be chosen to make sure that the
series value of capacitance added to the PCB track capacitance
adds up to the specified load capacitance of the crystal, usually
12 pF to 20 pF. Track capacitance values vary from 2 pF to 5 pF,
depending on board layout. When possible, choose capacitors
that have a very low temperature coefficient to ensure stable
frequency operation over all conditions.
The 3-bit R counter divides the reference input frequency by an
integer between 1 and 7. The divided-down signal is presented
as the reference clock to the phase frequency detector (PFD). The
divide ratio is set in R1_DB[4:6]. Maximizing the PFD frequency
reduces the N value. This reduces the noise multiplied at a rate of
20 log(N) to the output and reduces occurrences of spurious
components.
Register 1 defaults to R = 1 on power-up.
PFD [Hz] = XTAL/R
Using a TCXO Reference
Loop Filter
A single-ended reference (TCXO, VCXO, or OCXO) can also be
used with the ADF7021-N. This is recommended for applications
having absolute frequency accuracy requirements of <10 ppm, such
as applications requiring compliance with ARIB STD-T67 or
ETSI EN 300 220. The following are two options for interfacing
the ADF7021-N to an external reference oscillator.
The loop filter integrates the current pulses from the charge
pump to form a voltage that tunes the output of the VCO to the
desired frequency. It also attenuates spurious levels generated by
the PLL. A typical loop filter design is shown in Figure 34.
•
An oscillator with CMOS output levels can be applied to
OSC2. The internal oscillator circuit should be disabled by
setting R1_DB12 low.
An oscillator with 0.8 V p-p levels can be ac-coupled through
a 22 pF capacitor into OSC1. The internal oscillator circuit
should be enabled by setting R1_DB12 high.
Programmable Crystal Bias Current
Bias current in the oscillator circuit can be configured between 20
μA and 35 μA by writing to the XTAL_BIAS bits (R1_DB [13:14]).
Increasing the bias current allows the crystal oscillator to power
up faster.
CHARGE
PUMP OUT
VCO
07246-010
•
Figure 34. Typical Loop Filter Configuration
The loop should be designed so that the loop bandwidth (LBW)
is approximately 100 kHz. This provides a good compromise
between in-band phase noise and out-of-band spurious rejection.
Widening the LBW excessively reduces the time spent jumping
between frequencies, but it can cause insufficient spurious attenuation. Narrow-loop bandwidths can result in the loop taking long
periods to attain lock and can also result in a higher level of power
falling into the adjacent channel. The loop filter design on the
Rev. 0 | Page 22 of 64
ADF7021-N
EVAL-ADF7021-NDBxx should be used for optimum
performance.
The free design tool ADI SRD Design Studio™ can also
be used to design loop filters for the ADF7021-N (see the ADI
SRD Design Studio web site for details).
N Counter
The feedback divider in the ADF7021-N PLL consists of an
8-bit integer counter (R0_DB[19:26]) and a 15-bit, sigma-delta
(Σ-Δ) fractional_N divider (R0_DB[4:18]). The integer counter
is the standard pulse-swallow type that is common in PLLs. This
sets the minimum integer divide value to 23. The fractional divide
value provides very fine resolution at the output, where the output
frequency of the PLL is calculated as
Fractional _ N ⎞
XTAL ⎛
⎟⎟
× ⎜⎜ Integer _ N +
R
215
⎝
⎠
MUXOUT
The MUXOUT pin allows access to various digital points in the
ADF7021-N. The state of MUXOUT is controlled in Register 0
(R0_DB[29:31]).
REGULATOR_READY
REGULATOR_READY is the default setting on MUXOUT
after the transceiver is powered up. The power-up time of the
regulator is typically 50 μs. Because the serial interface is powered
from the regulator, the regulator must be at its nominal voltage
before the ADF7021-N can be programmed. The status of the
regulator can be monitored at MUXOUT. When the regulator
ready signal on MUXOUT is high, programming of the
ADF7021-N can begin.
When RF_DIVIDE_BY_2 (see the Voltage Controlled
Oscillator (VCO) section) is selected, this formula becomes
f OUT =
DVDD
Fractional _ N ⎞
XTAL
⎛
× 0.5 × ⎜ Integer_N +
⎟
R
2 15
⎠
⎝
REGULATOR_READY (DEFAULT)
FILTER_CAL_COMPLETE
DIGITAL_LOCK_DETECT
The combination of Integer_N (maximum = 255) and
Fractional_N (maximum = 32,768/32,768) gives a maximum
N divider of 255 + 1. Therefore, the minimum usable PFD is
PFD MIN [Hz ] =
RSSI_READY
Tx_Rx
MUX
MUXOUT
CONTROL
LOGIC_ZERO
TRISTATE
Maximum Required Output Frequency
LOGIC_ONE
(255 + 1)
For example, when operating in the European 868 MHz to
870 MHz band, PFDMIN = 3.4 MHz.
DGND
07246-009
f OUT =
voltage must be stabilized. Regulator status (CREG4) can be
monitored using the REGULATOR_READY signal from the
MUXOUT pin.
Figure 36. MUXOUT Circuit
REFERENCE IN
4\R
FILTER_CAL_COMPLETE
PFD/
CHARGE
PUMP
VCO
MUXOUT can be set to FILTER_CAL_COMPLETE. This signal
goes low for the duration of both a coarse IF filter calibration
and a fine IF filter calibration. It can be used as an interrupt to
a microcontroller to signal the end of the IF filter calibration.
4\N
THIRD-ORDER
Σ-Δ MODULATOR
INTEGER_N
07246-011
FRACTIONAL_N
DIGITAL_LOCK_DETECT
Figure 35. Fractional_N PLL
Voltage Regulators
The ADF7021-N contains four regulators to supply stable
voltages to the part. The nominal regulator voltage is 2.3 V.
Regulator 1 requires a 3.9 Ω resistor and a 100 nF capacitor in
series between CREG1 and GND, whereas the other regulators
require a 100 nF capacitor connected between CREGx and GND.
When CE is high, the regulators and other associated circuitry
are powered on, drawing a total supply current of 2 mA. Bringing
the CE pin low disables the regulators, reduces the supply current
to less than 1 μA, and erases all values held in the registers.
The serial interface operates from a regulator supply. Therefore,
to write to the part, the user must have CE high and the regulator
DIGITAL_LOCK_DETECT indicates when the PLL has locked.
The lock detect circuit is located at the PFD. When the phase
error on five consecutive cycles is less than 15 ns, lock detect is
set high. Lock detect remains high until a 25 ns phase error is
detected at the PFD.
RSSI_READY
MUXOUT can be set to RSSI_READY. This indicates that the
internal analog RSSI has settled and a digital RSSI readback can
be performed.
Tx_Rx
Tx_Rx signifies whether the ADF7021-N is in transmit or receive
mode. When in transmit mode, this signal is low. When in receive
mode, this signal is high. It can be used to control an external
Tx/Rx switch.
Rev. 0 | Page 23 of 64
ADF7021-N
To minimize spurious emissions, both VCOs operate at twice
the RF frequency. The VCO signal is then divided by 2 inside
the synthesizer loop, giving the required frequency for the
transmitter and the required local oscillator (LO) frequency for
the receiver. A further divide-by-2 (RF_DIVIDE_BY_2) is
performed outside the synthesizer loop to allow operation in
the 421 MHz to 458 MHz band (internal inductor VCO) and
the 80 MHz to 325 MHz band (external inductor VCO).
The VCO needs an external 22 nF capacitor between the CVCO
pin and the regulator (CREG1 pin) to reduce internal noise.
VCO_BIAS
R1_DB(19:22)
LOOP FILTER
VCO
MUX
÷2
TO PA
÷2
TO
N DIVIDER
RF_DIVIDE_BY_2
R1_DB18
07246-012
220µF
CVCO PIN
Figure 37. Voltage Controlled Oscillator (VCO)
Internal Inductor VCO
To select the internal inductor VCO, set R1_DB25 to Logic 0,
which is the default setting.
VCO bias current can be adjusted using R1_DB[19:22]. To
ensure VCO oscillation, the minimum bias current setting under
all conditions when using the internal inductor VCO is 0x8.
The VCO should be recentered, depending on the required
frequency of operation, by programming the VCO_ADJUST
bits (R1_DB[23:24]). This is detailed in Table 9.
External Inductor VCO
When using the external inductor VCO, the center frequency of the
VCO is set by the internal varactor capacitance and the combined
inductance of the external chip inductor, bond wire, and PCB track.
The external inductor is connected between the L2 and L1 pins.
750
700
650
fMAX
600
(MHz)
550
500
450
400
350
fMIN (MHz)
300
250
200
0
5
10
15
20
25
TOTAL EXTERNAL INDUCTANCE (nH)
30
07246-061
The ADF7021-N contains two VCO cores. The first VCO, the
internal inductor VCO, uses an internal LC tank and supports
842 MHz to 916 MHz and 421 MHz to 458 MHz operating
bands. The second VCO, the external inductor VCO, uses an
external inductor as part of its LC tank and supports the RF
operating band of 80 MHz to 650 MHz.
A plot of the VCO operating frequency vs. total external
inductance (chip inductor + PCB track) is shown in Figure 38.
FREQUENCY (MHz)
VOLTAGE CONTROLLED OSCILLATOR (VCO)
Figure 38. Direct RF Output vs. Total External Inductance
The inductance for a PCB track using FR4 material is approximately 0.57 nH/mm. This should be subtracted from the total
value to determine the correct chip inductor value.
Typically, a particular inductor value allows the ADF7021-N to
function over a range of ±6% of the RF operating frequency.
When the RF_DIVIDE_BY_2 bit (R1_DB18) is selected, this
range becomes ±3%. At 400 MHz, for example, an operating
range of ±24 MHz (that is, 376 MHz to 424 MHz) with a single
inductor (VCO range centered at 400 MHz) can be expected.
The VCO tuning voltage can be checked for a particular RF
output frequency by measuring the voltage on the VCOIN pin
when the part is fully powered up in transmit or receive mode.
The VCO tuning range is 0.2 V to 2 V. The external inductor
value should be chosen to ensure that the VCO is operating
as close as possible to the center of this tuning range. This is
particularly important for RF frequencies <200 MHz, where
the VCO gain is reduced and a tuning range of <±6 MHz exists.
The VCO operating frequency range can be adjusted by
programming the VCO_ADJUST bits (R1_DB[23:24]). This
typically allows the VCO operating range to be shifted up or
down by a maximum of 1% of the RF frequency.
To select the external inductor VCO, set R1_DB25 to Logic 1.
The VCO_BIAS should be set depending on the frequency of
operation (as indicated in Table 9).
Rev. 0 | Page 24 of 64
ADF7021-N
Table 9. RF Output Frequency Ranges for Internal and External Inductor VCOs and Required Register Settings
RF Frequency
Output (MHz)
870 to 916
842 to 870
440 to 458
421 to 440
450 to 650
200 to 450
80 to 200
VCO to
Be Used
Internal L
Internal L
Internal L
Internal L
External L
External L
External L
RF Divide
by 2
No
No
Yes
Yes
No
No
Yes
VCO_INDUCTOR
R1_DB25
0
0
0
0
1
1
1
CHOOSING CHANNELS FOR BEST SYSTEM
PERFORMANCE
An interaction between the RF VCO frequency and the
reference frequency can lead to fractional spur creation. When
the synthesizer is in fractional mode (that is, the RF VCO and
reference frequencies are not integer related), spurs can appear
on the VCO output spectrum at an offset frequency that
corresponds to the difference frequency between an integer
multiple of the reference and the VCO frequency.
Register Settings
RF_DIVIDE_BY_2
VCO_ADJUST
R1_DB18
R1_DB[23:24]
0
11
0
00
1
11
1
00
0
XX
0
XX
1
XX
VCO_BIAS
R1_DB[19:22]
8
8
8
8
4
3
2
These spurs are attenuated by the loop filter. They are more
noticeable on channels close to integer multiples of the reference
where the difference frequency may be inside the loop bandwidth;
thus, the name integer boundary spurs. The occurrence of these
spurs is rare because the integer frequencies are around multiples
of the reference, which is typically >10 MHz. To avoid having
very small or very large values in the fractional register, choose
a suitable reference frequency.
Rev. 0 | Page 25 of 64
ADF7021-N
TRANSMITTER
1
RF OUTPUT STAGE
2
3
4
...
8
...
16
DATA BITS
The power amplifier (PA) of the ADF7021-N is based on a
single-ended, controlled current, open-drain amplifier that has
been designed to deliver up to 13 dBm into a 50 Ω load at a
maximum frequency of 950 MHz.
The PA output current and consequently, the output power, are
programmable over a wide range. The PA configuration is shown
in Figure 39. The output power is set using R2_DB[13:18].
PA RAMP 0
(NO RAMP)
PA RAMP 1
(256 CODES PER BIT)
PA RAMP 2
(128 CODES PER BIT)
PA RAMP 3
(64 CODES PER BIT)
PA RAMP 4
(32 CODES PER BIT)
R2_DB(11:12)
PA RAMP 5
(16 CODES PER BIT)
2
PA RAMP 6
(8 CODES PER BIT)
R2_DB(13:18)
07246-014
PA RAMP 7
(4 CODES PER BIT)
RFOUT
Figure 40. PA Ramping Settings
R2_DB7
+
PA Bias Currents
R0_DB27
07246-013
RFGND
FROM VCO
Figure 39. PA Configuration
The PA is equipped with overvoltage protection, which makes it
robust in severe mismatch conditions. Depending on the application, users can design a matching network for the PA to exhibit
optimum efficiency at the desired radiated output power level
for a wide range of antennas, such as loop or monopole antennas.
See the LNA/PA Matching section for more information.
PA Ramping
The PA_BIAS bits (R2_DB[11:12]) facilitate an adjustment of
the PA bias current to further extend the output power control
range, if necessary. If this feature is not required, the default
value of 9 μA is recommended. If output power of greater than
10 dBm is required, a PA bias setting of 11 μA is recommended.
The output stage is powered down by resetting R2_DB7.
MODULATION SCHEMES
The ADF7021-N supports 2FSK, 3FSK, and 4FSK modulation.
The implementation of these modulation schemes is shown in
Figure 41.
When the PA is switched on or off quickly, its changing input
impedance momentarily disturbs the VCO output frequency.
This process is called VCO pulling, and it manifests as spectral
splatter or spurs in the output spectrum around the desired carrier
frequency. Some radio emissions regulations place limits on
these PA transient-induced spurs (for example, the ETSI EN 300 220
regulations). By gradually ramping the PA on and off, PA transient
spurs are minimized.
The ADF7021-N has built-in PA ramping configurability. As
Figure 40 illustrates, there are eight ramp rate settings, defined
as a certain number of PA setting codes per one data bit period.
The PA steps through each of its 64 code levels but at different
speeds for each setting. The ramp rate is set by configuring
R2_DB[8:10].
PFD/
CHARGE
PUMP
REF
TO
PA STAGE
LOOP FILTER
÷2
VCO
÷N
FRACTIONAL_N
THIRD-ORDER
Σ-Δ MODULATOR
INTEGER_N
Tx_FREQUENCY_
DEVIATION
2FSK
GAUSSIAN
OR
RAISED COSINE
FILTERING
If the PA is enabled/disabled by the PA_ENABLE bit (R2_DB7),
it ramps up and down. If it is enabled/disabled by the Tx/Rx bit
(R0_DB27), it ramps up and turns hard off.
Rev. 0 | Page 26 of 64
TxDATA
MUX
3FSK
4FSK
1 – D2 PR
SHAPING
PRECODER
4FSK
BIT SYMBOL
MAPPER
Figure 41. Transmit Modulation Implementation
07246-015
6
IDAC
ADF7021-N
Setting the Transmit Data Rate
3-Level Frequency Shift Keying (3FSK)
In all modulation modes except oversampled 2FSK mode, an
accurate clock is provided on the TxRxCLK pin to latch the data
from the microcontroller into the transmit section at the required
data rate. The exact frequency of this clock is defined by
In 3-level FSK modulation (also known as modified duobinary
FSK), the binary data (Logic 0 and Logic 1) is mapped onto
three distinct frequencies: the carrier frequency (fC), the carrier
frequency minus a deviation frequency (fC − fDEV), and the
carrier frequency plus the deviation frequency (fC + fDEV).
XTAL
DEMOD _ CLK _ DIVIDE × CDR _ CLK _ DIVIDE × 32
where:
XTAL is the crystal or TCXO frequency.
DEMOD_CLK_DIVIDE is the divider that sets the demodulator
clock rate (R3_DB[6:9]).
CDR_CLK_DIVIDE is the divider that sets the CDR clock rate
(R3_DB[10:17]).
A Logic 0 is mapped to the carrier frequency while a Logic 1 is
either mapped onto the fC − fDEV frequency or the fC + fDEV
frequency.
0
fC – fDEV
Refer to the Register 3—Transmit/Receive Clock Register
section for more programming information.
fC
fC + fDEV
RF FREQUENCY
Figure 42. 3FSK Symbol-to-Frequency Mapping
Setting the FSK Transmit Deviation Frequency
In all modulation modes, the deviation from the center
frequency is set using the Tx_FREQUENCY_DEVIATION bits
(R2_DB[19:27]).
The deviation from the center frequency in Hz is as follows:
For direct RF output,
Compared to 2FSK, this bits-to-frequency mapping results in a
reduced transmission bandwidth because some energy is removed
from the RF sidebands and transferred to the carrier frequency.
At low modulation index, 3FSK improves the transmit spectral
efficiency by up to 25% when compared to 2FSK.
Bit-to-symbol mapping for 3FSK is implemented using a linear
convolutional encoder that also permits Viterbi detection to be
used in the receiver. A block diagram of the transmit hardware
used to realize this system is shown in Figure 43. The convolutional encoder polynomial used to implement the transmit
spectral shaping is
For RF_DIVIDE_BY_2 enabled,
PFD × Tx _ FREQUENCY_ DEVIATION
216
where Tx_FREQUENCY_DEVIATION is a number from 1 to
511 (R2_DB[19:27]).
In 4FSK modulation, the four symbols (00, 01, 11, 10) are
transmitted as ±3 × fDEV and ±1 × fDEV.
Binary Frequency Shift Keying (2FSK)
Two-level frequency shift keying is implemented by setting the
N value for the center frequency and then toggling it with the
TxDATA line. The deviation from the center frequency is set
using the Tx_FREQUENCY_DEVIATION bits, R2_DB[19:27].
P(D) = 1 − D2
where:
P is the convolutional encoder polynomial.
D is the unit delay operator.
A digital precoder with transfer function 1/P(D) implements an
inverse modulo-2 operation of the 1 − D2 shaping filter in the
transmitter.
Tx DATA
0, 1
2FSK is selected by setting the MODULATION_SCHEME bits
(R2_DB[4:6]) to 000.
Minimum shift keying (MSK) or Gaussian minimum shift
keying (GMSK) is supported by selecting 2FSK modulation and
using a modulation index of 0.5. A modulation index of 0.5 is
set up by configuring R2_DB[19:27] for an fDEV = 0.25 ×
transmit data rate.
Rev. 0 | Page 27 of 64
PRECODER
1/P(D)
0, 1
CONVOLUTIONAL
ENCODER
P(D)
0, +1, –1
fC
FSK MOD
fC + fDEV
fC – fDEV
CONTROL
AND
DATA FILTERING
Figure 43. 3FSK Encoding
TO
N DIVIDER
07246-046
PFD × Tx _ FREQUENCY _ DEVIATION
f DEV [Hz] =
216
f DEV [Hz] = 0.5 ×
+1
–1
07246-057
DATA CLK =
ADF7021-N
The signal mapping of the input binary transmit data to the
3-level convolutional output is shown in Table 10. The
convolutional encoder restricts the maximum number of
sequential +1s or −1s to two and delivers an equal number of
+1s and −1s to the FSK modulator, thus ensuring equal spectral
energy in both RF sidebands.
The transmit clock from Pin TxRxCLK is available after writing
to Register 3 in the power-up sequence for receive mode. The
MSB of the first symbol should be clocked into the ADF7021-N
on the first transmit clock pulse from the ADF7021-N after
writing to Register 3. Refer to Figure 6 for more timing
information.
Table 10. 3-Level Signal Mapping of the Convolutional Encoder
Oversampled 2FSK
1
0
−1
In oversampled 2FSK, there is no data clock from the TxRxCLK
pin. Instead, the transmit data at the TxRxDATA pin is sampled
at 32 times the programmed rate.
Another property of this encoding scheme is that the transmitted
symbol sequence is dc-free, which facilitates symbol detection
and frequency measurement in the receiver. In addition, there is
no code rate loss associated with this 3-level convolutional encoder;
that is, the transmitted symbol rate is equal to the data rate
presented at the transmit data input.
This is the only modulation mode that can be used with the UART
mode interface for data transmission (refer to the Interfacing to
a Microcontroller/DSP section for more information).
TxDATA
Precoder Output
Encoder Output
1
0
1
0
+1 0
1
0
−1
1
1
+1
0
0
0
0
1
0
1
1
+1
0
1
0
0
1
0
3FSK is selected by setting the MODULATION_SCHEME bits
(R2_DB[4:6]) to 010. It can also be used with raised cosine
filtering to further increase the spectral efficiency of the transmit
signal.
4-Level Frequency Shift Keying (4FSK)
In 4FSK modulation, two bits per symbol spectral efficiency is
realized by mapping consecutive input bit-pairs in the Tx data
bit stream to one of four possible symbols (−3, −1, +1, +3). Thus,
the transmitted symbol rate is half of the input bit rate.
0
0
0
1
1
0
1
Gaussian frequency shift keying reduces the bandwidth occupied
by the transmitted spectrum by digitally prefiltering the transmit
data. The BT product of the Gaussian filter used is 0.5.
Gaussian filtering can only be used with 2FSK modulation. This
is selected by setting R2_DB[4:6] to 001.
1
Raised Cosine Filtering
f
+3fDEV
SYMBOL
FREQUENCIES
Gaussian or raised cosine filtering can be used to improve
transmit spectral efficiency. The ADF7021-N supports Gaussian
filtering (bandwidth time [BT] = 0.5) on 2FSK modulation.
Raised cosine filtering can be used with 2FSK, 3FSK, or 4FSK
modulation. The roll-off factor (alpha) of the raised cosine filter
has programmable options of 0.5 and 0.7. Both the Gaussian
and raised cosine filters are implemented using linear phase
digital filter architectures that deliver precise control over the
BT and alpha filter parameters, and guarantee a transmit spectrum
that is very stable over temperature and supply variation.
Gaussian Frequency Shift Keying (GFSK)
By minimizing the separation between symbol frequencies,
4FSK can have high spectral efficiency. The bit-to-symbol
mapping for 4FSK is gray coded and is shown in Figure 44.
Tx DATA
SPECTRAL SHAPING
+fDEV
–3fDEV
t
Figure 44. 4FSK Bit-to-Symbol Mapping
07246-016
–fDEV
Raised cosine filtering provides digital prefiltering of the transmit
data by using a raised cosine filter with a roll-off factor (alpha)
of either 0.5 or 0.7. The alpha is set to 0.5 by default, but the
raised cosine filter bandwidth can be increased to provide less
aggressive data filtering by using an alpha of 0.7 (set R2_DB30
to Logic 1). Raised cosine filtering can be used with 2FSK,
3FSK, and 4FSK.
Raised cosine filtering is enabled by setting R2_DB[4:6] as
outlined in Table 11.
The inner deviation frequencies (+fDEV and −fDEV) are set using
the Tx_FREQUENCY_DEVIATION bits, R2_DB[19:27]. The
outer deviation frequencies are automatically set to three times
the inner deviation frequency.
Rev. 0 | Page 28 of 64
ADF7021-N
MODULATION AND FILTERING OPTIONS
The various modulation and data filtering options are described
in Table 11.
Table 11. Modulation and Filtering Options
Modulation
BINARY FSK
2FSK
MSK1
OQPSK with Half Sine
Baseband Shaping2
GFSK
GMSK3
RC2FSK
Oversampled 2FSK
3-LEVEL FSK
3FSK
RC3FSK
4-LEVEL FSK
4FSK
RC4FSK
Data Filtering
R2_DB[4:6]
None
None
None
000
000
000
Gaussian
Gaussian
Raised cosine
None
001
001
101
100
None
Raised cosine
010
110
None
Raised cosine
011
111
Table 12. Bit/Symbol Latency in Transmit Mode for Various
Modulation Schemes
Modulation
2FSK
GFSK
RC2FSK, Alpha = 0.5
RC2FSK, Alpha = 0.7
3FSK
RC3FSK, Alpha = 0.5
RC3FSK, Alpha = 0.7
4FSK
RC4FSK, Alpha = 0.5
RC4FSK, Alpha = 0.7
Latency
1 bit
4 bits
5 bits
4 bits
1 bit
5 bits
4 bits
1 symbol
5 symbols
4 symbols
TEST PATTERN GENERATOR
The ADF7021-N has a number of built-in test pattern generators
that can be used to facilitate radio link setup or RF measurement.
A full list of the supported patterns is shown in Table 13. The
data rate for these test patterns is the programmed data rate set
in Register 3.
1
MSK is 2FSK modulation with a modulation index = 0.5.
Offset quadrature phase shift keying (OQPSK) with half sine baseband shaping
is spectrally equivalent to MSK.
3
GMSK is GFSK with a modulation index = 0.5.
2
The PN9 sequence is suitable for test modulation when carrying
out adjacent channel power (ACP) or occupied bandwidth
measurements.
Table 13. Transmit Test Pattern Generator Options
TRANSMIT LATENCY
Transmit latency is the delay time from the sampling of a
bit/symbol by the TxRxCLK signal to when that bit/symbol
appears at the RF output. The latency without any data filtering
is one bit. The addition of data filtering adds a further latency as
outlined in Table 12.
It is important that the ADF7021-N be left in transmit mode
after the last data bit is sampled by the data clock to account for
this latency. The ADF7021-N should stay in transmit mode for
a time equal to the number of latency bit periods for the applied
modulation scheme. This ensures that all of the data sampled by
the TxRxCLK signal appears at RF.
Test Pattern
Normal
Transmit Carrier
Transmit + fDEV Tone
Transmit − fDEV Tone
Transmit 1010 Pattern
Transmit PN9 Sequence
Transmit SWD Pattern Repeatedly
The figures for latency in Table 12 assume that the positive
TxRxCLK edge is used to sample data (default). If the TxRxCLK
is inverted by setting R2_DB[28:29], an additional 0.5 bit
latency can be added to all values in Table 12.
Rev. 0 | Page 29 of 64
R15_DB[8:10]
000
001
010
011
100
101
110
Error! Unknown document
property name.
ADF7021-N
RECEIVER SECTION
chosen as a compromise between interference rejection and
attenuation of the desired signal.
RF FRONT END
The ADF7021-N is based on a fully integrated, low IF receiver
architecture. The low IF architecture facilitates a very low
external component count and does not suffer from powerlineinduced interference problems.
Figure 45 shows the structure of the receiver front end. The
many programming options allow users to trade off sensitivity,
linearity, and current consumption to best suit their application.
To achieve a high level of resilience against spurious reception,
the low noise amplifier (LNA) features a differential input.
Switch SW2 shorts the LNA input when transmit mode is
selected (R0_DB27 = 0). This feature facilitates the design of a
combined LNA/PA matching network, avoiding the need for an
external Rx/Tx switch. See the LNA/PA Matching section for
details on the design of the matching network.
I (TO FILTER)
SW2
LNA
LNA_MODE
(R9_DB25)
LNA_BIAS
(R9_DB[26:27])
LO
IF Filter Bandwidth and Center Frequency Calibration
To compensate for manufacturing tolerances, the IF filter should be
calibrated after power-up to ensure that the bandwidth and
center frequency are correct. Coarse and fine calibration
schemes are provided to offer a choice between fast calibration
(coarse calibration) and high filter centering accuracy (fine
calibration). Coarse calibration is enabled by setting R5_DB4
high. Fine calibration is enabled by setting R6_DB4 high.
For details on when it is necessary to perform a filter
calibration, and in what applications to use either a coarse
calibration or fine calibration, refer to the IF Filter Bandwidth
Calibration section.
Q (TO FILTER)
RSSI/AGC
MIXER LINEARITY
(R9_DB28)
The RSSI is implemented as a successive compression log amp
following the baseband (BB) channel filtering. The log amp
achieves ±3 dB log linearity. It also doubles as a limiter to
convert the signal-to-digital levels for the FSK demodulator.
The offset correction circuit uses the BBOS_CLK_DIVIDE bits
(R3_DB[4:5]), which should be set between 1 MHz and 2 MHz.
The RSSI level is converted for user readback and for digitally
controlled AGC by an 80-level (7-bit) flash ADC. This level can
be converted to input power in dBm. By default, the AGC is on
when powered up in receive mode.
07246-017
LNA_GAIN
(R9_DB[20:21])
LNA/MIXER_ENABLE
(R8_DB6)
Figure 45. RF Front End
The LNA is followed by a quadrature downconversion mixer,
which converts the RF signal to the IF frequency of 100 kHz.
An important consideration is that the output frequency of the
synthesizer must be programmed to a value 100 kHz below the
center frequency of the received channel. The LNA has two
basic operating modes: high gain/low noise mode and low
gain/low power mode. To switch between these two modes, use
the LNA_MODE bit (R9_DB25). The mixer is also configurable
between a low current and an enhanced linearity mode using
the MIXER_LINEARITY bit (R9_DB28).
OFFSET
CORRECTION
1
A
IFWR
A
IFWR
A
IFWR
LATCH
IFWR
FSK
DEMOD
CLK
ADC
Based on the specific sensitivity and linearity requirements of
the application, it is recommended to adjust the LNA_MODE
bit and MIXER_LINEARITY bit as outlined in Table 15.
R
RSSI
07246-018
RFIN
Tx/Rx SELECT
(R0_DB27)
RFINB
If the AGC loop is disabled, the gain of the IF filter can be set to one
of three levels by using the FILTER_GAIN bits (R9_DB[22:23]).
The filter gain is adjusted automatically if the AGC loop is
enabled.
Figure 46. RSSI Block Diagram
RSSI Thresholds
The gain of the LNA is configured by the LNA_GAIN bits
(R9_DB[20:21]) and can be set by either the user or the
automatic gain control (AGC) logic.
IF FILTER
IF Filter Settings
Out-of-band interference is rejected by means of a fifth-order
Butterworth polyphase IF filter centered on a frequency of
100 kHz. The bandwidth of the IF filter can be programmed to
9 kHz, 13.5 kHz, or 18.5 kHz by R4_DB[30:31] and should be
When the RSSI is above AGC_HIGH_THRESHOLD
(R9_DB[11:17]), the gain is reduced. When the RSSI is
below AGC_LOW_THRESHOLD (R9_DB[4:10]), the gain
is increased. The thresholds default to 30 and 70 on power-up
in receive mode. A delay (set by AGC_CLK_DIVIDE,
R3_DB[26:31]) is programmed to allow for settling of the loop.
A value of 13 is recommended to give an AGC update rate of
7.7 kHz.
Rev. 0 | Page 30 of 64
ADF7021-N
The user has the option of changing the two threshold values
from the defaults of 30 and 70 (Register 9). The default AGC
setup values should be adequate for most applications. The
threshold values must be more than 30 apart for the AGC to
operate correctly.
By using the recommended setting for AGC_CLK_DIVIDE, the
total AGC settling time is
Offset Correction Clock
The worst case for AGC settling occurs when the AGC control
loop has to cycle through all five gain settings, which gives a
maximum AGC settling time of 650 μs.
In Register 3, the user should set the BBOS_CLK_DIVIDE bits
(R3_DB[4:5]) to give a baseband offset clock (BBOS CLK)
frequency between 1 MHz and 2 MHz.
AGC Settling Time [sec] =
AGC Update Rate [Hz]
RSSI Formula (Converting to dBm)
The RSSI formula is
BBOS CLK [Hz] = XTAL/(BBOS_CLK_DIVIDE)
Input Power [dBm] = −130 dBm + (Readback Code + Gain
Mode Correction) × 0.5
where BBOS_CLK_DIVIDE can be set to 4, 8, 16, or 32.
AGC Information and Timing
AGC is selected by default and operates by setting the appropriate
LNA and filter gain settings for the measured RSSI level. It is
possible to disable AGC by writing to Register 9 if the user wants to
enter one of the modes listed in Table 15. The time for the AGC
circuit to settle and, therefore, the time it takes to measure the RSSI
accurately, is typically 390 μs. However, this depends on how many
gain settings the AGC circuit has to cycle through. After each gain
change, the AGC loop waits for a programmed time to allow
transients to settle. This AGC update rate is set according to
AGC Update Rate [Hz] =
Number of AGC Gain Changes
SEQ _ CLK _ DIVIDE [Hz]
AGC _ CLK _ DIVIDE
where:
AGC_CLK_DIVIDE is set by R3_DB[26:31]. A value of 13 is
recommended.
SEQ_CLK_DIVIDE = 100 kHz (R3_DB[18:25]).
where:
Readback Code is given by Bit RV7 to Bit RV1 in the Register 7
readback register (see Figure 58 and the Readback Format
section).
Gain Mode Correction is given by the values in Table 14.
The LNA gain (LG2, LG1) and filter gain (FG2, FG1) values
are also obtained from the readback register, as part of an RSSI
readback.
Table 14. Gain Mode Correction
LNA Gain
(LG2, LG1)
H (1, 0)
M (0, 1)
M (0, 1)
M (0, 1)
L (0, 0)
Filter Gain
(FG2, FG1)
H (1, 0)
H (1, 0)
M (0, 1)
L (0, 0)
L (0, 0)
Gain Mode
Correction
0
24
38
58
86
An additional factor should be introduced to account for losses
in the front-end-matching network/antenna.
Table 15. LNA/Mixer Modes
Receiver Mode
High Sensitivity
Mode (Default)
Enhanced Linearity
High Gain
Medium Gain
Enhanced Linearity
Medium Gain
Low Gain
Enhanced Linearity
Low Gain
LNA_MODE
(R9_DB25)
0
LNA_GAIN
(R9_DB[20:21])
30
MIXER_LINEARITY
(R9_DB28)
0
Sensitivity (2FSK, DR =
4.8 kbps, fDEV = 4 kHz)
−118
Rx Current
Consumption (mA)
24.6
Input IP3
(dBm)
−24
0
30
1
−114.5
24.6
−20
1
1
10
10
0
1
−112
−105.5
22.1
22.1
−13.5
−9
1
1
3
3
0
1
−100
−92.3
22.1
22.1
−5
−3
Rev. 0 | Page 31 of 64
ADF7021-N
DEMODULATION, DETECTION, AND CDR
Correlator Demodulator
System Overview
The correlator demodulator can be used for 2FSK, 3FSK, and
4FSK demodulation. Figure 48 shows the operation of the
correlator demodulator for 2FSK.
The quadrature outputs of the IF filter are first limited and
then fed to either the correlator FSK demodulator or to the
linear FSK demodulator. The correlator demodulator is used
to demodulate 2FSK, 3FSK, and 4FSK. The linear demodulator
is used for frequency measurement and is enabled when the
AFC loop is active. The linear demodulator can also be used
to demodulate 2FSK.
Following the demodulator, a digital post demodulator filter
removes excess noise from the demodulator signal output.
Threshold/slicer detection is used for data recovery of 2FSK and
4FSK. Data recovery of 3FSK can be implemented using either
threshold detection or Viterbi detection.
An on-chip CDR PLL is used to resynchronize the received bit
stream to a local clock. It outputs the retimed data and clock on
the TxRxDATA and TxRxCLK pins, respectively.
FREQUENCY
CORRELATOR
MUX
LINEAR
DEMODULATOR
CLOCK
AND
DATA
RECOVERY
Q
IF – fDEV
IF
IF + fDEV
R4_DB9
Rx_INVERT
R4_DB(10:19)
DISCRIMINATOR_BW
R4_DB7
DOT_PRODUCT
Figure 48. 2FSK Correlator Demodulator Operation
The quadrature outputs of the IF filter are first limited and then
fed to a digital frequency correlator that performs filtering and
frequency discrimination of the 2FSK/3FSK/4FSK spectrum.
For 2FSK modulation, data is recovered by comparing the
output levels from two correlators. The performance of this
frequency discriminator approximates that of a matched filter
detector, which is known to provide optimum detection in the
presence of additive white Gaussian noise (AWGN). This
method of FSK demodulation provides approximately 3 dB to
4 dB better sensitivity than a linear demodulator.
MUX
VITERBI
DETECTION
3FSK
OUTPUT LEVELS:
2FSK = +1, –1
3FSK = +1, 0, –1
4FSK = +3, +1, –1, –3
LIMITERS
THRESHOLD
DETECTION
2/3/4FSK
TxRxDATA
TxRxCLK
I
07246-080
Q
POST
DEMOD FILTER
LIMITERS
I
FREQUENCY CORRELATOR
DISCRIM BW
07246-079
An overview of the demodulation, detection, and clock and
data recovery (CDR) of the received signal on the ADF7021-N
is shown in Figure 47.
Figure 47. Overview of Demodulation, Detection, and CDR Process
Rev. 0 | Page 32 of 64
ADF7021-N
Figure 49 shows a block diagram of the linear demodulator.
4FSK demodulation is implemented using the correlator
demodulator followed by the post demodulator filter and
threshold detection. The output of the post demodulation
filter is a 4-level signal that represents the transmitted symbols
(−3, −1, +1, +3). Threshold detection of 4FSK requires three
threshold settings, one that is always fixed at 0 and two that
are programmable and are symmetrically placed above and
below zero using the 3FSK/4FSK_SLICER_THRESHOLD bits
(R13_DB[4:10]).
LEVEL
IF
LIMITER
Q
FREQUENCY
LINEAR
DISCRIMINATOR
R4_DB(20:29)
+
SLICER
2FSK
2FSK RxDATA
RxCLK
FREQUENCY
READBACK
AND AFC LOOP
07246-073
I
ENVELOPE
DETECTOR
3FSK and 4FSK Threshold Detection
POST_DEMOD_
FILTER
Linear Demodulator
Figure 49. Block Diagram of Linear FSK Demodulator
A digital frequency discriminator provides an output signal that
is linearly proportional to the frequency of the limiter outputs.
The discriminator output is filtered and averaged using a combined
averaging filter and envelope detector. The demodulated 2FSK
data from the post demodulator filter is recovered by slicing against
the output of the envelope detector, as shown in Figure 49. This
method of demodulation corrects for frequency errors between
transmitter and receiver when the received spectrum is close to
or within the IF bandwidth. This envelope detector output is
also used for AFC readback and provides the frequency estimate
for the AFC control loop.
Post Demodulator Filter
A second-order, digital low-pass filter removes excess noise from
the demodulated bit stream at the output of the discriminator.
The bandwidth of this post demodulator filter is programmable
and must be optimized for the user’s data rate and received
modulation type. If the bandwidth is set too narrow, performance
degrades due to intersymbol interference (ISI). If the bandwidth
is set too wide, excess noise degrades the performance of the
receiver. The POST_DEMOD_BW bits (R4_DB[20:29]) set the
bandwidth of this filter.
2FSK Bit Slicer/Threshold Detection
2FSK demodulation can be implemented using the correlator
FSK demodulator or the linear FSK demodulator. In both cases,
threshold detection is used for data recovery at the output of the
post demodulation filter.
The output signal levels of the correlator demodulator are
always centered about zero. Therefore, the slicer threshold level
can be fixed at zero, and the demodulator performance is
independent of the run-length constraints of the transmit data
bit stream. This results in robust data recovery that does not
suffer from the classic baseline wander problems that exist in
the more traditional FSK demodulators.
When the linear demodulator is used for 2FSK demodulation,
the output of the envelope detector is used as the slicer threshold,
and this output tracks frequency errors that are within the IF
filter bandwidth.
3FSK demodulation is implemented using the correlator demodulator, followed by a post demodulator filter. The output of the
post demodulator filter is a 3-level signal that represents the
transmitted symbols (−1, 0, +1). Data recovery of 3FSK can be
implemented using threshold detection or Viterbi detection.
Threshold detection is implemented using two thresholds that
are programmable and are symmetrically placed above and
below zero using the 3FSK/4FSK_SLICER_THRESHOLD bits
(R13_DB[4:10]).
3FSK Viterbi Detection
Viterbi detection of 3FSK operates on a four-state trellis and is
implemented using two interleaved Viterbi detectors operating
at half the symbol rate. The Viterbi detector is enabled by
R13_DB11.
To facilitate different run length constraints in the transmitted
bit stream, the Viterbi path memory length is programmable
in steps of 4 bits, 6 bits, 8 bits, or 32 bits by setting the
VITERBI_PATH_MEMORY bits (R13_DB[13:14]). This
should be set equal to or longer than the maximum number
of consecutive 0s in the interleaved transmit bit stream.
When used with Viterbi detection, the receiver sensitivity
for 3FSK is typically 3 dB greater than that obtained using
threshold detection. When the Viterbi detector is enabled,
however, the receiver bit latency is increased by twice the
Viterbi path memory length.
Clock Recovery
An oversampled digital clock and data recovery (CDR) PLL is
used to resynchronize the received bit stream to a local clock
in all modulation modes. The oversampled clock rate of the PLL
(CDR CLK) must be set at 32 times the symbol rate (see the
Register 3—Transmit/Receive Clock Register section). The
maximum data/symbol rate tolerance of the CDR PLL is
determined by the number of zero-crossing symbol transitions
in the transmitted packet. For example, if using 2FSK with a
101010 preamble, a maximum tolerance of ±3.0% of the data
rate is achieved. However, this tolerance is reduced during
recovery of the remainder of the packet where symbol transitions may not be guaranteed to occur at regular intervals.
To maximize the data rate tolerance of the CDR, some form
of encoding and/or data scrambling is recommended that
guarantees a number of transitions at regular intervals.
Rev. 0 | Page 33 of 64
ADF7021-N
For 3FSK,
For example, using 2FSK with Manchester-encoded data
achieves a data rate tolerance of ±2.0%.
The CDR PLL is designed for fast acquisition of the recovered
symbols during preamble and typically achieves bit synchronization within 5-symbol transitions of preamble.
In 4FSK modulation, the tolerance using the +3, −3, +3, −3
preamble is ±3% of the symbol rate (or ±1.5% of the data rate).
However, this tolerance is reduced during recovery of the
remainder of the packet where symbol transitions may not be
guaranteed to occur at regular intervals. To maximize the
symbol/data rate tolerance, the remainder of the 4FSK packet
should be constructed so that the transmitted symbols retain close
to dc-free properties by using data scrambling and/or by inserting
specific dc balancing symbols that are inserted in the transmitted
bit stream at regular intervals such as after every 8 or 16 symbols.
In 3FSK modulation, the linear convolutional encoder scheme
guarantees that the transmitted symbol sequence is dc-free,
facilitating symbol detection. However, Tx data scrambling is
recommended to limit the run length of zero symbols in the
transmit bit stream. Using 3FSK, the CDR data rate tolerance is
typically ±0.5%.
RECEIVER SETUP
Correlator Demodulator Setup
To enable the correlator for various modulation modes, refer to
Table 16.
(DEMOD CLK × K )
400 ×10 3
where:
DEMOD CLK is as defined in the Register 3—Transmit/Receive
Clock Register section.
K is set for each modulation mode according to the following:
⎛ 100 × 10 3
K = Round ⎜⎜
⎝ f DEV
⎞
⎟
⎟
⎠
⎛ 100 ×103 ⎞
⎟
K = Round4 FSK ⎜⎜
⎟
⎝ 4 × f DEV ⎠
where:
Round is rounded to the nearest integer.
Round4FSK is rounded to the nearest of the following integers: 32,
31, 28, 27, 24, 23, 20, 19, 16, 15, 12, 11, 8, 7, 4, 3.
fDEV is the transmit frequency deviation in Hz. For 4FSK, fDEV is
the frequency deviation used for the ±1 symbols (that is, the
inner frequency deviations).
To optimize the coefficients of the correlator, R4_DB7 and
R4_DB[8:9] must also be assigned. The value of these bits
depends on whether K is odd or even. These bits are assigned
according to Table 17 and Table 18.
Table 17. Assignment of Correlator K Value for 2FSK and 3FSK
K
Even
Even
Odd
Odd
K/2
Even
Odd
N/A
N/A
(K + 1)/2
N/A
N/A
Even
Odd
R4_DB7
0
0
1
1
R4_DB[8:9]
00
10
00
10
R4_DB7
0
1
R4_DB[8:9]
00
00
Linear Demodulator Setup
To optimize receiver sensitivity, the correlator bandwidth must be
optimized for the specific deviation frequency and modulation
used by the transmitter. The discriminator bandwidth is
controlled by R4_DB[10:19] and is defined as
For 2FSK,
For 4FSK,
K
Even
Odd
DEMOD_SCHEME (R4_DB[4:6])
001
010
011
DISCRIMINATOR _ BW =
⎞
⎟
⎟
⎠
Table 18. Assignment of Correlator K Value for 4FSK
Table 16. Enabling the Correlator Demodulator
Received Modulation
2FSK
3FSK
4FSK
⎛ 100 × 10 3
K = Round ⎜⎜
⎝ 2 × f DEV
The linear demodulator can be used for 2FSK demodulation. To
enable the linear demodulator, set the DEMOD_SCHEME bits
(R4_DB[4:6]) to 000.
Post Demodulator Filter Setup
The 3 dB bandwidth of the post demodulator filter should be
set according to the received modulation type and data rate.
The bandwidth is controlled by R4_DB[20:29] and is given by
POST _ DEMOD _ BW =
2 11 × π × f CUTOFF
DEMOD CLK
where fCUTOFF is the target 3 dB bandwidth in Hz of the post
demodulator filter.
Table 19. Post Demodulator Filter Bandwidth Settings for
2FSK/3FSK/4FSK Modulation Schemes
Received
Modulation
2FSK
3FSK
4FSK
Rev. 0 | Page 34 of 64
Post Demodulator Filter Bandwidth,
fCUTOFF (Hz)
0.75 × data rate
1 × data rate
1.6 × symbol rate (= 0.8 × data rate)
ADF7021-N
3FSK Viterbi Detector Setup
3FSK Threshold Detector Setup
The Viterbi detector can be used for 3FSK data detection. This
is activated by setting R13_DB11 to Logic 1.
To activate threshold detection of 3FSK, R13_DB11 should be
set to Logic 0. The 3FSK/4FSK_SLICER_THRESHOLD bits
(R13_DB[4:10]) should be set as outlined in the 3FSK Viterbi
Detector Setup section.
The Viterbi path memory length is programmable in steps of 4,
6, 8, or 32 bits (VITERBI_PATH_MEMORY, R13_DB[13:14]).
The path memory length should be set equal to or greater than
the maximum number of consecutive 0s in the interleaved
transmit bit stream.
The Viterbi detector also uses threshold levels to implement the
maximum likelihood detection algorithm. These thresholds are
programmable via the 3FSK/4FSK_SLICER_THRESHOLD bits
(R13_DB[4:10]).
These bits are assigned as follows:
3FSK CDR Setup
In 3FSK, a transmit preamble of at least 40 bits of continuous
1s is recommended to ensure a maximum number of symbol
transitions for the CDR to acquire lock.
The clock and data recovery for 3FSK requires a number of
parameters in Register 13 to be set (see Table 20).
4FSK Threshold Detector Setup
The threshold for the 4FSK detector is set using the
3FSK/4FSK_SLICER_THRESHOLD bits (R13_DB[4:10]).
The threshold should be set according to
3FSK/4FSK_SLICER_THRESHOLD =
⎛ Transmit Frequency Deviation × K ⎞
⎟⎟
57 × ⎜⎜
100 × 10 3
⎝
⎠
3FSK/4FSK_SLICER_THRESHOLD =
⎛ 4FSK Outer Tx Deviation × K ⎞
⎟⎟
78 × ⎜⎜
100 × 10 3
⎝
⎠
where K is the value calculated for correlator discriminator
bandwidth.
where K is the value calculated for correlator discriminator
bandwidth.
Table 20. 3FSK CDR Settings
Parameter
PHASE_CORRECTION (R13_DB12)
3FSK_CDR_THRESHOLD (R13_DB[15:21])
3FSK_PREAMBLE_TIME_VALIDATE (R13_DB [22:25])
Recommended Setting
1
⎛ Transmit Frequency Deviation × K ⎞
62 × ⎜
⎟
100 × 10 3
⎝
⎠
where K is the value calculated for correlator
discriminator bandwidth.
15
Rev. 0 | Page 35 of 64
Purpose
Phase correction is on
Sets CDR decision threshold levels
Preamble detector time qualifier
ADF7021-N
DEMODULATOR CONSIDERATIONS
2FSK Preamble
The linear demodulator (AFC disabled) tracks frequency errors
in the receive signal when the receive signal is within the IF
filter bandwidth. For example, for a receive signal with an
occupied bandwith = 9 kHz, using the 18.5 kHz IF filter
bandwidth allows the linear demodulator to track the signal at
an error of ±4.75 kHz with no increase in bit errors or loss in
sensitivity.
The recommended preamble bit pattern for 2FSK is a dc-free
pattern (such as a 10101010… pattern). Preamble patterns with
longer run-length constraints (such as 11001100…) can also be
used but result in a longer synchronization time of the received
bit stream in the receiver. The preamble needs to allow enough
bits for AGC settling of the receiver and CDR acquisition. A
minimum of 16 preamble bits is recommended when using the
correlator demodulator and 48 bits when using the linear demodulator. When the receiver uses the internal AFC, the minimum
recommended number of preamble bits is 64.
Correlator Demodulator and Low Modulation Indices
The remaining fields that follow the preamble header do not
have to use dc-free coding. For these fields, the ADF7021-N can
accommodate coding schemes with a run length of greater than
eight bits without any performance degradation. Refer to
Application Note AN-915 for more information.
The receiver sensitivity performance and receiver frequency
tolerance can be maximized at low modulation index by
increasing the discriminator bandwidth of the correlator
demodulator. For modulation indices of less than 0.4, it is
recommended to double the correlator bandwidth by
calculating K as follows:
4FSK Preamble and Data Coding
The recommended preamble bit pattern for 4FSK is a repeating
00100010… bit sequence. This 2-level sequence of repeating
−3, +3, −3, +3 symbols is dc-free and maximizes the symbol
timing performance and data recovery of the 4FSK preamble in
the receiver. The minimum recommended length of the
preamble is 32 bits (16 symbols).
The remainder of the 4FSK packet should be constructed so
that the transmitted symbols retain close to a dc-free balance by
using data scrambling and/or by inserting specific dc balancing
symbols in the transmitted bit stream at regular intervals, such
as after every 8 or 16 symbols.
Demodulator Tolerance to Frequency Errors
Without AFC
The ADF7021-N has a number of options to combat frequency
errors that exist due to mismatches between the transmit and
receive crystals/TCXOs.
With AFC disabled, the correlator demodulator is tolerant to
frequency errors over a ±0.3 × fDEV range, where fDEV is the FSK
frequency deviation. For larger frequency errors, the frequency
tolerance can be increased by adjusting the value of K and thus
doubling the correlator bandwidth.
K should then be calculated as
⎛ 100 × 10 3
K = Round ⎜⎜
⎝ 2 × f DEV
⎞
⎟
⎟
⎠
The DISCRIMINATOR_BW setting in Register 4 should also be
recalculated using the new K value. Doubling the correlator
bandwidth to improve frequency error tolerance in this manner
typically results in a 1 dB to 2 dB loss in receiver sensitivity.
The modulation index in 2FSK is defined as
Modulation Index =
2 × f DEV
Data Rate
⎛ 100 3
K = Round ⎜⎜
⎝ 2 × f DEV
⎞
⎟⎟
⎠
The DISCRIMINATOR_BW in Register 4 should be recalculated
using the new K value. Figure 27 highlights the improved
sensitivity that can be achieved for 2FSK modulation, at low
modulation indices, by doubling the correlator bandwidth.
AFC OPERATION
The ADF7021-N also supports a real-time AFC loop that is
used to remove frequency errors due to mismatches between
the transmit and receive crystals/TCXOs. The AFC loop uses
the linear frequency discriminator block to estimate frequency
errors. The linear FSK discriminator output is filtered and
averaged to remove the FSK frequency modulation using a
combined averaging filter and envelope detector. In receive
mode, the output of the envelope detector provides an estimate
of the average IF frequency.
Two methods of AFC supported on the ADF7021-N are
external AFC and internal AFC.
External AFC
Here, the user reads back the frequency information through
the ADF7021-N serial port and applies a frequency correction
value to the fractional-N synthesizer-N divider.
The frequency information is obtained by reading the 16-bit
signed AFC readback, as described in the Readback Format
section, and by applying the following formula:
Frequency Readback [Hz] = (AFC READBACK × DEMOD
CLK)/218
Although the AFC READBACK value is a signed number, under
normal operating conditions, it is positive. In the absence of
frequency errors, the frequency readback value is equal to the
IF frequency of 100 kHz.
Rev. 0 | Page 36 of 64
ADF7021-N
Internal AFC
The ADF7021-N supports a real-time, internal, automatic
frequency control loop. In this mode, an internal control loop
automatically monitors the frequency error and adjusts the
synthesizer-N divider using an internal proportional integral
(PI) control loop.
The internal AFC control loop parameters are controlled in
Register 10. The internal AFC loop is activated by setting
R10_DB4 to 1. A scaling coefficient must also be entered, based
on the crystal frequency in use. This is set up in R10_DB[5:16]
and should be calculated using
⎛ 2 24 × 500 ⎞
⎟
AFC _ SCALING _ FACTOR = Round ⎜
⎜ XTAL ⎟
⎝
⎠
When AFC errors are removed using either the internal or
external AFC, further improvement in receiver sensitivity can
be obtained by reducing the IF filter bandwidth using the
IF_FILTER_BW bits (R4_DB[30:31]).
AUTOMATIC SYNC WORD DETECTION (SWD)
The ADF7021-N also supports automatic detection of the sync
or ID fields. To activate this mode, the sync (or ID) word must
be preprogrammed into the ADF7021-N. In receive mode, this
preprogrammed word is compared to the received bit stream.
When a valid match is identified, the external SWD pin is
asserted by the ADF7021-N on the next Rx clock pulse.
This feature can be used to alert the microprocessor that a
valid channel has been detected. It relaxes the computational
requirements of the microprocessor and reduces the overall
power consumption.
Maximum AFC Range
The maximum frequency correction range of the AFC loop is
programmable on the ADF7021-N. This is set by R10_DB[24:31].
The maximum AFC correction range is the difference in
frequency between the upper and lower limits of the AFC
tuning range. For example, if the maximum AFC correction
range is set to 10 kHz, the AFC can adjust the receiver LO
within the fLO ± 5 kHz range.
However, when RF_DIVIDE_BY_2 (R1_DB18) is enabled, the
programmed range is halved. The user should account for this
halving by doubling the programmed maximum AFC range.
The recommended maximum AFC correction range should be
≤1.5 × IF filter bandwidth. If the maximum frequency correction
range is set to be >1.5 × IF filter bandwidth, the attenuation of
the IF filter can degrade the AFC loop sensitivity.
The SWD signal can also be used to frame the received packet
by staying high for a preprogrammed number of bytes. The data
packet length can be set in R12_DB[8:15].
The SWD pin status can be configured by setting R12_DB[6:7].
R11_DB[4:5] are used to set the length of the sync/ID word, which
can be 12, 16, 20, or 24 bits long. A value of 24 bits is recommended
to minimize false sync word detection in the receiver that can
occur during recovery of the remainder of the packet or when a
noise/no signal is present at the receiver input. The transmitter
must transmit the sync byte MSB first and the LSB last to ensure
proper alignment in the receiver sync-byte-detection hardware.
An error tolerance parameter can also be programmed that
accepts a valid match when up to three bits of the word are
incorrect. The error tolerance value is assigned in R11_DB[6:7].
The adjacent channel rejection (ACR) performance of the
receivers can be degraded when AFC is enabled and the AFC
correction range is close to the IF filter bandwidth. However,
because the AFC correction range is programmable, the user
can trade off correction range and ACR performance.
Rev. 0 | Page 37 of 64
ADF7021-N
APPLICATIONS INFORMATION
Lower Tone Frequency (kHz)
IF FILTER BANDWIDTH CALIBRATION
The IF filter should be calibrated on every power-up in receive
mode to correct for errors in the bandwidth and filter center
frequency due to process variations. The automatic calibration
requires no external intervention once it is initiated by a write
to Register 5. Depending on numerous factors, such as IF filter
bandwidth, received signal bandwidth, and temperature variation,
the user must determine whether to carry out a coarse
calibration or a fine calibration.
The performance of both calibration methods is outlined in
Table 21.
1
Center Frequency
Accuracy1
100 kHz ± 2.5 kHz
100 kHz ± 0.6 kHz
Calibration
Time (Typ)
200 μs
8.2 ms
After calibration.
Calibration Setup
IF Filter calibration is initiated by writing to Register 5 and
setting the IF_CAL_COARSE bit (R5_DB4). This initiates a
coarse filter calibration. If the IF_FINE_CAL bit (R6_DB4) has
already been configured high, the coarse calibration is followed
by a fine calibration, otherwise the calibration ends.
Once initiated by writing to the part, the calibration is performed
automatically without any user intervention. Calibration time is
200 μs for coarse calibration and a few milliseconds for fine
calibration, during which time the ADF7021-N should not be
accessed. The IF filter calibration logic requires that the
IF_FILTER_DIVIDER bits (R5_DB[5:13]) be set such that
XTAL [Hz]
IF _ FILTER _ DIVIDER
IF_CAL_LOWER_TONE_DIVIDE × 2
Upper Tone Frequency (kHz)
XTAL
IF_CAL_UPPER_TONE_DIVIDE × 2
It is recommended to place the lower tone and upper tone as
outlined in Table 22.
Table 22. IF Filter Fine Calibration Tone Frequencies
Table 21. IF Filter Calibration Specifications
Filter Calibration
Method
Coarse Calibration
Fine Calibration
XTAL
= 50 kHz
The fine calibration uses two internally generated tones at
certain offsets around the IF filter. The two tones are attenuated
by the IF filter, and the level of this attenuation is measured
using the RSSI. The filter center frequency is adjusted to allow
equal attenuation of both tones. The attenuation of the two test
tones is then remeasured. This continues for a maximum of
10 RSSI measurements, at which stage the calibration algorithm
sets the IF filter center frequency to within 0.6 kHz of 100 kHz.
The frequency of these tones is set by the IF_CAL_LOWER_
TONE_DIVIDE (R6_DB[5:12]) and IF_CAL_UPPER_TONE_
DIVIDE (R6_DB[13:20]) bits, outlined in the following equations:
IF Filter
Bandwidth
9 kHz
13.5 kHz
18.5 kHz
Lower Tone
Frequency
78.1 kHz
79.4 kHz
78.1 kHz
Upper Tone
Frequency
116.3 kHz
116.3 kHz
119 kHz
Because the filter attenuation is slightly asymmetrical, it is
necessary to have a small offset in the filter center frequency to
give near equal rejection at the upper and lower adjacent
channels. The calibration tones given in Table 22 give this small
positive offset in the IF filter center frequency.
In some applications, an offset may not be required, and the
user may wish to center the IF filter exactly at 100 kHz. In this
case, the user can alter the tone frequencies from those given in
Table 22 to adjust the fine calibration result.
The calibration algorithm adjusts the filter center frequency
and measures the RSSI 10 times during the calibration. The
time for an adjustment plus RSSI measurement is given by
IF Tone Calibration Time =
IF_CAL_DWELL_TIME
SEQ CLK
It is recommended that the IF tone calibration time be at least
800 μs. The total time for the IF filter fine calibration is given by
IF Filter Fine Calibration Time = IF Tone Calibration Time × 10
When to Use Coarse Calibration
It is recommended to perform a coarse calibration on every
receive mode power-up. This calibration typically takes 200 μs.
The FILTER_CAL_COMPLETE signal from MUXOUT can be
used to monitor the filter calibration duration or to signal the
end of calibration. The ADF7021-N should not be accessed
during calibration.
Rev. 0 | Page 38 of 64
ADF7021-N
When to Use a Fine Calibration
LNA/PA MATCHING
In cases where the receive signal bandwidth is very close to the
bandwidth of the IF filter, it is recommended to perform a fine
filter calibration every time the unit powers up in receive mode.
The ADF7021-N exhibits optimum performance in terms of
sensitivity, transmit power, and current consumption, only if its
RF input and output ports are properly matched to the antenna
impedance. For cost-sensitive applications, the ADF7021-N is
equipped with an internal Rx/Tx switch that facilitates the use
of a simple, combined passive PA/LNA matching network.
Alternatively, an external Rx/Tx switch such as the ADG919 can
be used, which yields a slightly improved receiver sensitivity
and lower transmitter power consumption.
A fine calibration should be performed if
OBW + Coarse Calibration Variation > IF_FILTER_BW
where:
OBW is the 99% occupied bandwidth of the transmit signal.
Coarse Calibration Variation is 2.5 kHz.
IF_FILTER_BW is set by R4_DB[30:31].
Internal Rx/Tx Switch
When to Use Single Fine Calibration
In applications where the receiver powers up numerous times in
a short period, it is only necessary to perform a one-time fine
calibration on the initial receiver power-up.
After the initial coarse calibration and fine calibration, the result of
the fine calibration can be read back through the serial interface
using the FILTER_CAL_READBACK result (refer to the Filter
Bandwidth Calibration Readback section). On subsequent
power-ups in receive mode, the filter is manually adjusted using
the previous fine filter calibration result. This manual adjust is
performed using the IF_FILTER_ADJUST bits (R5_DB[14:19]).
Figure 50 shows the ADF7021-N in a configuration where
the internal Rx/Tx switch is used with a combined LNA/PA
matching network. This is the configuration used on the EVALADF7021-NDBxx evaluation board. For most applications, the
slight performance degradation of 1 dB to 2 dB caused by the
internal Rx/Tx switch is acceptable, allowing the user to take
advantage of the cost saving potential of this solution. The
design of the combined matching network must compensate for
the reactance presented by the networks in the Tx and the Rx
paths, taking the state of the Rx/Tx switch into consideration.
VBAT
C1
ANTENNA
ZOPT_PA
OPTIONAL
BPF OR LPF
ZIN_RFIN
CA
RFIN
LA
IF Filter Variation with Temperature
CB
If the receive signal occupied bandwidth is considerably less
than the IF filter bandwidth, the variation of filter center
frequency over the operating temperature range may not be
an issue. Alternatively, if the IF filter bandwidth is not wide
enough to tolerate the variation with temperature, a periodic
filter calibration can be performed or, alternatively, the on-chip
temperature sensor can be used to determine when a filter calibration is necessary by monitoring for changes in temperature.
PA_OUT
PA
This method should only be used if the successive power-ups in
receive mode are over a short duration, during which time there
is little variation in temperature (<15°C).
When calibrated, the filter center frequency can vary with changes
in temperature. If the ADF7021-N is used in an application where
it remains in receive mode for a considerable length of time, the
user must consider this variation of filter center frequency with
temperature. This variation is typically 1 kHz per 20°C, which
means that if a coarse filter calibration and fine filter calibration
are performed at 25°C, the initial maximum error is ±0.5 kHz,
and the maximum possible change in the filter center frequency
over temperature (−40°C to +85°C) is ±3.25 kHz. This gives a
total error of ±3.75 kHz.
L1
RFINB
LNA
ZIN_RFIN
ADF7021-N
07246-022
The FILTER_CAL_COMPLETE signal from MUXOUT (set by
R0_DB[29:31]) can be used to monitor the filter calibration
duration or to signal the end of calibration. A coarse filter
calibration is automatically performed prior to a fine filter
calibration.
Figure 50. ADF7021-N with Internal Rx/Tx Switch
The procedure typically requires several iterations until an
acceptable compromise has been reached. The successful implementation of a combined LNA/PA matching network for the
ADF7021-N is critically dependent on the availability of an
accurate electrical model for the PCB. In this context, the use of a
suitable CAD package is strongly recommended. To avoid this
effort, a small form-factor reference design for the ADF7021-N is
provided, including matching and harmonic filter components.
The design is on a 2-layer PCB to minimize cost. Gerber files
are available at www.analog.com.
Rev. 0 | Page 39 of 64
ADF7021-N
External Rx/Tx Switch
Figure 51 shows a configuration using an external Rx/Tx switch.
This configuration allows an independent optimization of the
matching and filter network in the transmit and receive path.
Therefore, it is more flexible and less difficult to design than the
configuration using the internal Rx/Tx switch. The PA is biased
through Inductor L1, while C1 blocks dc current. Together, L1
and C1 form the matching network that transforms the source
impedance into the optimum PA load impedance, ZOPT_PA.
IMAGE REJECTION CALIBRATION
VBAT
OPTIONAL
LPF
C1
L1
PA_OUT
PA
ANTENNA
ZOPT_PA
ZIN_RFIN
OPTIONAL CA
BPF
(SAW)
RFIN
ADG919
Rx/Tx – SELECT
CB
RFINB
LNA
ZIN_RFIN
ADF7021-N
07246-021
LA
Depending on the antenna configuration, the user may need a
harmonic filter at the PA output to satisfy the spurious emission
requirement of the applicable government regulations. The
harmonic filter can be implemented in various ways, for example, a
discrete LC pi or T-stage filter. The immunity of the ADF7021-N
to strong out-of-band interference can be improved by adding a
band-pass filter in the Rx path. Alternatively, the ADF7021-N
blocking performance can be improved by selecting one of the
enhanced linearity modes, as described in Table 15.
Figure 51. ADF7021-N with External Rx/Tx Switch
ZOPT_PA depends on various factors, such as the required
output power, the frequency range, the supply voltage range,
and the temperature range. Selecting an appropriate ZOPT_PA
helps to minimize the Tx current consumption in the application.
Application Note AN-764 and Application Note AN-859 contain a
number of ZOPT_PA values for representative conditions. Under
certain conditions, however, it is recommended to obtain a suitable
ZOPT_PA value by means of a load-pull measurement.
Due to the differential LNA input, the LNA matching network
must be designed to provide both a single-ended-to-differential
conversion and a complex, conjugate impedance match. The
network with the lowest component count that can satisfy these
requirements is the configuration shown in Figure 51, consisting
of two capacitors and one inductor.
The image channel in the ADF7021-N is 200 kHz below the
desired signal. The polyphase filter rejects this image with an
asymmetric frequency response. The image rejection performance
of the receiver is dependent on how well matched the I and Q
signals are in amplitude and how well matched the quadrature
is between them (that is, how close to 90° apart they are). The
uncalibrated image rejection performance is approximately
29 dB (at 450 MHz). However, it is possible to improve on this
performance by as much as 20 dB by finding the optimum I/Q
gain and phase adjust settings.
Calibration Using Internal RF Source
With the LNA powered off, an on-chip generated, low level RF
tone is applied to the mixer inputs. The LO is adjusted to make
the tone fall at the image frequency where it is attenuated by the
image rejection of the IF filter. The power level of this tone is then
measured using the RSSI readback. The I/Q gain and phase adjust
DACs (R5_DB[20:31]) are adjusted and the RSSI is remeasured.
This process is repeated until the optimum values for the gain
and phase adjust are found that provide the lowest RSSI readback
level, thereby maximizing the image rejection performance of
the receiver.
Rev. 0 | Page 40 of 64
ADF7021-N
ADF7021-N
RFIN
LNA
RFINB
GAIN ADJUST
POLYPHASE
IF FILTER
MUX
INTERNAL
SIGNAL
SOURCE
RSSI/
LOG AMP
7-BIT ADC
PHASE ADJUST
I
Q
FROM LO
SERIAL
INTERFACE
4
PHASE ADJUST
REGISTER 5
RSSI READBACK
4
GAIN ADJUST
REGISTER 5
MICROCONTROLLER
07246-072
I/Q GAIN/PHASE ADJUST AND
RSSI MEASUREMENT
ALGORITHM
Figure 52. Image Rejection Calibration Using the Internal Calibration Source and a Microcontroller
Using the internal RF source, the RF frequencies that can be
used for image calibration are programmable and are odd
multiples of the reference frequency.
IR_GAIN_ADJUST_I/Q bit (R5_DB30), whereas the
IR_GAIN_ADJUST_UP/DN bit (R5_DB31) sets whether
the gain adjustment defines a gain or an attenuation adjust.
Calibration Using External RF Source
The calibration results are valid over changes in the ADF7021-N
supply voltage. However, there is some variation with temperature.
A typical plot of variation in image rejection over temperature
after initial calibrations at −40°C, +25°C, and +85°C is shown in
Figure 53. The internal temperature sensor on the ADF7021-N
can be used to determine if a new IR calibration is required.
60
The IR calibration algorithm available from Analog Devices, Inc., is
based on a low complexity, 2D optimization algorithm that can
be implemented in an external microprocessor or microcontroller.
50
To enable the internal RF source, the IR_CAL_SOURCE_
DRIVE_LEVEL bits (R6_DB[28:29]) should be set to the
maximum level. The LNA should be set to its minimum gain
setting, and the AGC should be disabled if the internal source is
being used. Alternatively, an external RF source can be used.
IMAGE REJECTION (dB)
Calibration Procedure and Setup
The magnitude of the phase adjust is set by using the IR_PHASE_
ADJUST_MAG bits (R5_DB[20:23]). This correction can be
applied to either the I channel or Q channel, depending on the
value of the IR_PHASE_ADJUST_DIRECTION bit (R5_DB24).
The magnitude of the I/Q gain is adjusted by the IR_GAIN_
ADJUST_MAG bits (R5_DB[25:29]). This correction can be
applied to either the I or Q channel, depending on the value of
Rev. 0 | Page 41 of 64
CAL AT +25°C
40
CAL AT +85°C
CAL AT –40°C
30
20
10
VDD = 3.0V
IF BW = 25kHz
WANTED SIGNAL:
RF FREQ = 430MHz
MODULATION = 2FSK
DATA RATE = 9.6kbps,
PRBS9
fDEV = 4kHz
LEVEL= –100dBm
0
–60
–40
–20
0
INTERFERER SIGNAL:
RF FREQ = 429.8MHz
MODULATION = 2FSK
DATA RATE = 9.6kbps,
PRBS11
fDEV = 4kHz
20
40
60
80
100
TEMPERATURE (°C)
Figure 53. Image Rejection Variation with Temperature After Initial
Calibrations at −40°C, +25°C, and +85°C
07246-067
IR calibration can also be implemented using an external RF
source. The IR calibration procedure is the same as that used for
the internal RF source, except that an RF tone is applied to the
LNA input.
ADF7021-N
to a particular application, such as setting up sync byte
detection or enabling AFC. When going from Tx to Rx or vice
versa, the user needs to toggle the Tx/Rx bit and write only to
Register 0 to alter the LO by 100 kHz.
PACKET STRUCTURE AND CODING
PREAMBLE
SYNC
WORD
ID
FIELD
DATA FIELD
07246-023
The suggested packet structure to use with the ADF7021-N is
shown in Figure 54.
CRC
Figure 54. Typical Format of a Transmit Protocol
Refer to the Receiver Setup section for information on the
required preamble structure and length for the various modulation
schemes.
PROGRAMMING AFTER INITIAL POWER-UP
Table 23 lists the minimum number of writes needed to set up
the ADF7021-N in either Tx or Rx mode after CE is brought
high. Additional registers can also be written to tailor the part
Table 23. Minimum Register Writes Required for Tx/Rx Setup
Mode
Tx
Rx
Tx to Rx and Rx to Tx
Reg 1
Reg 1
Reg 0
Reg 3
Reg 3
Registers
Reg 0 Reg 2
Reg 0 Reg 5
Reg 4
The recommended programming sequences for transmit and
receive are shown in Figure 55 and Figure 56, respectively. The
difference in the power-up routine for a TCXO and XTAL
reference is shown in these figures.
Rev. 0 | Page 42 of 64
ADF7021-N
TCXO
REFERENCE
XTAL
REFERENCE
POWER-DOWN
CE LOW
CE HIGH
WAIT 10µs + 1ms
(REGULATOR POWER-UP + TYPICAL XTAL SETTLING)
CE HIGH
WAIT 10µs (REGULATOR POWER-UP)
WRITE TO REGISTER 1 (TURNS ON VCO)
WAIT 0.7ms (TYPICAL VCO SETTLING)
WRITE TO REGISTER 3 (TURNS ON Tx/Rx CLOCKS)
WRITE TO REGISTER 0 (TURNS ON PLL)
WAIT 40µs (TYPICAL PLL SETTLING)
WRITE TO REGISTER 2 (TURNS ON PA)
WAIT FOR PA TO RAMP UP (ONLY IF PA RAMP ENABLED)
Tx MODE
WAIT FOR Tx LATENCY NUMBER OF BITS
(REFER TO TABLE 12)
WRITE TO REGISTER 2 (TURNS OFF PA)
WAIT FOR PA TO RAMP DOWN
07246-086
CE LOW
POWER-DOWN
OPTIONAL. ONLY NECESSARY IF PA
RAMP DOWN IS REQUIRED.
Figure 55. Power-Up Sequence for Transmit Mode
Rev. 0 | Page 43 of 64
ADF7021-N
TCXO
REFERENCE
XTAL
REFERENCE
POWER-DOWN
CE LOW
CE HIGH
WAIT 10µs + 1ms
(REGULATOR POWER-UP + TYPICAL XTAL SETTLING)
CE HIGH
WAIT 10µs (REGULATOR POWER-UP)
WRITE TO REGISTER 1 (TURNS ON VCO)
WAIT 0.7ms (TYPICAL VCO SETTLING)
WRITE TO REGISTER 3 (TURNS ON Tx/Rx CLOCKS)
WRITE TO REGISTER 6 (SETS UP IF FILTER CALIBRATION)
OPTIONAL:
ONLY NECESSARY IF
IF FILTER FINE CAL IS REQUIRED.
WRITE TO REGISTER 5 (STARTS IF FILTER CALIBRATION)
WAIT 0.2ms (COARSE CAL) OR WAIT 8.2ms
(COARSE CALIBRATION + FINE CALIBRATION)
WRITE TO REGISTER 11 (SET UP SWD)
WRITE TO REGISTER 12 (ENABLE SWD)
OPTIONAL:
ONLY NECESSARY IF
SWD IS REQUIRED.
WRITE TO REGISTER 0 (TURNS ON PLL)
WAIT 40µs (TYPICAL PLL SETTLING)
WRITE TO REGISTER 4 (TURNS ON DEMOD)
WRITE TO REGISTER 10 (TURNS ON AFC)
OPTIONAL:
ONLY NECESSARY IF
AFC IS REQUIRED.
Rx MODE
07246-087
CE LOW
POWER-DOWN
OPTIONAL.
Figure 56. Power-Up Sequence for Receive Mode
Rev. 0 | Page 44 of 64
ADF7021-N
APPLICATIONS CIRCUIT
The ADF7021-N requires very few external components for
operation. Figure 57 shows the recommended application
circuit. Note that the power supply decoupling and regulator
capacitors are omitted for clarity.
For recommended component values, refer to the ADF7021-N
evaluation board data sheet and AN-859 application note
accessible from the ADF7021-N product page. Follow the
reference design schematic closely to ensure optimum
performance in narrow-band applications.
LOOP FILTER
VDD
TCXO
EXT VCO L*
CVCO
CAP
REFERENCE
VDD1
4
RFOUT
5
RFGND
6
RFIN
37
38
OSC2
39
OSC1
41
40
VDD3
CREG3
42
CPOUT
44
L2
43
VDD
45
47
46
L1
MUXOUT
3
CLKOUT 36
TxRxCLK 35
SWD 33
VDD2
32
ADF7021-N
ADCIN 30
RFINB
8
RLNA
GND2
29
9
VDD4
SCLK
28
10
RSET
SREAD
27
SDATA
26
CREG4
SLE
TEST_A
TO
MICROCONTROLLER
CONFIGURATION
INTERFACE
25
CE
24
23
FILT_Q
GND4
FILT_Q
GND4
22
21
20
FILT_I
RSET
RESISTOR
19
MIX_Q
FILT_I
18
17
13
RLNA
RESISTOR
MIX_Q
GND4
MIX_I
12
VDD
CREG2 31
7
11
TO
MICROCONTROLLER
Tx/Rx SIGNAL
INTERFACE
TxRxDATA 34
16
VDD
CREG1
14
T-STAGE LC
FILTER
VCOIN
2
MIX_I
VDD
1
15
MATCHING
GND
VDD
ANTENNA
CONNECTION
GND1
CVCO
48
VDD
CHIP ENABLE
TO MICROCONTROLLER
NOTES
1. PINS [13:18], PINS [20:21], AND PIN 23 ARE TEST PINS AND ARE NOT USED IN NORMAL OPERATION.
Figure 57. Typical Application Circuit (Regulator Capacitors and Power Supply Decoupling Not Shown)
Rev. 0 | Page 45 of 64
07246-084
*PIN 44 AND PIN 46 CAN BE LEFT FLOATING IF EXTERNAL INDUCTOR VCO IS NOT USED.
ADF7021-N
SERIAL INTERFACE
AFC Readback
The serial interface allows the user to program the 16-/32-bit
registers using a 3-wire interface (SCLK, SDATA, and SLE).
It consists of a level shifter, 32-bit shift register, and 16 latches.
Signals should be CMOS compatible. The serial interface is
powered by the regulator, and, therefore, is inactive when CE is low.
The AFC readback is valid only during the reception of FSK
signals with either the linear or correlator demodulator active.
The AFC readback value is formatted as a signed 16-bit integer
comprising Bit RV1 to Bit RV16 and is scaled according to the
following formula:
Data is clocked into the register, MSB first, on the rising edge of
each clock (SCLK). Data is transferred to one of 16 latches on the
rising edge of SLE. The destination latch is determined by the
value of the four control bits (C4 to C1); these are the bottom
4 LSBs, DB3 to DB0, as shown in Figure 2. Data can also be read
back on the SREAD pin.
FREQ RB [Hz] = (AFC_READBACK × DEMOD CLK)/218
In the absence of frequency errors, FREQ RB is equal to the IF
frequency of 100 kHz. Note that, for the AFC readback to yield
a valid result, the downconverted input signal must not fall outside
the bandwidth of the analog IF filter. At low input signal levels,
the variation in the readback value can be improved by averaging.
READBACK FORMAT
The readback operation is initiated by writing a valid control
word to the readback register and enabling the READBACK bit
(R7_DB8 = 1). The readback can begin after the control word
has been latched with the SLE signal. SLE must be kept high
while the data is being read out. Each active edge at the SCLK
pin successively clocks the readback word out at the SREAD
pin, as shown in Figure 58, starting with the MSB first. The data
appearing at the first clock cycle following the latch operation
must be ignored. An extra clock cycle is needed after the 16th
readback bit to return the SREAD pin to tristate. Therefore, 18
total clock cycles are needed for each read back. After the 18th
clock cycle, SLE should be brought low.
RSSI Readback
The format of the readback word is shown in Figure 58. It
comprises the RSSI-level information (Bit RV1 to Bit RV7), the
current filter gain (FG1, FG2), and the current LNA gain (LG1,
LG2) setting. The filter and LNA gain are coded in accordance
with the definitions in the Register 9—AGC Register section. For
signal levels below −100 dBm, averaging the measured RSSI values
improves accuracy. The input power can be calculated from the
RSSI readback value as outlined in the RSSI/AGC section.
READBACK VALUE
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
DB2
DB1
DB0
AFC READBACK
RV16
RV15
RV14
RV13
RV12
RV11
RV10
RV9
RV8
RV7
RV6
RV5
RV4
RV3
RV2
RV1
RSSI READBACK
X
X
X
X
X
LG2
LG1
FG2
FG1
RV7
RV6
RV5
RV4
RV3
RV2
RV1
BATTERY VOLTAGE/ADCIN/
TEMP. SENSOR READBACK
X
X
X
X
X
X
X
X
X
RV7
RV6
RV5
RV4
RV3
RV2
RV1
SILICON REVISION
RV16
RV15
RV14
RV13
RV12
RV11
RV10
RV9
RV8
RV7
RV6
RV5
RV4
RV3
RV2
RV1
FILTER CAL READBACK
0
0
0
0
0
0
0
0
RV8
RV7
RV6
RV5
RV4
RV3
RV2
RV1
Figure 58. Readback Value Table
Rev. 0 | Page 46 of 64
07246-029
READBACK MODE
ADF7021-N
Battery Voltage/ADCIN/Temperature Sensor Readback
Filter Bandwidth Calibration Readback
The battery voltage is measured at Pin VDD4. The readback
information is contained in Bit RV1 to Bit RV7. This also
applies to the readback of the voltage at the ADCIN pin and the
temperature sensor. From the readback information, the battery
or ADCIN voltage can be determined using
The filter calibration readback word is contained in Bit RV1 to
Bit RV8 (see Figure 58). This readback can be used for manual
filter adjust, thereby avoiding the need to do an IF filter
calibration in some instances. The manual adjust value is
programmed by R5_DB[14:19]. To calculate the manual adjust
based on a filter calibration readback, use the following formula:
VBATTERY = (BATTERY VOLTAGE READBACK)/21.1
IF_FILTER_ADJUST = FILTER_CAL_READBACK − 128
VADCIN = (ADCIN VOLTAGE READBACK)/42.1
The result should be programmed into R5_DB[14:19] as outlined
in the Register 5—IF Filter Setup Register section.
The temperature can be calculated using
Temp [°C] = −40 + (68.4 − TEMP READBACK) × 9.32
Silicon Revision Readback
The silicon revision readback word is valid without setting any
other registers. The silicon revision word is coded with four
quartets in BCD format. The product code (PC) is coded with
three quartets extending from Bit RV5 to Bit RV16. The revision
code (RC) is coded with one quartet extending from Bit RV1 to
Bit RV4. The product code for the ADF7021-N should read
back as PC = 0x211. The current revision code should read as
RC = 0x1.
Rev. 0 | Page 47 of 64
ADF7021-N
INTERFACING TO A MICROCONTROLLER/DSP
SPI Mode
Standard Transmit/Receive Data Interface
In SPI mode, the TxRxCLK pin is configured to input transmit
data in transmit mode. In receive mode, the receive data is available
on the TxRxDATA pin. The data clock in both transmit and receive
modes is available on the CLKOUT pin. In transmit mode, data is
clocked into the ADF7021-N on the positive edge of CLKOUT. In
receive mode, the TxRxDATA data pin should be sampled by
the microcontroller on the positive edge of the CLKOUT.
The standard transmit/receive signal and configuration interface
to a microcontroller is shown in Figure 59. In transmit mode,
the ADF7021-N provides the data clock on the TxRxCLK pin,
and the TxRxDATA pin is used as the data input. The transmit
data is clocked into the ADF7021-N on the rising edge of
TxRxCLK.
MICROCONTROLLER
ADF7021-N
MISO
TxRxDATA
MISO
SPI
MOSI
TxRxCLK
SCLOCK
MOSI
TxRxDATA
SCLK
CLKOUT
SS
SWD
SWD
P2.4
SREAD
P2.5
SLE
P2.6
SDATA
P2.7
SCLK
GPIO
SREAD
SLE
SDATA
07246-026
P3.2/INT0
GPIO
CE
CE
P3.7
ADF7021-N
TxRxCLK
07246-076
ADuC84x
SCLK
Figure 59. ADuC84x to ADF7021-N Connection Diagram
Figure 61. ADF7021-N (SPI Mode) to Microcontroller Interface
In receive mode, the ADF7021-N provides the synchronized
data clock on the TxRxCLK pin. The receive data is available on
the TxRxDATA pin. The rising edge of TxRxCLK should be
used to clock the receive data into the microcontroller. Refer to
Figure 4 and Figure 5 for the relevant timing diagrams.
To enable SPI interface mode, set R0_DB28 high and set
R15_DB[17:19] to 0x7. Figure 8 and Figure 9 show the relevant
timing diagrams for SPI mode, while Figure 61 shows the
recommended interface to a microcontroller using the SPI
mode of the ADF7021-N.
In 4FSK transmit mode, the MSB of the transmit symbol is
clocked into the ADF7021-N on the first rising edge of the data
clock from the TxRxCLK pin. In 4FSK receive mode, the MSB
of the first payload symbol is clocked out on the first negative
edge of the data clock after the SWD and should be clocked into
the microcontroller on the following rising edge. Refer to Figure 6
and Figure 7 for the relevant timing diagrams.
ADSP-BF533 interface
The suggested method of interfacing to the Blackfin® ADSPBF533 is given in Figure 62.
ADSP-BF533
SCK
MOSI
MISO
PF5
RSCLK1
In UART mode, the TxRxCLK pin is configured to input transmit
data in transmit mode. In receive mode, the receive data is available
on the TxRxDATA pin, thus providing an asynchronous data
interface. The UART mode can only be used with oversampled
2FSK. Figure 60 shows a possible interface to a microcontroller
using the UART mode of the ADF7021-N. To enable this UART
interface mode, set R0_DB28 high. Figure 8 and Figure 9 show
the relevant timing diagrams for UART mode.
ADF7021-N
MICROCONTROLLER
UART
TxDATA
TxRxCLK
RxDATA
TxRxDATA
CE
SWD
SREAD
SLE
SDATA
SCLK
07246-085
GPIO
Figure 60. ADF7021-N (UART Mode) to
Asynchronous Microcontroller Interface
Rev. 0 | Page 48 of 64
DT1PRI
SDATA
SREAD
SLE
TxRxCLK
TxRxDATA
DR1PRI
RFS1
PF6
SWD
CE
Figure 62. ADSP-BF533 to ADF7021-N Connection Diagram
07246-027
UART Mode
ADF7021-N
SCLK
ADF7021-N
DB20
N2
DB5
DB4
DB3
M2
M1
C4 (0)
DB6
DB7
M4
M3
DB8
DB12
M9
M5
DB13
M10
DB9
DB14
M11
DB10
DB15
M12
M6
DB16
M13
M7
DB17
M14
DB11
DB18
M8
DB19
N1
M15
M14
M13
...
M3
M2
M1
0
1
TRANSMIT
RECEIVE
0
0
0
.
.
.
1
1
1
1
0
0
0
.
.
.
1
1
1
1
0
0
0
.
.
.
1
1
1
1
...
...
...
...
...
...
...
...
...
...
0
0
0
.
.
.
1
1
1
1
0
0
1
.
.
.
0
0
1
1
0
1
0
.
.
.
0
1
0
1
UART_MODE
0
1
DISABLED
ENABLED
M3
M2
M1
MUXOUT
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
REGULATOR_READY (DEFAULT)
FILTER_CAL_COMPLETE
DIGITAL_LOCK_DETECT
RSSI_READY
Tx_Rx
LOGIC_ZERO
TRISTATE
LOGIC_ONE
N8
N7
N6
N5
N4
N3
N2
N1
INTEGER_N
DIVIDE RATIO
0
0
.
.
.
1
0
0
.
.
.
1
0
0
.
.
.
1
1
1
.
.
.
1
0
1
.
.
.
1
1
0
.
.
.
1
1
0
.
.
.
0
1
0
.
.
.
1
23
24
.
.
.
253
1
1
1
1
1
1
1
0
254
1
1
1
1
1
1
1
1
255
FRACTIONAL_N
DIVIDE RATIO
0
1
2
.
.
.
32764
32765
32766
32767
07246-030
U1
DB0
DB21
N3
M15
DB1
DB22
N4
Tx/Rx
C1 (0)
DB23
N5
TR1
C2 (0)
DB24
N6
ADDRESS
BITS
FRACTIONAL_N
DB2
DB25
N7
INTEGER_N
C3 (0)
DB26
N8
Tx/Rx
DB27
DB29
M1
DB28 UART_MODE
DB30
M2
U1
DB31
M3
MUXOUT
TR1
REGISTER 0—N REGISTER
Figure 63. Register 0—N Register Map
•
•
The RF output frequency is calculated by the following:
For the direct output
Fractional _ N ⎞
⎛
RFOUT = PFD × ⎜ Integer _ N +
⎟
2 15
⎠
⎝
For the RF_DIVIDE_BY_2 (R1_DB18) selected
•
Fractional _ N ⎞
⎛
RFOUT = PFD × 0.5 × ⎜ Integer _ N +
⎟
2 15
⎠
⎝
•
In UART/SPI mode, the TxRxCLK pin is used to input the
Tx data. The Rx Data is available on the TxRxDATA pin.
Rev. 0 | Page 49 of 64
FILTER_CAL_ COMPLETE in the MUXOUT map in
Figure 63 indicates when a coarse or coarse plus fine
IF filter calibration has finished. DIGITAL_
LOCK_DETECT indicates when the PLL has locked.
RSSI_READY indicates that the RSSI signal has settled
and an RSSI readback can be performed.
Tx_Rx gives the status of DB27 in this register, which
can be used to control an external Tx/Rx switch.
ADF7021-N
VB1
1
0
.
1
VCO_INDUCTOR
DB8
DB7
DB6
DB5
DB4
CL2
CL1
R3
R2
R1
INTERNAL L VCO
EXTERNAL L VCO
D1
0
1
CP2
0
0
1
1
CP1
RSET
0
1
0
1
ICP (mA)
3.6kΩ
0.3
0.9
1.5
2.1
DB0
DB9
CL3
VCO OFF
VCO ON
C1 (1)
DB10
CL4
0
1
DB1
DB11
D1
LOOP
CONDITION
CL1
0
1
0
.
.
.
1
DB2
DB12
X1
VE1
CL2
0
0
1
.
.
.
1
C2 (0)
DB13
XB1
3.75mA
CL3
0
0
0
.
.
.
1
C3 (0)
DB14
XB2
CL4
0
0
0
.
.
.
1
DB3
DB15
VCO_BIAS
CURRENT
0.25mA
0.5mA
C4 (0)
XOSC_
ENABLE
XTAL_
DOUBLER
DB16
OFF
ON
R2
0
1
.
.
.
1
RF R_COUNTER
R1 DIVIDE RATIO
1
1
0
2
.
.
.
.
.
.
1
7
CLKOUT_
DIVIDE RATIO
OFF
2
4
.
.
.
30
XTAL_
DOUBLER
DISABLE
ENABLED
X1 XOSC_ENABLE
0
OFF
1
ON
XB2 XB1
0
1
0
1
0
0
1
1
XTAL_
BIAS
20µA
25µA
30µA
35µA
07246-031
VB2
0
1
.
1
CP1
DB19
VB1
VB3
0
0
.
1
R3
0
0
.
.
.
1
RFD1 RF_DIVIDE_BY_2
0
1
ADDRESS
BITS
R_COUNTER
CP2
DB20
VB2
VB4
0
0
.
1
CLKOUT_
DIVIDE
DB17
DB21
VB3
NOMINAL
VCO ADJUST UP 1
VCO ADJUST UP 2
VCO ADJUST UP 3
XTAL_
BIAS
VE1
DB22
VB4
0
1
0
1
CP_
CURRENT
RF_DIVIDE_
BY_2
VCO_
ENABLE
DB23
VA1
VA1
0
0
1
1
RFD1 DB18
VCO_
ADJUST
DB24
VCO CENTER
FREQ ADJUST
VA2
VCL1
0
1
VCO_BIAS
VA2
VCL1 DB25
VCO_
INDUCTOR
REGISTER 1—VCO/OSCILLATOR REGISTER
Figure 64. Register 1—VCO/Oscillator Register Map
•
The R_COUNTER and XTAL_DOUBLER relationship is
as follows:
•
The VCO_BIAS bits should be set according to Table 9.
•
The VCO_ADJUST bits adjust the center of the VCO
operating band. Each bit typically adjusts the VCO band
up by 1% of the RF operating frequency (0.5% if
RF_DIVIDE_BY_2 is enabled).
•
Setting VCO_INDUCTOR to external allows the use of the
external inductor VCO, which gives RF operating
frequencies of 80 MHz to 650 MHz. If the internal
inductor VCO is being used for operation, set this bit low.
XTAL
If XTAL_DOUBLER = 0, PFD =
R _ COUNTER
If XTAL_DOUBLER =1, PFD =
XTAL × 2
R _ COUNTER
•
CLOCKOUT_DIVIDE is a divided-down and inverted
version of the XTAL and is available on Pin 36 (CLKOUT).
•
Set XOSC_ENABLE high when using an external crystal.
If using an external oscillator (such as TCXO) with CMOSlevel outputs into Pin OSC2, set XOSC_ENABLE low. If
using an external oscillator with a 0.8 V p-p clipped sine
wave output into Pin OSC1, set XOSC_ENABLE high.
Rev. 0 | Page 50 of 64
ADF7021-N
DI1
TxDATA_INVERT
0
0
1
1
0
1
0
1
NORMAL
INVERT CLK
INVERT DATA
INV CLK AND DATA
0
0
0
0
.
1
NRC1 R-COSINE_ALPHA
0.5 (Default)
0.7
...
...
...
...
...
...
0
0
0
0
.
1
0
0
1
1
.
1
0
1
0
1
.
1
0
1
PR2
PR1
PA_RAMP RATE
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
NO RAMP
256 CODES/BIT
128 CODES/BIT
64 CODES/BIT
32 CODES/BIT
16 CODES/BIT
8 CODES/BIT
4 CODES/BIT
.
P2
P1
POWER_
AMPLIFIER
0
0
0
0
.
.
1
.
.
.
.
.
.
1
.
.
.
.
.
.
.
0
0
1
1
.
.
1
0
1
0
1
.
.
1
0 (PA OFF)
1 (–16.0 dBm)
2
3
.
.
63 (13 dBm)
DB5
DB4
DB3
DB2
DB1
DB0
S2
S1
C4 (0)
C3 (0)
C2 (1)
C1 (0)
PA_
ENABLE
OFF
ON
0
0
0
0
1
1
1
1
.
DB6
DB9
PR2
PE1 PA_ENABLED
PR3
P6
ADDRESS
BITS
MODULATION_
SCHEME
S3
DB10
PR3
5µA
7µA
9µA
11µA
DB7
DB11
PA1
0
1
0
1
DB8
DB12
PA2
PA1 PA_BIAS
0
0
1
1
PE1
DB13
PA2
PR1
DB14
PA_RAMP
P1
PA_BIAS
P2
DB16
P4
0
1
2
3
.
511
DB15
DB17
P5
TFD3 TFD2 TFD1 fDEV
P3
DB18
P6
TFD1 DB19
TFD2 DB20
TFD3 DB21
TFD4 DB22
TFD5 DB23
POWER_AMPLIFIER
S3
S2
S1
MODULATION_SCHEME
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
2FSK
GAUSSIAN 2FSK
3FSK
4FSK
OVERSAMPLED 2FSK
RAISED COSINE 2FSK
RAISED COSINE 3FSK
RAISED COSINE 4FSK
07246-032
DI2
TFD9 ...
0
1
TFD6 DB24
TFD7 DB25
TFD8 DB26
DB28
DI1
Tx_FREQUENCY_DEVIATION
TFD9 DB27
DB29
DI2
TxDATA_
INVERT
NRC1 DB30
R-COSINE_
ALPHA
REGISTER 2—TRANSMIT MODULATION REGISTER
Figure 65. Register 2—Transmit Modulation Register Map
•
The 2FSK/3FSK/4FSK frequency deviation is expressed by
the following:
•
In the case of 4FSK, there are tones at ±3 × the frequency
deviation and at ±1 × the deviation.
Direct output
•
The power amplifier (PA) ramps at the programmed rate
(R2_DB[8:10]) until it reaches its programmed level
(R2_DB[13:18]). If the PA is enabled/disabled by the
PA_ENABLE bit (R2_DB7), it ramps up and down. If it is
enabled/disabled by the Tx/Rx bit (R0_DB27), it ramps up
and turns hard off.
•
R-COSINE_ALPHA sets the roll-off factor (alpha) of the
raised cosine data filter to either 0.5 or 0.7. The alpha is set
to 0.5 by default, but the raised cosine filter bandwidth can
be increased to provide less aggressive data filtering by
using an alpha of 0.7.
Frequency Deviation [Hz] =
Tx_FREQUENCY_DEVIATION × PFD
216
With RF_DIVIDE_BY_2 (R1_DB18) enabled
Frequency Deviation [Hz] =
Tx_FREQUENCY_DEVIATION × PFD
0.5 ×
216
where Tx_FREQUENCY_DEVIATION is set by
R2_DB[19:27] and PFD is the PFD frequency.
Rev. 0 | Page 51 of 64
ADF7021-N
GD5
GD4
GD3
GD2
GD1
AGC_CLK_DIVIDE
0
0
...
1
0
0
...
1
0
0
...
1
0
0
...
1
0
0
...
1
0
1
...
1
INVALID
1
...
63
DB5
BK2
DB0
DB6
OK1
C1 (1)
DB7
OK2
DB1
DB8
OK3
C2 (1)
DB9
OK4
DB2
DB10
FS1
DB3
DB11
FS2
C3 (0)
DB12
FS3
C4 (0)
DB13
FS4
DB4
DB14
BK1
DB15
FS5
ADDRESS
BITS
SK3
SK2
SK1
SEQ_CLK_DIVIDE
BK2
BK1
BBOS_CLK_DIVIDE
0
0
.
1
1
0
1
.
1
1
1
0
.
0
1
1
2
.
254
255
0
0
1
1
0
1
0
1
4
8
16
32
OK4 OK3 OK2 OK1
DEMOD_CLK_DIVIDE
0
0
...
1
INVALID
1
...
15
FS8
FS7
...
FS3
FS2
FS1
0
0
.
1
1
0
0
.
1
1
...
...
...
...
...
0
0
.
1
1
0
1
.
1
1
1
0
.
0
1
0
0
...
1
0
0
...
1
0
1
...
1
CDR_CLK_ DIVIDE
1
2
.
254
255
07246-033
GD6
FS6
...
...
...
...
...
...
DB16
0
0
.
1
1
DB17
SK7
0
0
.
1
1
FS7
DB22
SK5
DB18
DB23
SK6
SK8
FS8
DB24
SK7
SK1
DB25
SK8
DB19
DB26
GD1
DB20
DB27
GD2
SK2
DB28
GD3
DB21
DB29
GD4
SK3
DB30
GD5
DEMOD_CLK_
DIVIDE
CDR_CLK_DIVIDE
SK4
DB31
SEQ_CLK_DIVIDE
GD6
AGC_CLK_DIVIDE
BBOS_CLK_
DIVIDE
REGISTER 3—TRANSMIT/RECEIVE CLOCK REGISTER
Figure 66. Register 3—Transmit/Receive Clock Register Map
•
BBOS CLK =
•
XTAL
BBOS _ CLK _ DIVIDE
•
XTAL
DEMOD _ CLK _ DIVIDE
XTAL
SEQ _ CLK _ DIVIDE
The time allowed for each AGC step to settle is determined
by the AGC update rate. It should be set close to 8 kHz.
AGC Update Rate [Hz] =
For 2FSK/3FSK, the data/clock recovery frequency (CDR
CLK) needs to be within 2% of (32 × data rate). For 4FSK,
the CDR CLK needs to be within 2% of (32 × symbol rate).
CDR CLK =
The sequencer clock (SEQ CLK) supplies the clock to the
digital receive block. It should be as close to 100 kHz as
possible.
SEQ CLK =
Set the demodulator clock (DEMOD CLK) such that
2 MHz ≤ DEMOD CLK ≤ 15 MHz, where
DEMOD CLK =
•
•
Baseband offset clock frequency (BBOS CLK) must be
greater than 1 MHz and less than 2 MHz, where
DEMOD CLK
CDR _ CLK _ DIVIDE
Rev. 0 | Page 52 of 64
SEQ CLK
AGC _ CLK _ DIVIDE
ADF7021-N
DW10 .
POST_DEMOD_
DW6 DW5 DW4 DW3 DW2 DW1 BW
0
0
.
.
.
.
1
0
0
.
.
.
.
1
0
0
.
.
.
.
1
0
0
.
.
.
.
1
DB3
DB2
DB1
DB0
C3(1)
C2(0)
C1(0)
DB9
RI2
C4(0)
DB10
TD1
DB4
DB11
TD2
DB5
DB12
TD3
DS1
DB13
TD4
DS2
DB14
TD5
DOT_PRODUCT
DB6
DB15
TD6
DS3
DB16
TD7
NORMAL
INVERT CLK
INVERT DATA
INVERT CLK/DATA
DOT_PRODUCT
0
1
CROSS_PRODUCT
DOT_PRODUCTD
DS3 DS2 DS1
0
0
0
0
1
1
1
1
1
2
.
.
.
.
1023
TD10 .
TD6
TD5
TD4
TD3
TD2
TD1
DISCRIMINATOR_BW
0
0
.
.
.
.
1
0
0
.
.
.
.
0
0
0
.
.
.
.
1
0
0
.
.
.
.
0
0
0
.
.
.
.
1
0
1
.
.
.
.
0
1
0
.
.
.
.
0
1
2
.
.
.
.
660
.
.
.
.
.
.
.
DP1
Rx_INVERT
1
0
.
.
.
.
1
0
1
.
.
.
.
1
DB7
DB17
TD8
0
1
0
1
ADDRESS
BITS
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
DEMOD_SCHEME
2FSK LINEAR DEMODULATOR
2FSK CORRELATOR DEMODULATOR
3FSK DEMOD
4FSK DEMOD
RESERVED
RESERVED
RESERVED
RESERVED
07246-034
0
0
.
.
.
.
1
0
0
1
1
DB8
DB18
TD9
RI2 RI1
DEMOD_
SCHEME
RI1
DB19
DW1 DB20
TD10
DW2 DB21
IF_FILTER _
IFB2 IFB1 BW
0
0
9 kHz
1
0
13.5 kHz
0
1
18.5 kHz
1
1
INVALID
.
.
.
.
.
.
.
Rx_
INVERT
DISCRIMINATOR_BW
DW3 DB22
DW4 DB23
DW5 DB24
DW6 DB25
DW7 DB26
DW8 DB27
DW9 DB28
DB30
IFB1
DW10 DB29
DB31
IFB2
POST_DEMOD_BW
DP1
IF_FILTER_BW
REGISTER 4—DEMODULATOR SETUP REGISTER
Figure 67. Register 4—Demodulator Setup Register Map
•
where:
Round is rounded to the nearest integer.
Round4FSK is rounded to the nearest of the following integers:
32, 31, 28, 27, 24, 23, 20, 19, 16, 15, 12, 11, 8, 7, 4, 3.
fDEV is the transmit frequency deviation in Hz. For 4FSK,
fDEV is the frequency deviation used for the ±1 symbols
(that is, the inner frequency deviations).
To solve for DISCRIMINATOR_BW, use the following
equation:
DISCRIMINATOR_BW =
DEMOD CLK × K
400 × 10 3
where the maximum value = 660.
•
For 2FSK,
⎛ 100 × 10 3
K = Round ⎜⎜
⎝ f DEV
•
For 3FSK,
⎛ 100 × 10 3
K = Round ⎜⎜
⎝ 2 × f DEV
•
•
⎞
⎟
⎟
⎠
POST_DEMOD_BW =
⎞
⎟
⎟
⎠
211 × π × f CUTOFF
DEMOD CLK
where the cutoff frequency (fCUTOFF) of the post demodulator filter should typically be 0.75 × the data rate in
2FSK. In 3FSK, it should be set equal to the data rate, while
in 4FSK, it should be set equal to 1.6 × symbol rate.
For 4FSK,
⎛ 100 × 10 3
K = Round 4 FSK ⎜⎜
⎝ 4 × f DEV
Rx_INVERT (R4_DB[8:9]) and DOT_PRODUCT
(R4_DB7) need to be set as outlined in Table 17 and
Table 18.
⎞
⎟
⎟
⎠
Rev. 0 | Page 53 of 64
ADF7021-N
IF_CAL_COARSE
DB2
DB1
DB0
C3 (1)
C2 (0)
C1 (1)
DB3
DB5
IFD1
DB4
DB6
IFD2
CC1
DB7
IFD3
ADDRESS
BITS
C4 (0)
DB8
IFA6 DB19
IFD4
DB20
PM1
DB9
DB21
PM2
IFD5
DB22
PM3
IFD6 DB10
DB23
PM4
IFD7 DB11
IR_PHASE_
DB24 ADJUST_DIRECTION
PD1
IFD8 DB12
DB25
GM1
IFD9 DB13
DB26
GM2
IFA1 DB14
DB27
GM3
IFA2 DB15
DB28
GM4
IF_FILTER_DIVIDER
IF_FILTER_ADJUST
IFA3 DB16
DB29
GM5
IR_PHASE_
ADJUST_MAG
IFA4 DB17
IR GAIN_
ADJUST_I/Q
DB30
GQ1
IR_GAIN_
ADJUST_MAG
IFA5 DB18
IR_GAIN_
ADJUST_UP/DN
DB31
GA1
REGISTER 5—IF FILTER SETUP REGISTER
CC1 IF_CAL_COARSE
0
1
PM3
PM2
IR PHASE
PM1 PM1 ADJUST
0
0
0
.
1
0
0
0
.
1
0
0
1
.
1
0
1
2
...
15
0
1
0
.
1
IFD9 .
0
0
.
.
.
.
1
PD1 IR_PHASE_ADJUST_DIRECTION
0
1
0
0
0
.
1
0
0
0
.
1
GM3
IR_GAIN_
GM2 GM1 ADJUST_MAG
0
0
0
.
1
0
0
1
.
1
0
1
0
.
1
GQ1 IR_GAIN_ADJUST_I/Q
0
1
ADJUST I CH
ADJUST Q CH
GA1
IR_GAIN_ADJUST_UP/DN
0
1
GAIN
ATTENUATE
0
1
2
...
31
IFA6 IFA5 ...
0
0 ...
0
0 ...
0
0 ...
..
.. ...
0
1 ...
1
0 ...
1
0 ...
1
0 ...
1
. ...
1
1 ...
IF_FILTER_
IFD6 IFD5 IFD4 IFD3 IFD2 IFD1 DIVIDER
0
0
.
.
.
.
1
0
0
.
.
.
.
1
0
0
.
.
.
.
1
0
0
.
.
.
.
1
0
1
.
.
.
.
1
1
0
.
.
.
.
1
1
2
.
.
.
.
511
IFA2 IFA1 IF_FILTER_ADJUST
0
0
1
..
1
0
0
1
.
1
0
1
0
..
1
0
1
0
.
1
0
+1
+2
...
+31
0
–1
–2
...
–31
07246-035
GM5 GM4
ADJUST I CH
ADJUST Q CH
.
.
.
.
.
.
.
NO CAL
DO CAL
Figure 68. Register 5—IF Filter Setup Register Map
•
A coarse IF filter calibration is performed when the
IF_CAL_COARSE bit (R5_DB4) is set. If the IF_FINE_
CAL bit (R6_DB4) has been previously set, a fine IF filter
calibration is automatically performed after the coarse
calibration.
•
Set IF_FILTER_DIVIDER such that
XTAL
= 50 kHz
IF _ FILTER _ DIVIDER
•
IF_FILTER_ADJUST allows the IF fine filter calibration
result to be programmed directly on subsequent receiver
power-ups, thereby saving on the need to redo a fine filter
calibration in some instances. Refer to the Filter Bandwidth
Calibration Readback section for information about using
the IF_FILTER_ ADJUST bits.
•
R5_DB[20:31] are used for image rejection calibration. Refer
to the Image Rejection Calibration section for details on how
to program these parameters.
Rev. 0 | Page 54 of 64
ADF7021-N
IF_FINE_
CAL
DB4
DB0
DB5
LT1
FC1
C1 (0)
DB6
LT2
DB1
DB7
LT3
DB2
DB8
LT4
C2 (1)
DB9
LT5
C3 (1)
DB10
LT6
DB3
DB11
LT7
ADDRESS
BITS
C4 (0)
DB12
DB17
UT5
LT8
DB18
UT6
DB13
DB19
UT7
DB14
DB20
UT8
UT1
DB21
CD1
IF_CAL_LOWER_TONE_DIVIDE
UT2
DB22
CD2
DB15
DB23
CD3
UT3
DB24
CD4
DB16
DB25
CD5
UT4
DB26
CD6
IF_CAL_UPPER_TONE_DIVIDE
DB27
IF_CAL_DWELL_TIME
CD7
IRC2 DB29 IR_CAL_
SOURCE_
IRC1 DB28 DRIVE_LEVEL
IRD1 DB30 IR_CAL_
SOURCE ÷2
REGISTER 6—IF FINE CAL SETUP REGISTER
IRD1 IR_CAL_SOURCE ÷2
0
1
SOURCE ÷2 OFF
SOURCE ÷2 ON
UT8 UT7 ...
0
0
0
.
.
0
IR_CAL_SOURCE_
IRC2 IRC1 DRIVE_LEVEL
OFF
0
0
LOW
0
1
MED
1
0
HIGH
1
1
0
0
0
.
.
1
...
...
...
...
...
...
UT3
UT2
UT1
IF_CAL_UPPER_
TONE_DIVIDE
0
0
0
.
.
1
0
1
1
.
.
1
1
0
1
.
.
1
1
2
3
.
.
127
LT8 LT7 ... LT3
CD7
...
IF_CAL_
CD3 CD2 CD1 DWELL_TIME
0
0
0
.
.
1
...
...
...
...
...
...
0
0
0
.
.
1
1
0
1
.
.
1
1
2
3
.
.
127
0
0
0
.
.
1
...
...
...
...
...
...
0
0
0
.
.
1
IF_FINE_CAL
0
1
DISABLED
ENABLED
LT2
LT1
IF_CAL_LOWER_
TONE_DIVIDE
0
1
1
.
.
1
1
0
1
.
.
1
1
2
3
.
.
255
07246-036
0
1
1
.
.
1
0
0
0
.
.
1
FC1
Figure 69. Register 6—IF Fine Cal Setup Register Map
•
A fine IF filter calibration is set by enabling the
IF_FINE_CAL Bit (R6_DB4). A fine calibration is then
carried out only when Register 5 is written to and R5_DB4
is set.
•
The IF tone calibration time is the amount of time that is
spent at an IF calibration tone. It is dependent on the
sequencer clock. For best practice, is recommended to have
the IF tone calibration time be at least 500 μs.
IF Tone Calibration Time = IF _ CAL _ DWELL _ TIME
Lower Tone Frequency (kHz) =
SEQ CLK
XTAL
The total time for a fine IF filter calibration is
IF_CAL_LOWER_TONE_DIVIDE × 2
IF Tone Calibration Time × 10
Upper Tone Frequency (kHz) =
•
XTAL
IF_CAL_UPPER_TONE_DIVIDE × 2
It is recommended to place the lower tone and upper tone
as outlined in Table 24.
Table 24. IF Filter Fine Calibration Tone Frequencies
IF Filter
Bandwidth
9 kHz
13.5 kHz
18.5 kHz
Lower Tone
Frequency
78.1 kHz
79.4 kHz
78.1 kHz
Upper Tone
Frequency
116.3 kHz
116.3 kHz
119 kHz
Rev. 0 | Page 55 of 64
R6_DB[28:30] control the internal source for the image
rejection (IR) calibration. The IR_CAL_SOURCE_
DRIVE_LEVEL bits (R6_DB[28:29]) set the drive strength
of the source, whereas the IR_CAL_SOURCE_÷2 bit
(R6_DB30) allows the frequency of the internal signal
source to be divided by 2.
ADF7021-N
REGISTER 7—READBACK SETUP REGISTER
READBACK_
SELECT
CONTROL
BITS
ADC_
MODE
DB8
DB7
DB6
DB5
DB4
DB3
DB2
DB1
DB0
RB3
RB2
RB1
AD2
AD1
C4 (0)
C3 (1)
C2 (1)
C1 (1)
RB3 READBACK_SELECT
AD2 AD1 ADC_MODE
0
1
0
0
1
1
DISABLED
ENABLED
RB2 RB1 READBACK MODE
0
1
0
1
MEASURE RSSI
BATTERY VOLTAGE
TEMP SENSOR
TO EXTERNAL PIN
AFC WORD
ADC OUTPUT
FILTER CAL
SILICON REV
07246-037
0
0
1
1
0
1
0
1
Figure 70. Register 7—Readback Setup Register Map
•
•
Readback of the measured RSSI value is valid only in Rx
mode. Readback of the battery voltage, temperature sensor, or
voltage at the external pin is not valid in Rx mode.
•
To read back the battery voltage, the temperature sensor, or
the voltage at the external pin in Tx mode, users should
first power up the ADC using R8_DB8 because it is turned
off by default in Tx mode to save power.
Rev. 0 | Page 56 of 64
For AFC readback, use the following equations (see the
Readback Format section):
FREQ RB [Hz] = (AFC READBACK × DEMOD CLK)/218
VBATTERY = BATTERY VOLTAGE READBACK/21.1
VADCIN = ADCIN VOLTAGE READBACK/42.1
Temperature [°C] = −40 + (68.4 − TEMP READBACK) × 9.32
ADF7021-N
Tx/Rx_SWITCH_
ENABLE
LOG_AMP_
ENABLE
DEMOD_
ENABLE
ADC_
ENABLE
FILTER_
ENABLE
LNA/MIXER_
ENABLE
RESERVED
SYNTH_
ENABLE
DB15
PA_ENABLE_
Rx_MODE
COUNTER_
RESET
REGISTER 8—POWER-DOWN TEST REGISTER
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
PD7
SW1
LE1
PD6
Rx_RESET
DB14
DB13
CR1
PD4
PD5
PD3
PD1
CONTROL
BITS
DB3
CR1 COUNTER_RESET
0
1
DB2
DB1
DB0
C4 (1) C3 (0) C2 (0) C1 (0)
NORMAL
RESET
PD1
SYNTH_ENABLE
0
1
SYNTH OFF
SYNTH ON
CDR
RESET
DEMOD
RESET
PD7
PA (Rx MODE)
0
1
PA OFF
PA ON
SW1 Tx/Rx SWITCH
DEFAULT (ON)
OFF
LE1
LOG_AMP_ENABLE
0
1
LOG AMP OFF
LOG AMP ON
PD6
DEMOD_ENABLE
0
1
DEMOD OFF
DEMOD ON
LNA/MIXER_ENABLE
0
1
LNA/MIXER OFF
LNA/MIXER ON
PD4
FILTER_ENABLE
0
1
FILTER OFF
FILTER ON
PD5
ADC_ENABLE
0
1
ADC OFF
ADC ON
07246-038
0
1
PD3
Figure 71. Register 8—Power-Down Test Register Map
•
It is not necessary to write to this register under normal
operating conditions.
•
Rev. 0 | Page 57 of 64
For a combined LNA/PA matching network, R8_DB11
should always be set to 0, which enables the internal Tx/Rx
switch. This is the power-up default condition.
ADF7021-N
DB6
DB5
DB4
DB3
DB2
DB1
DB0
GL2
GL1
C4 (1)
C3 (0)
C2 (0)
C1 (1)
DB13
GH3
DB7
DB14
GH4
GL3
DB15
GH5
0
1
0
1
0
1
1
0
.
.
.
0
1
1
0
0
0
1
.
.
.
1
1
1
1
0
1
0
.
.
.
1
0
1
1
2
3
4
.
.
.
61
62
63
AGC_HIGH_
GH7 GH6 GH5 GH4 GH3 GH2 GH1 THRESHOLD
0
0
0
0
.
.
.
1
1
1
FG2 FG1 FILTER_GAIN
0
0
1
1
GL4
DB16
GH6
LOW
HIGH
DB8
DB17
GH7
0
1
GL5
DB18
GM1
FILTER_CURRENT
0
0
0
0
.
.
.
1
1
1
0
0
0
0
.
.
.
1
1
1
0
0
0
0
.
.
.
1
1
1
0
0
0
0
.
.
.
1
1
1
DEFAULT
REDUCED GAIN
FI1
DB9
DB19
GM2
AUTO AGC
MANUAL AGC
FREEZE AGC
RESERVED
LG1 LNA_MODE
0
1
DB10
DB20
LG1
800µA (DEFAULT)
GL6
DB21
LG2
LNA_BIAS
0
GL7
DB22
FG1
LI1
0
DB11
DB23
FG2
LI2
GH1
DB24
FI1
0
1
2
3
DB12
DB25
LG1
AGC_LOW_
GL7 GL6 GL5 GL4 GL3 GL2 GL1 THRESHOLD
AGC_MODE
DEFAULT
HIGH
ADDRESS
BITS
AGC_LOW_THRESHOLD
GH2
DB26
LI1
ML1 MIXER_LINEARITY
0
1
AGC_HIGH_THRESHOLD
DB27
FILTER_
CURRENT
LNA_MODE
AGC_
MODE
LI2
LNA_
GAIN
DB28
FILTER_
GAIN
ML1
LNA_
BIAS
MIXER_
LINEARITY
REGISTER 9—AGC REGISTER
8
24
72
INVALID
0
0
0
0
.
.
.
0
0
0
0
0
0
0
.
.
.
0
0
1
0
0
0
0
.
.
.
1
1
0
0
0
0
1
.
.
.
1
1
0
0
1
1
0
.
.
.
1
1
0
1
0
1
0
.
.
.
0
1
0
1
2
3
4
.
.
.
78
79
80
LG2 LG1 LNA_GAIN
0
1
0
1
3
10
30
INVALID
07246-039
0
0
1
1
Figure 72. Register 9—AGC Register Map
•
It is necessary to program this register only if AGC
settings, other than the defaults, are required.
•
AGC high and low settings must be more than 30 apart to
ensure correct operation.
•
In receive mode, AGC is set to automatic AGC by default
on power-up. The default thresholds are AGC_ LOW_
THRESHOLD = 30 and AGC_HIGH_ THRESHOLD = 70.
See the RSSI/AGC section for details.
•
An LNA gain of 30 is available only if LNA_MODE
(R9_DB25) is set to 0.
Rev. 0 | Page 58 of 64
ADF7021-N
...
MAX_AFC_
MA3 MA2 MA1 RANGE
0
0
0
0
.
.
.
1
1
1
...
...
...
...
...
...
...
...
...
...
0
0
0
1
.
.
.
1
1
1
1
0
1
0
.
.
.
1
0
1
DB7
DB6
DB5
M3
M2
M1
DB0
DB8
M4
C1 (0)
DB9
M5
DB1
DB10
M6
DB2
DB11
M7
C2 (1)
DB12
M8
C3 (0)
DB13
M9
DB3
DB14
M10
C4 (1)
DB15
M11
DB4
DB16
M12
AE1
DB17
KI1
DB18
ADDRESS
BITS
KI4
KI3
KI2
KI1
KI
AE1
AFC_EN
0
0
.
1
0
0
.
1
0
0
.
1
0
1
.
1
2^0
2^1
...
2^15
0
1
OFF
AFC ON
1
2
3
4
.
.
.
253
254
255
M12
...
M3
M2
M1
AFC_SCALING_
FACTOR
0
0
0
0
.
.
.
1
1
1
...
...
...
...
...
...
...
...
...
...
0
0
0
1
.
.
.
1
1
1
0
1
1
0
.
.
.
0
1
1
1
0
1
0
.
.
.
1
0
1
1
2
3
4
.
.
.
4093
4094
4095
07246-040
2^0
2^1
...
2^7
DB19
DB20
KI4
0
1
.
1
0
0
.
1
AFC_SCALING_FACTOR
KI2
DB21
KP1
KP1 KP
0
0
.
1
KI3
DB22
KP2
DB26
MA3
DB23
DB27
MA4
KP3
DB28
MA5
DB24
DB29
MA6
DB25
KP3 KP2
MA8
0
1
1
0
.
.
.
0
1
1
KI
MA1
DB30
MA7
KP
MA2
DB31
MA8
MAX_AFC_RANGE
AFC_EN
REGISTER 10—AFC REGISTER
Figure 73. Register 10—AFC Register Map
•
The AFC_SCALING_FACTOR can be expressed as
•
When the RF_DIVIDE_BY_2 (R1_DB18) is enabled, the
programmed AFC correction range is halved. The user
accounts for this halving by doubling the programmed
MAX_AFC_RANGE value.
•
Signals that are within the AFC pull-in range but outside
the IF filter bandwidth are attenuated by the IF filter. As a
result, the signal can be below the sensitivity point of the
receiver and, therefore, not detectable by the AFC.
⎛ 2 24 × 500 ⎞
⎟⎟
AFC _ SCALING _ FACTOR = Round ⎜⎜
⎝ XTAL ⎠
•
The settings for KI and KP affect the AFC settling time and
AFC accuracy. The allowable range of each parameter is
KI > 6 and KP < 7.
•
The recommended settings to give optimal AFC
performance are KI = 11 and KP = 4. To trade off between
AFC settling time and AFC accuracy, the KI and KP
parameters can be adjusted from the recommended settings
(staying within the allowable range) such that
AFC Correction Range = MAX_AFC_RANGE × 500 Hz
Rev. 0 | Page 59 of 64
ADF7021-N
DB0
C1 (1)
DB6
MT1
DB1
DB7
MT2
C2 (1)
DB8
SB1
DB2
DB9
SB2
DB3
DB10
SB3
C3 (0)
DB11
SB4
C4 (1)
DB12
SB5
DB4
DB13
SB6
PL1
DB14
SB7
DB5
DB15
SB8
CONTROL
BITS
PL2
MATCHING_
TOLERANCE
DB16
SB9
SB10 DB17
SB11 DB18
SB12 DB19
SB13 DB20
SB14 DB21
SB15 DB22
SB16 DB23
SB17 DB24
SB18 DB25
SB19 DB26
SB20 DB27
SB21 DB28
SB22 DB29
SB23 DB30
SYNC_BYTE_SEQUENCE
SB24 DB31
SYNC_BYTE_
LENGTH
REGISTER 11—SYNC WORD DETECT REGISTER
PL2
PL1
SYNC_BYTE_
LENGTH
0
0
1
1
0
1
0
1
12 BITS
16 BITS
20 BITS
24 BITS
0
0
1
1
0
1
0
1
ACCEPT 0 ERRORS
ACCEPT 1 ERROR
ACCEPT 2 ERRORS
ACCEPT 3 ERRORS
07246-041
MATCHING_
MT2 MT1 TOLERANCE
Figure 74. Register 11—Sync Word Detect Register Map
SWD_MODE
LOCK_
THRESHOLD_
MODE
REGISTER 12—SWD/THRESHOLD SETUP REGISTER
DB15
DB14
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
DB2
DB1
DB0
DP7
DP6
DP5
DP4
DP3
DP2
DP1
IL2
IL1
LM2
LM1
C4 (1)
C3 (1)
C2 (0)
C1 (0)
CONTROL
BITS
DP8
DATA_PACKET_LENGTH
DATA_PACKET_LENGTH
0
1
...
255
INVALID
1 BYTE
...
255 BYTES
SWD_MODE
0
1
2
3
SWD PIN LOW
SWD PIN HIGH AFTER NEXT SYNCWORD
SWD PIN HIGH AFTER NEXT SYNCWORD
FOR DATA PACKET LENGTH NUMBER OF BYTES
INTERRUPT PIN HIGH
3
THRESHOLD FREE RUNNING
LOCK THRESHOLD AFTER NEXT SYNCWORD
LOCK THRESHOLD AFTER NEXT SYNCWORD
FOR DATA PACKET LENGTH NUMBER OF BYTES
LOCK THRESHOLD
07246-042
LOCK_THRESHOLD_MODE
0
1
2
Figure 75. Register 12—SWD/Threshold Setup Register Map
Lock threshold locks the threshold of the envelope detector. This has the effect of locking the slicer in linear demodulation and locking
the AFC and AGC loops when using linear or correlator demodulation.
Rev. 0 | Page 60 of 64
ADF7021-N
REGISTER 13—3FSK/4FSK DEMOD REGISTER
...
0
0
0
0
.
.
1
...
...
...
...
...
...
...
0
0
0
0
.
.
1
OFF
1
2
3
.
.
127
VM2 VM1
0
0
1
1
0
1
0
1
DB2
DB1
DB0
C2 (0)
C1 (1)
DB4
ST1
C3 (1)
DB5
ST2
DB3
DB6
C4 (1)
DB7
ST3
CONTROL
BITS
ST4
DB8
ST5
DB9
DB12
PC1
0
1
0
1
.
.
1
DB10
DB13
VM1
0
0
1
1
.
.
1
ST6
DB14
VM2
3FSK_CDR_
THRESHOLD
ST7
DB15
VT1
VT1
DB11
DB16
VT2
VT2
3FSK/4FSK_
SLICER_THRESHOLD
VD1
DB17
VT3
PHASE_
CORRECTION
3FSK_VITERBI_
DETECTOR
DB18
DB19
VT4
DB20
VT6
VT7
VT3
VT5
DB21
3FSK_CDR_THRESHOLD
VT7
PTV1 DB22
PTV2 DB23
PTV4 DB25
PTV3 DB24
3FSK_PREAMBLE_
TIME_VALIDATE
VITERBI_
PATH_
MEMORY
Refer to the Receiver Setup section for information about programming these settings.
3FSK_VITERBI_
VD1 DETECTOR
0
DISABLED
1
ENABLED
PHASE_
PC1 CORRECTION
0
DISABLED
1
ENABLED
VITERBI_PATH _
MEMORY
4 BITS
6 BITS
8 BITS
32 BITS
ST7
...
ST3
ST2
ST1
SLICER
THRESHOLD
0
0
0
0
.
.
1
...
...
...
...
...
...
...
0
0
0
0
.
.
1
0
0
1
1
.
.
1
0
1
0
1
.
.
1
OFF
1
2
3
.
.
127
3FSK_PREMABLE_
PTV4 PTV3 PTV2 PTV1 TIME_VALIDATE
0
0
0
0
.
.
1
0
0
1
1
.
.
1
0
1
0
1
.
.
1
0
1
2
3
.
.
15
07246-043
0
0
0
0
.
.
1
Figure 76. Register 13—3FSK/4FSK Demod Register Map
Rev. 0 | Page 61 of 64
ADF7021-N
PULSE_EXTENSION
TEST_DAC_GAIN
0
1
2
3
0
1
...
15
NO PULSE EXTENSION
EXTENDED BY 1
EXTENDED BY 2
EXTENDED BY 3
TEST_
TDAC_EN
DB13
DB12
DB11
DB10
DB9
DB8
DB7
DB6
DB5
DB4
DB3
DB2
DB1
DB0
TO8
TO7
TO6
TO5
TO4
TO3
TO2
TO1
TE1
C4 (1)
C3 (1)
C2 (1)
C1 (0)
ADDRESS
BITS
TO9
TO10 DB14
TO11 DB15
TO12 DB16
TO13 DB17
TO14 DB18
TO15 DB19
DB21
DB23
TG3
TG1
DB24
TG4
TEST_DAC_OFFSET
TO16 DB20
DB25
ER1
DB22
DB26
LEAKAGE =
2^–8
2^–9
2^–10
2^–11
2^–12
2^–13
2^–14
2^–15
TG2
PULSE_
EXTENSION
DB27
EF1
ED_LEAK_FACTOR
0
1
2
3
4
5
6
7
TEST_DAC_GAIN
ER2
DB29
EF3
DB28
DB30
PE1
EF2
DB31
PE2
ED_LEAK_
FACTOR
ED_PEAK_
RESPONSE
REGISTER 14—TEST DAC REGISTER
NO GAIN
× 2^1
...
× 2^15
ED_PEAK_RESPONSE
FULL RESPONSE TO PEAK
0.5 RESPONSE TO PEAK
0.25 RESPONSE TO PEAK
0.125 RESPONSE TO PEAK
07246-044
0
1
2
3
Figure 77. Register 14—Test DAC Register Map
The demodulator tuning parameters, PULSE_EXTENSION,
ED_LEAK_FACTOR, and ED_PEAK_RESPONSE, can be
enabled only by setting R15_DB[4:7] to 0x9.
While the correlators and filters are clocked by DEMOD CLK,
CDR CLK clocks the test DAC. Note that although the test
DAC functions in regular user mode, the best performance is
achieved when the CDR CLK is increased to or above the
frequency of DEMOD CLK. The CDR block does not function
when this condition exists.
Using the Test DAC to Implement Analog FM DEMOD
and Measuring SNR
For detailed information about using the test DAC, see
Application Note AN-852.
The test DAC allows the post demodulator filter out for both
linear and correlator demodulators to be viewed externally. The
test DAC also takes the 16-bit filter output and converts it to a
high frequency, single-bit output using a second-order, error
feedback Σ-Δ converter. The output can be viewed on the SWD
pin. This signal, when filtered appropriately, can then be used to
do the following:
•
Monitor the signals at the FSK post demodulator filter
output. This allows the demodulator output SNR to be
measured. Eye diagrams of the received bit stream can also
be constructed to measure the received signal quality.
•
Provide analog FM demodulation.
Programming Register 14 enables the test DAC. Both the
linear and correlator/demodulator outputs can be multiplexed
into the DAC.
Register 14 allows a fixed offset term to be removed from the
signal (to remove the IF component in the ddt case). It also has
a signal gain term to allow the usage of the maximum dynamic
range of the DAC.
Rev. 0 | Page 62 of 64
ADF7021-N
DB20
DB19
DB18
DB17
DB16
DB15 PFD/CP_TEST_
MODES
DB14
DB13
DB12
PM1
CM3
CM2
CM1
PC3
PC2
PC1
SD3
SD2
DB11
DB10
DB9
DB8
DB7
DB6
DB5
SD1
TM3
TM2
TM1
RT4
RT3
RT2
DB0
DB21
PM2
DB1
DB22
PM3
C1 (1)
DB23
PM4
ADDRESS
BITS
DB2
DB24
AM1
Rx_TEST_
MODES
C2 (1)
DB25
AM2
Tx_TEST_
MODES
C3 (1)
DB26
AM3
Σ-Δ_TEST_
MODES
DB3
DB27
CLK_MUX
DB4
DB28
FH1
AM4
PLL_TEST_
MODES
RT1
DB29
RD1
ANALOG_TEST_
MODES
C4 (1)
REG 1_PD
FORCE_LD
HIGH
DB30
CO1
CAL_
OVERRIDE
DB31
CO2
REGISTER 15—TEST MODE REGISTER
CAL_OVERRIDE
0
1
2
3
AUTO CAL
OVERRIDE GAIN
OVERRIDE BW
OVERRIDE BW AND GAIN
PFD/CP_TEST_MODES
0
1
2
3
4
5
6
7
REG1_PD
0
1
NORMAL
PWR DWN
DEFAULT, NO BLEED
(+VE) CONSTANT BLEED
(–VE) CONSTANT BLEED
(–VE) PULSED BLEED
(–VE) PULSE BLD, DELAY UP?
CP PUMP UP
CP TRI-STATE
CP PUMP DN
FORCE_LD_HIGH
0
1
Σ-Δ_TEST_MODES
NORMAL
FORCE
0
1
2
3
4
5
6
7
ANALOG_TEST_MODES
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
BAND GAP VOLTGE
40µA CURRENT FROM REG4
FILTER I CHANNEL: STAGE 1
FILTER I CHANNEL: STAGE 2
FILTER I CHANNEL: STAGE 1
FILTER Q CHANNEL: STAGE 1
FILTER Q CHANNEL: STAGE 2
FILTER Q CHANNEL: STAGE 1
ADC REFERENCE VOLTAGE
BIAS CURRENT FROM RSSI 5µA
FILTER COARSE CAL OSCILLATOR O/P
ANALOG RSSI I CHANNEL
OSET LOOP +VE FBACK V (I CH)
SUMMED O/P OF RSSI RECTIFIER+
SUMMED O/P OF RSSI RECTIFIER–
BIAS CURRENT FROM BB FILTER
DEFAULT, 3RD ORDER SD, NO DITHER
1ST ORDER SD
2ND ORDER SD
DITHER TO FIRST STAGE
DITHER TO SECOND STAGE
DITHER TO THIRD STAGE
DITHER × 8
DITHER × 32
Tx_TEST_MODES
0
1
2
3
4
5
6
NORMAL OPERATION
Tx CARRIER ONLY
Tx +VE TONE ONLY
Tx –VE TONE ONLY
Tx "1010" PATTERN
Tx PN9 DATA, AT PROGRAMED RATE
Tx SYNC BYTE REPEATEDLY
Rx_TEST_MODES
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
PLL_TEST_MODES
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
NORMAL OPERATION
R DIV
N DIV
RCNTR/2 ON MUXOUT
NCNTR/2 ON MUXOUT
ACNTR TO MUXOUT
PFD PUMP UP TO MUXOUT
PFD PUMP DN TO MUXOUT
SDATA TO MUXOUT (OR SREAD?)
ANALOG LOCK DETECT ON MUXOUT
END OF COARSE CAL ON MUXOUT
END OF FINE CAL ON MUXOUT
FORCE NEW PRESCALER CONFIG.
FOR ALL N
TEST MUX SELECTS DATA
LOCK DETECT PERCISION
RESERVED
NORMAL
SCLK, SDATA -> I, Q
REVERSE I,Q
I,Q TO TxRxCLK, TxRxDATA
3FSK SLICER ON TxRxDATA
CORRELATOR SLICER ON TxRxDATA
LINEAR SLICER ON RXDATA
SDATA TO CDR
ADDITIONAL FILTERING ON I, Q
ENABLE REG 14 DEMOD PARAMETERS
POWER DOWN DDT AND ED IN T/4 MODE
ENVELOPE DETECTOR WATCHDOG DISABLED
RESERVED
PROHIBIT CALACTIVE
FORCE CALACTIVE
ENABLE DEMOD DURING CAL
CLK MUXES ON CLKOUT PIN
NORMAL, NO OUTPUT
DEMOD CLK
CDR CLK
SEQ CLK
BB OFFSET CLK
SIGMA DELTA CLK
ADC CLK
TxRxCLK
07246-045
0
1
2
3
4
5
6
7
Figure 78. Register 15—Test Mode Register Map
•
•
Analog RSSI can be viewed on the Test_A pin by setting
ANALOG_TEST_MODES to 11.
Tx_TEST_MODES can be used to enable test modulation.
•
Rev. 0 | Page 63 of 64
The CDR block can be bypassed by setting Rx_TEST_
MODES to 4, 5, or 6, depending on the demodulator used.
ADF7021-N
OUTLINE DIMENSIONS
7.00
BSC SQ
0.60 MAX
37
36
PIN 1
INDICATOR
TOP
VIEW
48
4.25
4.10 SQ
3.95
(BOTTOM VIEW)
25
24
12
13
0.25 MIN
5.50
REF
0.80 MAX
0.65 TYP
12° MAX
PIN 1
INDICATOR
1
EXPOSED
PAD
6.75
BSC SQ
0.50
0.40
0.30
1.00
0.85
0.80
0.30
0.23
0.18
0.60 MAX
0.05 MAX
0.02 NOM
0.50 BSC
SEATING
PLANE
0.20 REF
COPLANARITY
0.08
COMPLIANT TO JEDEC STANDARDS MO-220-VKKD-2
Figure 79. 48-Lead Lead Frame Chip Scale Package [LFCSP_VQ]
7 mm × 7 mm Body, Very Thin Quad
(CP-48-3)
Dimensions shown in millimeters
ORDERING GUIDE
Model
ADF7021-NBCPZ 1
ADF7021-NBCPZ-RL1
ADF7021-NBCPZ-RL71
ADF7021-NDF
EVAL-ADF70XXMBZ21
EVAL-ADF7021-NDBIZ1
EVAL-ADF7021-NDBEZ1
EVAL-ADF7021-NDBZ21
EVAL-ADF7021-NDBZ51
1
Temperature Range
−40°C to +85°C
−40°C to +85°C
−40°C to +85°C
−40°C to +85°C
Package Description
48-Lead Lead Frame Chip Scale Package [LFCSP_VQ]
48-Lead Lead Frame Chip Scale Package [LFCSP_VQ]
48-Lead Lead Frame Chip Scale Package [LFCSP_VQ]
Die on Film
Evaluation Platform Mother Board
426 MHz to 429 MHz Daughter Board
426 MHz to 429 MHz Daughter Board
860 MHz to 870 MHz Daughter Board
Matching Unpopulated Daughter Board
Z = RoHS Compliant Part.
©2008 Analog Devices, Inc. All rights reserved. Trademarks and
registered trademarks are the property of their respective owners.
D07246-0-2/08(0)
Rev. 0 | Page 64 of 64
Package Option
CP-48-3
CP-48-3
CP-48-3