NSC LMX2470

LMX2470
2.6 GHz Delta-Sigma Fractional-N PLL with 800 MHz
Integer-N PLL
General Description
Features
The LMX2470 is a low power, high performance delta-sigma
fractional-N PLL with an auxiliary integer-N PLL. The device
is fabricated using National Semiconductor’s advanced BiCMOS process.
With delta-sigma architecture, fractional spur compensation
is achieved with noise shaping capability of the delta-sigma
modulator and the inherent low pass filtering of the PLL loop
filter. Fractional spurs at lower frequencies are pushed to
higher frequencies outside the loop bandwidth. Unlike analog compensation, the digital feedback techniques used in
the LMX2470 are highly resistant to changes in temperature
and variations in wafer processing. With delta-sigma architecture, the ability to push close in spur and phase noise
energy to higher frequencies is a direct function of the modulator order. The higher the order, the more this energy can be
spread to higher frequencies. The LMX2470 has a programmable modulator up to order four, which allows the designer
to select the optimum modulator order to fit the phase noise,
spur, and lock time requirements of the system.
Programming is fast and simple. Serial data is transferred
into the LMX2470 via a three line MICROWIRE interface
(Data, Clock, Load Enable). Nominal supply voltage is 2.5 V.
The LMX2470 features a typical current consumption of 4.1
mA at 2.5 V. The LMX2470 is available in a 24 lead 3.5 X 4.5
X 0.6 mm package.
n
n
n
n
n
n
n
n
n
n
n
Low in-band phase noise and low fractional spurs
12 bit or 22 bit selectable fractional modulus
Up to 4th order programmable delta-sigma modulator
Enhanced Anti-Cycle Slip Fastlock Circuitry
Fastlock
Cycle slip reduction
Integrated timeout counters
Digital lock detect output
Prescalers allow wide range of N values
RF PLL: 16/17/20/21
IF PLL: 8/9 or 16/17
Crystal Reference Frequency up to 110 MHz
On-chip crystal reference frequency doubler.
Phase Comparison Frequency up to 30 MHz
Hardware and software power-down control
Ultra low consumption: ICC = 4.1 mA (typical)
Applications
n Cellular Phones and Base Stations
CDMA, WCDMA, GSM/GPRS, TDMA, EDGE, PDC
n Applications requiring fine frequency resolution
n Satellite and Cable TV Tuners
n WLAN Standards
Functional Block Diagram
20059301
© 2003 National Semiconductor Corporation
DS200593
www.national.com
LMX2470 2.6 GHz Delta-Sigma Fractional-N PLL with 800 MHz Integer-N PLL
April 2003
LMX2470
Connection Diagram
24-Pin CSP (SLE) Package
20059322
Pin Descriptions
Pin #
Pin
Name
I/O
1
CPoutRF
O
RF charge pump output.
2
GND
-
Ground
3
GND
-
RF Ground
4
GND
-
Ground for RF PLL digital circuitry.
5
FinRF
I
RF prescaler input. Small signal input from the VCO.
6
FinRF*
I
RF prescaler complimentary input. For single-ended operation, a bypass capacitor should
be placed as close as possible to this pin and be connected directly to the ground plane.
7
VccRF
8
EN
I
Chip enable input. High impedance CMOS input. When EN is high, the chip is powered
up, otherwise it is powered down.
9
ENOSC
I
This pin should be grounded for normal operation.
10
CLK
I
MICROWIRE Clock. High impedance CMOS Clock input. Data for the various counters is
clocked into the 24 bit shift register on the rising edge.
11
DATA
I
MICROWIRE Data. High impedance binary serial data input.
12
LE
13
Ftest/LD
O
Test frequency output / Lock Detect
14
VddIF
-
Digital power supply for IF PLL
15
VccIF
-
IF power supply voltage input. Must be equal to VccRF. Input may range from 2.25 V to
2.75 V. Bypass capacitors should be placed as close as possible to this pin and be
connected directly to the ground plane.
16
GND
-
Ground for RF PLL digital circuitry.
17
FinIF
I
IF prescaler input. Small signal input from the VCO.
18
GND
-
Digital ground for IF PLL
19
CPoutIF
O
IF PLL charge pump output
Pin Description
RF PLL power supply voltage input. Must be equal to VccIF . May range from 2.25V to
2.75V. Bypass capacitors should be placed as close as possible to this pin and be
connected directly to the ground plane.
MICROWIRE Load Enable. High impedance CMOS input. Data stored in the shift
registers is loaded into the internal latches when LE goes HIGH
20
FLoutIF
O
IF Fastlock Output. Also functions as Programmable TRI-STATE CMOS output.
21
OSCout*
I/O
Complementary reference input or oscillator output.
22
OSCin
I
Reference input
23
VddRF
-
Digital power supply for RF PLL
24
FLoutRF
O
RF Fastlock Output. Also functions as Programmable TRI-STATE CMOS output.
www.national.com
2
(Notes 1, 2)
Parameter
Value
Symbol
Power Supply Voltage
Min
Typ
Units
Max
VCC
-0.3
3.0
V
VDD
VCC
VCC
V
Vi
-0.3
VCC + 0.3
V
Storage Temperature Range
Ts
-65
+150
˚C
Lead Temperature (Solder 4 sec.)
TL
+260
˚C
Voltage on any pin with GND =VSS = 0V
Recommended Operating Conditions
Parameter
Value
Symbol
Power Supply Voltage (Note 1)
Operating Temperature
Min
Typ
Units
Max
VCC
2.25
2.75
VDD
VCC
VCC
V
TA
-40
+85
˚C
V
Note 1: “Absolute Maximum Ratings” indicate limits beyond which damage to the device may occur. "Recommended Operating Conditions" indicate conditions for
which the device is intended to be functional, but do not guarantee specific performance limits. For guaranteed specifications and test conditions, see the Electrical
Characteristics. The guaranteed specifications apply only for the test conditions listed. Note also that these maximum ratings imply that the voltage at all the power
supply pins of VccRF, VccIF, VddRF, and VddIF are the same. VCCwill be used to refer to the voltage at these pins.
Note 2: This Device is a high performance RF integrated circuit with an ESD rating < 2 kV and is ESD sensitive. Handling and assembly of this device should only
be done at ESD-free workstations.
Electrical Characteristics
Symbol
(VCC = 2.5V; -40˚C ≤ TA ≤ +85˚C unless otherwise specified)
Parameter
Conditions
Value
Min
Typ
Max
Units
Icc PARAMETERS
ICCRF
ICCIF
ICCTOTAL
ICCPD
Power Supply Current, RF
Synthesizer
IF PLL OFF
RF PLL ON
Charge Pump TRI-STATE
OSC=0
2.7
3.9
mA
Power Supply Current, IF
Synthesizer
IF PLL ON
RF PLL OFF
Charge Pump TRI-STATE
OSC=0
1.4
2.3
mA
Power Supply Current,
Entire Synthesizer
IF PLL ON
RF PLL ON
Charge Pump TRI-STATE
OSC=0
4.1
6.0
mA
Power Down Current
EN = ENOSC = 0V
CLK, DATA, LE = 0V
1
10
µA
RF SYNTHESIZER PARAMETERS
fFinRF
Operating Frequency
500
2600
MHz
pFinRF
Input Sensitivity
-15
0
dBm
fCOMP
Phase Detector Frequency
30
MHz
ICPoutRFSRCE
RF Charge Pump Source
Current
RF_CPG = 0
VCPoutRF=VCC /2
100
µA
RF_CPG = 1
VCPoutRF=VCC/2
200
µA
...
µA
1600
µA
...
RF_CPG = 15
VCPoutRF=VCC/2
3
www.national.com
LMX2470
Absolute Maximum Ratings
LMX2470
Electrical Characteristics (VCC = 2.5V; -40˚C ≤ TA ≤ +85˚C unless otherwise specified)
Symbol
Parameter
Conditions
(Continued)
Value
Min
Typ
Max
Units
RF SYNTHESIZER PARAMETERS
ICPoutRFSINK
RF Charge Pump Sink
Current
RF_CPG = 0
VCPoutRF=VCC/2
-100
µA
RF_CPG = 1
VCPoutRF=VCC/2
-200
µA
...
µA
-1600
µA
...
RF_CPG = 15
VCPoutRF=VCC/2
RF Charge Pump
TRI-STATE Current
Magnitude
0.4 ≤ VCPoutRF ≤ VCC -0.4
ICPoutRF%MIS
RF CP Sink vs. CP Source
Mismatch
ICPoutRF%V
ICPoutRF%TEMP
ICPoutRFTRI
2
10
nA
VCPoutRF = VCC/2
TA = 25˚C
3
10
%
RF CP Current vs. CP
Voltage
0.4 ≤ VCPoutRF ≤ VCC -0.4
TA = 25˚C
5
15
%
RF CP Current vs.
Temperature
VCPoutRF = VCC/2
8
%
IF SYNTHESIZER PARAMETERS
fFinIF
Operating Frequency
75
800
MHz
pFinIF
IF Input Sensitivity
-15
0
dBm
10
MHz
fCOMP
Phase Detector Frequency
ICPoutIFSRCE
IF Charge Pump Source
Current
ICPoutIFSINK
IF Charge Pump Sink
Current
IF_CPG = 0
VCPoutIF = VCC /2
1
mA
IF_CPG = 1
VCPoutIF = VCC/2
4
mA
IF_CPG = 0
VCPoutIF = VCC/2
-1
mA
IF_CPG = 1
VCPoutIF = VCC/2
-4
mA
IF Charge Pump
TRI-STATE Current
Magnitude
0.4 ≤ VCPoutIF ≤ VCC RF -0.4
ICPoutIF%MIS
IF CP Sink vs. CP Source
Mismatch
VCPoutIF = VCC/2
TA = 25˚C
3
ICPoutIF%V
IF CP Current vs. CP
Voltage
0.4 ≤ VCPoutIF ≤ VCC -0.4
TA = 25˚C
8
ICPoutIF%TEMP
IF CP Current vs.
Temperature
VCPoutIF = VCC/2
ICPoutIFTRI
2
10
nA
%
15
8
%
%
OSCILLATOR PARAMETERS
fOSCin
Oscillator Operating
Frequency
vOSCin
Oscillator Input Sensitivity
IOSCin
Oscillator Input Current
OSC2X = 0
5
OSC2X = 1
110
MHz
MHz
5
20
0.5
VCC
V
-100
100
µA
1.6
VCC
V
0.4
V
-1.0
1.0
µA
-1.0
1.0
µA
DIGITAL INTERFACE (DATA, CLK, LE, EN, ENRF, Ftest/LD, FLoutRF, FLoutIF)
VIH
High-Level Input Voltage
VIL
Low-Level Input Voltage
IIH
High-Level Input Current
IIL
Low-Level Input Current
VIL = 0 V
VOH
High-Level Output Voltage
IOH = -500 µA
www.national.com
VIH = VCC
VCC-0.4
4
V
Symbol
Parameter
Conditions
(Continued)
Value
Min
Typ
Max
Units
DIGITAL INTERFACE (DATA, CLK, LE, EN, ENRF, Ftest/LD, FLoutRF, FLoutIF)
VOL
Low-Level Output Voltage
IOL = 500 µA
0.4
V
MICROWIRE INTERFACE TIMING
TCS
Data to Clock Set Up Time
See Microwire Input Timing
50
ns
TCH
Data to Clock Hold Time
See Microwire Input Timing
10
ns
TCWH
Clock Pulse Width High
See Microwire Input Timing
50
ns
TCWL
Clock Pulse Width Low
See Microwire Input Timing
50
ns
TES
Clock to Load Enable Set
Up Time
See Microwire Input Timing
50
ns
TEW
Load Enable Pulse Width
See Microwire Input Timing
50
ns
RF Synthesizer Normalized
Phase Noise Contribution
(Note 3)
RF_CPG = 0
-200
dBc/Hz
RF_CPG = 3
-206
dBc/Hz
RF_CPG = 7
-208
dBc/Hz
RF_CPG = 15
-210
dBc/Hz
Applies to both low and high
current modes
-209
dBc/Hz
PHASE NOISE
LF1HzRF
LF1HzIF
IF Synthesizer Normalized
Phase Noise Contribution
(Note 3)
Note 3: Normalized Phase Noise Contribution is defined as: LN(f) = L(f) – 20log(N) – 10log(fCOMP) where L(f) is defined as the single side band phase noise
measured at an offset frequency, f, in a 1 Hz Bandwidth. The offset frequency, f, must be chosen sufficiently smaller than the PLL loop bandwidth, yet large enough
to avoid substantial phase noise contribution from the reference source. The offset chosen was 4 KHz.
MICROWIRE INPUT TIMING DIAGRAM
20059372
5
www.national.com
LMX2470
Electrical Characteristics (VCC = 2.5V; -40˚C ≤ TA ≤ +85˚C unless otherwise specified)
LMX2470
Typical Performance Characteristics : Sensitivity
(Note 4)
RF N Counter Sensitivity
TA = 25˚C
20059345
RF N Counter Sensitivity
VCC = 2.5 V
20059346
www.national.com
6
(Note 4)
LMX2470
Typical Performance Characteristics : Sensitivity
(Continued)
IF N Counter Sensitivity
TA = 25˚C
20059347
IF N Counter Sensitivity
VCC = 2.5 V
20059348
7
www.national.com
LMX2470
Typical Performance Characteristics : Sensitivity
(Note 4)
(Continued)
OSCin Counter Sensitivity
OSC=0
TA = 25˚C
20059349
OSCin Counter Sensitivity
OSC=0
VCC= 2.5 V
20059350
www.national.com
8
LMX2470
Typical Performance Characteristic : FinRF Input Impedance
(Note 4)
20059353
FinRF Input Impedance
Frequency (MHz)
Real (Ohms)
Imaginary (Ohms)
500
214
-255
600
175
-245
700
144
-230
800
118
-216
900
98
-203
1000
80
-189
1100
69
-177
1200
57
-165
1300
48
-153
1400
39
-141
1500
34
-130
1600
28
-119
1700
24
-110
1800
20
-101
1900
17
-94
2000
14
-87
2100
13
-82
2200
11
-77
2300
10
-72
2400
8
-67
2500
7
-62
2600
7
-56
2700
7
-53
2800
7
-46
2900
7
-41
3000
7
-39
9
www.national.com
LMX2470
Typical Performance Characteristic : FinIF Input Impedance
(Note 4)
20059354
FinIF Input Impedance
www.national.com
Freqeuncy (MHz)
Real (Ohms)
Imaginary (Ohms)
50
580
-313
100
500
-273
150
445
-250
200
410
-256
250
378
-259
300
349
-264
350
322
-267
400
297
-270
450
274
-271
500
253
-272
550
232
-272
600
214
-269
650
198
-265
700
184
-262
750
170
-259
800
157
-256
900
151
-253
1000
143
-250
10
LMX2470
Typical Performance Characteristic : OSCin Input Impedance
(Note 4)
20059355
OSCin Input Impedance
Frequency (MHz)
Real (Ohms)
Imaginary (Ohms)
Magnitude (Ohms)
50
2200
-4700
5189
10
710
-2700
2792
20
229
-1500
1517
30
133
-988
997
40
93
-752
758
50
74
-606
611
60
62
-505
509
70
53
-435
438
80
49
-382
385
90
45
-341
344
100
42
-309
312
110
40
-282
285
11
www.national.com
LMX2470
Typical Performance Characteristics : Currents
(Note 4)
Total Current Consumption
OSC=0
20059359
Powerdown Current
EN = LOW
20059361
www.national.com
12
(Note 4)
LMX2470
Typical Performance Characteristics : Currents
(Continued)
RF Charge Pump Current
VCC = 2.5 Volts
20059367
IF Charge Pump Current
VCC = 2.5 Volts
20059365
13
www.national.com
LMX2470
Typical Performance Characteristics : Currents
(Note 4)
(Continued)
Charge Pump Leakage
RF PLL
20059364
Charge Pump Leakage
IF PLL
20059363
Note 4: Typical performance characteristics do not imply any sort of guarantee. Guaranteed specifications are in the electrical characteristics section.
www.national.com
14
LMX2470
Bench Test Setups
20059369
Charge Pump Current Measurement Procedure
The above block diagram shows the test procedure for testing the RF and IF charge pumps. These tests include absolute current level, mismatch, and leakage. In order to measure the charge pump currents, a signal is applied to the high
frequency input pins. The reason for this is to guarantee that
the phase detector gets enough transitions in order to be
able to change states. If no signal is applied, it is possible
that the charge pump current reading will be low due to the
fact that the duty cycle is not 100%. The OSCin Pin is tied to
the supply. The charge pump currents can be measured by
simply programming the phase detector to the necessary
polarity. For instance, in order to measure the RF charge
pump current, a 10 MHz signal is applied to the FinRF pin.
The source current can be measured by setting the RF PLL
phase detector to a positive polarity, and the sink current can
be measured by setting the phase detector to a negative
polarity. The IF PLL currents can be measured in a similar
way. Note that the magnitude of the RF and IF PLL charge
pump currents are also controlled by the RF_CPG and IF_CPG bits. Once the charge pump currents are known, the
mismatch can be calculated as well. In order to measure
leakage currents, the charge pump current is set to a TRISTATE mode by enabling the counter reset bits. This is
RF_RST for the RF PLL and IF_RST for the IF PLL. The
table below shows a summary of the various charge pump
tests.
Current Test
RF_CPG
RF_CPP
RF_CPT
IF_CPG
IF_CPP
IF_CPT
RF Source
0 to 15
0
0
X
X
X
RF Sink
0 to 15
1
0
X
X
X
RF TRI-STATE
X
X
1
X
X
X
IF Source
X
X
X
0 to 1
0
0
IF Sink
X
X
X
0 to 1
1
0
IF TRI-STATE
X
X
X
X
X
1
15
www.national.com
LMX2470
Bench Test Setups
(Continued)
20059370
Frequency Input Pin
DC Blocking Capacitor
Corresponding Counter
OSCin
1000 pF
RF_R / 2
50
14
0
FinRF
100 pF
RF_N / 2
500
15
X
FinIF
100 pF
IF_N / 2
500
13
X
Sensitivity Measurement Procedure
Sensitivity is defined as the power level limits beyond which
the output of the counter being tested is off by 1 Hz or more
of its expected value. It is typically measured over frequency,
voltage, and temperature. In order to test sensitivity, the
MUX[3:0] word is programmed to the appropriate value. The
counter value is then programmed to a fixed value and a
frequency counter is set to monitor the frequency of this pin.
The expected frequency at the Ftest/LD pin should be the
signal generator frequency divided by twice the corresponding counter value. The factor of two comes in because the
LMX2470 has a flip-flop which divides this frequency by two
to make the duty cycle 50% in order to make it easier to read
with the frequency counter. The frequency counter input
impedance should be set to high impedance. In order to
www.national.com
Default Counter Value
MUX Value
OSC
perform the measurement, the temperature, frequency, and
voltage is set to a fixed value and the power level of the
signal is varied. Note that the power level at the part is
assumed to be 4 dB less than the signal generator power
level. This accounts for 1 dB for cable losses and 3 dB for the
pad. The power level range where the frequency is correct at
the Ftest/LD pin to within 1 Hz accuracy is recorded for the
sensitivity limits. The temperature, frequency, and voltage
can be varied in order to produce a family of sensitivity
curves. Since this is an open-loop test, the charge pump is
set to TRI-STATE and the unused side of the PLL (RF or IF)
is powered down when not being tested. For this part, there
are actually four frequency input pins, although there is only
one frequency test pin (Ftest/LD). The conditions specific to
each pin are show above.
16
LMX2470
Bench Test Setups
(Continued)
20059371
Input Impedance Measurement Procedure
The above block diagram shows the test procedure measuring the input impedance for the LMX2470. This applies to the
FinRF, FinIF, and OSCin pins. The basic test procedure is to
calibrate the network analyzer, ensure that the part is powered up, and then measure the input impedance. The network analyzer can be calibrated by using either calibration
standards or by soldering resistors directly to the evaluation
board. An open can be implemented by putting no resistor, a
short can be implemented by using a 0 ohm resistor, and a
short can be implemented by using two 100 ohm resistors in
parallel. Note that no DC blocking capacitor is used for this
test procedure. This is done with the PLL removed from the
PCB. This requires the use of a clamp down fixture that may
not always be generally available. If no clamp down fixture is
available, then this procedure can be done by calibrating up
to the point where the DC blocking capacitor usually is, and
then adding a 0 ohm resistor back for the actual measure-
ment. Once that the network analyzer is calibrated, it is
necessary to ensure that the PLL is powered up. This can be
done by toggling the power down bits (RF_PD and IF_PD)
and observing that the current consumption indeed increases when the bit is disabled. Sometimes it may be
necessary to apply a signal to the OSCin pin in order to
program the part. If this is necessary, disconnect the signal
once it is established that the part is powered up. It is useful
to know the input impedance of the PLL for the purposes of
debugging RF problems and designing matching networks.
Another use of knowing this parameter is make the trace
width on the PCB such that the input impedance of this trace
matches the real part of the input impedance of the PLL
frequency of operation. In general, it is good practice to keep
trace lengths short and make designs that are relatively
resistant to variations in the input impedance of the PLL.
17
www.national.com
LMX2470
frequencies, although this is application specific. The current
consumption may also slightly increase with higher phase
detector frequencies.
Functional Description
1.0 GENERAL
The basic phase-lock-loop (PLL) configuration consists of a
high-stability crystal reference oscillator, a frequency synthesizer such as the National Semiconductor LMX2470, a voltage controlled oscillator (VCO), and a passive loop filter. The
frequency synthesizer includes a phase detector, current
mode charge pump, as well as programmable reference [R]
and feedback [N] frequency dividers. The VCO frequency is
established by dividing the crystal reference signal down via
the R counter to obtain a frequency that sets the comparison
frequency. This comparison frequency, fCOMP, is input of a
phase/frequency detector and compared with another signal,
fN, the feedback signal, which was obtained by dividing the
VCO frequency down by way of the N counter and fractional
circuitry. The phase/frequency detector’s current source outputs a charge into the loop filter, which is then converted into
the VCO’s control voltage. The function of the phase/
frequency comparator is to adjust the voltage presented to
the VCO until the frequency and phase of the feedback
signal match that of the reference signal. When this ‘phaselocked’ condition exists, the VCO frequency will be N+F
times that of the comparison frequency, where N is the
integer component of the divide ratio and F is the fractional
component. Fractional synthesis allows the phase detector
frequency to be increased while maintaining the same frequency step size for channel selection. The division value N
is thereby reduced giving a lower phase noise referred to the
phase detector input, and the comparison frequency is increased allowing faster switching times.
1.2 OSCILLATOR
The LMX2470 provides maximum flexibility for choosing an
oscillator reference. One possible method is to use a singleended TCXO to drive the OSCin pin. The part can also be
configured to be driven differentially using the OSCin and
OSCout* pins. Note that the OSCin and OSCout* pins can
not be used as an inverter for a crystal. Selection between
these two modes does have a noticeable impact on phase
noise and sub-fractional spurs. Regardless of which mode is
used, the performance is generally best for higher oscillator
power levels.
1.3 POWER DOWN AND POWER UP MODES
The power down state of the LMX2470 is controlled by many
factors. The one factor that overrides all other factors is the
EN pin. If this pin is low, this guarantees the part will be
powered down. Asserting a high logic level on EN is necessary to power up the chip, however, there are other bits in the
programming registers that can override this and put the PLL
back in a power down state. Provided that the voltage on the
EN pin is high, programming the RF_PD and IF_PD bits to
zero guarantees that the part will be powered up. Programming either one of these bits to one will power down the
appropriate section of the synthesizer, provided that the
ATPU[1:0] ( Auto Power Up ) bits do not override this.
There are many different ways to power down this chip and
many different things that can be powered down. This section discusses how to power down the PLLs on the chip.
There are two terms that need to be defined first: synchronous power down and asynchronous power down. In the
case of synchronous power down, the PLL chip powers
down after the charge pump turns off. This is best to prevent
unwanted frequency glitches upon power up. However, in
certain cases where the charge pump is stuck on, such as
the case when there is no VCO signal applied, this type of
power down will not reliably work and asynchronous power
down is necessary. In the case of asynchronous power
down, the PLL powers down regardless of the status of the
charge pump. There are 4 factors that affect the power down
state of the chip: the EN pin, the power down bit, the TRISTATE bit, and writing to the RF N counter with the
RF_ATPU[1:0] bits enabled
1.1 PHASE DETECTOR OPERATING FREQUENCY
The maximum phase detector operating frequency for the
LMX2470 is 30 MHz. However, this is not possible in all
circumstances due to illegal divide ratios of the N counter.
The crystal reference frequency also limits the phase detector frequency. There are trade-offs in choosing what phase
detector frequency to operate at. If this frequency is run
higher, then phase noise will be lower, but lock time may be
increased due to cycle slipping. After this phase detector
frequency gets sufficiently high, then there are diminishing
returns for phase noise, and raising the charge pump current
has a greater impact on phase noise. This phase detector
frequency also has an impact on fractional spurs. In general,
the spur performance is better at higher phase detector
EN Pin
ATPU[1:0] Bits Enabled +
RF N Counter Written To
RF_PD Bit
RF_CPT Bit
PLL State
Low
X
X
X
Asynchronous Power Down
High
Yes
X
X
PLL is active with charge pump in the active state.
High
No
0
0
PLL is active with charge pump in the active state.
High
No
0
1
PLL is active, but charge pump is TRI-STATE.
High
No
1
0
Synchronous Power Down
High
No
1
1
Asynchronous Power Down
www.national.com
18
RF PLL, the digital lock detect circuitry does not function
reliably for comparison frequencies above 20 MHz.
The IF PLL digital lock detect circuitry works in a very similar
way as the RF PLL digitial lock circuitry, except that it uses a
delay of less than 15 nS for 5 reference cycles to determine
a locked condition and a delay of greater than 30 nS to
determine the IF PLL is unlocked. Note that if the MUX[3:0]
word is set such as to view lock detect for both PLLs, an
unlocked (LOW) condition is shown whenever either one of
the PLLs is determined to be out of lock. A flow chart of the
IF digital lock detect circuitry is shown below.
(Continued)
1.4 DIGITAL LOCK DETECT OPERATION
The RF PLL digital lock detect circuitry compares the difference between the phase of the inputs of the phase detector
to a RC generated delay of 10 nS. To enter the locked state
(Lock = HIGH) the phase error must be less than the 10nS
RC delay for 5 consecutive reference cycles. Once in lock
(Lock = HIGH), the RC delay is changed to approximately
20nS. To exit the locked state (Lock = LOW), the phase error
must become greater than the 20nS RC delay. When the
PLL is in the power down mode, Lock is forced LOW. For the
20059304
19
www.national.com
LMX2470
Functional Description
LMX2470
Functional Description
capacitor should be chosen at least to be 100 pF. It may turn
out that the frequency in this trace is above the self-resonant
frequency of the capacitor, but since the input impedance of
the PLL tends to be capacitive, it actually be a benefit to
exceed the self-resonant frequency. The pad and the DC
blocking capacitor should be placed as close to the PLL as
possible
(Continued)
1.5 PCB LAYOUT CONSIDERATIONS
Power Supply Pins For these pins, it is recommended that
these be filtered by taking a series 18 ohm resistor and then
placing two capacitors shunt to ground, thus creating a low
pass filter. Although it makes sense to use large capacitor
values in theory, the ESR ( Equivalent Series Resistance ) is
greater for larger capacitors. For optimal filtering minimize
the sum of the ESR and theoretical impedance of the capacitor. It is therefore recommended to provide two capacitors of very different sizes for the best filtering. 0.1 µF and
100 pF are typical values. The charge pump supply pins in
particular are vuvulnerablenerable to power supply noise.
High Frequency Input Pins, FinRF and FinIF The signal
path from the VCO to the PLL is the most sensitive and
challenging for board layout. It is generally recommended
that the VCO output go through a resistive pad and then
through a DC blocking capacitor before it gets to these high
frequency input pins. If the trace length is sufficiently short (
< 1/10th of a wavelength ), then the pad may not be necessary, but a series resistor of about 39 ohms is still recommended to isolate the PLL from the VCO. The DC blocking
Complimentary High Frequency Pin, FinRF* These inputs
may be used to drive the PLL differentially, but it is very
common to drive the PLL in a single ended fashion. A shunt
capacitor should be placed at the FinRF* pin. The value of
this capacitor should be chosen such that the impedance,
including the ESR of the capacitor, is as close to an AC short
as possible at the operating frequency of the PLL. 100 pF is
a typical value.
1.6 FASTLOCK AND CYCLE SLIP REDUCTION
The LMX2470 has enhanced features for Fastlock and cycle
slip operation. The next several sections discuss the the
benefits of using both of these features. There are four
possible combinations that are possible, and these are
shown in the table below:
Charge Pump Current
Keep Comparison
Frequency the Same
Decrease Comparison
Frequency (CSR)
(RF Side Only)
Increase Charge Pump Current
Classical Fastlock
Allows the loop bandwidth to be
increased. This has a frequency glitch
caused by switching the charge pump
currents, but there is no frequency glitch
caused by switching from fractional to
integer mode
CSR/Fastlock Combination
Engaging the CSR does decrease the
loop bandwidth during frequency
acquisition, but may be necessary to
reduce cycle slipping. By also
increasing the charge pump current, this
can compensate for the reduce loop
bandwidth due to the CSR
Keep Charge Pump Current the Same
Decrease Charge Pump Current
Operation with No Fastlock
CSR Only
This mode represents using no Fastlock This mode is not generally
recommended, but may reduce cycle
slipping in some applications. Although
the theoretical lock time is decreased,
due to the decreased loop bandwidth
during Fastlock, cycle slips can be
reduced or eliminated.
It never makes sense to use a lower
charge pump current during Fastlock
than in the steady state.
Note that if the charge pump current and cycle slip reduction circuitry are engaged in the same proportion, then it is not necessary
to switch in a Fastlock resistor and the loop filter will be optimized for both normal mode and Fastlock mode. For third and fourth
order filters which have problems with cycle slipping, this may prove to be the optimal choice of settings.
www.national.com
20
1.6.2 Determining the Theoretical Lock Time
Improvement and Fastlock Resistor, R2’
(Continued)
1.6.1 Determining the Loop Gain Multiplier, K
When using fastlock, it is necessary to switch in a resistor
R2’, in parallel with R2 in order to keep the loop filter optimized and maintain the same phase margin. After the PLL
has achieved a frequency that is sufficiently close to the
desired frequency, the resistor R2’ is disengaged and the
charge pump current is and comparison frequency are returned to normal. Of special concern is the glitch that is
caused when the resistor R2’ is disengaged. This glitch can
take up a significant portion of the lock time. The LMX2470
has enhanced switching circuitry to minimize this glitch and
therefore improve the lock time.
The loop bandwidth multiplier, K, is needed in order to determine the theoretical impact of fastlock/CSR on the loop
bandwidth and also which resistor should be switched in
parallel with the loop filter resistor R2. K = K_K · K_Fcomp
where K is the loop gain multiplier K_K is the ratio of the
Fastlock charge pump current to the steady state charge
pump current. Note that this should always be greater than
or equal to one. K_Fcomp is the ratio of the Fastlock comparison frequency to the steady state comparison frequency.
If this ratio is less than one, this implies that the CSR is being
used.
20059340
1.6.3 Using Fastlock and Cycle Slip Reduction (CSR)
to Avoid Cycle Slipping
The change in loop bandwidth is dependent upon the loop
gain multiplier, K, as determined in section 4. The theoretical
improvement in lock time is given below, but the actual
improvement will be less than this due to the glitch that is
caused by disengaging Fastlock. The theoretical improvement is given to show an upper bound on what improvement
is possible with Fastlock. In the case that K < 1, this implies
the CSR is being engaged and that the theoretical lock time
will be degraded. However, since this mode reduces or
eliminates cycle slipping, the actual lock time may be better
in cases where the loop bandwidth is small relative to the
comparison frequency. Realize that the theoretical lock time
multiplier does not account for the fastlock/CSR disengagement glitch, which is most severe for larger values of K.
In the case that the comparison frequency is very large ( i.e.
100 X ) of the loop bandwidth, cycle slipping may occur when
an instantaneous phase error is presented to the phase
detector. This can be reduced by increasing the loop bandwidth during frequency acquisition, decreasing the comparison frequency during frequency acquisition, or some combination of the these. If increasing the loop bandwidth during
frequency acquisition is not sufficient to reduce cycle slipping, the LMX2470 also has a routine to decrease the comparison frequency.
Loop Gain Loop Bandwidth
Lock Time
R2’ Value
Multiplier, K
Multiplier
Multiplier
1:8*
0.35
open
x 2.828
1:4*
0.50
open
x 2.000
1:2*
0.71
open
x 1.414
4:1
2.00
R2/1.00
x 0.500
8:1
2.83
R2/1.83
x 0.354
16:1
4.00
R2/3.00
x 0.250
32:1
5.66
R2/4.65
x 0.177
* These modes of operation are generally not recommended
21
www.national.com
LMX2470
Functional Description
LMX2470
Functional Description
(Continued)
1.6.4 RF PLL Fastlock Reference Table and Example
The table below shows most of the trade-offs involved in
choosing a steady-state charge pump current (RF_CPG),
Parameter
RF_CPG
the Fastlock charge pump current (RF_CPF), and the Cycle
Slip Reduction Factor (CSR).
Advantages to Choosing Smaller
Advantages to Choosing Larger
1. Allows capacitors in loop filter to be smaller
values making it easier to find physically smaller
components and components with better dielectric
properties.
Phase noise, especially within the loop bandwidth of
the system
will be slightly worse for lower charge pump
currents.
2. Allows a larger loop bandwidth multiplier for
fastlock, or a higher cycle slip reduction factor.
RF_CPF
The only reason not to always choose this to 1600
µA is to make it such that no resistor is required for
fastlock. For 3rd and 4th order filters, it is not
possible to keep the filter perfectly optimized by
simply switching in a resistor for fastlock.
This allows the maximum possible benefit for
fastlock.
CSR
Do not choose this any larger than necessary to
eliminate cycle slipping. Keeping this small allows a
larger loop bandwidth multiplier for fastlock.
This will eliminate cycle slips better.
The above table shows various combinations for using
RF_CPG, RF_CPF, and CSR. Although this table does not
show all possible combinations, it does show all the modes
that give the best possible performance. To use this table,
choose a CSR factor on the horizontal axis, then a fastlock
loop bandwidth multiplier on the vertical axis, and the table
will show all possible combinations of steady state current,
Fastlock current, and what resistor value (R2’) to use during
Fastlock. In order to better illustrate the cycle slipping and
Fastlock circuitry, consider the following example:
Crystal Reference
10 MHz
Comparison Frequency 10 MHz x 2 = 20 MHz (OSC2X = 1)
Output Frequency
1930 – 1990 MHz
PLL Loop Bandwidth
10 KHz
Loop Filter Order
4th ( i.e. 7 components )
The comparison frequency is 20 MHz and the loop bandwidth is 10 KHz. 20 MHz is a good comparison freqeuncy to
use because it yields the best phase noise performance.
This ratio of the comparison frequency to the loop bandwidth
is 2000, so cycle slipping will occur and degrade the lock
time, unless something is done to prevent it. Because the
filter is fourth order, it would be difficult to keep the loop filter
optimized if the loop gain multiplier, K was not one. For this
reason, choosing a loop gain multiplier of one makes sense.
One solution is to set the steady state current to be 100 µA,
and the fastlock current to be 1600 µA. The CSR factor could
be set to 1/16 and reduce this ratio to 2000/16 = 125.
However, using 100 µA charge pump current has phase
noise that is significantly worse than the higher charge pump
current modes. A better solution would be to use 200 µA
www.national.com
22
current and 1600 µA X2 ( using PDCP = X2 Fastlock ), since
the 200 µA mode will have better phase noise. Depending on
how important phase noise is, it could make sense to use a
higher steady state current. Using 800 µA steady state current provides much better phase noise than 200 uA ( about 5
dB ), but then the cycle slip reduction factor would need to be
reduced to 4. In general, it is good practice to use the PDCP
= X2 fastlock mode whenever cycle slip reduction is used, so
that the best phase noise can be achieved. If the 1⁄4 CSR
factor is used, then the ratio of comparison frequency to loop
bandwidth in fastlock is reduced to 250. There may be some
cycle slipping, but the phase noise benefit of using the higher
charge pump current may be worth it. If phase noise is even
more important, it might even make sense to have a steady
state current of 1600 µA and use a CSR factor of 1⁄2 and the
PDCP mode of X2 Fastlock. Another consideration is that
the comparison frequency could be lowered in the steady
state mode to reduce cycle slipping. This sacrifices phase
noise for lock time. In general, using Fastlock and CSR is not
the same for every application. There is a trade-off of lock
time vs. phase noise. It might be tempting to try to achieve
the best Fastlock benefit by using a K value of 32. Even if the
loop filter could be kept well optimized in Fastlock, this
hypothetical design would probably switch very fast when
the Fastlock was engaged, but then when Fastlock is disengaged, a large frequency glitch would appear, and the majority of the lock time would consist of waiting for this glitch to
settle out. Although this would definitely improve the lock
time, even accounting for the glitch, the same result could
probably be obtained by using a lower K value, like 8, and
having better phase noise instead.
fractional spurs. The second order modulator usually does
not have these sub-fractional spurs. The third order modulator will introduce them at 1⁄2 of the frequency where one
would expect to see a traditional fractional spur, thus the
name "sub-fractional spur". The fourth order modulator will
introduce these spurs at 1⁄2 and 1⁄4 of where a traditional
fractional spur would be. If the benefit of using a higher order
modulator seems significant enough, it may make sense to
try to compensate for them using the other two test bits, or
designing a higher order loop filter. Be aware that the impact
of the modulator order on the spurs may not be consistent
across tuning voltage. When the charge pump mismatch is
not so bad, the lower order modulators may seem to outperform the higher order modulators, but when the worst case
fractional spurs are considered over the whole range, often
the higher order modulator performs better.
(Continued)
1.6.5 Capacitor Dielectric Considerations for Lock
Time
The LMX2470 has a high fractional modulus and high
charge pump gain for the lowest possible phase noise. One
consideration is that the reduced N value and higher charge
pump may cause the capacitors in the loop filter to become
larger in value. For larger capacitor values, it is common to
have a trade-off between capacitor dielectric quality and
physical size. Using film capacitors or NP0/CG0 capacitors
yields the best possible lock times, where as using X7R or
Z5R capacitors can increase lock time by 0 – 500%. However, it is a general tendency that designs that use a higher
compare frequency tend to be less sensitive to the effects of
capacitor dielectrics. Although the use of lesser quality dielectric capacitors may be unavoidable in many circumstances, allowing a larger footprint for the loop filter capacitors, using a lower charge pump current, and reducing the
fractional modulus are all ways to reduce capacitor values.
Capacitor dielectrics have very little impact on phase noise
and spurs.
Second, adjust with the CPUD[2:0] bits. Setting this bit to
maximum tends to reduce the sub-fractional spurs the most,
however, it may degrade phase noise by up to 1 dB.
Third, experiment with the dithering. When dithering is enabled, it may increase phase noise by up to 2 dB. However,
enabling dithering may also reduce the sub-fractional spurs.
Also, sometimes both the fractional spurs and the subfractional spurs can be unpredictable with dithering disabled.
This is because the delta-sigma sequence is periodic, but
the starting point changes. Dithering takes these problems
away. When the fractional numerator is 0, enabling dithering
typically hurts spur performance, because it is trying to correct for spur that are not there.
Fourth, consider experimenting with the loop filter order and
comparison frequency. In general, higher order loop filters
are always better, but they require more components. Often,
the best spur performance is at higher comparison frequencies as well. The reason why this is the last step is not
because it has the least impact, but because it takes more
labor to do this than to change the FM[1:0], CPUD[2:0], and
DITH[1:0] bits.
Although general trends do exist, the optimal settings for test
bits may depend on the comparison frequency and loop
filter. Also the output frequency in important. In particular, the
charge pump tuning voltage is relevant. The recommended
way to do this is to test the spur levels at the low, middle, and
high range of the VCO, and use the worst case over these
three frequencies as a metric for performance. Also, it is
important to be aware that all the rules stated above have
counterexamples and exceptions. However, more often than
not, these rules apply.
1.7 FRACTIONAL SPUR AND PHASE NOISE
CONTROLS FOR THE LMX2470
The LMX2470 has several bits that have a large impact on
fractional spurs. These bits also have a lesser effect on
phase noise. The control words in question are CPUD[2:0],
FM[1:0], and DITH[1:0]. It is difficult to predict which settings
will be optimal for a particular application without testing
them, but the general recipe for using these bits can be
seen.
A good algorithm is to start with a 3rd order fractional modulator (FM=3) and dithering disabled. Then depending on
whether phase noise, fractional spurs, or sub-fractional
spurs are most important, optimize the settings. Integer
spurs and fractional spurs are nothing new, but subfractional spurs are something unique to delta-sigma PLLs.
These are spurs that occur at a fraction of the frequency of
where a fractional spur would appear.
First adjust the delta-sigma modulator order. Often increasing from a 2nd to a 3rd order modulator provides a large
benefit in spur levels. Increasing from a 3rd to a 4th order
modulator usually provides some benefit, but it is usually on
the order of a few dB. The modulator order by far has the
greatest impact on the main fractional spurs. If the loop
bandwidth is very wide, or the loop filter order is not high
enough, higher order modulators will introduce a lot of sub-
23
www.national.com
LMX2470
Functional Description
LMX2470
Programming Description
2.0 GENERAL PROGRAMMING INFORMATION
The descriptions below describe the 24-bit data registers loaded through the MICROWIRE Interface. These data registers are
used to program the R counter, the N counter, and the internal mode control latches. The data format of a typical 24-bit data
register is shown below. The control bits CTL [3:0] decode the register address. On the rising edge of LE, data stored in the shift
register is loaded into one of the appropriate latches (selected by address bits). Data is shifted in MSB first. Note that it is best
to program the N counter last, since doing so initializes the digital lock detector and Fastlock circuitry. Note that initialize means
it resets the counters, but it does NOT program values into these registers. Upon a cold power-up, it is necessary to program all
the registers. The exception is when 22-bit is not being used. In this case, it is not necessary to program the R7 register.
MSB
LSB
DATA [21:0]
CTL [3:0]
23
4
3
2
1
0
2.0.1 Register Location Truth Table
The control bits CTL [2:0] decode the internal register address. The table below shows how the control bits are mapped to the
target control register.
C3
C2
C1
C0
DATA Location
x
x
x
0
R0
0
0
0
1
R1
0
0
1
1
R2
0
1
0
1
R3
0
1
1
1
R4
1
0
0
1
R5
1
0
1
1
R6
1
1
0
1
R7
1
1
1
1
R8
2.0.2 Control Register Content Map
Because the LMX2470 registers are complicated, they are organized into two groups, basic and advanced. The first four registers
are basic registers that contain critical information necessary for the PLL to achieve lock. The last 5 registers are for features that
optimize spur, phase noise, and lock time performance. The next page shows these registers.
www.national.com
24
(Continued)
R8
0
0
R6
R7
0
R5
0
0
0
0
CSR[1:0]
IF_P
R4
IF_PD
R2
R3
RF_PD
R1
R0
1
0
R8
22
0
R7
23
0
R6
REGISTER
0
R5
0
0
0
0
0
0
0
0
1
0
17
16
15
14
13
12
11
1
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
1
0
0
9
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
RF_FD[11:0]
IF_R[14:0]
IF_N[16:0]
10
0
0
1
0
0
8
6
1
0
0
0
0
1
0
0
0
0
RF_FN[11:0]
7
1
0
0
0
0
5
21
0
0
0
19
0
0
1
0
0
DITH[1:0]
RF_FD2[9:0]
RF_CPT
0
0
RF_CPP
RF_N[10:0]
0
17
16
15
14
13
12
11
0
IF_CPT
0
0
IF_CPP
0
FDM
0
0
FM[1:0]
PDCP[1:0]
ATPU[1:0]
9
0
8
7
0
OSC
6
0
0
0
0
0
4
5
CPUD[2:0]
3
1
1
1
1
0
0
0
0
0
4
C3
MUX[3:0]
RF_FN[11:0]
RF_FN2[9:0]
OSC2X
IF_TOC[11:0]
RF_TOC[13:0]
IF_R[14:0]
IF_N[17:0]
10
RF_FD[11:0]
DATA[19:0] ( Except for the RF_N Register, which is [22:0] )
18
RF_R[5:0]
RF_CPF[3:0]
RF_CPG[3:0]
20
Complete Register Map The complete register map shows all the functionality of all registers, including the last five.
IF_CPG
0
0
0
0
0
0
0
R4
0
0
R3
RF_CPG[3:0]
RF_R[5:0]
IF_P
IF_CPG
18
IF_PD
1
19
DATA[19:0] ( Except for the RF_N Register, which is [22:0] )
RF_PD
20
R2
21
R1
22
RF_N[10:0]
23
R0
REGISTER
1
1
1
1
0
0
0
0
C3
3
1
1
0
0
1
1
0
0
C2
2
2
1
1
0
0
1
1
0
0
C2
1
0
1
0
1
0
1
0
1
1
0
1
0
1
0
1
0
C1
C1
1
1
1
1
1
1
1
1
1
0
C0
0
1
1
1
1
1
1
1
1
0
C0
0
Quick Start Register Map Although it is highly recommended that the user eventually take advantage of all the modes of the LMX2470, the quick start register map is shown in order
for the user to get the part up and running quickly using only those bits critical for basic functionality. The following default conditions for this programming state are a third order
delta-sigma modulator in 22-bit mode with no dithering and no Fastlock.
Programming Description
LMX2470
25
www.national.com
LMX2470
Programming Description
(Continued)
2.1 R0 REGISTER
Note that this register has only one control bit. The reason for this is that it enables the N counter value to be changed with a
single write statement to the PLL.
REGISTER
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
DATA[19:0] ( Except for the RF_N Register, which is [22:0] )
R0
RF_N[10:0]
4
3
2
1
0
C3
C2
C1
C0
FN[11:0]
0
2.1.1 RF_FN[11:0] -- Fractional Numerator for RF PLL
Refer to section 2.8.1 for a more detailed description of this control word.
2.1.2 RF_N[10:0] -- RF N Counter Value
The RF N counter contains a 16/17/20/21 prescaler. Because there is only one selection of prescaler, the value that is
programmed is simply the N counter value converted into binary form. However, because this counter does have a prescaler,
there are limitations on the divider values.
RF_N[10:0]
RF_N
RF_C
RF_B
RF_A
≤65
N values less than or equal to 65 are prohibited.
66
Possible only with a second order delta-sigma engine
67
Possible with a second or third order delta-sigma engine.
68
0
0
0
0
1
0
0
0
1
0
0
69
0
0
0
0
1
0
0
0
1
0
1
...
.
.
.
.
.
.
.
.
.
.
.
2040
1
1
1
1
1
1
1
1
0
0
0
2041-2044
Possible with a second or third order delta-sigma engine.
2045-2046
Possible only with a second order delta-sigma engine.
> 2046
N values above 2046 are prohibited.
www.national.com
26
LMX2470
Programming Description
(Continued)
2.2 R1 REGISTER
REGISTER
23
22
R1
RF_PD
1
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
DATA[19:0] ( Except for the RF_N Register, which is [22:0] )
RF_R[5:0]
RF_FD[11:0]
4
3
2
1
0
C3
C2
C1
C0
0
0
0
1
2.2.1 RF_FD[11:0] -- RF PLL Fractional Denominator
The function of these bits are described in section 2.8.2.
2.2.2 RF_R [5:0] -- RF R Divider Value
The RF R Counter value is determined by this control word. Note that this counter does allow values down to one.
R Value
RF_R[5:0]
1
0
0
0
0
0
...
.
.
.
.
.
1
.
63
1
1
1
1
1
1
2.2.3 RF_PD -- RF Power Down Control Bit
When this bit is set to 0, the RF PLL operates normally. When it is set to one, the RF PLL is powered down and the RF Charge
pump is set to a TRI-STATE mode. Because the EN pin and ATPU[1:0] word also controls power down functions, there may be
some conflicts. The order of precedence is as follows. First, if the EN pin is LOW, then the PLL will be powered down. Provided
this is not the case, the PLL will be powered up if the ATPU[1:0] word says to do so, regardless of the state of the RF_PD bit. After
the EN pin and the ATPU[1:0] word are considered, then the RF_PD bit then takes control of the power down function for the RF
PLL.
27
www.national.com
LMX2470
Programming Description
(Continued)
2.3 R2 REGISTER
REGISTER
23
22
21
20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4
3
DATA[19:0] ( Except for the RF_N Register, which is [22:0] )
R2
IF_PD IF_P IF_CPG
2
1
0
C3 C2 C1 C0
IF_N[17:0]
0
0
1
1
2.3.1 IF_N[16:0] -- IF N Divider Value
The IF N divider is a classical dual modulus prescaler with a selectable 8/9 or 16/17 modulus. The IF_N value is determined by
the IF_A , IF_B, and IF_P values. Note that the IF_P word can assume a value of 8 or 16. The RF_A and RF_B counter values
can be determined in accordance with the following equations.
B = N div P
A = N mod P
B≥A is required in order to have a legal N divider ratio
Here the div operator is defined as the division of two numbers with the remainder disregarded and the mod operator is defined
as the remainder as a result of this division. For the purposes of programming, it turns out that the register value is just the binary
representation of the N value, with the exception that the 4th LSB is not used and must be programmed to 0 when the 8/9
prescaler is used.
IF_N Programming with the 8/9 Prescaler
IF_N[16:0]
N Value
IF_B
< 24
IF_A
N Values Below 24 are prohibited since IF_B≥3 is required.
24-55
Legal divide ratios in this range are: 24-27, 32-36, 40-45, 48-54
56
0
0
0
0
0
0
0
0
0
0
1
1
1
0
0
0
...
.
.
.
.
.
.
.
.
.
.
.
.
.
0
.
.
0
.
65535
1
1
1
1
1
1
1
1
1
1
1
1
1
0
1
1
1
0
RF_N Programming with 16/17 Prescaler
IF_N[16:0]
N Value
IF_B
IF_A
≤47
N values less than or equal to 47 are prohibited because IF_B≥3 is required.
48-239
Legal divide ratios in this range are: 48-51, 64-68, 80-85, 96-102
240
0
0
0
0
0
0
0
0
0
0
1
1
1
0
0
0
...
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
131071
2.3.2 IF_CPG -- IF Charge Pump Gain
This bit determines the magnitude of the IF charge pump current
IF_CPG
IF Charge Pump Current (mA)
0
Low (1 mA)
1
High (4 mA)
2.3.3 IF_P -- IF Prescaler Value
This bit selects which prescaler will be used for the IF N counter.
IF_P
IF Prescaler Value
0
8 (8/9 Prescaler)
1
16 (16/17 Prescaler)
2.3.4 IF_PD -- IF Power Down Bit
When this bit is set to 0, the IF PLL operates normally. When it is set to 1, the IF PLL powers down and the output of the IF PLL
charge pump is set to a TRI-STATE mode. If the IF_CPT bit is set to 0, then the power down state is synchronous and will not
occur until the charge pump is off. If the IF_CPT bit is set to 1, then the power down will occur immediately regardless of the state
of the IF PLL charge pump.
www.national.com
28
LMX2470
Programming Description
(Continued)
2.4 R3 REGISTER
REGISTER
23
R3
0
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
DATA[19:0] ( Except for the RF_N Register, which is [22:0] )
RF_CPG[3:0]
IF_R[14:0]
3
2
1
0
C3
C2
C1
C0
0
1
0
1
2.4.1 IF_R[14:0] -- IF R Divider Value
For the IF R divider, the R value is determined by the IF_R[14:0] bits in the R3 register. The minimum value for IF_R is 3.
R Value
IF_R[14:0]
3
0
0
0
0
0
0
0
0
0
0
0
0
0
1
...
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
.
32767
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2.4.2 RF_CPG -- RF PLL Charge Pump Gain
This is used to control the magnitude of the RF PLL charge pump in steady state operation
RF_CPG[3:0]
RF Charge Pump Current (µA)
0
100
1
200
2
300
3
400
4
500
5
600
6
700
7
800
8
900
9
1000
10
1100
11
1200
12
1300
13
1400
14
1500
15
1600
29
www.national.com
LMX2470
Programming Description
(Continued)
2.5 R4 REGISTER
This register controls the conditions for the RF PLL in Fastlock.
REGISTER
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
DATA[19:0] ( Except for the RF_N Register, which is [22:0] )
R4
CSR[1:0]
RF_CPF[3:0]
RF_TOC[13:0]
3
2
1
0
C3
C2
C1
C0
0
1
1
1
2.5.1 RF_TOC -- RF Time Out Counter and Control for FLoutRF Pin
The RF_TOC[13:0] word controls the operation of the RF Fastlock circuitry as well as the function of the FLoutRF output pin.
When this word is set to a value between 0 and 3, the RF Fastlock circuitry is disabled and the FLoutRF pin operates as a general
purpose CMOS TRI-STATE I/O. When RF_TOC is set to a value between 4 and 16383, the RF Fastlock mode is enabled and
the FLoutRF pin is utilized as the RF Fastlock output pin. The value programmed into the RF_TOC[13:0] word represents four
times the number of phase detector comparison cycles the RF synthesizer will spend in the Fastlock state.
RF_TOC[13:0]
Fastlock Mode
Fastlock Period [CP events]
FLoutRF Pin Functionality
0
Disabled
N/A
High Impedance
1
Manual
N/A
Logic “0” State.
Forces all fastlock conditions
2
Disabled
N/A
Logic “0” State
3
Disabled
N/A
Logic “1” State
4
Enabled
4X2 = 8
Fastlock
5
Enabled
5X2 = 10
Fastlock
…
Enabled
…
Fastlock
16383
Enabled
16383X2=32766
Fastlock
2.5.2 RF_CPF -- RF PLL Fastlock Charge Pump Current
Specify the charge pump current for the Fastlock operation mode for the RF PLL. Note that the Fastlock charge pump current,
steady state current, and CSR control are all interrelated. Refer to section 4.0 for more details.
www.national.com
RF_CPF [3:0]
Fastlock Charge Pump Current (µA)
0000
100
0001
200
0010
300
0011
400
0100
500
0101
600
0110
700
0111
800
1000
900
1001
1000
1010
1100
1011
1200
1100
1300
1101
1400
1110
1500
1111
1600
30
(Continued)
2.5.3 RF_CSR[1:0] -- RF Cycle Slip Reduction
CSR controls the operation of the Cycle Slip Reduction Circuit. This circuit can be used to reduce the occurrence of phase
detector cycle slips. Note that the Fastlock charge pump current, steady state current, and CSR control are all interrelated. The
table below gives some rough guidelines. In the table below, fCOMP is the comparison frequency, and BW is the loop bandwidth
of the PLL system. The rough guideline gives an idea of when it makes sense to use this cycle slip reduction based on the
steady-state conditions of the PLL system.
CSR[1:0]
CSR State
Sample Rate Reduction Factor
Rough Guideline
0
Disabled
1
fCOMP < 100 x BW
1
Enabled
1/2
100 x BW < fCOMP < 200 x BW
2
Enabled
1/4
200 x BW < fCOMP < 400 x BW
3
Enabled
1/16
fCOMP > 400 x BW
2.6 R5 REGISTER
REGISTER
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
DATA[19:0] ( Except for the RF_N Register, which is [22:0] )
R5
0
0
0
0
0
0
0
0
7
6
5
4
3
2
1
0
C3
C2
C1
C0
1
0
0
1
IF_TOC[11:0]
2.6.1 IF_TOC[11:0] IF Timeout Counter for Fastlock
The IF_TOC word controls the operation of the IF Fastlock circuitry as well as the function of the FLoutIF output pin. When
IF_TOC is set to a value between 0 and 3, the IF Fastlock circuitry is disabled and the FLoutIF pin operates as a general purpose
CMOS TRI-STATE output. When IF_TOC is set to a value between 4 and 4095, the IF Fastlock mode is enabled and FLoutIF is
utilized as the IF Fastlock output pin. The value programmed into IF_TOC represents the number of phase comparison cycles that
the IF synthesizer will spend in the Fastlock state.
IF_TOC[11:0]
Fastlock Mode
Fastlock Period [Charge Pump Cycles]
FLoutIF Pin Functionality
0
Disabled
N/A
High Impedance
1
Manual
N/A
Logic “0” State
Forces IF charge pump current to 4 mA
2
Disabled
N/A
Logic “0” State
3
Disabled
N/A
Logic “1” State
4
Enabled
5
Fastlock
…
Enabled
…
Fastlock
4095
Enabled
4095
Fastlock
31
www.national.com
LMX2470
Programming Description
LMX2470
Programming Description
(Continued)
2.7 R6 REGISTER
REGISTER 23 22 21 20 19
18
17
16
15
14
13
12
11 10
9
8
7 6 5 4
DATA[19:0] ( Except for the RF_N Register, which is [22:0] )
R6
0
0
0
0
0
RF_ RF_ IF_ IF_ FDM
CPT CPP CPT CPP
FM[1:0]
ATPU
[1:0]
3
2
1
0
C3 C2 C1 C0
OSC OSC
2X
MUX
[3:0]
1
0
1
1
2.7.1 MUX[3:0] Frequency Out & Lock Detect MUX
These bits determine the output state of the Ftest/LD pin.
MUX[3:0]
Output Type
Output Description
0
0
0
0
High Impedance
Disabled
0
0
0
1
Push-Pull
General purpose output, Logical “High” State
0
0
1
0
Push-Pull
General purpose output, Logical “Low” State
0
0
1
1
Push-Pull
RF & IF Digital Lock Detect
0
1
0
0
Push-Pull
RF Digital Lock Detect
0
1
0
1
Push-Pull
IF Digital Lock Detect
0
1
1
0
Open Drain
RF & IF Analog Lock Detect
0
1
1
1
Open Drain
RF Analog Lock Detect
1
0
0
0
Open Drain
IF Analog Lock Detect
1
0
0
1
Push-Pull
RF & IF Analog Lock Detect
1
0
1
0
Push-Pull
RF Analog Lock Detect
1
0
1
1
Push-Pull
IF Analog Lock Detect
1
1
0
0
Push-Pull
IF R Divider divided by 2
1
1
0
1
Push-Pull
IF N Divider divided by 2
1
1
1
0
Push-Pull
RF R Divider divided by 2
1
1
1
1
Push-Pull
RF N Divider divided by 2
2.7.2 OSC -- Differential Oscillator Mode Enable
This bit selects between single-ended and differential mode for the OSCin and OSCout* pins. When this bit is set to 0, the RF R
and IF R counters are driven in a single-ended fashion through the OSCin pin. Note that the OSCin and OSCout* pin can not be
used to drive a crystal. When this bit is set to 1, the OSCin and OSCout* pins are used to drive these R counters differentially.
In some cases, spur performance may be better when this is set to differential mode, even if the R counters are being driven in
a single-ended fashion. Current consumption in differential mode is slightly higher than when in single-ended mode.
2.7.3 OSC2X -- Oscillator Doubler Enable
When this bit is set to 0, the oscillator doubler is disabled TCXO frequency presented to the IF R counter is unaffected. Phase
noise added by the doulber is negligible.
2.7.4 ATPU -- PLL Automatic Power Up
This word enables the PLLs to be automatically powered up when their respective registers are written to. Note that since the IF
Powerdown bit is in the IF register, there is no need to have an ATPU function activated by the R2 word.
ATPU
RF PLL
IF PLL
0
No auto power up
No auto power up
1
Powers up when R0 is written to
No auto power up
2
Powers up when R0 is written to
3
www.national.com
Powers up when R0 is written to
Reserved
32
(Continued)
2.7.5 FM[1:0] -- Fractional Mode
Determines the order of the delta-sigma modulator. Higher order delta-sigma modulators reduce the spur levels closer to the
carrier by pushing this noise to higher frequency offsets from the carrier. In general, the order of the loop filter should be at least
one greater than the order of the delta-sigma modulator in order to allow for sufficient roll-off.
FM
Function
0
Fractional PLL mode with a 4th order delta-sigma modulator
1
Disable the delta-sigma modulator. Recommended for test use only.
2
Fractional PLL mode with a 2nd order delta-sigma modulator
3
Fractional PLL mode with a 3rd order delta-sigma modulator
2.7.6 FDM -- Fractional Denominator Mode
When this bit is set to 0, the part operates with a 12- bit fractional denominator. For most applications, 12-bit mode should be
adequate, but for those applications requiring ultra fine tuning resolution, there is 22-bit mode. Note that the PLL may consume
slightly more current when it is in 22-bit mode.
FDM
Bits for Fractional Denominator/Numerator
Maximum Size of Fractional Denominator/Numerator
0
12-bit
4095
1
22-bit
4194303
2.7.7 IF_CPP -- IF PLL Charge Pump Polarity
When this bit is set to 1, the phase detector polarity for the IF PLL charge pump is positive. Otherwise set this bit to 0 for a
negative phase detector polarity
2.7.8 IF_CPT -- IF PLL Charge Pump TRI-STATE Mode
This bit enables the user to put the charge pump in a TRI-STATE ( high impedance ) condition. Note that if there is a conflict, the
ATPU bit overrides this bit.
RF_CPT
Charge Pump State
0
ACTIVE
1
TRI-STATE
2.7.9 RF_CPP -- RF PLL Charge Pump Polarity
For a positive phase detector polarity, which is normally the case, set this bit to 1. Otherwise set this bit to 0 for a negative phase
detector polarity.
2.7.10 RF_CPT -- RF PLL Charge Pump TRI-STATE Mode
This bit enables the user to put the charge pump in a TRI-STATE ( high impedance) condition. Note that if there is a conflict, the
ATPU bit overrides this bit.
RF_CPT
Charge Pump State
0
Active
1
TRI-STATE
33
www.national.com
LMX2470
Programming Description
LMX2470
Programming Description
(Continued)
2.8 R7 REGISTER
REGISTER
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
DATA[19:0]
R7
RF_FD2[9:0]
RF_FN2[9:0]
3
2
1
0
C3
C2
C1
C0
1
1
0
1
2.8.1 Fractional Numerator Determination { RF_FN2[9:0], RF_FN[11:0], FDM }
In the case that the FDM bit is 0, then the part operates in 12-bit fractional mode, and the RF_FN2 bits become don’t care bits.
When the FDM is set to 1, the part operates in 22-bit mode and the fractional numerator is expanded from 12 to 22-bits.
Fractional
RF_FN2[9:0]
Numerator
( These bits only apply in 22- bit mode)
RF_FN[11:0]
0
In 12- bit mode, these are don’t care.
In 22- bit mode, for N < 4096,
these bits should be all set to 0.
1
...
4095
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
.
.
.
.
.
.
.
.
.
.
.
.
1
1
1
1
1
1
1
1
1
1
1
1
0
4096
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
...
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4194303
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2.8.2 Fractional Denominator Determination { RF_FD2[9:0], RF_FD[11:0], FDM }
In the case that the FDM bit is 0, then the part is operates in the 12-bit fractional mode, and the RF_FD2 bits become don’t care
bits. When the FDM is set to 1, the part operates in 22-bit mode and the fractional denominator is expanded from 12 to 22-bits.
Fractional
RF_FD2[9:0]
Denominator
( These bits only apply in 22- bit mode)
0
In 12- bit mode, these are don’t care.
In 22- bit mode, for N < 4096,
these bits should be all set to 0.
1
...
RF_FD[11:0]
4095
4096
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
.
.
.
.
.
.
.
.
.
.
.
.
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
...
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4194303
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
www.national.com
34
LMX2470
Programming Description
(Continued)
2.9 R8 REGISTER
23
22
21
20
0
0
0
0
19
18
17
16
0
0
15
14
13
12
11
10
9
8
0
0
7
6
5
4
DATA[19:0]
R8
DITH
[1:0]
0
0
0
0
PDCP
[1:0]
CPUD
[2:0]
3
2
1
0
C3
C2
C1
C0
1
1
1
1
0
The R8 Register controls some additional bits that may be useful in optimizing phase noise, lock time, and spurs.
2.9.1 CPUD[2:0] -- Charge Pump User Definition
This bit allows the user to choose from several different modes in the charge pump. The charge pump current is unaffected, but
the fractional spurs and phase noise are impacted by a few dB. In some designs, particularly if the loop bandwidth is wide and
a 4th order delta-sigma engine is used, small spurs may appear at a fraction of where the first fractional spur should appear. In
other designs, these sub-fractional spurs are not present. The user needs to use this adjustment to make these sub-fractional
spurs go away, while still getting the best phase noise possible.
CPUD
Mode Name
Phase Noise
Sub-Fractional Spurs
0
Reserved
N/A
N/A
1
Reserved
N/A
N/A
2
Minimum
Best
Worst
3
Maximum
Worst
Best
4
Reserved
N/A
N/A
5
Reserved
N/A
N/A
6
Reserved
N/A
N/A
7
Nominal
Medium
Medium
2.9.2 PDCP[1:0] -- Power Drive for Charge Pump
If this bit is enabled, the Fastlock current can be doubled during Fastlock. The charge pump current in steady state is unaffected.
States 0 and 1 should never be used.
PDCP
Fastlock Charge Pump Current
0
Reserved
1
Reserved
2
Double Fastlock Current
3
Disabled
2.9.3 DITH[1:0] -- Dithering Control
Dithering is a technique used to spread out the spur energy. Enabling dithering can reduce the main fractional spurs, but can also
give rise to a family of smaller spurs. Whether dithering helps or hurts is application specific. Enabling the dithering may also
increase the phase noise. In most cases where the fractional numerator is zero, dithering usually degrades performance.
Determining tends to be most beneficial in applications where there is insufficient filtering of the spurs. This often occurs when the
loop bandwidth is very wide or a higher order delta-sigma modulator is used. Dithering tends not to impact the main fractional
spurs much, but has a much larger inpact on the sub-fractional spurs. If it is decided that dithering will be used, best results will
be obtained when the fractional denominator is at least 1000.
DITH
Dithering Mode Used
0
Dithering Enabled
1
Reserved
2
Reserved
3
Dithering Disabled
35
www.national.com
LMX2470 2.6 GHz Delta-Sigma Fractional-N PLL with 800 MHz Integer-N PLL
Physical Dimensions
inches (millimeters) unless otherwise noted
Ultra Thin Chip Scale Package (SLE)
For Tape and Reel (2500 Units per Reel)
Order Number LMX2470SLEX
NS Package Number SLE24A
LIFE SUPPORT POLICY
NATIONAL’S PRODUCTS ARE NOT AUTHORIZED FOR USE AS CRITICAL COMPONENTS IN LIFE SUPPORT
DEVICES OR SYSTEMS WITHOUT THE EXPRESS WRITTEN APPROVAL OF THE PRESIDENT AND GENERAL
COUNSEL OF NATIONAL SEMICONDUCTOR CORPORATION. As used herein:
1. Life support devices or systems are devices or
systems which, (a) are intended for surgical implant
into the body, or (b) support or sustain life, and
whose failure to perform when properly used in
accordance with instructions for use provided in the
labeling, can be reasonably expected to result in a
significant injury to the user.
National Semiconductor
Americas Customer
Support Center
Email: [email protected]
Tel: 1-800-272-9959
www.national.com
National Semiconductor
Europe Customer Support Center
Fax: +49 (0) 180-530 85 86
Email: [email protected]
Deutsch Tel: +49 (0) 69 9508 6208
English Tel: +44 (0) 870 24 0 2171
Français Tel: +33 (0) 1 41 91 8790
2. A critical component is any component of a life
support device or system whose failure to perform
can be reasonably expected to cause the failure of
the life support device or system, or to affect its
safety or effectiveness.
National Semiconductor
Asia Pacific Customer
Support Center
Fax: +65-6250 4466
Email: [email protected]
Tel: +65-6254 4466
National Semiconductor
Japan Customer Support Center
Fax: 81-3-5639-7507
Email: [email protected]
Tel: 81-3-5639-7560
National does not assume any responsibility for use of any circuitry described, no circuit patent licenses are implied and National reserves the right at any time without notice to change said circuitry and specifications.