LTC6948 - Ultralow Noise 0.37GHz to 6.39GHz Fractional-N Synthesizer with Integrated VCO

LTC6948
Ultralow Noise 0.37GHz to
6.39GHz Fractional-N Synthesizer
with Integrated VCO
Description
Features
Low Noise Fractional-N PLL with Integrated VCO
n No ∆-∑ Modulator Spurs
n 18-Bit Fractional Denominator
n–226 dBc/Hz Normalized In-Band Phase Noise Floor
n–274 dBc/Hz Normalized In-Band 1/f Noise
n–157 dBc/Hz Wideband Output Phase Noise Floor
n Excellent Integer Boundary Spurious Performance
n Output Divider (1 to 6, 50% Duty Cycle)
n Output Buffer Muting
n Reference Input Frequency Up to 425MHz
n Fast Frequency Switching
n FracNWizard™ Software Design Tool Support
n
n
n
n
The fractional divider uses an advanced, 4th order Δ∑
modulator which provides exceptionally low spurious
levels. This allows wide loop bandwidths, producing extremely low integrated phase noise values.
The programmable VCO output divider, with a range of 1
through 6, extends the output frequency range.
Output Frequency Options
Applications
n
The LTC®6948 is a high performance, low noise, 6.39GHz
phase-locked loop (PLL) with a fully integrated VCO,
including a reference divider, phase-frequency detector
(PFD), ultralow noise charge pump, fractional feedback
divider, and VCO output divider.
LTC6948-1
Wireless Basestations (LTE, WiMAX, W-CDMA, PCS)
Microwave Data Links
Military and Secure Radio
Test and Measurement
L, LT, LTC, LTM, Linear Technology and the Linear logo are registered trademarks and
FracNWizard is a trademark of Linear Technology Corporation. All other trademarks are the
property of their respective owners.
LTC6948-2
LTC6948-3
LTC6948-4
O_DIV = 1
2.240 to 3.740 3.080 to 4.910 3.840 to 5.790 4.200 to 6.390
O_DIV = 2
1.120 to 1.870 1.540 to 2.455 1.920 to 2.895 2.100 to 3.195
O_DIV = 3
0.747 to 1.247 1.027 to 1.637 1.280 to 1.930 1.400 to 2.130
O_DIV = 4
0.560 to 0.935 0.770 to 1.228 0.960 to 1.448 1.050 to 1.598
O_DIV = 5
0.448 to 0.748 0.616 to 0.982 0.768 to 1.158 0.840 to 1.278
O_DIV = 6
0.373 to 0.623 0.513 to 0.818 0.640 to 0.965 0.700 to 1.065
Typical Application
6.3GHz Wideband Receiver
1.6nF
10µH
76.8Ω
5V
15Ω
0.1µF 3.3V
1µF
LTC6948-4 Phase Noise,
fRF = 6236MHz
76.8Ω
2.4nF
–90
UNUSED OUTPUT
AVAILABLE FOR
OTHER USE
50Ω
3.3V
0.1µF
GND
CP
VCP+
VREF+
VVCO+
VRF+
3.3V
0.1µF
BB
3.3V
RF+
0.1µF
LTC6948-4
RF–
SPI BUS
GND
51.1Ω
GND
REF–
1µF
REF+
STAT
CS
SCLK
SDI
SDO
LDO
V D+
BVCO
GND
CMA
CMB
CMC
GND
TB
TUNE
3.3V
68nH
68nH 100pF
–100
0.01µF
1µF
MUTE
100MHz
56nF
0.01µF
2.2µF
1µF
R = 2, fPFD = 50MHz
N = 84 TO 127.8
LBW = 186kHz
O=1
PHASE NOISE (dBc/Hz)
1µF
–110
–120
–130
RMS NOISE = 0.412°
–140 RMS JITTER = 183fs
fPFD = 50MHz
–150 LOOP BW = 186kHz
INTN = 0
CPLE = 1
–160
1M
1k
100
10k
100k
OFFSET FREQUENCY (Hz)
10M 40M
6946 TA01b
0.01µF
100pF
fLO = 4200MHz TO 6390MHz IN 190.7Hz STEPS
RF INPUT SIGNAL
LO IF
RF
IF OUTPUT
6948 TA01a
6948f
For more information www.linear.com/LTC6948
1
LTC6948
Absolute Maximum Ratings
Pin Configuration
(Note 1)
VVCO+
GND
VCP+
CP
REF–
VREF+
TOP VIEW
Supply Voltages
V+ (VREF+, VRF+, VD+) to GND................................3.6V
VCP+, V VCO+ to GND..............................................5.5V
Voltage on CP Pin..................GND – 0.3V to VCP+ + 0.3V
Voltage on all other Pins............GND – 0.3V to V+ + 0.3V
Operating Junction Temperature Range, TJ (Note 2)
LTC6948I............................................ –40°C to 105°C
Junction Temperature, TJMAX................................. 125°C
Storage Temperature Range................... –65°C to 150°C
28 27 26 25 24 23
REF+ 1
22 BVCO
STAT 2
21 GND
CS 3
20 CMA
19 CMB
SCLK 4
29
GND
SDI 5
18 CMC
SDO 6
17 GND
LDO 7
16 TB
VD+ 8
15 TUNE
BB
VRF
+
RF+
RF
–
GND
MUTE
9 10 11 12 13 14
UFD PACKAGE
28-LEAD (4mm × 5mm) PLASTIC QFN
TJMAX = 125°C, θJCbottom = 7°C/W
EXPOSED PAD (PIN 29) IS GND, MUST BE SOLDERED TO PCB
Order Information
LEAD FREE FINISH
TAPE AND REEL
PART MARKING
PACKAGE DESCRIPTION
JUNCTION TEMPERATURE RANGE
LTC6948IUFD-1#PBF
LTC6948IUFD-1#TRPBF
69481
28-Lead (4mm × 5mm) Plastic QFN
–40°C to 105°C
LTC6948IUFD-2#PBF
LTC6948IUFD-2#TRPBF
69482
28-Lead (4mm × 5mm) Plastic QFN
–40°C to 105°C
LTC6948IUFD-3#PBF
LTC6948IUFD-3#TRPBF
69483
28-Lead (4mm × 5mm) Plastic QFN
–40°C to 105°C
LTC6948IUFD-4#PBF
LTC6948IUFD-4#TRPBF
69484
28-Lead (4mm × 5mm) Plastic QFN
–40°C to 105°C
Consult LTC Marketing for parts specified with wider operating temperature ranges.
Consult LTC Marketing for information on nonstandard lead based finish parts.
For more information on lead free part marking, go to: http://www.linear.com/leadfree/
For more information on tape and reel specifications, go to: http://www.linear.com/tapeandreel/
Available Options
PACKAGE STYLE
OUTPUT FREQUENCY RANGE vs OUTPUT DIVIDER SETTING (GHz)
VCO FREQUENCY
RANGE (GHz)
QFN-28 (UFD28)
0 DIV = 6
0 DIV = 5
0 DIV = 4
0 DIV = 3
0 DIV = 2
0 DIV = 1
2.240 to 3.740
LTC6948IUFD-1
0.373 to 0.623
0.448 to 0.748
0.560 to 0.935
0.747 to 1.247
1.120 to 1.870
2.240 to 3.740
3.080 to 4.910
LTC6948IUFD-2
0.513 to 0.818
0.616 to 0.982
0.770 to 1.228
1.027 to 1.637
1.540 to 2.455
3.080 to 4.910
3.840 to 5.790
LTC6948IUFD-3
0.640 to 0.965
0.768 to 1.158
0.960 to 1.448
1.280 to 1.930
1.920 to 2.895
3.840 to 5.790
4.200 to 6.390
LTC6948IUFD-4
0.700 to 1.065
0.840 to 1.278
1.050 to 1.598
1.400 to 2.130
2.100 to 3.195
4.200 to 6.390
Overlapping Frequency Bands
6948f
2
For more information www.linear.com/LTC6948
LTC6948
Electrical Characteristics
The l denotes the specifications which apply over the full operating
junction temperature range, otherwise specifications are at TA = 25°C. VREF+ = VD+ = VRF+ = 3.3V, VCP+ = VVCO+ = 5V unless otherwise
specified (Note 2). All voltages are with respect to GND.
SYMBOL
PARAMETER
CONDITIONS
MIN
TYP
MAX
UNITS
425
MHz
Reference Inputs (REF+, REF–)
fREF
Input Frequency
VREF
Input Signal Level
Single-Ended, 1µF AC-Coupling Capacitors
Input Slew Rate
l
10
l
0.5
l
20
l
1.65
1.85
2.25
V
l
5.8
8.4
11.6
kΩ
Input Duty Cycle
2
2.7
50
Self-Bias Voltage
Input Resistance
Differential
Input Capacitance
Differential
fVCO
Frequency Range
LTC6948-1 (Note 3)
LTC6948-2 (Note 3)
LTC6948-3 (Note 3)
LTC6948-4 (Note 3)
l
l
l
l
KVCO
Tuning Sensitivity
LTC6948-1 (Notes 3, 4)
LTC6948-2 (Notes 3, 4)
LTC6948-3 (Notes 3, 4)
LTC6948-4 (Notes 3, 4)
l
l
l
l
VP-P
V/µs
%
14
pF
VCO
2.24
3.08
3.84
4.20
3.74
4.91
5.79
6.39
4.7 to 7.2
4.7 to 7.0
4.0 to 6.0
4.5 to 6.5
GHz
GHz
GHz
GHz
%Hz/V
%Hz/V
%Hz/V
%Hz/V
RF Output (RF+, RF–)
fRF
Output Frequency
O
Output Divider Range
All Integers Included
l
0.373
l
1
Output Duty Cycle
PRF-SE
6.39
GHz
6
50
%
Output Resistance
Single-Ended, Each Output to VRF
+
l
100
136
175
Ω
Output Power, Single-Ended, fRF = 900MHz
RFO[1:0] = 0, RZ = 50Ω, LC Match
RFO[1:0] = 1, RZ = 50Ω, LC Match
RFO[1:0] = 2, RZ = 50Ω, LC Match
RFO[1:0] = 3, RZ = 50Ω, LC Match
l
l
l
l
–9
–6.1
–2.9
0.1
–7.3
–4.5
–1.4
1.5
–5.5
–2.8
0.2
3
dBm
dBm
dBm
dBm
Output Power, Muted, fRF = 900MHz
RZ = 50Ω, Single-Ended, O = 2 to 6
l
–80
dBm
Mute Enable Time
l
110
ns
Mute Disable Time
l
170
ns
l
100
MHz
l
l
l
l
l
76.1
66.3
56.1
45.9
34.3
MHz
MHz
MHz
MHz
MHz
11.2
mA
±6
%
±3.5
±2
%
%
Phase/Frequency Detector
fPFD
Input Frequency
Integer mode
Fractional mode
LDOEN = 0
LDOV = 3, LDOEN = 1
LDOV = 2, LDOEN = 1
LDOV = 1, LDOEN = 1
LDOV = 0, LDOEN = 1
Charge Pump
ICP
Output Current Range
8 Settings (See Table 6)
Output Current Source/Sink Accuracy
All Settings, V(CP) = VCP+/2
Output Current Source/Sink Matching
ICP = 1.0mA to 2.8mA, V(CP) = VCP+/2
ICP = 4.0mA to 11.2mA, V(CP) = VCP+/2
Output Current vs Output Voltage Sensitivity (Note 5)
Output Current vs Temperature
V(CP) = VCP+/2
Output Hi-Z Leakage Current
ICP = 11.2mA, CPCLO = CPCHI = 0 (Note 5)
1
l
0.2
l
170
1
ppm/°C
%/V
0.03
nA
6948f
For more information www.linear.com/LTC6948
3
LTC6948
Electrical Characteristics
The l denotes the specifications which apply over the full operating
junction temperature range, otherwise specifications are at TA = 25°C. VREF+ = VD+ = VRF+ = 3.3V, VCP+ = VVCO+ = 5V unless otherwise
specified (Note 2). All voltages are with respect to GND.
SYMBOL
PARAMETER
CONDITIONS
VCLMP-LO
Low Clamp Voltage
CPCLO = 1
0.84
V
VCLMP-HI
High Clamp Voltage
CPCHI = 1, Referred to VCP+
–0.96
V
0.48
V/V
VMID
MIN
+ – GND)
Mid-Supply Output Bias Ratio
Referred to (VCP
TYP
MAX
UNITS
Reference (R) Divider
R
All Integers Included
l
1
31
Counts
All Integers Included, Integer Mode
All Integers Included, Fractional Mode
l
l
32
35
1023
1019
Counts
Counts
Numerator Range
All Integers Included
l
1
262143
Counts
Output Voltage
LDO Enabled, Four Values
LDO Disabled
External Pin Capacitance
Required for LDO Stability
l
0.047
1.55
Divide Range
VCO (N) Divider
N
Divide Range
Fractional ∆∑ Modulator
Modulator LDO
V
V
1.7 to 2.6
VD+
0.1
1
µF
Digital Pin Specifications
VIH
High Level Input Voltage
MUTE, CS, SDI, SCLK
l
VIL
Low Level Input Voltage
MUTE, CS, SDI, SCLK
l
VIHYS
Input Voltage Hysteresis
MUTE, CS, SDI, SCLK
Input Current
MUTE, CS, SDI, SCLK
V
0.8
250
+ – 400mV
IOH
High Level Output Current
SDO and STAT, VOH = VD
l
IOL
Low Level Output Current
SDO and STAT, VOL = 400mV
l
SDO Hi-Z Current
mV
±1
l
–3.3
2.0
–1.9
3.4
µA
mA
mA
±1
l
V
µA
Digital Timing Specifications (See Figure 6 and Figure 7)
tCKH
SCLK High Time
l
25
ns
tCKL
SCLK Low Time
l
25
ns
tCSS
CS Setup Time
l
10
ns
tCSH
CS High Time
l
10
ns
tCS
SDI to SCLK Setup Time
l
6
ns
tCH
SDI to SCLK Hold Time
l
6
tDO
SCLK to SDO Time
To VIH/VIL/Hi-Z with 30pF Load
ns
l
16
ns
Power Supply Voltages
VREF+ Supply Range
l
3.15
3.3
3.45
V
VD
+ Supply Range
l
3.15
3.3
3.45
V
VRF+ Supply Range
l
3.15
3.3
3.45
V
l
4.75
5.0
5.25
V
l
4.0
5.25
V
+ Supply Range
VVCO
VCP
+ Supply Range
Power Supply Currents
IDD
VD+ Supply Current
Digital Inputs at Supply Levels, Integer Mode l
l
Digital Inputs at Supply Levels, Fractional
Mode, fPFD = 66.3MHz, LDOV[1:0] = 3
18.2
1500
22
µA
mA
ICC(5V)
Sum VCP+, VVCO+ Supply Currents
ICP = 11.2mA
ICP = 1.0mA
PDALL = 1
48
26
450
60
35
1000
mA
mA
µA
l
l
l
6948f
4
For more information www.linear.com/LTC6948
LTC6948
Electrical Characteristics
The l denotes the specifications which apply over the full operating
junction temperature range, otherwise specifications are at TA = 25°C. VREF+ = VD+ = VRF+ = 3.3V, VCP+ = VVCO+ = 5V unless otherwise
specified (Note 2). All voltages are with respect to GND.
SYMBOL
ICC(3.3V)
PARAMETER
CONDITIONS
+, V
Sum VREF
+ Supply Currents
MIN
RF Muted, OD[2:0] = 1
RF Enabled, RFO[1:0] = 0, OD[2:0] = 1
RF Enabled, RFO[1:0] = 3, OD[2:0] = 1
RF Enabled, RFO[1:0] = 3, OD[2:0] = 2
RF Enabled, RFO[1:0] = 3, OD[2:0] = 3
RF Enabled, RFO[1:0] = 3, OD[2:0] = 4 to 6
PDALL = 1
RF
l
l
l
l
l
l
l
TYP
MAX
UNITS
70.4
81.1
91.3
109.2
114.8
119.6
53
80
95
105
125
135
140
250
mA
mA
mA
mA
mA
mA
µA
Phase Noise and Spurious
LVCO
VCO Phase Noise (LTC6948-1, fVCO = 3.0GHz, 10kHz Offset
fRF = 3.0GHz, OD[2:0] = 1 (Note 6))
1MHz Offset
40MHz Offset
–80
–130
–157
dBc/Hz
dBc/Hz
dBc/Hz
VCO Phase Noise (LTC6948-2, fVCO = 4.0GHz, 10kHz Offset
fRF = 4.0GHz, OD[2:0] = 1 (Note 6))
1MHz Offset
40MHz Offset
–77
–127
–156
dBc/Hz
dBc/Hz
dBc/Hz
VCO Phase Noise (LTC6948-3, fVCO = 5.0GHz, 10kHz Offset
fRF = 5.0GHz, OD[2:0] = 1 (Note 6))
1MHz Offset
40MHz Offset
–75
–126
–155
dBc/Hz
dBc/Hz
dBc/Hz
VCO Phase Noise (LTC6948-4, fVCO = 6.0GHz, 10kHz Offset
fRF = 6.0GHz, OD[2:0] = 1 (Note 6))
1MHz Offset
40MHz Offset
–73
–123
–154
dBc/Hz
dBc/Hz
dBc/Hz
VCO Phase Noise (LTC6948-3, fVCO = 5.0GHz, 10kHz Offset
fRF = 2.50GHz, OD[2 :0] = 2 (Note 6))
1MHz Offset
40MHz Offset
–81
–132
–155
dBc/Hz
dBc/Hz
dBc/Hz
VCO Phase Noise (LTC6948-3, fVCO = 5.0GHz, 10kHz Offset
fRF = 1.667GHz, OD[2 :0] = 3 (Note 6))
1MHz Offset
40MHz Offset
–84
–135
–156
dBc/Hz
dBc/Hz
dBc/Hz
VCO Phase Noise (LTC6948-3, fVCO = 5.0GHz, 10kHz Offset
fRF = 1.25GHz, OD[2 :0] = 4 (Note 6))
1MHz Offset
40MHz Offset
–87
–138
–156
dBc/Hz
dBc/Hz
dBc/Hz
VCO Phase Noise (LTC6948-3, fVCO = 5.0GHz, 10kHz Offset
fRF = 1.00GHz, OD[2 :0] = 5 (Note 6))
1MHz Offset
40MHz Offset
–89
–140
–157
dBc/Hz
dBc/Hz
dBc/Hz
VCO Phase Noise (LTC6948-3, fVCO = 5.0GHz, 10kHz Offset
fRF = 0.833GHz, OD[2 :0] = 6 (Note 6))
1MHz Offset
40MHz Offset
–90
–141
–158
dBc/Hz
dBc/Hz
dBc/Hz
LNORM(INT)
Integer Normalized In-Band Phase Noise
Floor
INTN = 1, ICP = 5.6mA (Notes 7, 8, 10)
–226
dBc/Hz
LNORM(FRAC)
Fractional Normalized In-Band Phase Noise
Floor
INTN = 0, CPLE = 1, ICP = 5.6mA
(Notes 7, 8, 10)
–225
dBc/Hz
L1/f
Normalized In-Band 1/f Phase Noise
ICP = 11.2mA (Notes 7, 11)
–274
dBc/Hz
In-Band Phase Noise Floor
Fractional Mode, CPLE = 1
(Notes 7, 9, 10, 12)
–113
dBc/Hz
Integrated Phase Noise from 100Hz to
40MHz
Fractional Mode, CPLE = 1 (Notes 9, 12)
0.14
°RMS
Spurious
Fractional Mode, fOFFSET = fPFD, PLL locked
(Notes 8, 13, 14)
–98
dBc
Note 1: Stresses beyond those listed under Absolute Maximum Ratings
may cause permanent damage to the device. Exposure to any Absolute
Maximum Rating condition for extended periods may affect device
reliability and lifetime.
Note 2: The LTC6948I is guaranteed to meet specified performance limits
over the full operating junction temperature range of –40°C to 105°C.
6948f
For more information www.linear.com/LTC6948
5
LTC6948
Electrical Characteristics
Note 3: Valid for 1.60V ≤ V(TUNE) ≤ 2.85V with part calibrated after a
power cycle or software power-on reset (POR).
Note 4: Based on characterization.
Note 5: For 0.9V < V(CP) < (VCP+ – 0.9V).
Note 6: Measured outside the loop bandwidth, using a narrowband loop,
RFO[1:0] = 3.
Note 7: Measured inside the loop bandwidth with the loop locked.
Note 8: Reference frequency supplied by Wenzel 501-04516,
fREF = 100MHz, PREF = 10dBm.
Note 9: Reference frequency supplied by Wenzel 500-23571,
fREF = 61.44MHz, PREF = 10dBm.
Note 10: Output phase noise floor is calculated from normalized phase
noise floor by LOUT = LNORM + 10log10 (fPFD) + 20log10 (fRF/fPFD).
Note 11: Output 1/f noise is calculated from normalized 1/f phase noise by
LOUT(1/f) = L1/f + 20log10 (fRF) – 10log10 (fOFFSET).
Note 12: ICP = 5.6mA, fPFD = 61.44MHz, FILT[1:0] = 0, Loop BW = 180kHz;
fRF = 2377.7MHz, fVCO = 4755.4MHz (LTC6948-3)
Note 13: ICP = 5.6mA, fPFD = 25MHz, FILT[1:0] = 0, Loop BW = 73kHz;
fRF = 891.85MHz, fVCO = 2675Hz (LTC6948-1), fVCO = 4459MHz
(LTC6948‑2), fVCO = 5351MHz (LTC6948-3, LTC6948-4).
Note 14: Measured using DC1959.
Typical
Performance Characteristics
5V, INTN = 0, DITHEN = 1, CPLE = 1, RFO[1:0] = 3, unless otherwise noted.
CP Hi-Z Current vs Voltage,
Temperature
SENSITIVITY (dBm)
–20
–25
–30
–35
5
3
2
–40
–45
–50
0
1
–3
–3
–4
–4
0
–1
5
5
4
4
3
3
2
2
1
1
0
–1
–2
–2
–3
–3
–4
–4
–5
–5
0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
OUTPUT VOLTAGE (V)
6948 G04
0
ICP = 11.2mA
CPLE = 0
105°C
25°C
–40°C
0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
OUTPUT VOLTAGE (V)
6948 G05
0.5
1.0
1.5
2.0
VOLTAGE (V)
2.5
3.0
6948 G03
Charge Pump Source Current
Error vs Voltage, Output Current
ERROR (%)
1
–5
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
OUTPUT VOLTAGE (V)
Charge Pump Sink Current Error
vs Voltage, Temperature
ERROR (%)
ERROR (%)
2
0
–1
6948 G02
CPLE = 0
1mA
5.6mA
11.2mA
3
1
–2
Charge Pump Sink Current Error
vs Voltage, Output Current
4
2
–2
6948 G01
5
3
–1
0
105°C
25°C
–40°C
4
0
–5
50 100 150 200 250 300 350 400 450
FREQUENCY (MHz)
5
ICP = 11.2mA
CPRST = 1
CPLE = 0
105°C
25°C
–40°C
4
CURRENT (nA)
BST = 1
FILT = 0
105°C
25°C
–40°C
TUNE Current vs Voltage,
Temperature
CURRENT (nA)
REF Input Sensitivity vs
Frequency
–15
TA = 25°C. VREF+ = VD+ = VRF+ = 3.3V, VCP+ = VVCO+ =
CPLE = 0
1mA
5.6mA
11.2mA
0
–1
–2
–3
–4
–5
0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
OUTPUT VOLTAGE (V)
6948 G06
6948f
6
For more information www.linear.com/LTC6948
LTC6948
Typical
Performance Characteristics
TA = 25°C. VREF+ = VD+ = VRF+ = 3.3V, VCP+ = VVCO+ =
5V, INTN = 0, DITHEN = 1, CPLE = 1, RFO[1:0] = 3, unless otherwise noted.
2.0
1.5
3
1.0
2
0.5
0
–1
–1.0
–1.5
–5
0
LTC6948-4
LC = 68nH
CS = 100pF
105°C
25°C
–40°C
–2.0
–2.5
–3.0
–3.5
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
OUTPUT VOLTAGE (V)
0.5
1.5
6948 G07
O=6
–10
O=1
–35 LTC6948-4
LC = 68nH
–40 CS = 100pF
fRF = fVCO/O
–45
4.2
4.6
5.0
6.2
–75
O=4
4.6
5.0
5.4
fVCO (MHz)
6948 G10
LTC6948-1 VCO Tuning Sensitivity
5.8
2.7
2.6
2.5
O=5
2.4
O=6
2.3
6.2
2.2
6.6
LTC6948-2 VCO Tuning Sensitivity
7.5
6.5
7.0
7.0
6.0
6.5
6.0
5.5
5.0
4.5
5.5
5.0
4.0
3.5
3.5
3.0
2.8 3.0 3.2 3.4
FREQUENCY (GHz)
3.6
3.8
6948 G13
3.0
25
4.5
4.0
2.6
20
5.0
4.0
2.4
10
15
TIME (µs)
5.5
4.5
2.2
5
7.0
3.5
3.0
0
LTC6948-3 VCO Tuning Sensitivity
7.5
6.5
–5
6948 G12
8.0
6.0
126MHz STEP
fPFD = 61.44MHz
fCAL = 1.28MHz
LOOP BW = 180kHz
MTCAL = 0
O=2
6948 G11
8.0
KVCO (%/V)
KVCO (%/V)
O=2
O=3
–95
4.2
6.6
CALIBRATION TIME
2.9
2.8
LTC6948-4, LC = 68nH
CS = 100pF, fRF = fVCO/O
–85
5.4
5.8
fVCO (GHz)
6948 G09
3.0
–55
–65
6.6
LTC6948-3 Frequency Step
Transient
KVCO (%/V)
HD3 (dBc)
O=4
–25
–30
6.5
O=1
O=2
–20
5.5
–45
O=5
O=3
–15
2.5
3.5
4.5
FREQUENCY (GHz)
MUTE Output Power vs fVCO and
Output Divide
(Single-Ended on RF+)
POUT AT fVCO/O (dBm)
–5
O=4
–40
O=6
–50 LTC6948-4, L = 68nH,
C
CS = 100pF , fRF = fVCO/O
–55
4.6
5.4
6.2
5.0
5.8
4.2
fVCO (GHz)
6948 G08
RF Output HD3 vs Output Divide
(Single-Ended on RF+)
O=2
O=1
–35
–45
FREQUENCY (GHz)
–4
–30
–0.5
ICP = 11.2mA
CPLE = 0
105°C
25°C
–40°C
–3
O=5
O=3
–25
HD2 (dBc)
1
–20
0
POUT (dBm)
ERROR (%)
5
4
–2
RF Output HD2 vs Output Divide
(Single-Ended on RF+)
RF Output Power vs Frequency
(Single-Ended on RF+)
Charge Pump Source Current
Error vs Voltage, Temperature
3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8 5.0
FREQUENCY (GHz)
6948 G14
2.5
3.8
4.2
4.6
5.0
FREQUENCY (GHz)
5.4
5.8
6948 G15
6948f
For more information www.linear.com/LTC6948
7
LTC6948
Typical
Performance Characteristics
TA = 25°C. VREF+ = VD+ = VRF+ = 3.3V, VCP+ = VVCO+ =
5V, INTN = 0, DITHEN = 1, CPLE = 1, RFO[1:0] = 3, unless otherwise noted.
–222
PHASE NOISE FLOOR (dBc/Hz)
KVCO (%/V)
6.0
5.5
5.0
4.5
4.0
–222
ICP = 5.6mA
CPLE = 1
7.0
–223
PHASE NOISE FLOOR (dBc/Hz)
7.5
6.5
Normalized In-Band Phase Noise
Floor vs CP Current
Normalized In-Band Phase Noise
Floor vs fVCO
LTC6948-4 VCO Tuning Sensitivity
–224
FRACTIONAL-N
–225
INTEGER-N
–226
fVCO = 5GHz
CPLE = 1
–223
–224
FRACTIONAL-N
–225
–226
INTEGER-N
3.5
3.0
4.3
4.7
5.1
5.5
5.9
FREQUENCY (GHz)
6.3
–227
6.7
2
3
LTC6948-1 VCO Phase Noise
–110
–130
PHASE NOISE (dBc/Hz)
–70
PHASE NOISE (dBc/Hz)
–80
–90
–100
–110
–120
–130
–90
–110
–120
–130
–140
–140
–150
–150
–150
–160
–160
10k
100k
1M
10M 40M
OFFSET FREQUENCY (Hz)
10k
100k
1M
10M 40M
OFFSET FREQUENCY (Hz)
1k
–75
fRF = fVCO = 6GHz
–50
PHASE NOISE (dBc/Hz)
–80
–90
–100
–110
–120
–130
fRF = fVCO/O
O=3
–90
–95
O=6
O=4
1k
10k
100k
1M
10M 40M
OFFSET FREQUENCY (Hz)
6948 G22
–105
2.2
2.4
2.6
fRF = fVCO/O
O=1
–80
O=2
–85
O=3
O=4
–90
–95
O=5
2.8 3.0 3.2
fVCO (GHz)
LTC6948-2 VCO Phase Noise vs
fVCO, Output Divide
(fOFFSET = 10kHz)
–75
O=2
–150
–160
–70
O=1
–85
–100
–140
10k
100k
1M
10M 40M
OFFSET FREQUENCY (Hz)
1k
6948 G21
LTC6948-1 VCO Phase Noise vs
fVCO, Output Divide
(fOFFSET = 10kHz)
–80
–60
–70
–160
6948 G20
LTC6948-4 VCO Phase Noise
–40
3.4
6948 G18
–100
–140
1k
11
–80
PHASE NOISE (dBc/Hz)
PHASE NOISE (dBc/Hz)
–60
–70
–120
9
fRF = fVCO = 5GHz
–50
–60
–100
7
ICP (mA)
LTC6948-3 VCO Phase Noise
–70
–90
5
–40
–60
–80
3
1
6948 G17
fRF = fVCO = 4GHz
–50
6948 G19
PHASE NOISE (dBc/Hz)
–227
6
LTC6948-2 VCO Phase Noise
–40
fRF = fVCO = 3GHz
–50
5
fVCO (GHz)
6948 G16
–40
4
3.6
3.8
6948 G23
–100
O=6
3.0
3.3
3.6
O=5
3.9
4.2
fVCO (GHz)
4.5
4.8
6948 G24
6948f
8
For more information www.linear.com/LTC6948
LTC6948
Typical
Performance Characteristics
TA = 25°C. VREF+ = VD+ = VRF+ = 3.3V, VCP+ = VVCO+ =
5V, INTN = 0, DITHEN = 1, CPLE = 1, RFO[1:0] = 3, unless otherwise noted.
–70
fRF = fVCO/O
–75
O=1
–80
O=2
–85
O=3
O=4
–90
–95
3.8
4.2
O=3
–80
4.6
5.0
fVCO (GHz)
fRF = fVCO/O
O=4
–85
O=5
O=6
5.4
–95
5.8
4.3
4.7
5.1
6948 G25
5.5
5.9
fVCO (GHz)
6.3
–120
O=3
O=4
O=6
3.3
3.6
3.9
4.2
fVCO (GHz)
O=5
4.5
O=1
–130
O=2
–135
–145
4.8
2.2
O=3
O=4
fRF = fVCO/O
4.2
O=6
O=5
O=6
3.8
O=4
4.6
5.0
fVCO (GHz)
5.4
–145
5.8
4.3
1k
1M
10k
100k
OFFSET FREQUENCY (Hz)
–120
–140
–150
10M 40M
6948 G31
PHASE NOISE (dBc/Hz)
PHASE NOISE (dBc/Hz)
–110
RMS NOISE = 0.141°
fPFD = 61.44MHz
O=2
INTN = 0
CPLE = 1
LOOP BW = 180kHz
NOTES 9, 12
–160
100
1k
1M
10k
100k
OFFSET FREQUENCY (Hz)
6948 G32
5.5
5.9
fVCO (GHz)
6.3
6.7
6948 G30
–110
–120
–130
–140
–150
10M 40M
5.1
O=5
Closed-Loop Phase Noise,
LTC6948-2 fRF = 3646.464MHz
–100
–130
4.7
6948 G29
–100
–160
100
6948 G27
O=1
–135
–100
–150
3.8
O=3
–90
–140
3.6
O=2
–90
RMS NOISE = 0.074°
fPFD = 25MHz
O=3
INTN = 0
CPLE = 1
fIB-SPUR = 572kHz
LOOP BW = 73kHz
NOTES 8, 13
3.4
–130
–90
–130
2.8 3.0 3.2
fVCO (GHz)
LTC6948-4 VCO Phase Noise vs
fVCO, Output Divide
(fOFFSET = 1MHz)
Closed-Loop Phase Noise,
LTC6948-3 fRF = 2377.728MHz
–120
2.6
–140
6948 G28
–110
2.4
–125
Closed-Loop Phase Noise,
LTC6948-1 fRF = 891.857MHz
PHASE NOISE (dBc/Hz)
O=6
O=5
–120
fRF = fVCO/O
–140
3.0
O=4
6948 G26
PHASE NOISE (dBc/Hz)
PHASE NOISE (dBc/Hz)
PHASE NOISE (dBc/Hz)
O=2
–145
O=3
–140
–150
6.7
LTC6948-3 VCO Phase Noise vs
fVCO, Output Divide
(fOFFSET = 1MHz)
O=1
–130
–140
O=2
–135
–145
–125
–135
O=1
–130
O=2
O=5
LTC6948-2 VCO Phase Noise vs
fVCO, Output Divide
(fOFFSET = 1MHz)
–125
fRF = fVCO/O
O=1
–90
O=6
LTC6948-1 VCO Phase Noise vs
fVCO, Output Divide
(fOFFSET = 1MHz)
–125
fRF = fVCO/O
–75
PHASE NOISE (dBc/Hz)
PHASE NOISE (dBc/Hz)
–70
LTC6948-4 VCO Phase Noise vs
fVCO, Output Divide
(fOFFSET = 10kHz)
PHASE NOISE (dBc/Hz)
LTC6948-3 VCO Phase Noise vs
fVCO, Output Divide
(fOFFSET = 10kHz)
RMS NOISE = 0.224°
fPFD = 61.44MHz
O=1
INTN = 0
CPLE = 1
LOOP BW = 178kHz
NOTE 9
–160
100
1k
1M
10k
100k
OFFSET FREQUENCY (Hz)
10M 40M
6948 G33
6948f
For more information www.linear.com/LTC6948
9
LTC6948
Typical
Performance Characteristics
TA = 25°C. VREF+ = VD+ = VRF+ = 3.3V, VCP+ = VVCO+ =
5V, INTN = 0, DITHEN = 1, CPLE = 1, RFO[1:0] = 3, unless otherwise noted.
POUT (dBm)
–40
–20
–40
–60
–80
–100dBc
–103dBc
–100
–120
–140
RBW = 100Hz
VBW = 100Hz
INTN = 0
CPLE = 1
O=2
NOTES 9, 14
–80
–91dBc
–78dBc
–40
–78dBc
–96dBc
–120
–120
3.3V CURRENT (mA)
O=2
O=3
O=4
O=5
O=6
89
36
88
35
87
34
86
33
32
84
83
–40
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5
OUTPUT FREQUENCY (GHz)
O = 1, MUTE = 0
RFO = 3, ICP = 5.6mA
–20
0
6948 G37
20
14
12
LDOV = 2
10
8
LDOV = 1
6
LDOV = 0
4
25
16
30
6948 G38
LDOV = 3, 65MHz
14
12
LDOV = 2, 55MHz
10
LDOV = 1, 45MHz
8
6
LDOEN = 0
15
100
LDOEN = 0, 75MHz
18
SUPPLY CURRENT (mA)
16
5
80
20
LDOV = 3
2
60
VD
Temperature
(INTN = 0, PDFN = 0, fPFD Noted)
TC = 25°C
18
20
40
TJ (°C)
31
+ Supply Current vs LDOV,
VD+ Supply Current vs LDOV, fPFD
(INTN = 0, PDFN = 0)
SUPPLY CURRENT (mA)
37
85
–75
0
38
EXCLUDES VD+
90
5V CURRENT (mA)
IB SPUR LEVEL (dBc)
39
91
O=1
–70
–80
6948 G36
92
SPUR IN BAND
–45 fPFD = 50MHz
CPLE = 1
–65
–200 –150 –100 –50 0 50 100 150 200
FREQUENCY OFFSET (MHz IN 10kHz SEGMENTS)
LTC6948-4 Supply Current
–40
–60
–140
–70dBc
–72dBc
6948 G35
Integer Boundary Spur Power vs
Output Frequency, LTC6948-3
–55
RBW = 100Hz
VBW = 100Hz
INTN = 0
CPLE = 1
O=1
NOTES 8, 14
–80
–100
–246 –184 –123–61.4 0 61.4 123 184 246
FREQUENCY OFFSET (MHz IN 10kHz SEGMENTS)
LTC6948-4 Spurious Response
fRF = 6236MHz, fREF = 100MHz,
fPFD = 50MHz, Loop BW = 152kHz
–60
–100
6948 G34
–50
0
–20
–60
–140
–100 –75 –50 –25 0 25 50 75 100
FREQUENCY OFFSET (MHz IN 10kHz SEGMENTS)
LTC6948-3 Spurious Response
fRF = 2377.73MHz, fREF = 61.44MHz,
fPFD = 61.44MHz, Loop BW = 180k
POUT (dBm)
–20
0
RBW = 10Hz
VBW = 10Hz
INTN = 0
CPLE = 1
O=3
NOTES 8, 14
POUT (dBm)
0
LTC6948-1 Spurious Response
fRF = 891.85MHz, fREF = 100MHz,
fPFD = 25MHz, Loop BW = 74kHz
35 45 55
fPFD (MHz)
65
75
4
–40
LDOV = 0, 30MHz
–20
6948 G39
0
20
40
TJ (°C)
60
80
100
6948 G40
6948f
10
For more information www.linear.com/LTC6948
LTC6948
Pin Functions
REF+, REF– (Pins 1, 28): Reference Input Signals. This
differential input is buffered with a low noise amplifier,
which feeds the reference divider. They are self-biased and
must be AC-coupled with 1µF capacitors. If used singleended with VREF+ ≤ 2.7VP-P, bypass REF– to GND with a
1µF capacitor. If used single-ended with VREF+ > 2.7VP-P,
bypass REF– to GND with a 47pF capacitor.
STAT (Pin 2): Status Output. This signal is a configurable
logical OR combination of the UNLOK, LOK, ALCHI, ALCLO,
THI, and TLO status bits, programmable via the STATUS
register. See the Operation section for more details.
CS (Pin 3): Serial Port Chip Select. This CMOS input initiates a serial port communication burst when driven low,
ending the burst when driven back high. See the Operation
section for more details.
SCLK (Pin 4): Serial Port Clock. This CMOS input clocks
serial port input data on its rising edge. See the Operation
section for more details.
SDI (Pin 5): Serial Port Data Input. The serial port uses
this CMOS input for data. See the Operation section for
more details.
SDO (Pin 6): Serial Port Data Output. This CMOS threestate output presents data from the serial port during a
read communication burst. Optionally attach a resistor of
>200k to GND to prevent a floating output. See the Applications Information section for more details.
LDO (Pin 7): Δ∑ Modulator LDO Bypass Pin. This pin
should be bypassed directly to the ground plane using a
low ESR (<0.8Ω) 0.1µF ceramic capacitor as close to the
pin as possible.
VD+ (Pin 8): 3.15V to 3.45V Positive Supply Pin for Serial
Port and Δ∑ Modulator Circuitry. This pin should be bypassed directly to the ground plane using a 0.1µF ceramic
capacitor as close to the pin as possible.
MUTE (Pin 9): RF Mute. The CMOS active-low input mutes
the RF± differential outputs while maintaining internal bias
levels for quick response to deassertion.
GND (Pins 10, 17, 21, Exposed Pad Pin 29): Negative
Power Supply (Ground). These pins should be tied directly
to the ground plane with multiple vias for each pin. The
package exposed pad must be soldered directly to the PCB
land. The PCB land pattern should have multiple thermal
vias to the ground plane for both low ground inductance
and also low thermal resistance.
RF–, RF+ (Pins 11, 12): RF Output Signals. The VCO
output divider is buffered and presented differentially on
these pins. The outputs are open-collector, with 136Ω
(typical) pull-up resistors tied to VRF+ to aid impedance
matching. If used single-ended, the unused output should
be terminated to 50Ω. See the Applications Information
section for more details on impedance matching.
VRF+ (Pin 13): 3.15V to 3.45V Positive Supply Pin for
RF circuitry. This pin should be bypassed directly to the
ground plane using a 0.01µF ceramic capacitor as close
to the pin as possible.
BB (Pin 14): RF Reference Bypass. This output has a 2.5k
resistance and must be bypassed with a 1µF ceramic capacitor to GND. Do not couple this pin to any other signal.
TUNE (Pin 15): VCO Tuning Input. This frequency control
pin is normally connected to the external loop filter. See
the Applications Information section for more details.
6948f
For more information www.linear.com/LTC6948
11
LTC6948
Pin Functions
TB (Pin 16): VCO Bypass. This output has a 2k resistance
and must be bypassed with a 2.2µF ceramic capacitor to
GND. It is normally connected to CMA, CMB, and CMC with
a short trace. Do not couple this pin to any other signal.
CMA, CMB, CMC (Pins 18, 19, 20): VCO Bias Inputs. These
inputs are normally connected to TB with a short trace
and bypassed with a 2.2µF ceramic capacitor to GND. Do
not couple these pins to any other signal. For best phase
noise performance, DO NOT place a trace between these
pads underneath the package.
BVCO (Pin 22): VCO Bypass Pin. This output must be
bypassed with a 1µF ceramic capacitor to GND. Do not
couple this pin to any other signal.
VVCO+ (Pin 23): 4.75V to 5.25V Positive Supply Pin for
VCO Circuitry. This pin should be bypassed directly to the
ground plane using a 0.01µF ceramic capacitor as close
to the pin as possible.
GND (24): Negative Power Supply (Ground). This pin is
attached directly to the die attach paddle (DAP) and should
be tied directly to the ground plane.
VCP+ (Pin 25): 4V to 5.25V Positive Supply Pin for Charge
Pump Circuitry. This pin should be bypassed directly to
the ground plane using a 0.1µF ceramic capacitor as close
to the pin as possible.
CP (Pin 26): Charge Pump Output. This bidirectional current
output is normally connected to the external loop filter.
See the Applications Information section for more details.
VREF+ (Pin 27): 3.15V to 3.45V Positive Supply Pin for
Reference Input Circuitry. This pin should be bypassed
directly to the ground plane using a 0.1µF ceramic capacitor as close to the pin as possible.
6948f
12
For more information www.linear.com/LTC6948
LTC6948
Block Diagram
1
28
REF+
2
3
4
5
6
8
7
REF–
STAT
≤425MHz
CS
27
25
VREF+
24
VCP+
≤100MHz
R_DIV
GND
1mA TO
11.2mA
PFD
÷1 TO 31
SERIAL
PORT
GND 21
B_DIV
SDI
CMA 20
÷32 TO 1023
N_DIV
SDO
VD+
CAL, ALC
CONTROL
373MHz TO 6.39GHz
1.7V TO 2.6V
+
–
LDO
REGULATOR
29
MUTE
9
GND
10
GND 17
÷1 TO 6, 50%
TUNE
O_DIV
15
2.24GHz TO 3.74GHz (LTC6948-1) OR
3.08GHz TO 4.91GHz (LTC6948-2) OR
3.84GHz TO 5.79GHz (LTC6948-3) OR
4.20GHz TO 6.390GHz (LTC6948-4)
MUTE
EXPOSED
PAD
CMB 19
CMC 18
∆∑
LDO
26
VVCO+ 23
LOCK
SCLK
CP
RF– RF+ VRF+
11
12
13
BB
TB
BVCO
14
16
22
6948 BD
6948f
For more information www.linear.com/LTC6948
13
LTC6948
Operation
The LTC6948 is a high-performance fractional-N PLL
complete with a low noise VCO available in four different
frequency range options. The output frequency range
may be further extended by utilizing the output divider
(see Available Options table for more details). The device
is able to achieve superior integrated phase noise by the
combination of its extremely low in-band phase noise
performance and excellent VCO noise characteristics.
The fractional-N feedback divider uses an advanced Δ∑
modulator, resulting in virtually no discrete modulator
spurious tones. The modulator may be disabled if integer-N
feedback is required.
Reference Input Buffer
The PLL’s reference frequency is applied differentially on
pins REF+ and REF–. These high impedance inputs are
self-biased and must be AC-coupled with 1µF capacitors
(see Figure 1 for a simplified schematic). Alternatively, the
inputs may be used single-ended by applying the reference frequency at REF+ and bypassing REF– to GND with
a 1µF capacitor. If the single-ended signal is greater than
2.7VP-P, then use a 47pF capacitor for the GND bypass.
VREF+
1
REF+
4.2k
LOWPASS
4.2k
FILT[1:0]
28
REF –
Table 1. FILT[1:0] Programming
FILT[1:0]
fREF
3
<20MHz
2
NA
1
20MHz to 50MHz
0
>50MHz
The BST bit should be set based upon the input signal level
to prevent the reference input buffer from saturating. See
Table 2 for recommended settings and the Applications
Information section for programming examples.
Table 2. BST Programming
BST
VREF
1
<2VP-P
0
≥2VP-P
VREF+
BIAS
1.9V
Additional options are available through serial port register
h0B to further refine the application. Bits FILT[1:0] control
the reference input buffer’s lowpass filter, and should be
set based upon fREF to limit the reference’s wideband
noise. The FILT[1:0] bits must be set correctly to reach the
LNORM normalized in-band phase noise floor. See Table 1
for recommended settings.
6948 F01
Reference (R) Divider
A 5-bit divider, R_DIV, is used to reduce the frequency
seen at the PFD. Its divide ratio R may be set to any integer from 1 to 31, inclusive. Use the RD[4:0] bits found in
registers h06 to directly program the R divide ratio. See
the Applications Information section for the relationship
between R and the fREF, fPFD, fVCO, and fRF frequencies.
BST
Figure 1. Simplified REF Interface Schematic
A high quality signal must be applied to the REF± inputs
as they provide the frequency reference to the entire PLL.
To achieve the part’s in-band phase noise performance,
apply a CW signal of at least 6dBm into 50Ω, or a square
wave of at least 0.5VP-P with slew rate of at least 40V/µs.
Phase/Frequency Detector (PFD)
The phase/frequency detector (PFD), in conjunction with
the charge pump, produces source and sink current pulses
proportional to the phase difference between the outputs
of the R and N dividers. This action provides the necessary
feedback to phase-lock the loop, forcing a phase align-
6948f
14
For more information www.linear.com/LTC6948
LTC6948
Operation
ment at the PFD’s inputs. The PFD may be disabled with
the CPRST bit which prevents UP and DOWN pulses from
being produced. See Figure 2 for a simplified schematic
of the PFD.
D
Q
UP
R DIV
RST
CPRST
DELAY
D
Q
DOWN
6948 F02
N DIV
RST
Figure 2. Simplified PFD Schematic
Table 3. LKWIN[2:0] Fractional Mode Programming
LKWIN[2:0]
tLWW
fVCO (CPLE = 1)
fVCO (CPLE = 0)
0
5.0ns
≥2.97GHz
≥1.35GHz
1
7.35ns
≥2.00GHz
≥919MHz
2
10.7ns
≥1.39GHz
≥632MHz
3
15.8ns
≥941MHz
≥428MHz
4
23.0ns
≥646MHz
≥294MHz
5
34.5ns
≥431MHz
≥196MHz
6
50.5ns
≥294MHz
≥134MHz
7
76.0ns
≥196MHz
≥89MHz
When using the device as an integer-N synthesizer (integer
mode), the phase difference seen at the PFD is minimized
by the feedback of the PLL and no longer depends upon
fVCO. Table 4 contains recommended settings for different
fPFD frequencies. See the Applications Information section
for examples.
Table 4. LKWIN[2:0] Integer Mode Programming
Lock Indicator
LKWIN[2:0]
tLWW
fPFD
The lock indicator uses internal signals from the PFD to
measure phase coincidence between the R and N divider
output signals. It is enabled by programming LKCT[1:0]
in the serial port register h0C (see Table 5), and produces
both LOCK and UNLOCK status flags, available through
both the STAT output and serial port register h00.
0
5.0ns
>6.8MHz
1
7.35ns
≤6.8MHz
The user sets the phase difference lock window time tLWW
for a valid LOCK condition with the LKWIN[2:0] bits. When
using the device as a fractional-N synthesizer (fractional
mode), the Δ∑ modulator changes the instantaneous phase
seen at the PFD on every R_DIV and N_DIV cycle. The
maximum allowable time difference in this case depends
upon both the VCO frequency fVCO and also the charge
pump linearization enable bit CPLE (see the Charge Pump
Linearizer section for an explanation of this function). Table
3 contains recommended settings for LKWIN[2:0] when
using the device in fractional mode. See the Applications
Information section for examples.
2
10.7ns
≤4.7MHz
3
15.8ns
≤3.2MHz
4
23.0ns
≤2.2MHz
5
34.5ns
≤1.5MHz
6
50.5ns
≤1.0MHz
7
76.0ns
≤660kHz
The PFD phase difference must be less than tLWW for the
COUNTS number of successive counts before the lock
indicator asserts the LOCK flag. The LKCT[1:0] bits found
in register h0C are used to set COUNTS depending upon
the application. Set LKCT to 0 to disable the lock indicator.
See Table 5 for LKCT[1:0] programming and the Applications Information section for examples.
6948f
For more information www.linear.com/LTC6948
15
LTC6948
Operation
Table 5. LKCT[1:0] Programming
VCP+
LKCT[1:0]
COUNTS
0
Lock Indicator Disabled
1
32
2
256
3
2048
VCP+
+
–
UP
CPUP
CPMID
0.9V
–
+
ICP
VCP+/2
When the PFD phase difference is greater than tLWW, the
lock indicator immediately asserts the UNLOCK status
flag and clears the LOCK flag, indicating an out-of-lock
condition. The UNLOCK flag is immediately de-asserted
when the phase difference is less than tLWW. See Figure 3
for more details.
Note that fREF must be present for the LOCK and UNLOCK
flags to properly assert and clear.
THI
–
+
DOWN
CPDN
+
–
CP
TLO
0.9V
VCP+
CP LINEARIZER
CONTROL
CP[2:0]
ND[9:0]
CPINV
ILIN
ENABLE
INTN
CPLE
+tLWW
PHASE
DIFFERENCE
AT PFD
26
6948 F04
Figure 4. Simplified Charge Pump Schematic
0
to the lower impedance of the loop filter components,
although currents larger than 5.6mA typically cause worse
spurious performance. See Table 6 for programming
specifics and the Applications Information section for
loop filter examples.
–tLWW
UNLOCK FLAG
Table 6. CP[2:0] Programming
t = COUNTS/fPFD
LOCK FLAG
6948 F03
Figure 3. UNLOCK and LOCK Timing
Charge Pump
The charge pump, controlled by the PFD, forces sink
(DOWN) or source (UP) current pulses onto the CP pin,
which should be connected to an appropriate loop filter.
See Figure 4 for a simplified schematic of the charge pump.
The output current magnitude ICP may be set from 1mA to
11.2mA using the CP[2:0] bits found in serial port register
h0C. A larger ICP can result in lower in-band noise due
CP[2:0]
ICP
0
1.0mA
1
1.4mA
2
2.0mA
3
2.8mA
4
4.0mA
5
5.6mA
6
8.0mA
7
11.2mA
The CPINV bit found in register h0D should be set for applications requiring signal inversion from the PFD, such as
for complex external loops using an inverting op amp. A
passive loop filter as shown in Figure 13 requires CPINV = 0.
6948f
16
For more information www.linear.com/LTC6948
LTC6948
Operation
Charge Pump Functions
The charge pump contains additional features to aid in
system startup. See Table 7 below for a summary.
Charge Pump Linearizer
Table 7. Charge Pump Function Bit Descriptions
BIT
CPCHI
DESCRIPTION
Enable High Voltage Output Clamp
CPCLO
Enable Low Voltage Output Clamp
CPDN
Force Sink Current
CPINV
Invert PFD Phase
CPLE
Linearizer Enable
CPMID
Enable Mid-Voltage Bias
CPRST
Reset PFD
CPUP
Force Source Current
CPWIDE
Extend Current Pulse Width
THI
High Voltage Clamp Flag
TLO
Low Voltage Clamp Flag
The CPWIDE bit extends the charge pump output current
pulse width by increasing the PFD reset path’s delay value
(see Figure 2). CPWIDE is normally set to 0.
When the LTC6948 is operated in fractional mode, the
charge pump’s current output versus its phase stimulus
(its gain linearity) must be extremely accurate. The CP
gain linearizer automatically adds a correction current ILIN
to minimize the charge pump’s impact on in-band phase
noise and spurious products during fractional operation.
The CP gain linearizer is enabled by setting CPLE = 1. It is
automatically disabled when in integer mode. CPLE should
be set to 0 if CPRST or CPMID are asserted to prevent the
linearizer from producing unintended currents.
VCO
The CPCHI and CPCLO bits found in register h0D enable
the high and low voltage clamps, respectively. When CPCHI
is enabled and the CP pin voltage exceeds approximately
VCP+ – 0.9V, the THI status flag is set, and the charge
pump sourcing current is disabled. Alternately, when
CPCLO is enabled and the CP pin voltage is less than
approximately 0.9V, the TLO status flag is set, and the
charge pump sinking current is disabled. See Figure 4 for
a simplified schematic.
The CPMID bit also found in register h0D enables a resistive VCP+/2 output bias which may be used to pre-bias
troublesome loop filters into a valid voltage range. When
using CPMID, it is recommended to also assert the CPRST
bit, forcing a PFD reset. Both CPMID and CPRST must be
set to 0 for normal operation.
The CPUP and CPDN bits force a constant ICP source or
sink current, respectively, on the CP pin. The CPRST bit
may also be used in conjunction with the CPUP and CPDN
bits, allowing a precharge of the loop to a known state,
if required. CPUP, CPDN, and CPRST must be set to 0 to
allow the loop to lock.
The integrated VCO is available in one of four frequency
ranges. The output frequency range may be further extended by utilizing the output divider (see Available Options
table, for more details). The wide frequency range of the
VCO, coupled with the output divider capability, allows the
LTC6948 to cover an extremely wide range of continuously
selectable frequencies.
The BB and TB pins are used to bias internal VCO circuitry. The BB pin has a 2k output resistance and should
be bypassed with a 1µF ceramic capacitor to GND, giving
a time constant of 2ms. The TB pin has a 2.5k output
resistance and should be bypassed with a 2.2µF ceramic
capacitor to GND, resulting in a time constant of 5.5ms.
Stable bias voltages are achieved after approximately 3
time constants following power-up.
VCO Calibration
The VCO must be calibrated each time its frequency is
changed by changing any of fREF, the R divider value, the
N divider value, or Δ∑ modulator fractional value, but not
the O divider (see the Applications Information section
6948f
For more information www.linear.com/LTC6948
17
LTC6948
Operation
for the relationship between R, N, NUM, O, and the fREF,
fPFD, fVCO, and fRF frequencies). The output frequency is
then stable over the LTC6948’s entire temperature range,
regardless of the temperature at which it was calibrated,
until the part is reset due to a power cycle or software
power-on reset (POR).
The output of the B divider is used to clock digital calibration circuitry as shown in the Block Diagram. The B value,
programmed with bits BD[3:0], and fPFD determine the
calibration frequency fCAL. Calculate the B value using
Equation 1.
fPFD
Once the RD[4:0], ND[9:0], NUM[17:0], and BD[3:0] bits
are written and the reference frequency fREF is present and
stable at the REF± inputs, the VCO must be calibrated by
setting CAL = 1 (the bit is self-clearing). The calibration
cycle takes between 12 and 14 clocks of the B divider
output. The MTCAL bit may be set to mute the RF output
during the calibration.
Note that the fREF frequency and TB and BB voltages must
be stable for proper calibration. Stable bias voltages are
achieved after approximately 3 time constants following
power-up.
The maximum calibration frequency fCAL-MAX for each part
option is shown in Table 8.
Setting AUTOCAL = 1 causes the CAL bit to be set automatically whenever any of serial port registers h06 through
h0A is written. When AUTOCAL is enabled, there is no
need for a separate register write to set the CAL bit. See
Table 10 for a summary of the VCO bits.
Table 8. Maximum Calibration Frequency
Table 10. VCO Bit Descriptions
B≥
fCAL-MAX
(1)
BIT
DESCRIPTION
PART
fCAL-MAX (MHz)
LTC6948-1
1.0
LTC6948-2
1.33
CAL
Start VCO Calibration (Auto Clears)
LTC6948-3
1.7
MTCAL
Mute RF Output During Calibration
LTC6948-4
1.8
The relationship between bits BD(3:0) and the B value is
shown in Table 9.
Table 9. BD[3:0] Programming
BD[3:0]
B DIVIDE VALUE
AUTOCAL Calibrate VCO Whenever Registers h06 to h0A Are Written
VCO Automatic Level Control (ALC)
The VCO uses an internal automatic level control (ALC)
algorithm to maintain an optimal amplitude on the VCO
resonator, and thus optimal phase noise performance. The
user has several ALC configuration and status reporting
options as seen in Table 11.
0
8
1
12
2
16
3
24
4
32
5
48
6
64
7
96
8
128
9
192
ALCMON Enable ALC Monitoring for Status Flags Only; Does Not Enable
Amplitude Control
10
256
ALCULOK Auto Enable ALC when PLL Unlocked
11
384
12 to 15
Invalid
Table 11. ALC Bit Descriptions
BIT
DESCRIPTION
ALCCAL
Auto Enable ALC during CAL operation
ALCEN
Always Enable ALC (Overrides ALCCAL, ALCMON, and
ALCULOK)
ALCHI
ALC too High Flag (Resonator Amplitude too High)
ALCLO
ALC too Low Flag (Resonator Amplitude too Low)
6948f
18
For more information www.linear.com/LTC6948
LTC6948
Operation
Changes in the internal ALC output can cause extremely
small jumps in the VCO frequency. These jumps may be
acceptable in some applications but not in others. Use the
above table to choose when the ALC is active. The ALCHI
and ALCLO flags, valid only when the ALC is active or the
ALCMON bit is set, may be used to monitor the resonator
amplitude.
The ALC must be allowed to operate during or after a
calibration cycle. At least one of the ALCCAL, ALCEN, or
ALCULOK bits must be set.
VCO (N) Divider
The 10-bit N divider provides the feedback from the VCO to
the PFD. Its divide ratio N is restricted to any integer from
35 to 1019, inclusive, when in fractional mode. The divide
ratio may be programmed from 32 to 1023, inclusive, when
in integer mode. Use the ND[9:0] bits found in registers
h06 and h07 to directly program the N divide ratio. See
the Applications Information section for the relationship
between N and the fREF, fPFD, fVCO, and fRF frequencies.
Δ∑ Modulator
The Δ∑ modulator changes the N divider’s ratio each PFD
cycle to achieve an average fractional divide ratio. The
fractional numerator NUM[17:0] is programmable from
1 to 262143, or 218 – 1. The fractional denominator is
fixed at 262144 (or 218), with the resulting fractional ratio
F given by Equation 4. See the Applications Information
section for the relationship between NUM, F, and the fREF,
fPFD, fVCO, and fRF frequencies.
The Δ∑ modulator uses digital signal processing (DSP)
techniques to achieve an average fractional divide ratio.
The modulator is clocked at the fPFD rate. This process
produces output modulation noise known as quantization
noise with a highpass frequency response. The external
lowpass loop filter is used to filter this quantization noise to
a level beneath the phase noise of the VCO. This prevents
the noise from contributing to the overall phase noise of
the system. The loop filter must be designed to adequately
filter the quantization noise.
The oversampling ratio OSR is defined as the ratio of the
Δ∑ modulator clock frequency fPFD to the loop bandwidth
BW of the PLL (see Equation 11). See the Applications
Information section for guidelines concerning the OSR
and the loop filter.
When the desired output frequency is such that the needed
NUM value is 0, the LTC6948 should be operated in integer
mode (INTN = 1). In integer mode, the modulator is placed
in standby, with all blocks still powered up, thus allowing
it to resume fractional operation immediately.
Enable numerator dither mode (DITHEN = 1) to further
reduce spurious produced by the modulator. Dither has
no measurable impact on in-band phase noise, and is
enabled by default. See Table 12 for a complete list of
modulator bit descriptions.
Modulator Reset
To achieve consistent spurious performance, the modulator
DSP circuitry should be re-initialized by setting RSTFN = 1
whenever NUM[17:0] is changed. Setting AUTORST = 1
causes the RSTFN bit to be set automatically whenever
any of serial port registers h05 through h0A are written.
When AUTORST is enabled, there is no need for a separate register write to set the RSTFN bit. See Table 12 for
a summary of the modulator bits.
Table 12. Fractional Modulator Bit Descriptions
BIT
AUTORST
DITHEN
INTN
DESCRIPTION
Automatically Reset Modulator when Registers h05 to h0A
Are Written
Enable Fractional Numerator Dither
Integer Mode; Fractional Modulator Placed in Standby
RSTFN
Reset Modulator (Auto Clears)
SEED
Seed Value for Pseudorandom Dither Algorithm
6948f
For more information www.linear.com/LTC6948
19
LTC6948
Operation
LDO Regulator
The adjustable low dropout (LDO) regulator supplies power
to the Δ∑ modulator. The regulator requires a low ESR
ceramic capacitor (ESR < 0.8Ω) connected to the LDO pin
(pin 7) for stability. The capacitor value may range from
0.047µF to 1µF.
The LDO voltage is set using the LDOV[1:0] bits, and should
be chosen based upon the fPFD frequency to minimize
power and spurious. The regulator is disabled by setting
the LDOEN bit to 0. When disabled by using either the
LDOEN or PDFN bits, the LDO pin is connected directly
to VD+ using a low impedance switch, and the regulator
is powered down. See Table 13 for programming details.
Table 13. LDOV[1:0] and LDOEN Programming
LDOV[1:0]
LDOEN
VLDO
fPFD
0
1
1.7V
≤34.3MHz
1
1
2.0V
≤45.9MHz
2
1
2.3V
≤56.1MHz
3
1
2.6V
≤66.3MHz
X
0
VD+
≤76.1MHz
Output (O) Divider
be combined externally, or used individually. Terminate
any unused output with a 50Ω resistor to VRF+.
Table 14. RFO[1:0] Programming
RFO[1:0}
PRF (DIFFERENTIAL)
PRF (SINGLE-ENDED)
0
–4.3dBm
–7.3dBm
1
–1.5dBm
–4.5dBm
2
1.6dBm
–1.4dBm
3
4.5dBm
1.5dBm
Each output is open-collector with 136Ω pull-up resistors
to VRF+, easing impedance matching at high frequencies.
See Figure 5 for circuit details and the Applications Information section for matching guidelines. The buffer may be
muted with either the OMUTE bit, found in register h02,
or by forcing the MUTE input low.
VRF+
136Ω
136Ω
RF+
RF–
9
The 3-bit O divider can reduce the frequency from the VCO
to extend the output frequency range. Its divide ratio O
may be set to any integer from 1 to 6, inclusive, outputting
a 50% duty cycle even with odd divide values. Use the
OD[2:0] bits found in register h0B to directly program the
O divide ratio. See the Applications Information section
for the relationship between O and the fREF, fPFD, fVCO,
and fRF frequencies.
VRF+
MUTE
OMUTE
12
11
MUTE
RFO[1:0]
6948 F05
Figure 5. Simplified RF Interface Schematic
Serial Port
The SPI-compatible serial port provides control and
monitoring functionality. A configurable status output
STAT gives additional instant monitoring.
Communication Sequence
RF Output Buffer
The low noise, differential output buffer produces a differential output power of –4.3dBm to +4.5dBm, settable
with bits RFO[1:0] according to Table 14. The outputs may
The serial bus is comprised of CS, SCLK, SDI, and SDO.
Data transfers to the part are accomplished by the serial bus master device first taking CS low to enable the
LTC6948’s port. Input data applied on SDI is clocked on
6948f
20
For more information www.linear.com/LTC6948
LTC6948
Operation
the rising edge of SCLK, with all transfers MSB first. The
communication burst is terminated by the serial bus master
returning CS high. See Figure 6 for details.
Single Byte Transfers
The serial port is arranged as a simple memory map, with
status and control available in 15 byte-wide registers. All
data bursts are comprised of at least two bytes. The seven
most significant bits of the first byte are the register address,
with an LSB of 1 indicating a read from the part, and LSB
of 0 indicating a write to the part. The subsequent byte,
or bytes, is data from/to the specified register address.
See Figure 8 for an example of a detailed write sequence,
and Figure 9 for a read sequence.
Data is read from the part during a communication burst
using SDO. Readback may be multidrop (more than one
LTC6948 connected in parallel on the serial bus), as SDO
is three-stated (Hi-Z) when CS = 1, or when data is not
being read from the part. If the LTC6948 is not used in
a multidrop configuration, or if the serial port master is
not capable of setting the SDO line level between read
sequences, it is recommended to attach a high value
resistor of greater than 200k between SDO and GND to
ensure the line returns to a known level during Hi-Z states.
See Figure 7 for details.
Figure 10 shows an example of two write communication
bursts. The first byte of the first burst sent from the serial
bus master on SDI contains the destination register ad-
MASTER–CS
tCSS
tCKL
tCSS
tCKH
tCSH
MASTER–SCLK
tCS
MASTER–SDI
tCH
DATA
DATA
6948 F06
Figure 6. Serial Port Write Timing Diagram
MASTER–CS
8TH CLOCK
MASTER–SCLK
tDO
LTC6948–SDO
tDO
Hi-Z
tDO
tDO
DATA
DATA
Hi-Z
6948 F07
Figure 7. Serial Port Read Timing Diagram
MASTER–CS
16 CLOCKS
MASTER–SCLK
7-BIT REGISTER ADDRESS
MASTER–SDI
LTC6948–SD0
8 BITS OF DATA
A6 A5 A4 A3 A2 A1 A0 0 D7 D6 D5 D4 D3 D2 D1 D0
Hi-Z
0 = WRITE
6948 F08
Figure 8. Serial Port Write Sequence
6948f
For more information www.linear.com/LTC6948
21
LTC6948
Operation
dress (Addr0) and an LSB of 0 indicating a write. The next
byte is the data intended for the register at address Addr0.
CS is then taken high to terminate the transfer. The first
byte of the second burst contains the destination register
address (Addr1) and an LSB indicating a write. The next
byte on SDI is the data intended for the register at address
Addr1. CS is then taken high to terminate the transfer.
byte and its data in the second byte as before, but continues
sending bytes destined for subsequent registers. Byte 1’s
address is Addr0+1, Byte 2’s address is Addr0+2, and so
on. If the register address pointer attempts to increment
past 14 (h0E), it is automatically reset to 0.
An example of an auto-increment read from the part is
shown in Figure 12. The first byte of the burst sent from
the serial bus master on SDI contains the destination register address (Addr0) and an LSB of 1 indicating a read.
Once the LTC6948 detects a read burst, it takes SDO out
of the Hi-Z condition and sends data bytes sequentially,
beginning with data from register Addr0. The part ignores
all other data on SDI until the end of the burst.
Multiple Byte Transfers
More efficient data transfer of multiple bytes is accomplished by using the LTC6948’s register address autoincrement feature as shown in Figure 11. The serial port
master sends the destination register address in the first
MASTER–CS
16 CLOCKS
MASTER–SCLK
7-BIT REGISTER ADDRESS
1 = READ
A6 A5 A4 A3 A2 A1 A0 1
MASTER–SDI
8 BITS OF DATA
Hi-Z
LTC6948–SDO
X D7 D6 D5 D4 D3 D2 D1 D0 DX
Hi-Z
6948 F09
Figure 9. Serial Port Read Sequence
MASTER–CS
ADDR0 + Wr
MASTER–SDI
LTC6948–SDO
BYTE 0
ADDR1 + Wr
BYTE 1
Hi-Z
6948 F10
Figure 10. Serial Port Single Byte Write
MASTER–CS
ADDR0 + Wr
MASTER–SDI
LTC6948–SDO
BYTE 0
BYTE 1
BYTE 2
Hi-Z
6948 F11
Figure 11. Serial Port Auto-Increment Write
MASTER–CS
ADDR0 + Rd
MASTER–SDI
LTC6948–SDO
Hi-Z
DON’T CARE
BYTE 0
BYTE 1
Hi-Z
BYTE 2
6948 F12
Figure 12. Serial Port Auto-Increment Read
6948f
22
For more information www.linear.com/LTC6948
LTC6948
Operation
Multidrop Configuration
STAT Output
Several LTC6948s may share the serial bus. In this multidrop configuration, SCLK, SDI, and SDO are common
between all parts. The serial bus master must use a separate
CS for each LTC6948 and ensure that only one device has
CS asserted at any time. It is recommended to attach a
high value resistor to SDO to ensure the line returns to a
known level during Hi-Z states.
The STAT output pin is configured with the x[5:0] bits
of register h01. These bits are used to bit-wise mask, or
enable, the corresponding status flags of status register
h00, according to Equation 2. The result of this bit-wise
Boolean operation is then output on the STAT pin.
STAT = OR (Reg00[5:0] AND Reg01[5:0])(2)
or, expanded,
Serial Port Registers
STAT = (UNLOCK AND x[5]) OR
The memory map of the LTC6948 may be found below in
Table 15, with detailed bit descriptions found in Table 16.
The register address shown in hexadecimal format under
the ADDR column is used to specify each register. Each
register is denoted as either read-only (R) or read-write
(R/W). The register’s default value on device power-up or
after a reset is shown at the right.
The read-only register at address h00 is used to determine
different status flags. These flags may be instantly output
on the STAT pin by configuring register h01. See STAT
Output section that follows for more information.
(ALCHI AND x[4]) OR
(ALCLO AND x[3]) OR
(LOCK AND x[2]) OR
(THI AND x[1]) OR
(TLO AND x[0])
For example, if the application requires STAT to go high
whenever the ALCHI, ALCLO, or THI flags are set, then
x[4], x[3], and x[1] should be set to 1, giving a register
value of h1A.
The read-only register at address h0E is a ROM byte for
device identification.
Table 15. Serial Port Register Contents
ADDR
MSB
[6]
[5]
[4]
[3]
[2]
[1]
LSB
R/W
h00
*
*
UNLOCK
ALCHI
ALCLO
LOCK
THI
TLO
R
DEFAULT
h01
*
*
x[5]
x[4]
x[3]
x[2]
x[1]
x[0]
R/W
h02
PDALL
PDPLL
PDVCO
PDOUT
PDFN
MTCAL
OMUTE
POR
R/W
h06
h03
ALCEN
ALCMON
ALCCAL
ALCULOK
AUTOCAL
AUTORST
DITHEN
INTN
R/W
h3E
h04
h04
BD[3]
BD[2]
BD[1]
BD[0]
CPLE
LDOEN
LDOV[1]
LDOV[0]
R/W
h47
h05
SEED[7]
SEED[6]
SEED[5]
SEED[4]
SEED[3]
SEED[2]
SEED[1]
SEED[0]
R/W
h11
h06
RD[4]
RD[3]
RD[2]
RD[1]
RD[0]
*
ND[9]
ND[8]
R/W
h08
h07
ND[7]
ND[6]
ND[5]
ND[4]
ND[3]
ND[2]
ND[1]
ND[0]
R/W
hFA
h08
*
*
NUM[17]
NUM[16]
NUM[15]
NUM[14]
NUM[13]
NUM[12]
R/W
h3F
h09
NUM[11]
NUM[10]
NUM[9]
NUM[8]
NUM[7]
NUM[6]
NUM[5]
NUM[4]
R/W
hFF
h0A
NUM[3]
NUM[2]
NUM[1]
NUM[0]
*
*
RSTFN
CAL
R/W
hF0
h0B
BST
FILT[1]
FILT[0]
RFO[1]
RFO[0]
OD[2]
OD[1]
OD[0]
R/W
hF9
h0C
LKWIN[2]
LKWIN[1]
LKWIN[0]
LKCT[1]
LKCT[0]
CP[2]
CP[1]
CP[0]
R/W
h4F
h0D
CPCHI
CPCLO
CPMID
CPINV
CPWIDE
CPRST
CPUP
CPDN
R/W
hE4
REV[3]
REV[2]
REV[1]
REV[0]
PART[3]
PART[2]
PART[1]
PART[0]
R
hxx†
h0E
*unused
†varies depending on version
6948f
For more information www.linear.com/LTC6948
23
LTC6948
Operation
Block Power-Down Control
The LTC6948’s power-down control bits are located in
register h02, described in Table 16. Different portions of
Table 16. Serial Port Register Bit Field Summary
BITS
DESCRIPTION
the device may be powered down independently. Care must
be taken with the LSB of the register, the POR (power-onreset) bit. When written to a 1, this bit forces a full reset
of the part’s digital circuitry to its power-up default state.
BITS
DEFAULT
LDOEN
DESCRIPTION
LDO Enable
DEFAULT
1
ALCCAL
Auto Enable ALC During CAL Operation
1
LDOV[1:0] LDO Voltage
h3
ALCEN
Always Enable ALC (Override)
1
LKCT[1:0]
h1
ALCHI
ALC too Hi Flag
ALCLO
ALC too Low Flag
PLL Lock Cycle Count
LKWIN[2:0] PLL Lock Indicator Window
LOCK
ALCMON
Enable ALC Monitor for Status Flags Only
0
ALCULOK
Enable ALC When PLL Unlocked
0
AUTOCAL
Calibrate VCOs Whenever Registers h06 to
h0A Are Written
1
AUTORST
Reset Modulator Whenever Registers h05
to h0A Are Written
1
BD[3:0]
Calibration B Divider Value
h3
BST
REF Buffer Boost Current
1
CAL
Start VCO Calibration (Auto Clears)
0
CP[2:0]
CP Output Current
h7
CPCHI
Enable Hi-Voltage CP Output Clamp
1
CPCLO
Enable Low-Voltage CP Output Clamp
1
PLL Lock Indicator Flag
MTCAL
Mute RF Output During Calibration
ND[9:0]
N Divider Value (ND[9:0] ≥ 32)
NUM[17:0] Fractional Numerator Value
OD[2:0]
Output Divider Value (0 < OD[2:0] < 7)
OMUTE
Mutes RF Output
PART[3:0]
h2
Part Code (h01 for -1, h02 for -2, h03 for
-3, h04 for -4 version)
1
h0FA
h3FFF
h1
1
h01, h02,
h03, h04
PDALL
Full Chip Powerdown
0
PDFN
Powers Down LDO and Modulator Clock
0
PDOUT
Powers Down O_DIV, RF Output Buffer
0
PDPLL
Powers Down REF, R_DIV, PFD, CPUMP,
N_DIV
0
Powers Down VCO, N_DIV
0
Force Power-On-Reset
0
CPDN
Force CP Pump Down
0
PDVCO
CPINV
Invert CP Phase
0
POR
CPLE
CP Linearizer Enable
0
RD[4:0]
R Divider Value (RD[4:0] > 0)
CPMID
CP Bias to Mid-Rail
1
REV[3:0]
Rev Code
h1
CPRST
CP Tri-State
1
RFO[1:0]
RF Output Power
h3
CPUP
Force CP Pump Up
0
RSTFN
Force Modulator Reset (Auto Clears)
0
CPWIDE
Extend CP Pulse Width
0
DITHEN
Enable Fractional Numerator Dither
1
THI
CP Clamp High Flag
FILT[1:0]
REF Input Buffer Filter
h3
TLO
CP Clamp Low Flag
Integer Mode; Fractional Modulator Placed
in Standby
0
UNLOK
PLL Unlock Flag
x[5:0]
STAT Output OR Mask
INTN
SEED[7:0] Modulator Dither Seed Value
h001
h11
h04
6948f
24
For more information www.linear.com/LTC6948
LTC6948
Applications Information
Introduction
where the fractional value F is given by Equation 4:
A PLL is a complex feedback system that may conceptually be considered a frequency multiplier. The system
multiplies the frequency input at REF± and outputs a higher
frequency at RF±. The PFD, charge pump, N divider, VCO
and external loop filter form a feedback loop to accurately
control the output frequency (see Figure 13).
fPFD =
fREF
R fVCO = fPFD • (N + F)(6)
The output frequency fRF produced at the output of the O
divider is given by Equation 7:
fRF =
fVCO
O fSTEP(MIN) =
fREF
(8)
R •O• 218 (3)
LOOP FILTER
(FOURTH ORDER)
LTC6948
REF±
(7)
Using the above equations, the minimum output frequency
resolution fSTEP(MIN) produced by a unit change in the
fractional numerator NUM while in fractional mode is
given by Equation 8:
When the loop is locked, the frequency fVCO (in Hz)
produced at the output of the VCO is determined by the
reference frequency fREF, the R and N divider values, and
the fractional value F, given by Equation 3:
(fREF)
(5)
and fVCO may be alternatively expressed as:
Output Frequency
(4)
The PFD frequency fPFD is given by the following equation:
The R and O divider and input frequency fREF are used to
set the output frequency resolution. When in fractional
mode, the Δ∑ modulator changes the N divider’s ratio
each PFD cycle to produce an average fractional divide
ratio. This achieves a much smaller frequency resolution
for a given fPFD as compared to integer mode.
fREF • (N + F )
R
NUM
218 NUM is programmable from 1 to 262143, or 218 – 1. When
using the LTC6948 in integer mode, F = 0.
The external loop filter is used to set the PLL’s loop
bandwidth BW. Lower bandwidths generally have better
spurious performance and lower Δ∑ modulator quantization
noise. Higher bandwidths can have better total integrated
phase noise.
fVCO =
F=
R_DIV
fPFD
KPFD
ICP
CP
26
÷R
C2
L1
R1
÷(N + F)
RZ
CP
CI
N_DIV
LF(s)
∆∑
RF±
(fRF)
O_DIV
÷O
TUNE
fVCO
15
KVCO
6948 F13
Figure 13. PLL Loop Diagram
6948f
For more information www.linear.com/LTC6948
25
LTC6948
Applications Information
Alternatively, to calculate the numerator step size NUMSTEP
needed to produce a given frequency step fSTEP(FRAC), use
Equation 9:
NUM STEP =
fSTEP(FRAC) •R •O • 218
fREF
(9)
The output frequency resolution fSTEP(INT) produced by a
unit change in N while in integer mode is given by Equation 10:
f
fSTEP(INT) = REF
R •O (10)
A stable loop, both in integer and fractional mode,
requires that the OSR is greater than or equal to 10.
Further, in fractional mode, OSR must be high enough
to allow the loop filter to reduce modulator quantization
noise to an acceptable level.
Choosing a higher order loop filter when using the Δ∑
modulator allows for a smaller OSR, and thus a larger
loop bandwidth. Linear Technology’s FracNWizard helps
choose the appropriate OSR and BW values.
3)Select loop filter component RZ and charge pump current ICP based on BW and the VCO gain factor, KVCO.
BW (in Hz) is approximated by the following equation:
BW ≅
Loop Filter Design
A stable PLL system requires care in designing the external
loop filter. The Linear Technology FracNWizard application, available from www.linear.com, aids in design and
simulation of the complete system.
The loop design should use the following algorithm:
1)Determine the output frequency fRF and frequency step
size fSTEP based on application requirements. Using
Equations 3, 5, 7, and 8, change fREF, N, R, and O until
the application frequency constraints are met. Use the
minimum R value that still satisfies the constraints.
Then calculate B using Equation 1 and Tables 8 and 9.
2)Select the open loop bandwidth BW constrained by
fPFD and oversampling ratio OSR. The OSR is the ratio
of fPFD to BW (see Equation 11):
f
OSR = PFD
BW
(11)
fPFD
OSR where BW and fPFD are in Hz.
(12)
or
RZ =
2 • π • BW • N
ICP • K VCO
where KVCO is in Hz/V, ICP is in Amps, and RZ is in
Ohms. KVCO is obtained from the VCO Tuning Sensitivity in the Electrical Characteristics. Use ICP = 5.6mA to
lower in-band noise unless component values force a
lower setting.
4)Select loop filter components CI and CP based on BW
and RZ. A reliable second-order loop filter design can
be achieved by using the following equations for the
loop capacitors (in Farads).
or
BW =
ICP • RZ • K VCO
2 • π •N
CI =
3.5
2 • π • BW • RZ (13)
1
7 • π • BW • RZ (14)
CP =
Use FracNWizard to aid in the design of higher order
loop filters.
6948f
26
For more information www.linear.com/LTC6948
LTC6948
Applications Information
Design and Programming Example
This programming example uses the DC1959 with the
LTC6948-2. Assume the following parameters of interest:
fREF = 100MHz at 7dBm into 50Ω
A calibration cycle takes 12 to 14 clock cycles of fCAL. This
gives a VCO calibration time of approximately:
fSTEP = 50kHz
tCAL ≅
14
B
= 14 •
= 13.4µs
fCAL
fPFD
Selecting Filter Type and Loop Bandwidth
fRF = 1921.650MHz
The next step in the algorithm is choosing the open loop
bandwidth. Select the minimum bandwidth resulting from
the below constraints.
From the Electrical Characteristics table:
fVCO = 3.080GHz to 4.910GHz
1)The OSR must be at least 10 (sets absolute maximum BW).
KVCO% = 4.7%Hz/V to 7%Hz/V
Determining Divider Values
2)The integrated phase noise due to thermal noise should
be minimized, neglecting any modulator noise.
Following the loop filter design algorithm, first determine
all the divider values. The maximum fPFD while in fractional
mode is less than 100MHz, so R must be greater than 1.
3)The loop bandwidth must be narrow enough to adequately filter the modulator’s quantization noise.
R=2
Then, using Equations 5 and 7, calculate the following
values:
O=2
FracNWizard reports the thermal noise optimized loop
bandwidth is 211kHz. Filter 3 (fourth order response) has
a quantization noise constrained BW of 150kHz, making
it a good choice. Select Filter 3 and use the smaller of the
two bandwidths (150kHz) for optimal integrated phase
noise. Use Equation 11 to calculate OSR:
fPFD = 50MHz
Then using Equation 6:
N+F =
FracNWizard reports the loop bandwidths resulting from
each of the above constraints. The quantization noise constrained results vary according to the shape of the external
loop filter. FracNWizard reports an optimal bandwidth for
several filter types.
2 •1921.650MHz
= 76.866
50MHz
Therefore:
N = 76
F = 0.866
OSR =
50MHz
= 333.3
150kHz
Loop Filter Component Selection
Then, from Equation 4,
NUM = 0.866 • 218 = 227017
Also, from Equation 1 and Tables 8 and 9 determine B:
Now set loop filter resistor RZ and charge pump current
ICP. Because the KVCO varies over the VCO’s frequency
range, using the KVCO geometric mean gives good results:
K VCO = 3.843 • 10 9 • 0.047 • 0.07
B = 48 and BD[3:0] = h5
= 220.4MHz / V
6948f
For more information www.linear.com/LTC6948
27
LTC6948
Applications Information
Using an ICP of 5.6mA, the FracNWizard uses Equation 12
to determine RZ:
the Δ∑ modulator modes (DITHEN = 1, INTN = 0) at the
same time:
2 • π •150k • 76
5.6m • 220.4M
R
=
58.0Ω
Z
Reg03 = h7E
RZ =
For the 4th order Filter 3, FracNWizard uses modified
Equations 13 and 14 to calculate CI, CP:
4.5
= 82.3nF
2 • π •150k • 58
1
CP =
= 3.5nF
10.5 • π •150k • 58
CI =
The ALC will only be active during a calibration cycle or
when the loop is unlocked, but the ALCHI and ALCLO status
conditions will be monitored continuously. The VCO will
be calibrated and the Δ∑ modulator will be reset at the
end of the SPI write communication burst (assuming an
auto-increment write is used to write all registers).
LDO Programming
FracNWizard calculates R1, L1, and C2 to be:
R1 = 58.0Ω
Use Table 13 and fPFD = 50MHz to determine V(LDO) and
LDOV[1:0]:
V(LDO) = 2.3V and LDOV[1:0] = 2
Use LDOV[1:0], LDOEN = 1 to enable the LDO, and the
previously determined BD[3:0] value to set Reg04. CPLE
should be set to 1 to reduce in-band noise and spurious
due to the Δ∑ modulator:
C2 = 2.3nF
L1 = 7.8uH
Status Output Programming
Reg04 = h5E
This example will use the STAT pin to alert the system
whenever the LTC6948 generates a fault condition. Program x[5], x[4], x[3], x[1], x[0] = 1 to force the STAT pin
high whenever any of the UNLOCK, ALCHI, ALCLO, THI,
or TLO flags asserts:
Reg01 = h3B
SEED Programming
The SEED[7:0] value is used to initialize the Δ∑ modulator
dither circuitry. Use the default value:
Reg05 = h11
R and N Divider and Numerator Programming
Power Register Programming
For correct PLL operation all internal blocks should be
enabled. OMUTE may remain asserted (or the MUTE pin
held low) until programming is complete. For OMUTE = 1:
Program registers Reg06 to Reg0A with the previously
determined R and N divider and numerator values. Because
the AUTORST and AUTOCAL bits were previously set to
1, CAL and RSTFN do not need to be set:
Reg02 = h02
Reg06 = h10
VCO ALC, AUTOCAL, and AUTORST Programming
Reg07 = h4C
Set the ALC options (ALCMON = ALCULOK = ALCCAL = 1),
the auto reset options (AUTOCAL = AUTORST = 1), and
Reg08 = h37
Reg09 = h6C
Reg0A = h90
6948f
28
For more information www.linear.com/LTC6948
LTC6948
Applications Information
Reference Input Settings and Output Divider
Programming
From Table 1, FILT = 0 for a 100MHz reference frequency.
Next, convert 7dBm into VP–P. For a CW tone, use the
following equation with R = 50:
(dBm – 21)/20
VP-P ≅ R • 10
(15)
This gives VP-P = 1.41V, and, according to Table 2, set
BST = 1.
RF±
output
Now program Reg0B, assuming maximum
power (RFO[1:0] = 3 according to Table 14) and OD[2:0] = 2:
Reg0B = h9A
Lock Detect and Charge Pump Current Programming
Next, determine the lock indicator window from fPFD. From
Table 3 we see that LKWIN[1:0] = 0 with a tLWW of 5ns for
CPLE = 1 and fVCO = 3.843GHz. The LTC6948 will consider
the loop locked as long as the phase coincidence at the
PFD is within 90°, as calculated below.
phase = 360° • tLWW • fPFD = 360 • 5n • 50M
= 90°
Choosing the correct COUNTS value depends upon the
OSR. Smaller ratios dictate larger COUNTS values, although
application requirements will vary. A COUNTS value of 32
will work for the OSR of 333. From Table 5, LKCT[1:0] = 1
for 32 counts.
Using Table 6 with the previously selected ICP of 5.6mA
gives CP[3:0] = 7. This gives enough information to program Reg0C:
Reg0C = h0D
Charge Pump Function Programming
This example uses the additional voltage clamp features
to allow the monitoring of fault conditions by setting
CPCHI = 1 and CPCLO = 1. If something occurs and the
system can no longer lock to its intended frequency, the
charge pump output will move toward either GND or
VCP+, thereby setting either the TLO or THI status flags,
respectively. Disable all the other charge pump functions
(CPMID, CPINV, CPRST, CPUP, and CPDN), allowing the
loop to lock:
Reg0D = hC0
The loop should now lock. Now un-mute the output by
setting OMUTE = 0 (assumes the MUTE pin is high).
Reg02 = h04
Reference Source Considerations
A high quality signal must be applied to the REF± inputs as
they provide the frequency reference to the entire PLL. As
mentioned previously, to achieve the part’s in-band phase
noise performance, apply a CW signal of at least 6dBm
into 50Ω, or a square wave of at least 0.5VP-P with slew
rate of at least 40V/µs.
The LTC6948 may be driven single-ended to CMOS levels
(greater than 2.7VP-P ). Apply the reference signal at REF+,
and bypass REF– to GND with a 47pF capacitor. The BST
bit must also be set to 0, according to guidelines given
in Table 2.
The LTC6948 achieves an integer mode in-band normalized
phase noise floor LNORM(INT) = –226dBc/Hz typical, and
a fractional mode phase noise floor LNORM(FRAC) = –225
dBc/Hz typical. To calculate its equivalent input phase
noise floor LIN, use the following Equation 16.
LIN = LNORM + 10 • log10 (fREF)(16)
For example, using a 10MHz reference frequency in integer
mode gives an input phase noise floor of –156dBc/Hz.
The reference frequency source’s phase noise must be at
least 3dB better than this to prevent limiting the overall
system performance.
6948f
For more information www.linear.com/LTC6948
29
LTC6948
Applications Information
In-Band Output Phase Noise
–90
PHASE NOISE (dBc/Hz)
The in-band phase noise floor LOUT produced at fRF may
be calculated by using Equation 17.
LOUT= LNORM + 10 • log10 (fPFD)(17)
+ 20 • log10 (fRF/fPFD)
or
TOTAL NOISE
fPFD = 3MHz
–100
TOTAL NOISE
fPFD = 100MHz
–110
–120
1/f NOISE
CONTRIBUTION
LOUT≈ LNORM + 10 • log10 (fPFD)
+ 20 • log10 (N/O)
–130
where LNORM is –226dBc/Hz for integer mode and
–225dBc/Hz for fractional mode. See the Typical Performance Characteristics section for graphs showing LNORM
variation versus ICP and fVCO.
As can be seen from Equation 17 for a given PFD frequency
fPFD, the output in-band phase noise increases at a 20dBper-decade rate with the N divider count. So, for a given
output frequency fRF, fPFD should be as large as possible
(or N should be as small as possible) while still satisfying
the application’s frequency step size requirements.
Output Phase Noise Due to 1/f Noise
In-band phase noise at very low offset frequencies may
be influenced by the LTC6948’s 1/f noise, depending upon
fPFD. Use the normalized in-band 1/f noise L1/f of –274dBc/
Hz with Equation 18 to approximate the output 1/f phase
noise at a given frequency offset fOFFSET.
LOUT(1/f) (fOFFSET) = L1/f + 20 • log10 (fRF)(18)
– 10 • log10 (fOFFSET)
Unlike the in-band noise floor LOUT, the 1/f noise LOUT(1/f)
does not change with fPFD, and is not constant over offset
frequency. See Figure 14 for an example of integer mode
in-band phase noise for fPFD equal to 3MHz and 100MHz.
The total phase noise will be the summation of LOUT and
LOUT(1/f).
10
1k
10k
100
OFFSET FREQUENCY (Hz)
100k
6948 F14
Figure 14. Theoretical Integer Mode In-Band
Phase Noise, fRF = 2500MHz
Integer Boundary Spurs
Integer boundary spurs are caused by intermodulation
between harmonics of the PFD frequency fPFD and the
VCO frequency fVCO. The coupling between the frequency
source harmonics can occur either on- or off-chip. The
spurs are located at offset frequencies defined by the beat
frequency between the reference harmonics and the VCO
frequency, and are attenuated by the loop filter. The spurs
only occur while in fractional mode.
Integer boundary spurs are most commonly seen when
the fractional value F approaches either zero or one such
that the VCO frequency offset from an integer frequency
is within the loop bandwidth:
fPFD • F ≤ BW
or
fPFD • (1 – F) ≤ BW
The spur will have a relatively constant power in-band,
and be attenuated by the loop out-of-band. An example
integer boundary spur measurement is shown in Figure 15.
RF Output Matching
The RF± outputs may be used in either single-ended or differential configurations. Using both RF outputs differentially
6948f
30
For more information www.linear.com/LTC6948
LTC6948
Applications Information
Single-ended impedance matching is accomplished using
the circuit of Figure 16, with component values found in
Table 18. Using smaller inductances than recommended
can cause phase noise degradation, especially at lower
center frequencies.
–40
IB SPUR LEVEL (dBc)
–50
–60
–70
–80
VRF+
fVCO ≈ 3GHz
–90 fPFD = 50MHz
LOOP BW = 130kHz
–100 CPLE = 1
N = 60
SWEPT NUM
–110
1k
10k
100k
1M
10M
FREQUENCY OFFSET FROM 3GHz (Hz)
LC
RF+(–)
CS
50Ω
6948 F15
Figure 15. Integer Boundary Spur Power vs Frequency
Offset from Boundary, LTC6948-1
will result in approximately 3dB more output power than
single-ended. Impedance matching to an external load in
both cases requires external chokes tied to VRF+. Measured
RF± S-parameters are shown below in Table 17 to aid in
the design of impedance matching networks.
VRF+
LC
RF–(+)
CS
TO 50Ω
LOAD
6948 F16
Figure 16. Single-Ended Output Matching Schematic
Table 18. Suggested Single-Ended Matching Component Values
Table 17. Single-Ended RF Output Impedance
FREQUENCY (MHz)
IMPEDANCE (Ω)
S11 (dB)
100
133.0 – j16.8
–6.7
500
110.8 – j46.1
–6.8
1000
74.9 – j57.0
–6.9
1500
49.0 – j51.3
–6.7
2000
34.4 – j41.4
–6.5
2500
27.0 – j32.1
–6.5
3000
23.2 – j24.1
–6.6
3500
21.6 – j15.9
–7.1
4000
20.9 – j7.7
–7.5
4500
20.1 – j0.2
–7.4
5000
18.1 + j7.4
–6.4
5500
16.7 + j12.5
–5.6
6000
17.1 + j16.1
–5.5
6500
20.2 + j20.1
–6.2
7000
26.9 + j24.6
–7.6
7500
38.8 + j32.3
–8.8
8000
52.9 + j43.1
–8.2
fRF (MHz)
LC (nH)
CS (pF)
350 to 1500
180
270
1000 to 5800
68
100
Return loss measured on the DC1959 using the above
component values is shown in Figure 17. A broadband
match is achieved using an {LC, CS} of either {68nH, 100pF}
or {180nH, 270pF}. However, for maximum output power
and best phase noise performance, use the recommended
component values of Table 18. LC should be a wirewound
inductor selected for maximum Q factor and SRF, such
as the Coilcraft HP series of chip inductors.
The LTC6948’s differential RF± outputs may be combined
using an external balun to drive a single-ended load. The
advantages are approximately 3dB more output power than
each output individually and better 2nd order harmonic
performance.
6948f
For more information www.linear.com/LTC6948
31
LTC6948
Applications Information
0
The listed TL baluns do not provide input-to-output DC
isolation and must be AC-coupled at the output. Figure 19
displays RF± connections using these baluns.
68nH, 100pF
180nH, 270pF
–2
–4
S11 (dB)
–6
VRF+
–8
–10
–12
RF+ 12
–14
–16
3
LTC6948
0
1
2
3
4
5
FREQUENCY (GHz)
6
7
RF–
2
1
TO 50Ω
LOAD
BALUN
11
4
5
6
6948 F18
6947 F17
Figure 17. RF Single-Ended Return Loss
For lower frequencies, transmission line (TL) baluns such as
the M/A-COM MABACT0065 and the TOKO #617DB-1673
provide good results. At higher frequencies, surface mount
(SMT) baluns such as those produced by TDK, Anaren,
and Johanson Technology, can be attractive alternatives.
See Table 19 for recommended balun part numbers versus
frequency range.
The listed SMT baluns contain internal chokes to bias RF±
and also provide input-to-output DC isolation. The pin
denoted as GND or DC FEED should be connected to the
VRF+ voltage. Figure 18 shows a surface mount balun’s
connections with a DC FEED pin.
Table 19. Suggested Baluns
BALUN PIN CONFIGURATION
1
UNBALANCED PORT
2
GND OR DC FEED
3
BALANCED PORT
4
BALANCED PORT
5
GND
6
NC
Figure 18. Example SMT Balun Connection
VRF+
TO 50Ω
LOAD
RF+ 12
LTC6948
RF–
PRI
11
SEC
6948 F19
Figure 19. Example TL Balun Connection
fRF (MHz)
PART NUMBER
MANUFACTURER
TYPE
350 to 900
#617DB-1673
TOKO
TL
400 to 600
HHM1589B1
TDK
SMT
600 to 1400
BD0810J50200
Anaren
SMT
600 to 3000
MABACT0065
M/A-COM
TL
1000 to 2000
HHM1518A3
TDK
SMT
1400 to 2000
HHM1541E1
TDK
SMT
1900 to 2300
2450BL15B100E
Johanson
SMT
2000 to 2700
HHM1526
TDK
SMT
3700 to 5100
HHM1583B1
TDK
SMT
4000 to 6000
HHM1570B1
TDK
SMT
Supply Bypassing and PCB Layout Guidelines
Care must be taken when creating a PCB layout to minimize power supply decoupling and ground inductances.
All power supply V+ pins should be bypassed directly to
the ground plane using a 0.1µF ceramic capacitor as close
to the pin as possible. Multiple vias to the ground plane
should be used for all ground connections, including to
the power supply decoupling capacitors.
6948f
32
For more information www.linear.com/LTC6948
LTC6948
Applications Information
The package’s exposed pad is a ground connection, and
must be soldered directly to the PCB land pattern. The
PCB land pattern should have multiple thermal vias to the
ground plane for both low ground inductance and also low
thermal resistance (see Figure 20 for an example). See
QFN Package Users Guide, page 8, on Linear Technology
website’s Packaging Information page for specific recommendations concerning land patterns and land via solder
masks. A link is provided below.
http://www.linear.com/designtools/packaging
Reference Signal Routing, Spurious, and
Phase Noise
The charge pump operates at the PFD’s comparison
frequency fPFD. The resultant output spurious energy is
small and is further reduced by the loop filter before it
modulates the VCO frequency.
However, improper PCB layout can degrade the LTC6948’s
inherent spurious performance. Care must be taken to
prevent the reference signal fREF from coupling onto the
VCO’s tune line, or into other loop filter signals. Example
suggestions are the following.
1)Do not share power supply decoupling capacitors
between same-voltage power supply pins.
2)Use separate ground vias for each power supply decoupling capacitor, especially those connected to VREF+,
VD+, LDO, VCP+, and VVCO+.
3)Physically separate the reference frequency signal from
the loop filter and VCO.
6948 F20
Figure 20. Example Exposed Pad Land Pattern
4)Do not place a trace between the CMA, CMB, and CMC
pads underneath the package, as worse phase noise
could result.
6948f
For more information www.linear.com/LTC6948
33
LTC6948
Typical Applications
Driving a Modulator LO for High Image Rejection and Low Noise Floor
1.8nF
10µH
66.5Ω
5V
15Ω
0.1µF 3.3V
66.5Ω
2.2nF
1µF
68nF
1µF
0.1µF
3.3V
GND
CP
VCP+
VREF+
VVCO+
VRF+
3.3V
BB
3.3V
RF+
0.1µF
LTC6948-3
RF–
SPI BUS
GND
51.1Ω
GND
REF–
1µF
REF+
STAT
CS
SCLK
SDI
SDO
LDO
VD+
MUTE
61.44MHz
0.01µF
0.01µF
1µF
CMB
CMC
GND
TB
TUNE
2.2µF
0.1µF
100pF
100pF
UNUSED OUTPUT
AVAILABLE FOR
OTHER USE
1Ω
4.7µF
LPF
AVAGO
VMMK-2503
50Ω
3.3V
1µF
0.01µF
10nH
68nH 100pF
R = 1, fPFD = 61.44MHz
N = 73.24 TO 84.64
LBW = 180kHz
O=2
5V
3.3V
68nH
0.1µF
BVCO
GND
CMA
3.3V
1nF
BASEBAND
I-CHANNEL
1.3Ω
MINI-CIRCUITS
LFCN-2600+
1nF
GND
GNDRF
BBPI
GND
LOP
VCC2
LOM
GNDRF
GND
RF
LTC5588-1
NC
1nF
6.8pF
RF OUTPUT, 1500MHz TO 2600MHz CARRIER
NC
0.2pF
GNDRF
LINOPT
NC
GND
GNDRF
GND
BBPQ
BBMQ
GND
100nF
GNDRF
fLO = 1500MHz TO 2600MHz IN 117.2Hz STEPS
PLO = 13.5dBm, ~ <–40dBc ODD HARMONIC CONTENT
~ <–30dBc EVEN HARMONIC CONTENT
GND
BBMI
50Ω 1nF
VCC1
4.7µF
EN
BASEBAND
Q-CHANNEL
6948 TA02a
Measured Image Rejection and LO
Leakage Ratio vs Output Frequency
–20
Measured Noise Floor at 70MHz
Offset vs RF Output Power
–148
BASEBAND = 100kHz AT 500mVPK
–150
–30
NOISE FLOOR (dBm/Hz)
LO LEAKAGE RATIO
(dBc)
–40
–50
–60
–70
–80
1500
–152
–154
–156
–158
–160
UNADJUSTED
IMAGE REJECTION
1700 1900 2100 2300
OUTPUT FREQUENCY (MHz)
fLO = 2100MHz
BASEBAND = 2kHz SINE
2500
–162
–20
6948 TA02b
–15
–10
–5
0
OUTPUT POWER (dBm)
5
6948 TA02c
6948f
34
For more information www.linear.com/LTC6948
LTC6948
Package Description
Please refer to http://www.linear.com/designtools/packaging/ for the most recent package drawings.
UFD Package
28-Lead Plastic QFN (4mm × 5mm)
(Reference LTC DWG # 05-08-1712 Rev B)
0.70 ±0.05
4.50 ±0.05
3.10 ±0.05
2.50 REF
2.65 ±0.05
3.65 ±0.05
PACKAGE OUTLINE
0.25 ±0.05
0.50 BSC
3.50 REF
4.10 ±0.05
5.50 ±0.05
RECOMMENDED SOLDER PAD PITCH AND DIMENSIONS
APPLY SOLDER MASK TO AREAS THAT ARE NOT SOLDERED
4.00 ±0.10
(2 SIDES)
0.75 ±0.05
R = 0.05
TYP
PIN 1 NOTCH
R = 0.20 OR 0.35
× 45° CHAMFER
2.50 REF
R = 0.115
TYP
27
28
0.40 ±0.10
PIN 1
TOP MARK
(NOTE 6)
1
2
5.00 ±0.10
(2 SIDES)
3.50 REF
3.65 ±0.10
2.65 ±0.10
(UFD28) QFN 0506 REV B
0.25 ±0.05
0.200 REF
0.50 BSC
0.00 – 0.05
BOTTOM VIEW—EXPOSED PAD
NOTE:
1. DRAWING PROPOSED TO BE MADE A JEDEC PACKAGE OUTLINE MO-220 VARIATION (WXXX-X).
2. DRAWING NOT TO SCALE
3. ALL DIMENSIONS ARE IN MILLIMETERS
4. DIMENSIONS OF EXPOSED PAD ON BOTTOM OF PACKAGE DO NOT INCLUDE
MOLD FLASH. MOLD FLASH, IF PRESENT, SHALL NOT EXCEED 0.15mm ON ANY SIDE
5. EXPOSED PAD SHALL BE SOLDER PLATED
6. SHADED AREA IS ONLY A REFERENCE FOR PIN 1 LOCATION
ON THE TOP AND BOTTOM OF PACKAGE
6948f
Information furnished by Linear Technology Corporation is believed to be accurate and reliable.
However, no responsibility is assumed for its use. Linear Technology Corporation makes no representation that the interconnection
of its circuits
as described
herein will not infringe on existing patent rights.
For more
information
www.linear.com/LTC6948
35
LTC6948
Typical Applications
Generating a 12GHz LO Signal
315.5MHz DATACONVERTER CLOCK
16.5Ω
3.3V
3.3V
0.1µF
3.3V
UNUSED OUTPUT
AVAILABLE FOR
OTHER USE
50Ω
GND
CP
VCP+
BVCO
GND
CMA
CMB
CMC
GND
TB
TUNE
3.3V
68nH
–100
–110
–120
RMS NOISE = 0.825°
RMS JITTER = 191fs
fPFD = 63.1MHz
LOOP BW = 218kHz
INTN = 0
–130
–140
2.2µF
–150
1k
10k
100k
1M
10M
OFFSET FREQUENCY (Hz)
100M
6948 TA03b
1µF
0.01µF
68nH 100pF
0.1µF
VVCO+
VRF+
0.1µF
LTC6948-4
BB
SPI BUS
RF+
51.1Ω
RF–
REF+
STAT
CS
SCLK
SDI
SDO
LDO
VD+
VREF+
1µF
GND
16.5Ω
–90
0.01µF
1µF
–80
71.5Ω
4.7nF
47nF
0.01µF
GND
REF–
16.5Ω
315.5MHz, 15dBm
2.2nF
1µF
1µF
MUTE
CRYSTEK
CCSO-914X
5V
15Ω
0.1µF 3.3V
System Phase Noise
fRF = 12.024GHz
10µH
71.5Ω
PHASE NOISE (dBc/Hz)
1.5nF
5V
1nF
1µF
100pF
MA-COM
XX1002-QH
X2
R = 5, fPFD = 63.1MHz
N = 66.6 TO 95.1
LBW = 218kHz
O=1
fLO = 8400MHz TO 12000MHz IN 481.4Hz STEPS
PLO = 14dBm TO 16dBm
6948 TA03a
Related Parts
PART NUMBER
DESCRIPTION
COMMENTS
LTC6946-x
Ultralow Noise and Spurious Integer-N
Synthesizer with Integrated VCO
370MHz to 6.4GHz, –226dBc/Hz Normalized In-Band Phase Noise Floor, –157dBc/Hz
Wideband Output Phase Noise Floor
LTC6945
Ultralow Noise and Spurious Integer-N
Synthesizer
350MHz to 6GHz, –226dBc/Hz Normalized In-Band Phase Noise Floor, –157dBc/Hz
Wideband Output Phase Noise Floor
LTC6947
Ultralow Noise and Spurious Fractional-N
Synthesizer
350MHz to 6GHz, –226dBc/Hz Normalized In-Band Phase Noise Floor, –157dBc/Hz
Wideband Output Phase Noise Floor
LTC6957
Low Phase Noise, Dual Output Buffer/Driver/
Logic Converter
Optimized Conversion of Sine Waves to Logic Levels, LVPECL/LVDS/CMOS Outputs,
DC-300MHz, 45fsRMS additive jitter (LVPECL)
LTC5540/LTC5541/ High Dynamic Range Downconverting Mixers
LTC5542/LTC5543
8dB Conversion Gain, 26.4dBm IIP3, 9.6dB NF, 600MHz to 4GHz
LTC5590/LTC5591/ High Linearity Dual Mixers
LTC5592/LTC5593
600MHz to 4.5GHz, 8.5dB Gain, 26.2dBm IIP3, 9.9dB NF
LTC5569
Broadband Dual Mixer
300MHz to 4GHz, 26.8dBm IIP3, 2dB Gain, 11.7dB NF, 600mW Power
LTC5588-1
Ultrahigh OIP3 I/Q Modulator
200MHz to 6GHz, 31dBm OIP3, –160.6dBm/Hz Noise Floor
LT®5575
Direct Conversion I/Q Demodulator
800MHz to 2.7GHz, 22.6dBm IIP3, 60dBm IIP2, 12.7dB NF
6948f
36 Linear Technology Corporation
1630 McCarthy Blvd., Milpitas, CA 95035-7417
For more information www.linear.com/LTC6948
(408) 432-1900 ● FAX: (408) 434-0507
●
www.linear.com/LTC6948
LT 0614 • PRINTED IN USA
 LINEAR TECHNOLOGY CORPORATION 2014