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