LMX2541 Family Ultra Low Noise Frequency Synthesizer with Integrated VCO Evaluation Board Operating Instructions National Semiconductor Corporation High Speed Signal Path Division Precision Timing Devices 12-4-2009 TABLE OF CONTENTS LOOP FILTER VALUES ...................................................................................6 CODELOADER SETUP....................................................................................7 SPURS .......................................................................................................10 LMX2541SQ2060E DEFAULT SETUP AND MEASURED PERFORMANCE ..........13 LMX2541SQ2380E DEFAULT SETUP AND MEASURED PERFORMANCE ..........16 LMX2541SQ2690E DEFAULT SETUP AND MEASURED PERFORMANCE ..........19 LMX2541SQ3030E DEFAULT SETUP AND MEASURED PERFORMANCE ..........22 LMX2541SQ3030E VCO PHASE NOISE.....................................................24 LMX2541SQ3320E DEFAULT SETUP AND MEASURED PERFORMANCE ..........25 LMX2541S3320E SPURS FOR WORST CASE CHANNELS ..........................26 LMX2541SQ3320E VCO PHASE NOISE.....................................................27 LMX2541SQ3740E DEFAULT SETUP AND MEASURED PERFORMANCE ..........28 LMX2541S3740E SPURS FOR WORST CASE CHANNELS ..........................29 LMX2541SQ3740E VCO PHASE NOISE.....................................................30 BILL OF MATERIALS.....................................................................................31 NO COMPONENTS SETUP.SCHEMATIC ARE ASSEMBLED ON THE BOTTOM LAYER IN THE DEFAULT ......................................................................................33 SCHEMATIC ................................................................................................34 BOARD LAYER STACKUP .............................................................................35 QUICK SETUP Do Not Connect 3.4 V LMX2541 Phase Noise Analyzer RFout uWire Laptop or PC RFout Connect to a spectrum analyzer or phase noise analyzer. The Agilent E5052A was used for these instructions. Vcc Connect to a 3.4 volt low noise power supply uWire Connect to a computer with CodeLoader software ExtVCOin This is not used in the default setup, but is included to support the use of an external VCO. In Full Chip Mode, this device has an on-chip VCO. Ftest/LD The LED is to ensure that the part is locked. This output can be very useful for diagnostic purposes OSCin/OSCin* This is not used in the default setup, but is included to support the use of an external OSCin signal. The board includes a 100 MHz TCXO, which has two varieties. The OSCin signal is absolutely critical for the phase noise and spur performance of the LMX2541. Manufacturer Model Frequency Connor Winfield CWX-813 100 MHz Crystek CVPD920 100 MHz LMK01000 National LMK02000 Any Semiconductor LMK03000 LMK04000 * * * * * * Comments Good phase noise Low Cost Drifts a lot. Good Phase Noise Drifts much less High Cost * * * Eliminate drift issues Potentially the best phase noise Best for fractional spurs Be very aware of the TCXO drift and the contribution that it can have to phase noise. Termination of the TCXO has a large impact on fractional spurs as well. The best results can be achieved by driving this board with an LMK01000 LVPECL output. Doing so results in about a 4 dB spur improvement. Loop Filter Values Parameter VCO Frequency (MHz) VCO Gain (MHz/V) Charge Pump Gain (mA) Phase Detector Frequency (MHz) OSCin Frequency (MHz) Loop Bandwidth (kHz) Phase Margin (deg) C1_LF (nF) C2_LF (nF) R2_LF (Kohm) C3_LF (Internal) (nF) C4_LF (Internal) (nF) R3_LF (Internal) (Kohm) R4_LF (Internal) (Kohm) LMX2541 SQ2060E 1990 - 2240 LMX2541 LMX2541 LMX2541 LMX2541 SQ2380E SQ2690E SQ3030E SQ3320E 2200 2490 - 2810 3130 2530 2865 3230 3600 LMX2541 SQ3470E 3480 4000 13-23 16-30 17-32 20-37 21-37 27 - 42 3.2 3.2 3.2 3.2 3.2 3.2 25 25 25 25 25 25 100 100 100 100 100 100 37.3 - 54.6 52.7 - 52.8 2.2 22 0.47 40.8 - 61.7 53.152 2.2 22 0.47 38.7 - 58.6 52.9 - 52.4 2.2 22 0.47 40.0 - 59.9 53.0 - 52.2 2.2 22 0.47 38.1 - 54.7 52.8 - 52.8 2.2 22 0.47 43.1 - 55.7 53.2 - 52.7 2.2 22 0.47 0.02 0.02 0.02 0.02 0.02 0.02 0.1 0.1 0.1 0.1 0.1 1 1 1 1 1 1 0.2 0.2 0.2 0.2 0.2 0.2 0.1 * Note that the VCO gain does change a fair amount. Although not demonstrated in these instructions, the charge pump gain could be adjusted to account for this variation. CodeLoader Setup Select the part. In this case, it is the LMX2541SQ3320E. Choose the correct startup mode. This is determined by the part option. Load the part. You can load it from the menu or also press Cntrl + L. On the Port Setup tab, the user may select the type of communication port (USB or Parallel) that will be used to program the device on the evaluation board. If parallel port is selected, the user should ensure that the correct port address is entered. The Pin Configuration field is hardware dependent and normally SHOULD NOT be changed by the user. The evaluation board is typically shipped with a parallel port cable that is used to interconnect the board to a PC LPT port, enabling the board to be programmed. Separately available is a USB2UWIRE-IFACE board which simplifies evaluation by enabling the user to establish a USB connection from the Codeloader 4 software to the evaluation board. http://www.national.com/store/view_item/index.html?nsid=USB2UWIRE-IFACE To view the function of any bit on the CodeLoader configuration tabs, place the cursor over the desired bit register label and click the right mouse button on it for a description. This Bits/Pins configuration is common to all options of the LMX2541 evaluation board. -70 1975 2000 2025 2050 2075 2100 2125 2150 2175 2200 2225 2250 2275 2300 2325 2350 2375 2400 2425 2450 2475 2500 2525 2550 2575 2600 2625 2650 2675 2700 2725 2750 2775 2800 2825 2850 2875 2900 2925 2950 2975 3000 3025 3050 3075 3100 3125 3150 3175 3200 3225 3250 3275 3300 3325 3350 3375 3400 3425 3450 3475 3500 3525 3550 3575 3600 3625 3650 3675 3700 3725 3750 3775 3800 3825 3850 3875 3900 3925 3950 3975 4000 1975 2000 2025 2050 2075 2100 2125 2150 2175 2200 2225 2250 2275 2300 2325 2350 2375 2400 2425 2450 2475 2500 2525 2550 2575 2600 2625 2650 2675 2700 2725 2750 2775 2800 2825 2850 2875 2900 2925 2950 2975 3000 3025 3050 3075 3100 3125 3150 3175 3200 3225 3250 3275 3300 3325 3350 3375 3400 3425 3450 3475 3500 3525 3550 3575 3600 3625 3650 3675 3700 3725 3750 3775 3800 3825 3850 3875 3900 3925 3950 3975 4000 Spurs Oscillator Spurs Oscillator spurs occur at the oscillator frequency (100 MHz) offset from the carrier. They can be largely impacted by the board layout. These were taken in 25 MHz increments. g p -70 -75 -80 -85 -90 Option -95 -100 2060E 2380E 2690E 3030E 3320E 3740E -105 -110 -115 -120 Frequency Reference (Phase Detector) Spurs Reference spurs occur at a multiple of the phase detector frequency (25 MHz) from the carrier g p -75 -80 -85 -90 Option -95 -100 2060E 2380E 2690E 3030E 3320E 3740E -105 -110 -115 -120 Frequency Fractional Spurs In-band spurs occur inside the loop bandwidth. These spurs were measured with a WORST CASE fraction of 1/5000. The primary fractional spurs are at 5 kHz and the sub-fractional spurs are at 2.5 kHz. The actual frequency is the frequency g p in the chart plus 5 kHz. -30 -35 Option Spur Order -40 2060E - Primary 2060E - SubFractional 2380E - Primary 2380E - SubFractional 2690E - Primary 2690E - SubFractional 3030E - Primary 3030E - SubFractional 3320E - Primary 3320E - SubFractional 3740E - Primary 3740E - SubFractional -45 -50 -55 -60 -65 1975 2000 2025 2050 2075 2100 2125 2150 2175 2200 2225 2250 2275 2300 2325 2350 2375 2400 2425 2450 2475 2500 2525 2550 2575 2600 2625 2650 2675 2700 2725 2750 2775 2800 2825 2850 2875 2900 2925 2950 2975 3000 3025 3050 3075 3100 3125 3150 3175 3200 3225 3250 3275 3300 3325 3350 3375 3400 3425 3450 3475 3500 3525 3550 3575 3600 3625 3650 3675 3700 3725 3750 3775 3800 3825 3850 3875 3900 3925 3950 3975 4000 -70 Frequency Out-band spurs occur inside the loop bandwidth. These spurs were measured with a WORST CASE fraction of 1/100. The primary fractional spurs are at 250 kHz and the sub-fractional spurs are at 125 kHz. The actual frequency isg thep frequency in the chart plus 250 kHz. -30 -35 Option Spur Order -40 2060E - Primary 2060E - SubFractional 2380E - Primary 2380E - SubFractional 2690E - Primary 2690E - SubFractional 3030E - Primary 3030E - SubFractional 3320E - Primary 3320E - SubFractional 3740E - Primary 3740E - SubFractional -45 -50 -55 -60 -65 1975 2000 2025 2050 2075 2100 2125 2150 2175 2200 2225 2250 2275 2300 2325 2350 2375 2400 2425 2450 2475 2500 2525 2550 2575 2600 2625 2650 2675 2700 2725 2750 2775 2800 2825 2850 2875 2900 2925 2950 2975 3000 3025 3050 3075 3100 3125 3150 3175 3200 3225 3250 3275 3300 3325 3350 3375 3400 3425 3450 3475 3500 3525 3550 3575 3600 3625 3650 3675 3700 3725 3750 3775 3800 3825 3850 3875 3900 3925 3950 3975 4000 -70 Frequency Minimizing Fractional Spurs * * * * Both fractional and sub-fractional spurs are highly sensitive to the OSCin signal. Higher slew rates are desired. Also, the termination makes a big difference. For this evaluation board, a series 120 ohm resistor had a large impact. The best results have been achieved when driving the part differentially with an LVPECL output of the LMK01000/2000/3000/400 series of clock conditioner devices from National Semiconductor. The spurs on this evaluation board are relatively high because the loop bandwidth is very wide. This wide loop bandwidth takes advantage of the close-in phase noise, but it does expose the fractional spurs more. The fractional spurs can be reduced by orders of magnitude by reducing the loop bandwidth. This requires a re-design of the loop filter. To eliminate the SubFractional spurs entirely, choose a fractional denominator with no factors of 2 or 3. For this 100 MHz TCXO and 250 kHz channel spacing, a phase detector frequency of 6.25 MHz and a fractional denominator of 25 would work. However, the higher N value does degrade the phase noise. An ideal scenario would be to use a TCXO frequency of something like 125 MHz. Then the sub-fractional spurs are eliminated if the phase detector is chosen to be 30.25 MHz and the fractional denominator is chosen to be 125. In the plot below, one was taken with the default 100 MHz TCXO and another was taken with a 125 MHz signal. The phase detector frequency was changed to 31.25 MHz, but the charge pump gain was reduced to 26X to compensate for this. This is the same part, board, and frequency (3030.25 MHz). Although the fraction is different, notice that the fractional denominator of 125 has no sub fractional spurs at 125 kHz, 375 kHz, and so on. An LMK01000 evaluation board driven by a 1250 MHz signal was used to produce this 125 MHz signal. Phase Noise (dBc/Hz) -60 -80 -100 -120 -140 -160 1.E+04 1.E+05 1.E+06 1.E+07 Offset (Hz) FDEN = 125 FDEN = 100 1.E+08 LMX2541SQ2060E Default Setup and Measured Performance LMX2541SQ2060E Spurs for WORST CASE Channels (Fraction = 1/100) 0 -10 1975.25 MHz -20 2000.25 MHz 2025.25 MHz Power (dBm) -30 2050.25 MHz 2075.25 MHz -40 2100.25 MHz 2125.25 MHz -50 2150.25 MHz 2175.25 MHz -60 2200.25 MHz 2225.25 MHz -70 2250.25 MHz -80 Offset from Carrier (kHz) This graph has been normalized to the average carrier power. 300 250 200 150 100 50 0 -50 -100 -150 -200 -250 -300 -90 LMX2541SQ2060E VCO Phase Noise The plots show the VCO phase noise at low, middle, and high frequency. To measure the VCO phase noise, a simple technique is to lock the PLL to the desired frequency and set the charge pump state to "Tri-State", by clicking this on the PLL tab on Codeloader. If the phase noise analyzer can track the signal, a reasonable measurement can be made. To ensure that this measurement is of the VCO noise, omit the the spurs and disconnect programming cable. the microwire LMX2541SQ2380E Default Setup and Measured Performance LMX2541SQ2380E Spurs for WORST CASE Channels (Fraction = 1/100) 0 2200.25 MHz -10 2225.25 MHz 2250.25 MHz -20 2275.25 MHz 2300.25 MHz -30 Power (dBm) 2325.25 MHz 2350.25 MHz -40 2375.25 MHz 2400.25 MHz -50 2425.25 MHz 2450.25 MHz -60 2475.25 MHz 2500.25 MHz -70 2525.25 MHz 2550.25 MHz -80 300 250 200 150 100 50 0 -50 -100 -150 -200 -250 -300 -90 Offset from Carrier (kHz) This graph has been normalized to the average carrier power. LMX2541SQ2380E VCO Phase Noise The plots show the VCO phase noise at low, middle, and high frequency. To measure the VCO phase noise, a simple technique is to lock the PLL to the desired frequency and set the charge pump state to "Tri-State", by clicking this on the PLL tab on Codeloader. If the phase noise analyzer can track the signal, a reasonable measurement can be made. To ensure that this measurement is of the VCO noise, omit the the spurs and disconnect programming cable. the microwire LMX2541SQ2690E Default Setup and Measured Performance LMX2541SQ2690E Spurs for WORST CASE Channels (Fraction = 1/100) 0 2475.25 MHz 2500.25 MHz -10 2525.25 MHz 2550.25 MHz -20 2575.25 MHz 2600.25 MHz -30 Power (dBm) 2625.25 MHz 2650.25 MHz -40 2675.25 MHz 2700.25 MHz -50 2725.25 MHz 2750.25 MHz -60 2775.25 MHz 2800.25 MHz -70 2825.25 MHz 2850.25 MHz -80 2875.25 MHz Offset (kHz) This graph has been normalized to the average carrier power. 300 250 200 150 100 50 0 -50 -100 -150 -200 -250 -300 -90 LMX2541SQ2690E VCO Phase Noise The plots show the VCO phase noise at low, middle, and high frequency. To measure the VCO phase noise, a simple technique is to lock the PLL to the desired frequency and set the charge pump state to "Tri-State", by clicking this on the PLL tab on Codeloader. If the phase noise analyzer can track the signal, a reasonable measurement can be made. To ensure that this measurement is of the VCO noise, omit the the spurs and disconnect programming cable. the microwire LMX2541SQ3030E Default Setup and Measured Performance LMX2541S3030E Spurs for WORST CASE Channels (Fraction = 1/100) 0 2800.25 MHz 2825.25 MHz -10 2850.25 MHz 2875.25 MHz -20 2900.25 MHz 2925.25 MHz -30 Power (dBm) 2950.25 MHz 2975.25 MHz -40 3000.25 MHz -50 3025.25 MHz 3050.25 MHz -60 3075.25 MHz 3100.25 MHz -70 3125.25 MHz 3150.25 MHz -80 3175.25 MHz 3200.25 MHz Offset (kHz) This graph has been normalized the average carrier power. 300 250 200 150 100 50 3225.25 MHz 0 -50 -100 -150 -200 -250 -300 -90 3250.25 MHz LMX2541SQ3030E VCO Phase Noise The plots show the VCO phase noise at low, middle, and high frequency. To measure the VCO phase noise, a simple technique is to lock the PLL to the desired frequency and set the charge pump state to "Tri-State", by clicking this on the PLL tab on Codeloader. If the phase noise analyzer can track the signal, a reasonable measurement can be made. To ensure that this measurement is of the VCO noise, omit the the spurs and disconnect programming cable. the microwire LMX2541SQ3320E Default Setup and Measured Performance LMX2541S3320E Spurs for WORST CASE Channels (Fraction = 1/100) 0 3125.25 MHz 3150.25 MHz -10 3175.25 MHz 3200.25 MHz -20 3225.25 MHz 3250.25 MHz Power (dBm) -30 3275.25 MHz 3300.25 MHz -40 3325.25 MHz 3350.25 MHz -50 3375.25 MHz 3400.25 MHz -60 3425.25 MHz 3450.25 MHz -70 3475.25 MHz 3500.25 MHz -80 3525.25 MHz 3550.25 MHz 300 250 200 150 100 50 0 -50 -100 -150 -200 -250 -300 -90 Offset (kHz) Note that the graphs have all been normalized the average carrier power. 3575.25 MHz 3600.25 MHz LMX2541SQ3320E VCO Phase Noise The plots show the VCO phase noise at low, middle, and high frequency. To measure the VCO phase noise, a simple technique is to lock the PLL to the desired frequency and set the charge pump state to "Tri-State", by clicking this on the PLL tab on Codeloader. If the phase noise analyzer can track the signal, a reasonable measurement can be made. To ensure that this measurement is of the VCO noise, omit the the spurs and disconnect programming cable. the microwire LMX2541SQ3740E Default Setup and Measured Performance LMX2541S3740E Spurs for WORST CASE Channels (Fraction = 1/100) 3475.25 MHz 0 3500.25 MHz 3525.25 MHz -10 3550.25 MHz 3575.25 MHz -20 3600.25 MHz 3625.25 MHz Power (dBm) -30 3650.25 MHz 3675.25 MHz -40 3700.25 MHz 3725.25 MHz -50 3750.25 MHz 3775.25 MHz -60 3800.25 MHz 3825.25 MHz -70 3850.25 MHz 3875.25 MHz -80 3900.25 MHz 3925.25 MHz 300 250 200 150 100 50 0 -50 -100 -150 -200 -250 -300 -90 Offset (kHz) Note that the graphs have all been normalized the average carrier power. 3950.25 MHz 3975.25 MHz 4000.25 MHz LMX2541SQ3740E VCO Phase Noise The plots show the VCO phase noise at low, middle, and high frequency. To measure the VCO phase noise, a simple technique is to lock the PLL to the desired frequency and set the charge pump state to "Tri-State", by clicking this on the PLL tab on Codeloader. If the phase noise analyzer can track the signal, a reasonable measurement can be made. To ensure that this measurement is of the VCO noise, omit the the spurs and disconnect programming cable. the microwire Bill of Materials Revision 9/30/2009 Part Manufacturer Part Number Qty Identifier 100 pF Kemet C0603C101J5GAC 4 C1, C5, C33, C35 2.2 nF Kemet C0603C222J5GAC 1 C3_LF 22 nF Kemet C0603C223K5RAC 1 C2_LF 0.1 uF Kemet C0603C104K5RAC 19 C2, C6, C9, C10, C11, C12, C13, C17, C18, C21, C24, C25, C26, C27, C28, C32, C36, C40, C49 1 uF Kemet C0603C105K8VAC 10 C3, C7, C16, C19, C31, C37, C41, C42, C43, C50 Capacitors 4.7 uF Kemet C0603C475K9PAC 4 C14, C15, C20, C34 10 uF Kemet C0805C106K9PAC 6 C4, C8, C30, C38, C39, C48 Resistors 0 ohm Vishay/Dale CRCW06030000Z0EA 7 4.7 ohm Vishay/Dale CRCW06034R7JNEA 2 R3_LF, R12, R21, R39, R41, R42, R46 R49, R50 10 ohm Vishay/Dale CRCW060310R0JNEA 3 R6, R47, R48 18 ohm Vishay/Dale CRCW060318R0JNEA 1 R2 51 ohm Vishay/Dale CRCW060351R0JNEA 2 R10, R14 120 ohm Vishay/Dale CRCW0603120RJNEA 3 R7, R8, R51 180 ohm Vishay/Dale CRCW0603180RJNEA 1 R36 330 ohm Vishay/Dale CRCW0603330RJNEA 2 R1, R3 470 ohm Vishay/Dale CRCW0603470RJNEA 1 R2_LF 2.2 k Vishay/Dale CRCW06032K20JNEA 1 R35 15 k Vishay/Dale CRCW060315K0JNEA 3 R30, R32, R34 27 k Vishay/Dale CRCW060327K0JNEA 5 R29, R31, R33, R37, R38 100 k Vishay/Dale CRCW0603100KJNEA 2 R4, R5 Ferrite Digikey 490-1015-1-ND 7 L1, L2, L3, L4, L5, L6, L7 3.3 V zener Comchip CZRU52C3V3 1 D2 HEADER_2X5 (POLARIZED) FCI Electronics 52601-S10-8 1 uWire Green LED Lumex SML-LX2832GC-TR 1 D1 POWER_SMALL Weidmuller Johnson Components 1594540000.0 1 P1 142-0701-851 4 OSCin, OSCin*, RFout, Vcc Connor-Winfield CWX813 - 100 MHz 1 Y1 Fairchild BSS138 1 U5 CRYSTEK CVPD-920 - 100 MHz 1 U4 LMX2541SQxxxx 1 U1 4 Place in Holes at the corners of the board. Other SMA TCXO - 100 MHz NFET VCXO - 100 MHz LMX2541 Standoffs National Semiconductor SPC Technology SPCS-6 Open Open - Open Open Capacitor 14 C1_LF, C2pLF, C4_LF, C22, C23, C29, C44, C45, C46, C47 Open Capacitor 6 Open - Open Resistor 22 Open - Open Resistor 9 Open - Open Crystal 1 C100, C101, C102, C103, C104, C105 R2pLF,R4_LF, R9, R11, R13, R15, R16, R17, R18, R19, R20, R22, R23, R24, R25, R26, R27, R28, R40, R43, R44, R45 R100, R101, R102, R103, R104, R105, R106, R107, R108, R109 Y100 Open - Open IC 3 U2, U3, U100 Open - Open SMA 3 ExtVCOin, Ftest/LD, VccAux Top Assembly Diagram 234 32 Bottom Assembly Diagram No Components are assembled on the bottom layer in the default setup. 33 Schematic VccAux OSCin SMA 100 k OSCin/OSCin* Options 120 ohm R16 Open 3 R15 R12 R14 0.1 uF 0 ohm 51 ohm Vcc Vcont Loop Filter Options T CXO 0.1 uF Open C101 Open VccPlane - DATA R101 Open Monitor theFtest/LD pin Provisions for Analog LockDetect Supplypower for an op-ampfor an active filter Allowtheuser to applyFSKmodulation Monitor CPout Pin Monitor or Drive theVtunepin R21 Ftest/LD Open VccAux 0 ohm R20 Open C29 Open R100 Open VccPlane VccPlane2 1 Ferrite C28 C104 Open D1 R2_LF 470 ohm C27 VccCP2 0.1 uF VccBias FLout Bypass CPout VccDiv Vtune R2pLF 18 DATA CLK 32 33 34 DATA VccCP1 CLK VccPLL1 LE ExtVCOin R104 Open R3_LF NC CE C3 0.1 uF 1 uF C33 C4 10 uF 100 pF C32 C31 0.1 uF 1 uF Open R107 Vtune C3_LF VccPlane2 Ferrite VccPlane C5 100 pF Open C6 0.1 uF C7 1 uF C35 100 pF C8 10 uF C26 0.1 uF 5 6 7 8 GND Mod GND GND C25 0.1 uF U2 Open GND GND Vcc GND C47 16 15 14 13 5 VccAux R47 10 ohm C24 0.1 uF C46 6 open R43 R44 Open Open 7 Open 8 VccPlaneR49 4.7 ohm C38 10 uF OUTPUT C44 4 INPUT ADJ GND NC NC SHUTDOWN 3 Open 2 C45 1 Bypass U3 Open C23 Open 10 R27 Open R45 C22 R24 Open R23 Open R26 R28 Open Open Open ExtVCOin Open R25 Open C21 1 2 VccPlane R48 10 ohm 1 uF Open CE Open GND C37 C36 0.1 uF 1 2 VrefVCO 11 10 uF VccPlane R4_LF Open C4_LF Open 9 VregVCO VccVCO 8 7 L2 6 Lmid 5 4 VccRFout 3 L1 GND VregRFout GND RFout 0 0.1 uF R3 330 ohm 36 2 18 ohm R1 330 ohm C9 DAP R2 RFout SMA 1 35 C30 2.2 nF L6 14 Ferrite VccPlane C2 VccPlane Open 16 L7 VccPlane C1 100 pF VccPlane C102 15 12 V- Vout R106 Open R105 Open 13 C103 Open C105 Open 9 10 11 12 LE 31 1 uF R102 0 ohm VccPlane 2 1 C1_LF Open C2_LF 22 nF Open RFoutEN C43 R38 27 k R109 Open U100 Open CPout_Net Open C2pLF 17 R37 27 k EN_RFOUT 4 3 2 1 19 VccPLL2 21 20 OSCin Ftest/LD 22 OSCin* 23 VccOSCin 25 26 24 RFoutEN L5 Ferrite GND GND Vtune GND 30 Green LED GND Fout GND GND C18 0.1 uF VccDig VccFRAC 27 GND 28 VccPlane 29 VregFRAC C17 0.1 uF U1 LMX2541 Readback CPout_Net Open 0.1 uF L1 Ferrite uW ire HEADER_2X5(POLARIZED) R108 R51 120 ohm V+ EN_RFOUT VccAux U5 NFET 5 VccPlane2 L4 Gate Source 3 L3 Ferrite R103 Open 2 L2 Ferrite 4.7 ohm C16 1 uF Drain Readback R50 VccPlane 1. Internal VCO Donot useanybottomcomponents for this. 2. External VCOwithPassive Filter Disconnecting Vtune isolates Vtune. Might not benecessary 3. External VCOwithActiveFilter Usethe components on theback. OpAMPcanbe suppliedwithFtest/LD pin or with VccAuxvoltage. UseC1_LF and backside resistors for slowslewratedesign. 3rd Poleis formedwithC3_LF and resistor backto top of board. 4thPole is alreadyon topof board. 4. Monitor CPout through Ftest/LD Disconnect Vtune pin for moreaccurate results. 5. Monitor or Drive Vtune Usetest point or Ftest/LD SMAand backlayer resistors. Remove R3_LF or Tri-state charge pumptodrivevoltage. Ftest/LD Pin Options R22 Open VccPlane2 CE 1 2 4 6 8 10 4 C34 4.7 uF 1 3 5 7 9 R19 Open R18 C100 Open LE 2 IN+ Open GND C12 Y100 Open R17 Out 4 C11 0.1 uF CE D2 R36 C42 180 ohm 1 uF R34 15 k Open R13 Open R11 Open R33 27 k VccAux IN- R10 51 ohm R32 15 k Y1 C13 Open R31 27 k R30 15 k R35 2.2 k LE 0 C10 0.1 uF R9 3 120 ohm R29 27 k DATA 3.3 V zener R8 CLK GND Vs C14 4.7 uF R7 VccAux 1. TCXOMode TerminateOSCin* if this is done 2. VCXOMode Connect VCXOand tiepower dupplyto Vcc/2. Driveandterminatedifferentially. 3. Differential Mode Use100ohmor 2x51 ohmresistors for termination. 4. External Crystal Mode Use backside of board. Mayhave tobend leads of crystal to fit. 5. Use external signal source likeLMK04000 evaluationboard todrive OSCin/OSCin* 6. Special provisions on GND plane GND planeisolated belowOSCin and connectedbya single trace on the bottomlayer. Cut thistracetoexperiment further withOSCinisolation R6 10 ohm DATA R5 100 k 6 5 4 RF RF* U4 VCXO - 100 MHz C15 4.7 uF Vtune NC GND 3 2 1 R4 CLK OSCin* SMA 0.1 uF P1 POW ER_SMALL VccPlane VccPlane C19 1 uF C20 4.7 uF FRAME VregRFout Series resistance on VregRFout improves the noisefloor whenusedin dividedmode. R40 Pad and External VCO Options Tousewitha VCO, use 3components to make a T-splitter and don't usethePi-Pad. For sensitivitytesting, put components in the Pi-Pad andzero out resistors inthe T-Pad. Also, for fractional spurs outsidethe loop bandwidth with the external VCOoption, thereis some evidence that putting a resistive pad canimprove fractional spurs outside the loopbandwidth. SMA_FRAME R42 Open 0 ohm Vcc SMA VccPlane2 R39 C50 1 uF C49 0.1 uF 0 ohm C48 10 uF R41 VccAuxTP 0 ohm VccAux R46 C41 1 uF Additional Comments - Anypart withdesignator 100 or higher is on the bottomlayer andnot assembled bydefault. 34 C40 0.1 uF 0 ohm C39 10 uF VccAux Open Board Layer Stackup Board Material Number of Layers Board Thickness Copper Weight Finish Solder Mask Color Top Layer 1oz thick RO4003 (r = 3.38, Tand = 0.0022) CONTROLLED THICKNESS of 16 mils thick GND1 Layer FR4 (r = ~4.6) ?? mils thick, but thinner is preferable VccPlane Layer FR4 ?? mils thick Mid Layer 1 FR4 (r = ~4.6) ?? mils thick, but thinner is preferable GND2 Layer FR4 (r = ~4.6) = ~4.6) ?? mils thick Bottom Layer 35 62 mils thick total Testing Rogers RO4003 6 0.062" 1 oz Finished Immersion Gold Green/Gloss 100% Electrical Testing Top Layer and Silkscreen 36 GND1 Layer Beneath the TCXO, the ground plane is separated. It is believed that this may improve spurs, especially at offset frequencies equal to the TCXO frequency. These planes are connected on the bottom layer by a small trace. 37 VccPlane Layer Beneath the TCXO, the power plane is removed to minimize the chance of any noise getting onto this plane. 38 MidLayer 1 Certain pins like VccFRAC and the TCXO supply pins could potentially be sources of noise. These traces were put on this separate layer to try to isolate them more from the 39 GND2 Layer Beneath the TCXO, the ground plane is separated. It is believed that this may improve spurs, especially at offset frequencies equal to the TCXO frequency. These planes are connected on the bottom layer by a small trace. 40 Bottom Layer and Silkscreen This layer has the small trace that connects the grounded pieces in the GND1 layer. It has options for an active filter and crystal as well. 41