SCAS871H February   2009  – January 2016 CDCM61004

PRODUCTION DATA.  

  1. Features
  2. Applications
  3. Description
  4. Revision History
  5. Description (Continued)
  6. Pin Configuration and Functions
    1. 6.1 Pin Characteristics
  7. Specifications
    1. 7.1  Absolute Maximum Ratings
    2. 7.2  ESD Ratings
    3. 7.3  Recommended Operating Conditions
    4. 7.4  Thermal Information
    5. 7.5  Electrical Characteristics
    6. 7.6  Typical Output Phase Noise CharacteristicsCorrected units for tRJIT (RMS phase jitter); changed to fs, RMS from ps, RMS
    7. 7.7  Typical Output Jitter Characteristics
    8. 7.8  Crystal Characteristics
    9. 7.9  Dissipation Ratings
    10. 7.10 Typical Characteristics
  8. Parameter Measurement Information
  9. Detailed Description
    1. 9.1 Overview
    2. 9.2 Functional Block Diagram
    3. 9.3 Feature Description
      1. 9.3.1  Phase-Locked Loop (PLL)
      2. 9.3.2  Configuring the PLL
      3. 9.3.3  Crystal Input Interface
      4. 9.3.4  Phase Frequency Detector (PFD)
      5. 9.3.5  Charge Pump (CP)
      6. 9.3.6  On-Chip PLL Loop Filter
      7. 9.3.7  Prescaler Divider and Feedback Divider
      8. 9.3.8  On-Chip VCO
      9. 9.3.9  LVCMOS Input Interface
      10. 9.3.10 Output Divider
      11. 9.3.11 Output Buffer
    4. 9.4 Device Functional Modes
  10. 10Application and Implementation
    1. 10.1 Application Information
      1. 10.1.1 Start-Up Time Estimation
      2. 10.1.2 Output Termination
      3. 10.1.3 LVPECL Termination
      4. 10.1.4 LVDS Termination
      5. 10.1.5 LVCMOS Termination
      6. 10.1.6 Interfacing Between LVPECL and HCSL
    2. 10.2 Typical Application
      1. 10.2.1 Design Requirements
      2. 10.2.2 Detailed Design Procedure
        1. 10.2.2.1 Device Selection
          1. 10.2.2.1.1 Calculation Using LCM
        2. 10.2.2.2 Device Configuration
      3. 10.2.3 Application Curve
  11. 11Power Supply Recommendations
    1. 11.1 Power Considerations
    2. 11.2 Thermal Management
    3. 11.3 Power-Supply Filtering
  12. 12Layout
    1. 12.1 Layout Guidelines
    2. 12.2 Layout Example
  13. 13Device and Documentation Support
    1. 13.1 Device Support
      1. 13.1.1 Third-Party Products Disclaimer
    2. 13.2 Community Resources
    3. 13.3 Trademarks
    4. 13.4 Electrostatic Discharge Caution
    5. 13.5 Glossary
  14. 14Mechanical, Packaging, and Orderable Information

10 Application and Implementation

NOTE

Information in the following applications sections is not part of the TI component specification, and TI does not warrant its accuracy or completeness. TI’s customers are responsible for determining suitability of components for their purposes. Customers should validate and test their design implementation to confirm system functionality.

10.1 Application Information

10.1.1 Start-Up Time Estimation

The CDCM61004 start-up time can be estimated based on the parameters defined in Table 9 and graphically shown in Figure 16.

Table 9. Start-Up Time Dependencies

PARAMETER DEFINITION DESCRIPTION FORMULA/METHOD OF DETERMINATION
tREF Reference clock period The reciprocal of the applied reference frequency in seconds. CDCM61004 q_tref_cas869.gif
tpul Power-up time (low limit) Power-supply rise time to low limit of Power On Reset (POR) trip point Time required for power supply to ramp to 2.27 V
tpuh Power-up time (high limit) Power supply rise time to high limit of POR trip point Time required for power supply to ramp to 2.64 V
trsu Reference start-up time After POR releases, the Colpits oscillator is enabled. This start-up time is required for the oscillator to generate the requisite signal levels for the delay block to be clocked by the reference input. 500 μs best-case and 800 μs worst-case
tdelay Delay time Internal delay time generated from the reference clock. This delay provides time for the reference oscillator to stabilize. tdelay= 16384 × tref
tVCO_CAL VCO calibration time VCO Calibration Time generated from the reference clock. This process selects the operating point for the VCO based on the PLL settings. tVCO_CAL= 550 × tref
tPLL_LOCK PLL lock time Time required for PLL to lock within ±10 ppm of fREF Based on the 400-kHz loop bandwidth, the PLL settles in 5τ or 12.5 μs.
CDCM61004 ai_tim_startup_cas869.gif Figure 16. Start-up Time Dependencies

The CDCM61004 start-up time limits, tMAX and tMIN, can be calculated as follows in Equation 3 and Equation 4:

Equation 3. tMAX = tpuh + trsu + tdelay + tVCO_CAL + tPLL_LOCK
Equation 4. tMIN = tpul + trsu + tdelay + tVCO_CAL + tPLL_LOCK

10.1.2 Output Termination

The CDCM61004 is a 3.3-V clock driver with the following output options: LVPECL, LVDS, or LVCMOS.

10.1.3 LVPECL Termination

The CDCM61004 is an open emitter for LVPECL outputs. Therefore, proper biasing and termination are required to ensure correct operation of the device and to minimize signal integrity. The proper termination for LVPECL is 50 Ω to (VCC–2) V, but this DC voltage is not readily available on most PCBs. Thus, a Thevenin equivalent circuit is worked out for the LVPECL termination in both direct-coupled (DC) and ac-coupled (AC) cases, as shown in Figure 17 and Figure 18. TI recommends placing all resistive components close to either the driver end or the receiver end. If the supply voltage of the driver and receiver are different, ac-coupling is required.

CDCM61004 ai_lvpecl_dc_output_cas871.gif Figure 17. LVPECL Output DC Termination
CDCM61004 ai_lvpecl_ac_output_cas871.gif Figure 18. LVPECL Output AC Termination

10.1.4 LVDS Termination

The proper LVDS termination for signal integrity over two 50-Ω lines is 100 Ω between the outputs on the receiver end. Either direct-coupled termination or ac-coupled termination can be used for LVDS outputs, as shown in Figure 19 and Figure 20. TI recommends placing all resistive components close to either the driver end or the receiver end. If the supply voltage of the driver and the receiver are different, ac-coupling is required.

CDCM61004 ai_lvds_dc_output_cas871.gif Figure 19. LVDS Output DC Termination
CDCM61004 ai_lvds_ac_output_cas871.gif Figure 20. LVDS Output AC Termination

10.1.5 LVCMOS Termination

Series termination is a common technique used to maintain the signal integrity for LVCMOS drivers, if connected to a receiver with a high-impedance input with a pullup or a pulldown resistor. For series termination, a series resistor (RS) is placed close to the driver, as shown in Figure 21. The sum of the driver impedance and RS should be close to the transmission line impedance, which is usually 50 Ω. Because the LVCMOS driver in the CDCM61004 has an impedance of 30 Ω, RS is recommended to be 22 Ω to maintain proper signal integrity.

CDCM61004 ai_lvcmos_output_cas871.gif Figure 21. LVCMOS Output Termination

10.1.6 Interfacing Between LVPECL and HCSL

Because the LVPECL common-mode voltage is different from the HCSL common-mode voltage, ac-coupled termination is used. The 150-Ω resistor ensures proper biasing of the CDCM61004 LVPECL output stage, while the 471-Ω and 56-Ω resistor network biases the HCSL receiver input stage, as shown in Figure 22.

CDCM61004 ai_lvpecl_hcsl_if_cas871.gif Figure 22. LVPECL to HCSL Interface

10.2 Typical Application

CDCM61004 ethernet_switch_scas871.gif Figure 23. Ethernet Switch

10.2.1 Design Requirements

Consider a typical wired communications application, like a top-of-rack switch, which needs to clock 1-Gbps or 3.125-Gbps Ethernet PHYs. For such asynchronous systems, the reference input can be a crystal. In such systems, the clocks are expected to be available upon power up without the need for any device-level programming. An example of clock input and output requirements is shown below:

  • Clock Input:
    • 25-MHz crystal
  • Clock Outputs:
    • 2× 156.25 MHz clock for uplink 3.125 Gbps, LVPECL
    • 2× 156.25 MHz clock for downlink 3.125 Gbps, LVPECL

The section below describes the detailed design procedure to generate the required output frequencies for the above scenario using CDCM61004.

10.2.2 Detailed Design Procedure

Design of all aspects of the CDCM61004 is quite involved and software support is available to assist in part selection and phase noise simulation. This design procedure will give a quick outline of the process.

  1. Device Selection
    • The first step is to calculate the VCO frequency given the required output frequency. The device must be able to produce the VCO frequency that can be divided down to the required output frequency.
    • The WEBENCH Clock Architect Tool from TI will aid in the selection of the right device that meets the customer's output frequencies and format requirements.
  2. Device Configuration
    • The WEBENCH Clock Architect Tool attempts to maximize the phase detector frequency, use smallest dividers, and maximizes PLL bandwidth.

10.2.2.1 Device Selection

Use the WEBENCH Clock Architect Tool. Enter the required frequencies and formats into the tool. To use this device, find a solution using the CDCM61004.

10.2.2.1.1 Calculation Using LCM

In this example, the valid VCO frequency for CDCM61004 is 1.875 GHz.

10.2.2.2 Device Configuration

For this example, when using the WEBENCH Clock Architect Tool, the reference would have been manually entered as 25 MHz according to input frequency requirements. Enter the desired output frequencies and click on Generate Solutions. Select CDCM61004 from the solution list.

From the simulation page of the WEBENCH Clock Architect Tool, it can be seen that to maximize phase detector frequencies, the N divider is set to 25 and prescaler divider is set to 3. This results in a VCO frequency of
1.875
GHz. The output divider is set to 4. At this point the design meets all input and output frequency requirements and it is possible to design a loop filter for system and simulate performance on the clock outputs. Figure 24 shows the typical phase noise plot of the 156.25 MHz LVPECL output.

10.2.3 Application Curve

CDCM61004 lan3_out0_caa100.gif Figure 24. Typical Phase Noise Plot of 156.25 MHz LVPECL Output