SBAA661 February   2025 LMX1205

 

  1.   1
  2.   Abstract
  3.   Trademarks
  4. 1Flicker Noise, Noise Floor, and Total Noise
    1. 1.1 Flicker Noise
    2. 1.2 Noise Floor
    3. 1.3 Total Noise
  5. 2Finding the Slew Rate
    1. 2.1 Finding the Slew Rate With an Oscilloscope
    2. 2.2 Calculating the Slew Rate From Power and Frequency
  6. 3Impact of Slew Rate on Phase Noise
    1. 3.1 Modeling of Input Clock Slew Rate, Inherent Device Noise, and Output Jitter
    2. 3.2 Slew Rate Impact on Flicker Noise and Noise Floor
  7. 4Application of Slew Rate Rules to PLL Synthesizers
    1. 4.1 PLL Flicker Noise
    2. 4.2 PLL Figure of Merit
    3. 4.3 Other Areas in PLLs Where Slew Rate has an Impact on Performance
    4. 4.4 Improving PLL Slew Rate for Better Performance
  8. 5Application of Slew Rate Rules to Data Converters
  9. 6Summary
  10. 7References
  11.   Appendix A: Relating Slew Rate, Power, and Frequency
  12.   Appendix B: Relating Slew Rate, Frequency, Jitter, and Phase Noise
  13.   Appendix C: Equations for Data Converters
    1. 8.1 Relating Sampled Signal Slew Rate to SNR
    2. 8.2 Justification That SNR Decreases 1dB per 1dB With Input Power for Slew Rate Limited Case
  14.   Appendix D: Calculations for Data Converter Example

PLL Flicker Noise

A PLL can have an input buffer that can limit the phase noise in some cases. The close in 1/f noise behaves very much like a buffer with the exception that the noise is multiplied up by a factor of 20×log(N), where N is the feedback divider value. The flicker noise is typically normalized to 1GHz carrier frequency and 10kHz offset, although there are other ways to create a metric. Figure 4-1 shows the PLL flicker noise degradation as a function if input power for the Texas Instruments LMX2615-SP PLL using a constant 100MHz input frequency. Note that this behaves very similar to a buffer.

 Impact of Clock Input Power on
                    Flicker Noise for LMX2615-SP PLL Figure 4-1 Impact of Clock Input Power on Flicker Noise for LMX2615-SP PLL

The model assumes for low input power, the trend is 1dB degradation per 1dB reduction in output power. Around 3dB input power, a knee is reached where the slew rate limited noise and non-slew rate limited noise are equally contributing to the total noise.