Digital PLL and TDC Detailed Explanation

Jun 27, 2024

Digital PLL Lecture Notes

Overview

  • Previous Session: Simplified block diagram of a digital PLL
  • This Session: Detailed function of digital PLL components, with a focus on TDC (Time to Digital Converter)

Simplified Block Diagram of Digital PLL

  • Components:
    • TDC (Time to Digital Converter): Converts phase error to digital bits
    • Digital Loop Filter:
      • Proportional path gain (Kp)
      • Integral path gain (Ki)
    • Accumulator (z⁻¹ models the delay)
    • Digital to Analog Converter (DAC): Converts digital control word to control voltage
    • Oscillator controlled by voltage from DAC
    • Feedback divider for frequency multiplication
  • Signals:
    • Reference frequency (Input)
    • Output frequency (Controlled by V-control)
    • Symbols:
      • d<sub>tdc</sub> (Output of TDC)
      • d<sub>p</sub> (Output of proportional path)
      • d<sub>i</sub> (Output of integral path)
      • d<sub>f</sub> (Combined digital control word)

Time to Digital Converter (TDC)

  • TDC Function: Measures phase error between reference (R) and feedback (V) clocks, converting it into a digital word (d<sub>tdc</sub>)
  • Digital Word: x-bit representation of phase error
  • Error Measurement:
    • TDC Range: Typically counts phase error from -T<sub>ref</sub> to +T<sub>ref</sub> (2T<sub>ref</sub> total)
    • Digitizing Phase Error: Phase error (∆t) converted to binary equivalent using TDC resolution (∆t<sub>LSB</sub>)

Key Equations and Concepts

  • Analog to Digital Conversion: A<sub>error</sub> = (Voltage/Reference Voltage)
  • Digital Output Calculation: For phase error ∆t
    • ∆t digitized to control word d<sub>tdc</sub>
    • ∆t<sub>LSB</sub> (Resolution):
      • ∆t<sub>LSB</sub> = 2T<sub>ref</sub>/2ⁿ
    • d<sub>tdc</sub> Calculation: d<sub>tdc</sub> = (∆t/∆t<sub>LSB</sub>)
    • Gain: K<sub>tdc</sub> = (2ⁿ/(4π)) × φ<sub>error</sub>

Example

  • N-bit TDC with resolution ∆t<sub>LSB</sub> for 10-bit word:
    • N = 10, ∆t<sub>LSB</sub> = (2T<sub>ref</sub>) / 2¹⁰
    • Digital Equivalent Calculation:
      • For phase error of 496: d<sub>tdc</sub> ≈ 0100000000 (binary), 496 (decimal)
      • Phase error φ<sub>error</sub> = (4π × 496) / 1024

Implementation of TDC

  • Simple Delay Line Implementation:
    • Delay line with 2ⁿ delay cells (Each cell delays signal by ∆t<sub>LSB</sub>)
    • Sampled outputs by clocked registers
    • Registers store values at rising edge of reference clock
    • Sum of register values gives phase error
  • Delay Line Setup:
    • Ensure delay line = 2T<sub>ref</sub> for accurate error measurement
    • Each cell delay = 2T<sub>ref</sub>/2ⁿ = ∆t<sub>LSB</sub>
  • Resolution and Range:
    • Range of TDC: 2T (or T)
    • Number of delay cells ∝ Reference Clock Period
  • Quantization Noise:
    • Limited by ±(∆t<sub>LSB</sub>/2)
    • Actual phase error: φ<sub>error</sub> = (phase error + quantization noise)

TDC Small Signal Model

  • Concept: Measure φ<sub>error</sub>, digitize via TDC to output d<sub>tdc</sub>
  • Gain of TDC: K<sub>tdc</sub> = (2ⁿ/4π)
  • Model:
    • φ<sub>error</sub> = φ<sub>ref</sub> - φ<sub>fb</sub>
    • d<sub>tdc</sub> = [φ<sub>error</sub> × K<sub>tdc</sub>] + quantization noise (q<sub>tdc</sub>)

Conclusion

  • Key Takeaways:
    • TDC converts phase error between reference and feedback to digital control word
    • Accurate measurement hinges on precise delays (∆t<sub>LSB</sub>) and number of delay cells
    • Quantization noise affects accuracy but is limited to ±(∆t<sub>LSB</sub>/2)
    • Small signal model includes both gain and quantization noise

Note: Multiple TDC implementations exist; refer to literature for more complex designs.