SLAA870 February   2019 AFE7422 , AFE7444

 

  1.   Evaluating the frequency hopping capability of the AFE74xx
    1.     Trademarks
    2. 1 Introduction
    3. 2 Phase Coherency vs Phase Continuity
    4. 3 AFE74xx Architecture
      1. 3.1 AFE74xx Receivers: Multiband DDC
      2. 3.2 AFE74xx Transmitters: Multiband DUC
      3. 3.3 Numerically Controlled Oscillator (NCO)
        1. 3.3.1 Programming the NCO frequency
          1. 3.3.1.1 Example: Programming NCO to 1700MHz
        2. 3.3.2 Direct Digital Synthesis (DDS) Mode
    5. 4 Frequency Hopping Methods
      1. 4.1 Maintaining Phase Continuity
        1. 4.1.1 Phase Continuous Hop Time
          1. 4.1.1.1 Serial Peripheral Interface (SPI)
          2. 4.1.1.2 Test Setup
          3. 4.1.1.3 Software Configuration
          4. 4.1.1.4 Test Results
      2. 4.2 Maintaining Phase Coherency
        1. 4.2.1 TX NCO Hopping Using SPI
          1. 4.2.1.1 TX NCO Switch Using SPI Hop Time
            1. 4.2.1.1.1 Software Configuration
            2. 4.2.1.1.2 Test Results
          2. 4.2.1.2 AFE74xx DAC Settling Time
            1. 4.2.1.2.1 Hardware Setup
            2. 4.2.1.2.2 Software Configuration
            3. 4.2.1.2.3 Test Results
        2. 4.2.2 RX NCO Hopping Using the GPIO Pins
          1. 4.2.2.1 Test Setup
          2. 4.2.2.2 Software Configuration
          3. 4.2.2.3 Test Results
    6. 5 NCO Frequency Resolution Versus Hop Time
    7. 6 Fast Frequency Hopping With the Load and Switch
    8. 7 Register Addresses
    9. 8 References

5 NCO Frequency Resolution Versus Hop Time

There is a trade-off between NCO RF step size (resolution) and hop time. As mentioned previously, a 32-bit accumulator word determines the output frequency of the NCO, and there are four designated NCO registers that each store a byte of the total 32-bit accumulator word. Thus, a single SPI write is required to program each byte of the total accumulator word. Each SPI write adds about 600 ns to the overall hop time. Therefore, reducing the NCO resolution from 32 bits to 24 bits by programming only the first three NCO registers bypasses the NCO register containing the 8-LSB bits, and ultimately reduces overall hop time by 600 ns. The trade-off is that reducing NCO resolution from 32 bits to 24 bits decreases the NCO resolution. A 32-bit NCO can move in frequency steps of approximately 2 Hz. A 24-bit NCO can save roughly 600 ns of hop time compared to a 32-bit NCO, but NCO resolution decreases from 2 Hz to roughly 527.35 Hz. If an RF step-size of 135 kHz is suitable for an application, only the 16 most significant bits of the NCO accumulator word must be programmed, reducing the number of required SPI writes, and ultimately reducing the overall hop time by approximately 1.2 µs compared to the full 32-bit NCO. Table 3 shows the trade-off between NCO resolution and overall hop time.

Table 3. Trade-Off Between NCO Resolution and Overall Hop Time

NCO Word MSB Frequency Step Size (Resolution) Total SPI Write Time (Update NCO Frequency + NCO Reset) Settling Time Total Hop Time When Updating 1 NCO Total Hops per Second
32 bits of NCO word 2.06 Hz 4.2 µs 90 ns 4.29 µs 1 / (4.29 µs + Dwell time)
24 MSB of NCO word 527.35 Hz 3.6 µs 90 ns 3.69 µs 1 / (3.69 µs + Dwell time)
16 MSB of NCO word 135.00274 kHz 2.4 µs 90 ns 2.49 µs 1 / (2.49 µs + Dwell time)
8 MSB of NCO word 34.43 MHz 1.0 µs 90 ns 1.09 µs 1 / (1.09 µs + Dwell time)