AN-1154: Optimizing Phase Noise and Spur Performance of the ADF4157 and ADF4158 PLLs Using Constant Negative Bleed

by Robert Brennan Download PDF

Introduction

The phase noise (PN) and integer boundary spur (IBS) performance of the ADF4157 and the ADF4158 can be improved by activating a constant negative bleed current. The biggest improvement is achieved at frequencies at, or close to, integer multiples of the phase frequency detector (PFD) frequency. It is more visible for loop bandwidths greater than 60 kHz; however, it is recommended to use constant negative bleed for all PLL loop bandwidths.

Constant negative bleed current works by adding a constant offset to the charge pump (equivalent to a phase offset in the PLL loop). This has the effect of linearizing the charge pump by moving away from the nonlinear area near the origin (sometimes referred to as the charge pump dead zone). Figure 1 illustrates this phenomenon.

Figure 1. The Effect of Bleed Current on Charge Pump.

Without this constant current offset, sigma-delta quantization (dQ) noise can fold back in-band and cause excessive noise or spurs. This aliasing of sigma-delta noise in-band only happens for high resolution sigma-delta modulators (equivalent to high value of modulus) as used on the ADF4157 and ADF4158. These parts require activation of a constant negative bleed current to achieve optimal phase noise and spur performance. This current is not needed for other Fractional-N PLLs with lower value of modulus or Integer-N PLLs.

The constant negative bleed on the ADF4157 and ADF4158 is activated by setting bits DB[24:23] in Register 4 to 0b11.

The Effect of Bleed Current on Phase Noise and Spurs

The use of a constant negative bleed only improves phase noise and integer boundary spurs for a certain range of charge pump currents (ICP). For some values of ICP, phase noise and spurs can degrade. This phenomenon was measured for two PFD frequencies, 12.5 MHz and 25 MHz. Each PFD frequency was tested near two consecutive integer channels. The loop filter configuration for each PFD frequency is show in the Appendix.

How Measurements Were Recorded

The measurements were recorded on an EV-ADF4157SD1Z evaluation board. The loop filter was modified for each PFD frequency.

  1. The loop was locked at 5800.001 MHz using a PFD frequency of 25 MHz.
  2. The charge pump current was set to the minimum value (0.31 mA).
  3. Negative bleed was disabled.
  4. The phase noise at a 5 kHz offset and the integer boundary spur at 1 kHz were recorded.
  5. Negative bleed was enabled.
  6. The phase noise at a 5 kHz offset and the integer boundary spur at 1 kHz were recorded.
  7. Step 3 to Step 6 were repeated for every charge pump current setting up to 5 mA.
  8. Step 2 to Step 7 were repeated with the loop locked at 5825.001 MHz.
  9. The loop filter was modified for a PFD frequency of 12.5 MHz, the loop was locked at 5800.001 MHz using a PFD frequency of 12.5 MHz. Step 2 to Step 8 were repeated.

Results

PFD Frequency = 25 MHz


Output frequency = 5800.001 MHz

Figure 2. Phase Noise and IBS at 5800.001 MHz with a PFD Frequency = 25 MHz.

Output frequency = 5825.001 MHz

Figure 3. Phase Noise and IBS at 5825.001 MHz with a PFD Frequency = 25 MHz.

PFD Frequency = 12.5 MHz


Output frequency = 5800.001 MHz

Figure 4. Phase Noise and IBS at 5800.001 MHz with a PFD Frequency = 12.5 MHz.

Output frequency = 5825.001 MHz

Figure 5. Phase Noise and IBS at 5825.001 MHz with a PFD Frequency = 12.5 MHz.

Analysis of Results

In Figure 2, it can be seen that, for a PFD frequency of 25 MHz, using a charge pump current between 3.13 and 3.75, is the best option for optimum PN and IBS. This is consistent with Figure 3, which shows the values between 3.13 and 3.75 are optimum for both frequencies.

From Figure 4 and Figure 5, it is clear that, for a PFD frequency of 12.5 MHz, no value of charge pump current improves PN, but using a charge pump current of 4.06, and higher, results in a considerable improvement of IBS without too much degradation of PN.

Conclusion

For some PFD frequencies, using negative bleed with a particular charge pump current results in improved integer boundary spurs and phase noise.

At other PFD frequencies, using negative bleed does not result in any improvement to phase noise, but can give significant improvement in integer boundary spurs. In this situation, the tradeoff between optimum integer boundary spurs or optimum phase noise depends on the application.

It may be necessary to repeat the measurement in this application note with a specific application’s PFD frequency, to find the optimum charge pump current.

Appendix

Table 1. Constant Negative Bleed vs. Charge Pump Current Scaling
CP Current (mA) Bleed (µA % Bleed
0 0.3125 100 32
1 0.625 200 32
2 0.9375 200 21
3 1.25 300 24
4 1.5625 600 38
5 1.875 700 37
6 2.1875 700 32
7 2.5 800 32
8 2.8125 100 4
9 3.125 200 6
10 3.4375 200 6
11 3.75 300 8
12 4.0625 600 15
13 4.375 700 16
14 4.6875 700 15
15 5.00 700 14

Loop Filters


Loop filter configuration for PFD frequency = 25 MHz. Charge pump current = 2.5 mA.

Loop bandwidth      107 kHz
Phase Margin        45°
C1                  560 pF
R1                  680 Ω
C2                  6.8 nF
R2                  1.2 kΩ
C3                  220 pF

Loop filter configuration for PFD frequency = 12.5 MHz. Charge pump current = 2.5 mA.

Loop bandwidth      101 kHz
Phase Margin        47°
C1                  220 pF
R1                  1.2 kΩ
C2                  3.3 nF
R2                  2.7 kΩ
C3                  100 pF

Author

Robert Brennan

Robert Brennan

Robert Brennan joined Analog Devices in 2010 after graduating with a bachelor’s in electronic engineering from University of Limerick, Ireland. He worked as an RF applications engineer at the Analog Devices Limerick office for several years before relocating to the U.S. He is currently a marketing engineer in the RF and Microwave Group based in Massachusetts. His primary focus is on PLLs, VCOs, and integrated PLL/VCOs. He is currently studying for a Masters of Science in Engineering Management at Tufts University.