• State
  • Locked Locked
  • Replies 1 reply
  • Subscribers 21 subscribers
  • Views 125 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Original question:

LMX2594: LMX2594 Phase Noise Prediction vs Measurement

LMX2594: LMX2594 Phase Noise

?? ?
Prodigy 10 points
Part Number:

Tool/software:

Hi, TI:

I have a question regarding phase noise performance for the LMX2594.

          The following figure shows the GUI interface settings of Case1, Case2, and Case3, respectively. The difference between case 1 and case 3 is that Fosc is 10 MHz and 100 MHz. Case 2 is the fine-tuning result of Case 1 (Doubler: X2 and Pump Charge 15mA).

          The final figure shows the measurement result of SA, and the yellow line(case 1) is the worst case of the three cases. Since the actual application requires the setting of Case1, I would like to know if there are other settings that can reduce the noise when the doubler is fixed to X1 and the charge pump is fixed to 15mA.

 

Case1:

Case2:

Case3:

 

Thanks for your help.

Brian Chiou

  • +1
    TI__Guru** 107030 points

    Hi Brian,

    Couples of thing that will affect PLL inband noise.

    1. Input clock slew rate (or edge rate), higher slew rate is better. If you have to use 10MHz, do not use sine wave clock, use square wave clock such as CMOS or differential clock.

    2. Input clock frequency. With 10MHz input, fpd is limited to 20MHz. However your VCO frequency is 12GHz, as a result, N is very big. N divider adds noise to the PLL, the bigger the N value, the more added noise to the PLL.

    3. Loop bandwidth is limited by fpd. Again your max. fpd is just 20MHz, if you use 10MHz input. Therefore, the max. possible loop bandwidth is limited. You cannot design a loop filter with high loop bandwidth. As a result, you cannot filter more close-in VCO noise. 

    Case 2 configuration is the best you can get from a 10MHz input, if you were using a sine wave clock, you can get better result with a square wave clock. If you do not have a good square wave clock, you can use our LMK1C clock buffer to convert 10MHz sine wave to square wave clock. See below appnote for details.

    https://www.ti.com/lit/pdf/snaa411