• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 21 subscribers
  • Views 434 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.

LMX2820: Clock Reference conditioner

Aizkibel Salcedo
Prodigy 180 points
Part Number: LMX2820


Hi,

In a new design we are going to use the LMX2820 as PLL. We must use as reference an external 50 mhz sinewave signal, which power level goes from -10 to 15 dBm.

 To avoid burning the PLL, we will use an HBT amplifier with low phase noise working in saturation, but we have many doubts:

    -Is the approach correct?

    -If the answer is yes, shall we filter the signal?

    -If the approach is non-correct, instead of using an amplifier, What can we use? Maybe a clock cleaner?

    -What is the minimum input level that the LMX2820 can support without increase the phase noise?

Thanks and best regards,

Aizkibel

  • +1
    TI__Guru** 110430 points

    Hi Aizkibel,

    Both phase noise and slew rate of the reference clock are important to a PLL.

    If 50MHz is the max. input frequency you can provide, I suggest make its power equal to 15dBm so as to maximize its slew rate.

    whenever you use an active device to massage the reference clock, make sure the close-in phase noise (less than the desired loop bandwidth of LMX2820) will not get degraded. What kind of active device is not important. 

    Input level is not a factor for phase noise performance. For example, we could get good phase noise from LMX2820 with a 300mV LVDS reference clock. However, we may get bad phase noise with a low quality CMOS XO reference clock.

  • 0 in reply to Noel Fung
    Prodigy 180 points

    Thanks for the answer Noel, 

    15 dBm=3.5Vpp, more than the limit of the LMX2820, but I understand your commentary about the slew rate.

    The problem is that we cannot choose the input level. In some situations it will be -10 dBm( 150 mVpp) and in other situations it will be 10 dBm (2Vpp), so sometimes we will be below the minimum input level of the LMX2820 (0.3mVpp). To avoid this we will need the active component, as you say, however, if we need an Vpp lower than 3.3V at any case we must saturate the amplifier, so our sinewave signal  will be a """"square""" signal (tone + harmonics). We don't know if after this step it is better to filter the signal to remove the harmonics or to use the square signal directly.

    Thanks and best regards,

    Aizkibel

  • 0 in reply to Aizkibel Salcedo
    TI__Guru** 110430 points

    Hi Aizkibel,

    The internal OSCin buffer will also create harmonics, a square wave reference clock is fine, no need to do filtering.