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.

LMX2572: Fail to lock to specific frequency

Part Number: LMX2572

We are using part to sweep through 500 frequencies continuously. We are letting the device handle calibration at each freq without any assist. We read back lock status via SPI. Sometimes it fails to lock. The failure is always when we go from 2671.5MHz to 2684.5MHz. This happens to be the freq where device goes from VCO4 to VCO5.
We have checked and rechecked clocks, power etc. When it is in this unlocked state we can read back all correct values in its registers. Writing R0 a second time does succeed. Is this a known issue?

  • Hi David,

    These two frequencies fall into VCO5, and you only need to program PLL_NUM + R0 to change the frequency from one to the other. Did you mean you need to program PLL_NUM + R0 + R0 to get it lock?

    Could you share you register setting?

  • What do you mean they fall into VCO5? If you are referring to the datasheet, it is not consistent with the parts behavior when you let it do its own calibration (I have verified this using eval board and TICS Pro).

    We are writing a subset of registers to configure next freq then starting calibration with R0 (no assist), waiting for lock, then polling lock status and VCO with R110. 

    As shown below, moving the reference frequency up or down slightly (without changing the programming of the part) seems to prevent the failure.

    Our nominal reference is 104MHz:

    With reference of 103.96MHz, freqs are 2670.4725MHz (5340.945MHz VCO4) to 2683.4675MHz (5366.935MHz VCO5) always locks successfully.

    With reference of 104MHz, freqs are 2671.5MHz (5343MHz VCO VCO4) to 2684.5MHz (5369MHz VCO5) sometimes fails to lock.

    With reference of 104.04MHz, freqs are 2672.5275MHz (5345.055 MHz VCO4) to 2685.5325 MHz (5371.065 MHz VCO5) always locks successfully

    I have attached two files. First text file is our initialization of all registers. Second text file is register values written to configure the two freqs in question. 

    2766.lmx2572RegisterInit.cfg

    7752.register configs.txt
    Values are in hex, first byte is address.
    
    Register values at first freq
    4B0800
    2DC600
    2C12A3
    2B0000
    2AC000
    250305
    240066
    220010
    E1820
    C5001
    BB028
    A10F8
    90004
    10808
    2098
    
    Register values at second freq (where it sometimes fails to lock)
    4B0800
    2DC600
    2C12A3
    2B0000
    2A4000
    250305
    240067
    220010
    E1820
    C5001
    BB028
    A10F8
    90004
    10808
    2098
    

  • Hi David,

    Could you try with PLL_DEN = 10400 (0x2A0000, 0x2B1E78)?

    Could you also read back R111?

  • Noel,

    Can you let me know what you are getting at with this specific denominator?

    We have tried with both denominator at 0xFFFF_FFFF and 0xC000_0000, problem still occurs.

    Do you mean shift the VCO freq by changing only the denominator, or maintain same VCO freq by adjusting the numerator appropriately?

    Dave 

  • Hi Dave,

    Since you can get exact frequencies from the synthesizer with PLL_DEN = 10400, I am wondering if it is necessary to make PLL=DEN = 2^32. In addition, I am not quite convinced that VCO4 is picked up by calibration, so I want to make PLL_DEN = 10400 to see if there is any difference. Reading back the value from R111 can help me make a better guess.