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.

LMX2592: Problem About LMX2592 Output Power with Corresponding Register Setting

Part Number: LMX2592

Hi there:

Now we have a problem about LMX2592 using:

LMX2592 is used to generate two 3GHz sampling clocks connected to two ADC12DJ3200s, which are designed as below:

Where, the LMX2592 outputs are named as ADC_LMX2592_REFOUTAP/M and ADC_LMX2592_REFOUTBP/M. Behind the AC-coupling capacitors the two differential-ended clocks are connected to ADC12DJ3200s.

The measured Vpp of LMX2592 output is about 660mV using KEYSIGHT DSO90404A and the attached differential strobe, with the LMX2592 OUTA_POW register(REG46[13:8]) set to 0(decimal). The output the result is as below:

The 660mVpp (approximately -3.46dBm on 100ohm characteristic impedance differential lines) result is far smaller than the nominal rating in the datasheet: 8dBm@3GHz output frequency (Page7: Chapter 6.5 Electrical Characteristics). So, we set the register to a larger value, which was meant to make the output power larger, but the output power didnot increase as the datasheet says "increase power from 0 to 31" in the page 27:"Table41.R46 Register Field Descriptions".

For example: when set to 8(decimal): about 70mVpp

When set to 31(decimal): about 40mVpp

Even when set to 63(decimal): about 30mVpp

Result from the screenshots above, we can see the output power didnot increase with the register value got larger; Inversely the output power seemed to be decreasing with the register increasing.

So, I wanna know where maybe the problem exists, or what is the correct method  to increase the output power of LMX2592.

Looking forward to your answer, thanks a lot~~~!

P.S. the register settings are as below: (register values are generated with the official tool "TICSPro")

        0x002216, ////RESET
        0x002214,

        0x4000AF,
        0x3F00A8,
        0x3E0000,
        0x3D0001,
        0x3B0000,
        0x3003FC,
        0x2F00C0, ///0x2F00C0,///0x2F00CF, power enhance
        0x2E00A3,//0x2E0023,///0x2E0F23,  power enhance

        0x26003C,
        0x2D0000,
        0x2C0000,
        0x2B0000,
        0x2A0000,
        0x2903E8,///0x290001,///Denominator LSB of N-divider fraction
        0x280000,
        0x278204,////17

        0x254000,
        0x240418,////0x240C18,
        0x23101B,
        0x22C3EA,
        0x214210,
        0x204210,
        0x1F0601,
        0x1E0034,
        0x1D0084,
        0x1C2924,
        0x190000,
        0x180509,
        0x178842,
        0x162300,
        0x14012C,
        0x130965,
        0x0E018C,
        0x0D4000,
        0x0C7001,
        0x0B0018,
        0x0A10D8,
        0x090302,
        0x081084,
        0x0728B2,
        0x041943,
        0x020500,
        0x010808,
        0x00221C

  • Hello,

    Are you able to probe at different points along the transmission line? I'm wondering if there are some reflections that are changing location based on your power setting that are resulting in the reduced amplitude.

    Have you done any SI modeling of your trace?

    73,
    Timothy
  • Hi,

    You might want to try these things;
    1. Try this at lower frequency like 1 GHz to rule out transmission line effects
    2. Can you send your raw TICSPro file? This is a lot easier to check.
    3. Also, realize that output power in the range of 32 to 47 is actually LESS than for 31.


    Regards,
    Dean