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.

LMX2581: phase noise between measure and simulation

Part Number: LMX2581
Other Parts Discussed in Thread: CODELOADER, , LMX2582, LMK61E2, LMK04208

Dear Specialists,

My customer is testing LMX2581EVM.

He confirmed phase noise, the data is worse than webench result.

He couldn't achieve target spec.

Could you please advise about workaround?

Please see attached files.

(1)codeloader setting.pdf

(2)WEBENCH R2=1.8k.pdf

(3)webench R2=6.8k.pdf

(4)test data R2=1.8k.bmp

(5)test data R2=6.8k.bmp

(6)phase noise comparison

The customer has changed the parts from the default .

Ref in 10MHz >12MHz

Phase noise -140dBc/Hz@10kHz, -145dBc/Hz@100kHz, -145dBc/Hz@1MHz

R6 >NC

R13>47ohm

R14>NC

L3> 0ohm

Each outputs are terminated by 50ohm.

I appreciate your great help.

Best regards,

Shinichi

①codeloader setting.pdf

②webench_R2=1.8k.pdf

③webench_R2=6.8k.pdf

⑥phase noise.pdf

  • Hi Shinichi,

    If you are trying to meet the following specs:

    Phase noise -140dBc/Hz@10kHz, -145dBc/Hz@100kHz, -145dBc/Hz@1MHz

    It will not be possible to meet this with LMX2581.

    So the options you have:

    1)Have more flexibility with these specs

    2) Use LMX2582 and change the OSCin (ref in frequency) from 10 MHz to 100 MHz, using a high performance reference oscillator.

    I encourage the customer to play with PLLatinum Sim, our software used for design and simulation.

    http://www.ti.com/tool/pllatinumsim-sw

    Below is a simulation of LMX2581 assuming no phase noise contribution from the reference oscillator, and there might be slight variations with practical measurement. This simulation fails to meet the spec at 10 kHz and 100kHz offsets.

    I recommend using LMX2582 which has better PLL and VCO performance. Below there is a simulation showing that it meets the specs:

    As you can see it meets the specs but the reference oscillator had to be changed from 10MHz --> 100 MHz.

    -142.4 dBc/Hz @ 10k Hz offset

    -148.7 dBc/Hz @ 100 kHz offset

    -158.8 dBc/Hz @ 1MHz offset

    Remember this is a simulation and it's not accounting for reference oscillator phase noise contribution which typically affect it for lower offsets (1-10kHZ)

    I would recommend a high quality oscillator to be able to achieve this performance for LMX2582:

    1) VC-708 fixed 100 MHz frequency from Vectron

    2) LMK61E2 Programmable oscillator from TI. (Can output from 10 MHz to 1GHz)

    Hope this helps!

    Thanks,

    Julian

  • Dear Julian

    Thank you for your precise answer.

    I understand LMX2581 is not enough to meet the customer's requirements.

    I'll send the answer to him and make him consider.

    If he has an additional question, I will consult you.

    I appreciate your great help.

    Best regards,

    Shinichi

  • Dear Julian

    I explained your answer to the customer.

    He considers LMX2582, too.

    He wants to use for FPGA input differential clock, so I also recommended LMK04208.
    This device is VCO outside device, it has a possibility to improve phase noise.

    Which device to use, depend on the system requirement.

    I appreciate your great help.

    Best regards,
    Shinichi
  • Hi Shinichi,

    Great to hear that the customer is considering LMX2582. Let us know if he has any more questions.

    Thanks,
    Julian