• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 22 subscribers
  • Views 680 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.

LMX2541 Stabilize the output frequency.--Urgent

vengadesh uthirapathy
Intellectual 770 points

Other Parts Discussed in Thread: LMX2541

Hi,

This is vengat here from Avnet, one of my customer they are using ‘LMX2541SQ2690E’ in our board.  They are trying to set output frequency to 1333MHz, but we are getting 1332.985830MHz. Synthesizer register values to be programmed are taken from code loader. Apart from this, output frequency is not stable also, but they are getting ‘LOCK’, It is varying as below,

1332.985830MHz  at T0 state (power on state)

1332.985600MHz  at ‘T0 + 30seconds’

1332.985400MHz  at ‘T0 + 1minutes’

1332.985200MHz  at ‘T0 + 2minutes’

Deviation from set frequency is OK for their  requirement, but it should be stable one. We have observed same behavior in evaluation board (‘LMX2541SQ2060E’) also. Can you please sugges the solution to stabilize the output frequency ?

Regards/

Vengat

 

  • 0
    TI__Mastermind 32575 points

    Hi Vengat,

    The LMX2541 output frequency drift is likely due to the frequency drift of the reference oscillator.  Assuming the LMX2541 PLL loop response is well designed, it should precisely track the frequency accuracy of the reference oscillator.  What reference oscillator model are they using on their board?

    My comments on their frequency measurements:

    • At time T0 (power up), the frequency error is -10.6 ppm relative to the nominal center frequency; as you said, this error is OK for their requirement.
    • At T0 + 2 minutes, the frequency error has drifted to -11.1 ppm (-0.5 ppm relative to the frequency at T0).  This seems within the normal variation of most standard XOs (not temperature-compensated or oven-controlled oscillators), which will likely drift more over temperature and voltage variation.

    Bottom line: 
    If they need tighter frequency tolerance, they may want to try a TCXO or even OCXO as the reference oscillator, depending on their requirements.  TCXOs usually have frequency tolerance specs as low as +/-0.25 ppm and up to about +/-5 ppm across the specified temperature range.

    Note that the LMX2541 EVM uses a standard XO (CWX-813) as the reference oscillator, not a TCXO as stated in the EVM user guide.

    Best regards,
    Alan

     

  • 0 in reply to Alan O
    Intellectual 770 points

    Hi Alan ,

    Thanks for your support ,

    They are using VCXO oscillator “CVHD-950X-100” from Crystek.  Based on your feedback customer is working on that i will get back to you.

    Regards/

    Vengat.

  • 0 in reply to vengadesh uthirapathy
    TI__Mastermind 32575 points

    Hi Vengat,

    Check that the voltage control (Vc) input is biased to mid-rail using a voltage divider and not floating or tied to the rails.  However, the VCXO frequency can still be pulled according to the VCXO's tuning sensitivity due to Vc / supply voltage variation.

    Regards,
    Alan

  • 0
    TI__Mastermind 28031 points

    Vengat,


    The LMX2541 locks to the PLL with 0 ppm error to the input reference. 

    However, if the input reference drifts, then so does the PLL.

    It would seem a trivial task to find an TCXO that had (1) good phase noise, (2)  good stability, (3)  Easily available (I can get on digikey), (4)  Standard package, (5) <$20 in low volume.  I could find no such TCXO.

    The LMX2541 board uses the Connor Winfield CWX-813 XO (not TCXO).   This oscillator has good phase noise, is easily available, <$20, and has a standard package.   The only issue is that it drifts a lot, which I htink is what you are measuring.  On our evaluation board, you have to let this warm up and it will still drift, but not as much.

    If this drifting is bothering you, they you can use a signal generator and flip the resistor on the OSCin pin for that.  Just be aware that signal generators are noisy and can add noise.  If you do this, try using a higher frequency on the signal generator and dividing this down more.