• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 21 subscribers
  • Views 1569 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.

LMX2592: Apparent problem with SEG2 Divider at /8

Raymond Harker
Prodigy 150 points
Part Number: LMX2592

I have a custom application utilizing the LMX2592 in a broadband configuration.  I'm working on calibrating my output power, but I've come across 5 frequency ranges with issues (424-444MHz, 212-222MHz, 106-112MHz, 53-56MHz, and 26-28MHz).   When I run the part without the dividers enabled, it covers the full VCO range (3.55-7.1GHz) without issue.  When I enable the dividers, everything works perfectly except in these 5 ranges.  Clearly these correspond to the high end of the VCO range (~7.04GHz) when used with the dividers.  When I enter into these problem ranges, my SEG1 * SEG2 divider value apparently switches from /16 to /20 (setting the output to 440MHz results in a 352MHz output, 7040MHz VCO frequency, 220MHz to 176MHz, etc).  If I toggle between the channel divider and the VCO buffer, my output goes to 7.04GHz  so my N value doesn't change.  The device achieves and maintains lock at all output frequencies, so the problem appears to be on the divider side of the signal chain.  I've matched my registers to the TICS Pro Software registers in the same configuration, but the problem persists.  Any ideas what could cause this?

  • 0
    TI__Expert 8505 points

    Raymond,

    You are right on with your understanding. Let me give you additional information to help.

    Information from the LMX2592 datasheet:

    If you see the channel divider table (table 8). You see that if you just use SEG1 (divide by 2 or 3) there is no limitations, you can use the whole range. After that, the guideline from the table is to use the smallest divide combination until you reach the bottom of the VCO 3550MHz (rounded to the nearest integer) and then using the next divide combination. As you see, this will use the lowest frequency portion of the VCO as much as possible.

    In depth look at the divider segments:

    Now the question is what VCO frequency can you really use each SEG up to. This is just a guideline but will help you have a better understanding. Within SEG2, "/2" is good up to around 7100MHz, "/4" is good up to around 7000MHz, "/6" and "/8" is good up to around 6600MHz.

    TICS Pro:

    I took a look, TICS Pro does not have the equations correct to reflect the datasheet procedure, thus was giving you wrong recommendations. Sorry for the inconvenience. For now, use the datasheet as the correct reference, and we will get the TICS Pro formulas fixed.

    Thanks and let us know if anything else!

    Regards,

    Brian Wang

  • 0 in reply to Brian Wang
    Prodigy 150 points
    Thanks Brian, good to know it's a device limitation and not a board-level error on my part. I was working off the Feb. 2017 revision of the datasheet which doesn't include this information, so I'll update my copy ASAP.
  • 0 in reply to Brian Wang
    Prodigy 150 points

    While implementing the above fix, I've come across another anomaly.  It seems that there is a fairly large (~5dBm) discrepancy between different settings for the 1183-2366MHz region.  If I follow the table and use the SEG1 /3 setting for this range, my output power is about 5dBm lower than if I implement SEG1 and SEG2 in a /4 configuration.  This is at a power setting of 31.  If I drop down to 0, then the discrepancy is only about 2dBm.  I've already checked to make sure I have the proper buffers and muxes enabled/disabled.  Strangely, I don't see this discrepancy if I have SEG2 or SEG3 enabled while SEG1 is at /3.  It only appears when SEG1 is the only enabled divider.  Any suggestions on where I'm going wrong?

  • 0 in reply to Raymond Harker
    TI__Expert 8505 points

    Raymond,

    It is probably due to the waveform. When only using SEG1 (just /2 or /3), the divider circuitry is alittle different. The /2 is a normal 50% duty cycle square wave, the /3 however is more like a 33% duty cycle wave form. So the measured RMS output power may be different even at the same output frequency. When you use SEG2 and 3 though, as you saw it doesn't matter what you had in SEG1, and that is because now your last stage uses a normal /2 type divider again, which produces outputs with 50% duty cycle. I did a measurement with an EVM as well, but did not see that big a difference though, I get about 0.3dBm difference if using power setting of 15 and about 1dBm difference with a setting of 31. What kind of measurement equipment do you use? And do you have other stages after the LMX2592 output (a balun/ transformer)?

    Regards,

    Brian Wang

  • 0 in reply to Brian Wang
    Prodigy 150 points

    I found some funkiness with my register handling, after fixing that I'm seeing about a 1dB difference, which is manageable.  I appreciate the help.

  • 0 in reply to Raymond Harker
    TI__Expert 8505 points

    Raymond,

    Great to hear! Just like what I see here.

    Regards,

    Brian Wang