• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Answers 1 answer
  • Subscribers 21 subscribers
  • Views 937 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.

LMK04828-EP: Minimum Output Frequency Configuration

Lawrence Trevor1
Prodigy 90 points
Part Number: LMK04828-EP
Other Parts Discussed in Thread: LMK04208, LMX2594

Hi Forum,

I'm investigating substituting an LMK04208 in our system for an LMK04828-EP, simply for the expanded temperature range.

Our LMK04208 configuration was producing a clock output of 25.6 MHz for an LMX2594. The product page for the LMK04828-EP says the minimum frequency is 0.001 MHz so I thought this would be an easy substitution. But looking at the minimum VCO frequency and output clock dividers, I can only see getting an output of 2450 MHz / 32 = 76.5625 MHz. Having a peek at the LMK04828B page, I see that you have stated the minimum output frequency as 0.289 MHz, which is the minimum VCO frequency divided by the maximum SYSREF frequency (2370 / 8191 = 0.289...). I'm led to believe that, as for the 04828B, this very low output frequency for the EP model is in fact the lowest SYSREF frequency, and not that of the regular clock output paths.

Is this assumption about the minimum frequency correct? If so, am I limited by the 76.5625 MHz frequency on the clock paths? Is this 0.001 MHz output presuming an external VCO that can operate at far lower frequencies? Am I missing something here?

Also, if its too difficult to make this substitution, is there another LMK product that has the temperature range and a better low output frequency, comparable to the 04208?

Thanks,

Lawrence

  • 0
    TI__Mastermind 28560 points

    Hello Lawrence,

    Lawrence Trevor1 said:
    Is this 0.001 MHz output presuming an external VCO that can operate at far lower frequencies? Am I missing something here?

    So what you may be missing is the minimum output frequency is dependent on the mode in which you are using the device.  In reality, if you provided a 1 kHz input to the device, used it in distribution mode and the SYSREF divider to divide 1 kHz, you could then produce 1 kHz / 8191 = 0.1221... Hz.  So unfortunately the minimum frequency of 1 kHz seems incorrect for any of the operation modes...  Unless you are operating with an external VCO that has a minimum frequency of 8.191 MHz.

    In general, there is no trouble with operating the output buffers at lower frequencies.  The thing you would need to be aware of however if AC coupling the AC coupling capacitor would need to be sized appropriately for your output frequency.

    It is true that the LMK04828-EP's clock output divider is limited to /32 so you would not be able to achieve 25.6 MHz.  However you can use the SYSREF divider, run the SYSREF outputs in continuous mode and produce VCO frequency / 8191 which will be less than 25.6 MHz and meet your requirement.  On the LMK04828-EP only the odd outputs can output the SYSREF clock, so you would be limited to 6 outputs at 25.6 MHz which is on par with the number of outputs you can get from the LMK04208.

    Are you operating other frequencies on the LMK04208?  If so keep in mind an active SYSREF clock on an odd output clock will have strong crosstalk to its shared even output.  For example OUT8 and OUT9 would crosstalk strongly together.

    Be aware that the LMK04828-EP has a lead finish of SnPb.

    73,
    Timothy

  • 0 in reply to Timothy T
    Prodigy 90 points

    Hey Timothy, thanks for the detailed answer.

    To expand on my use case, we're also syncing across multiple LMX2594 (or potentially LMX2694) for a Xilinx RFSoC application, so we want to still have the LMK produce the SYNC/SYSREF signal at around 8 MHz, and we need three of these as outputs. Ideally (I presume) we would use the SYSREF dividers to achieve this, and then have one other clock output drive the clock of the LMX. It seems there would be plenty of outputs to avoid cross-talk then.

    A question on distribution mode: This bypasses any and all VCO & PLLs correct? There would not be any clock 'cleaning' going on so OSCin would need to be clean, correct?

    Anyway, I think we can surmount the change in output frequency if needed. I just wanted to understand the other modes of the LMK to see if we could get away without that change.

    Regards,

    Lawrence