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.

LMK04208: Why 0-day mode can't work?

Part Number: LMK04208

Tool/software:

Hi,

I want LMK04208 work at Dual PLL, Int VCO, 0-delay mode on my board. It will feedback CLKout0 to PLL1 in the internal path as below figure shown. Clock setup is listed below:

CLKin1: external 7.68MHz sine clock from signal generator

CLKout0: 7.68MHz

CLKout1: 7.68MHz

CLKout2: 122.88MHz

CLKout4: 122.88MHz

Test with oscilloscope, I find that CLKout0 and CLKout1 couldn't be aligned, and also don' t have fixed phase relation over power cycle. I think this clock plan or setup conform to two rule of Zero-delay mode totally. But CLKout1 couldn't share the same phase relation with CLKout0 against CLKin1, why?

attach the .tcs file for your inspection.

122M88_PL_122M88_SYSREF_7M68_clk5_12M8_MCS_No_Sync_V4.tcs

Thanks in advance!

Best regards!

Jason

  • Hello Jason, 
    Are you sync the dividers to ensure all the outputs are synchronized? 
    Since you are utilizing zero-delay dual PLL mode? 
    The idea becomes you feed CLKOUT0 as reference to the N divider in PLL1 which now ensure CLKINx and CLKOUTx are aligned. 
    Given that sync synchronizes your outputs - and one output is in sync with the input now all the outputs are in sync with the input.

    You should have deterministic delay every power cycle after performing sync. 

    Best regards, 

    Vicente

  • Hi, Vicente:

    I only use zero-delay mode (feedback CLKout0 to PLL1) , and don't perform sync. Your meaning is that performing sync is also be needed and together with zero-delay mode to align phase of all CLKouts, right??? 

    Application Report named Multi-Clock Synchronization mentioned  two rule of Zero delay mode. 

    It also mention that all the other outputs can share the phase determinism by synchronization with the D2 divider in the feedback.

    Here, the synchronization is the meaning of performing sync???

    My clock plan listed below is conform to the rule totally. Test also find that CLKout0 align CLKin1 determinately, but CLKout1 don't have fixed phase relation with CLKout0

    CLKin1: external 7.68MHz sine clock from signal generator

    CLKout0: 7.68MHz(internal feedback to PLL1)

    CLKout1: 7.68MHz

    CLKout2: 122.88MHz

    CLKout4: 122.88MH

    Thanks in advance!

    Best regards!

    Jason

  • Hi Jason,

    correct you still need to sync.

    You have confirmed that CLKINx & CLKOUT0 have deterministic phase. 
    now if you sync - all the outputs are phase aligned but given that CLKOUT0 is aligned with CLKINx - now all the outputs are aligned with CLKINx as well.

    best regards,

    vicente 

  • Hi, vicente:

    Let's continue this thread again, cause I find a new case

    During test, when all NO_SYNC_CLKoutX bit =0, I found that all output clocks have been aligned each other after configuring the registers file into LMK04208. But, at this moment I didn't apply external SYNC signal or toggle the SYNC_POL_INV bit yet. I check the methods of Generating SYNC, and programing PLL2_N in the Register R30 will generate a SYNC event automatically as shown in the Figure.1. So, during configuring registers file after power on,  the configuration sequence is from R0 to R31 as shown in the Figure.2.

    When R30 was programmed with all NO_SYNC_CLKoutX bit =0, will it generate a SYNC to align the phase of all output clocks ?

    Figure.1

    Figure.2

    Thanks in advance!

    Best regards!

    Jason

  • Hi Jason, 
    If the NO_SYNC_CLKOUTx bit is set to zero - this would result in a sync event as you're stating you do want to have that specific x output sync. 
    Best regards, 

    Vicente