• State TI Thinks Resolved
  • Locked Locked
  • Replies 15 replies
  • Answers 4 answers
  • Subscribers 21 subscribers
  • Views 769 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

LMX2572: full assist mode

Kailyn Chen
Genius 9560 points
Part Number: LMX2572

Hello，

I'd like to confirm the full assist mode of LMX2572, according to the datasheet, it describes: 

The user forces the VCO core (VCO_SEL), band (VCO_CAPCTRL), and amplitude (VCO_DACISET) and manually sets the value. No VCO calibration will be performed. To force the set values, set VCO_SEL_FORCE, VCO_CAPCTRL_FORCE, and VCO_DACISET_FORCE equal 1. First do a VCO calibration and then read back the values to obtain the set values.

I am not clear about the " First do a VCO calibration and then read back the values to obtain the set values." 

My understanding about the configuration sequence of full assist as this: VCO calibration---Set register value----VCO_SEL_FORCE, VCO_CAPCTRL_FORCE=1-----------read back set value

But PLL couldn't lock.

So could you please tell me the right configuration sequence about the full assist mode?

Best regards

Kailyn 

  • 0
    TI__Mastermind 28540 points

    Hi Kailyn,

    What this is saying is don't use the full assist mode and lock the device.  Then readback the values to see what was set by the auto calibration, then you can use the force mode with the values you've recorded.

    73,
    Timothy

  • 0 in reply to Timothy T
    Prodigy 80 points

    The operation is as follows

    (1) In the automatic locking state, read these three values:VCO_SEL,VCO_CAPCTRL,VCO_DACISET;

    (2) Configuration sequence R0--R75--R62--R45--R44--R43--R42--R39--R38--R37--R36--R20(VCO_SEL & VCO_SEL_FORCE)--R19(VCO_CAPCTRL)--R16(VCO_DACISET)--R8(VCO_CAPCTRL_FORCE & VCO_DACISET_FORCE)  --R0,then device can be locked,however the total locked time maybe 140us;

    (3) Configuration sequence R0--R75--R62--R45--R44--R43--R42--R39--R38--R37--R36--R20(VCO_SEL & VCO_SEL_FORCE)--R19(VCO_CAPCTRL)--R16(VCO_DACISET)--R8(VCO_CAPCTRL_FORCE & VCO_DACISET_FORCE)  then the device can not be locked;

    my question is :

    What is the correct configuration order for the full assist mode of the device ？

  • 0 in reply to gxian l
    TI__Guru* 81740 points

    Hi Gxian,

    Were you enabling double buffering in R62? If this is the case, all the double buffered registers will be freeze until R0 is programmed. In (3), there is no R0 at the end of the sequence, this is probably the reason for unlock. 

  • 0 in reply to Noel Fung
    Prodigy 80 points

    R62 Value：24'h3E00AF（default）

  • 0 in reply to Noel Fung
    Prodigy 80 points

    It is described in datasheet of lmx2572 as follows ：“The fast calibration algorithm allows the user to change frequencies faster than 20 µs.”  My application requires that the locking time of the chip must be less than 50us.since,I want to know how the quick calibration operation is carried out.It's best to give me the recommended configuration sequence for registers of lmx2572.

    The configuration value of R62 is 24'h3e00af（default）

  • 0 in reply to gxian l
    TI__Guru* 81740 points

    Hi Gxian,

    For details of VCO calibration and full assist, please read SNAA336.

    When switching frequency, we should program those registers that really need to refresh. For example, if PLL_DEN is not changing, then we don't need to program it again. 

  • 0 in reply to Noel Fung
    Prodigy 80 points

    According to P16 in the manual of SNAA336, the device is locked, but the locking time is still very long, about 240us.Frequency switching is from 1700MHz to 1800MHz or from 1800 to 1700MHz. Please tell me what the problem is.

    Thank you!

  • 0 in reply to gxian l
    TI__Guru* 81740 points

    Hi Gxian,

    Below is the programming sequence of full assist.

    Assuming your fpd = 200MHz, switching between 1700MHz and 1800MHz, PLL_N will need to change from 51 to 54. So to switch the frequency, 

    1. program PLL_N (=54)

    2. program VCO_SEL

    3. program VCO_CAPCTRL

    4. program VCO_DACISET

  • 0 in reply to Noel Fung
    Prodigy 80 points

    our application as follows：

    fpd =  38.4Mhz，

    1700MHz parameter：

    CH_div = 2

    PLL_N = 88,

    PLL_NUM = 2326440618，

    PLL_DEN = 4294967295

    1800MHz parameter:

    CH_div = 2

    PLL_N =93,

    PLL_NUM = 3221225471，

    PLL_DEN = 4294967295

    I configure it according to the above figure.The device can be locked in full assist mode,but locked time is very too long.

    Please help me analyze the existing problems.

  • 0 in reply to gxian l
    TI__Guru* 81740 points

    Hi gxian,

    suggested programming sequence as follows:

    1. program PLL_N

    2. program PLL_NUM

    3. program VCO parameters

    What is your loop filter bandwidth and phase margin? Full assist will bypass VCO calibration, but the loop filter characteristic will remain affect the lock time. 

  • 0 in reply to Noel Fung
    Prodigy 80 points

    THank you very much!  According to your suggestion, I adjusted the configuration order, and the locking time is still very long. Refer to the following figure for my experimental data.Please help me analyze the problem.

    configration suquence：

    counter @122.88MHz

  • 0 in reply to gxian l
    TI__Guru* 81740 points

    Hi gxian,

    were you using the MUXout pin to determine lock time? what was register LD_DLY setting?

  • 0 in reply to Noel Fung
    Prodigy 80 points

     I  use the MUXout pin to determine lock time.LD_DLY:3E8

  • 0 in reply to gxian l
    TI__Guru* 81740 points

    Hi Gxian,

    Could you make LD_DLY = 0?

  • 0 in reply to Noel Fung
    Prodigy 80 points

    Think you very much.The problem has been solved,then this case can be closed.