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: PLL unlocked for several frequencies for FCAL_FAST=1

Part Number: LMX2592

hallo,

i am testing the LMX2592 with the following conditions:

frefin=100MHz; Fpdf=6.25MHz, Fout = 3.0GHz ... 5.0GHz with ACAL_FAST=1 and FCAL_FAST=1; Auto calibration (no VCO start settings and no manual Amp.settings)

The PLL works fine with an aprox. Locktime of 600us. But near Fout of 4.4GHz (aprox. 4.3GHz...4.5GHz) there are several frequency (band) settings, which can not be locked (LD-Out do not goes high) and no output signal can be seen on RFoutA.

Over and below this band the PLL works well. I have tried out several settings.

With FCAL_FAST=0 this frequencies are locked but the Locktime increases 2 to 4 times than in FAST-Mode.

In the datasheet i can not find any settings corresponding to this problem.

Do you have some idea to lock all the frequencies in FAST-Mode?

best regards

Eddi

  • Hi Eddi,

    Can you send me the saved .tcs file of your current setup? Get from File-->Save in TICS Pro.

    Regards,

    Brian Wang

  • LMX2592_01_170515.7z hallo Brian,

    here is the tics-file (please rename the LMX2592_01_170515.zip extension to tcs). The Fpdf and the corresponding settings are changed to 2MHz in the TCS-File. In the first post i tested with Fpdf=6,25MHz, but the problem looks like equal.

    regards

  • Eddi,

    The files was not readable after renaming. maybe try sending a zip file with the tcs file within in.

    In the meantime can you check if AJUMP_SIZE is like below. previously it was 4 bits, but there was an update, the LSB should be 1.

  • hallo Brian,

    thanks for reply and the hint.

    Here another try.

    R64 has the Value: 0x40037F : according to DS LMX2592 Revision D: -> A_JUMP_SIZE=3;   FJUMP_SIZE = 15; F_CAL_FAST=1; ACAL_FAST=1

    Is this correct with the constant data Bit4=1?

    There seems to be a discrepance between TICSpro and the DS?

    But i used it according to DS.

    regards6237.LMX2592_Config_01_170513.zip

  • hallo Brian,

    i have checked the settings of R64. It is according to DS Rev.D and ok. The problem ist still there.
    Do you think i have to change the A_JUMP_SIZE ?
    Which values are stable?

    regards

  • hi Eddi,

    I tried your settings they are ok. R64[4]=1 is the key. Basically, AJUMP_SIZE used to be R64[7:4], which changed to R64[7:5] and locking R64[4] to 1, this means you are setting always odd jump values. Once you do this should lock every time. This will work for any AJUMP_SIZE value. What indicator do you use to determine unlock and what do you see on the specAn is there some tone anywhere?

    Regards,

    Brian Wang

  • hallo brian,

    thanks for reply.

    The indicator for unlocked state is at first the LD-Signal, which is measured by the uC. The uC measures the locktime. In the band with unlocked frequency settings the LD-Signal does not go high and with a spektum analyzer i see there is now Signal on the RFoutA. I measured the LD- and the VCO-Tune-Signal additional with a scope, which confim the unlocked state. But until now I have not analyzed the whole spectrum according to the range of the analyzer (Fmax= 6GHz).
    At last i have tested A_JUMP_SIZE=2 with FCAL_FAST=1 and the problem is changed. With this settings there is an unlocked state in a lower band (<2GHz) and afterwards there is no lock on every frequency until a Reinit sequence is done on the LMX2592.

    But for this new problem i have to make more tests to get further detail infos about it.

    regards

    Eddi

  • Eddi,

    Great to see you are using three methods to check for ultimate confirmation. I see now, with 6G above you cannot use the spectrum analyzer method. So it sounds like for a different set of settings you saw unlock at different frequencies? We have done sweeps before that covered all frequencies and all jump sizes with no problem. But if you let me know which specific settings (registers, OSCin, RFout frequency), I can check to see if anything abnormal going on. Also, it's a good thing you did to reinit after you see an unlock, always want to make sure you go back to a good fresh state.

  • hallo Brian,

    i have tried out to reproduces the first mailfunction with unlocked gaps around 4,3GHz. But i cannot see this problem again. With your hints i have corrected and improved the LMX-driver in the last weeks and the LMX2592 locks over the band without problems with FCAL_FAST=1.
    I think the problem is fixed and i take a look at it and the changes of my driver.
    Thank you for your support. This thread may be closed.

    best regards
    Eddi