• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 20 subscribers
  • Views 455 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.

LMX2492: Modulation before locking on Reference Frequency Change

user5926428
Prodigy 80 points
Part Number: LMX2492

Dear all,

I am using a LMX2492 EVM board to test the performance of the PLL on my application. I am using an external reference clock instead of the provided clock to lock the PLL and an external VCO. I have followed the EVM reference manual instructions for connecting an external reference and care has been taken to disconnect every power supply for the onboard VCO and the Vtune input and RFout. The RFout SMA has been directly connected to the PLL Fin pin. I have mounted an active 2nd order type B filter on the board and the output of the filter is connected to the Vtune pin of an external VCO through a wire. The parameters for the filter and VCO are as follow:

Loop Filter: Type B 2nd Order (1.9 MHz BW, 40 deg Phase Margin) [Component Topological names according to PLLatinum SIM]
OpAmp: AD8065 -> Vs+ = 18V, Vs- = GND
C1 = 27 pF
C2 = 0.18 nF
C3 = 1 nF
R2 = 1.2 kOhm
R3 = 15 Ohm

VCO:
K = 280 MHz/V
Cin = 15pF

I have programmed the PLL N value to 60. I am changing the external reference frequency (generated by a high end bench RF synthesizer with a very clean, low spurious output) between 160 and 200 MHz to make the PLL lock to different frequencies (9.6 GHz to 12 GHz) and it is working like a charm except for a few specific frequencies on the Ref pin. These frequencies are 196, 197, 198 and 200 MHz. What happens in these frequencies is that the PLL makes a wideband not uniform FM modulation and after a few seconds (the exact time is not constant between tries) it finally locks on to the frequency it should be locked on to. In the case of 197 MHz specifically (11.82 GHz) I cannot get it to lock, it just outputs an FM modulation. According to basic PLL theory, I don't understand the mechanism behind this behaviour, as it is not an oscillating, unstable response, but a temporary unstability followed by correct locking, and it also just depends on the PFD frequency and not its amplitude or any other factor that may affect its response.

Also, this FM modulation and eventual locking may or may not happen (and if it happens it takes an arbitrary amount of time for the system to eventually lock) on any of the working frequencies for Ref IN if I switch off and on the Ref IN frequency.

I hope you can help me find the cause for this strange behaviour.

Kind Regards,

Alejandro

  • 0
    TI__Guru* 81230 points

    Hi Alejandro,

    So the fpd is also varying between 160 and 200MHz? If this is the case, you have problem when fpd gets higher. You can try reducing the fpd by a half with the R-counter. If the test result is positive, we may conclude the problem is due to fpd being too high. 

    When FM modulation appear, can you probe the Vtune pin to see if it is still a flat line? Is the Vtune closed to the supply voltage rail?

    Will the problem go away if you set FRAC_ORDER to "integer mode"?

  • 0 in reply to Noel Fung
    Prodigy 80 points

    Dear Noel,

    The FPD is varying between 160 and 200 MHz and it should be fine as it is stated in the datasheet, actually, I have tried 201 and 202 MHz and the PLL has locked just fine, even if it is above specifications.

    I have been doing some more testing and I have actually achieved locking with no FM modulation by slightly disconnecting the SMA that connects my VCO to the feedback path into the PLL! The VCO is connected through a connectorized coupler back into the PLL board (through the RF_out connector). The path from the RF_out connector back into the PLL Fin pin consists on three 0 ohm resistors in series and a 10 pF 0402 capacitor (the one that comes with the eval board). The RF power that goes into the PLL board is around 0 dBm.

    As slightly disengaging the SMA connector from the coupler back into the PLL board seems to help with correct locking and no FM modulation, I measured the S11 parameter on the Evaluation Board RF_out connector (as it is right now, with just 0 ohm resistors in series and a 10 pF 0402 NP0 capacitor and the adaptation is pretty bad in my opinion: -5 dB S11 around 12 GHz. This arises some more questions: What is the expected adaptation on the Fin pin for the LMX2492? Is the 10 pF capacitor the cause for such a bad adaptation in this port? Can this be the cause for the sudden FM modulation when the SMA connector is correctly engaged?

    I have also measured my coupler and it is perfectly fine and I have reproduced the experiment of engaging and slightly disconnecting the SMA connector and it seems like it is the cause for modulation before locking (i.e. I see a wideband modulation, I slightly touch the SMA connector and it locks instantly after doing so).

  • 0 in reply to user5926428
    TI__Guru* 81230 points

    Hi Alejandro,

    I guess there might be a matching issue between the VCO and Fin, however, I cannot explain why the mismatch will create FM modulation.

    Can you try put a 50Ω shunt between VCO and Fin?

  • 0 in reply to Noel Fung
    Prodigy 80 points

    Hi again Noel,

    Sorry for the late response. I tried to use a shunt between the VCO and Fin but it didn't seem to help so much. However, I tried to use a 3 dB attenuator between the VCO and Fin and that fixed the problem completely! My theory is that the impedance mismatch might be causing a standing wave in the Fin port that is causing sensitivity issues in this port and might be causing loss of lock for certain frequencies. The attenuator surely helps with the matching and reducing the SW in the line.

    Kind Regards,