How to tune a loop filter for ADF4150HV operated with HMC739

Hi, dear

I used ADIsimPLL to help designing a loop filter, but the output signal is unsatisfying.

The reference clock frequency is 25MHz, and the PDF is 5MHz for which R=5. 

My desired output is 22.4GHz, so here is the spectrum result.

Maybe there are some wrong usage when I using ADIsimPLL, but I've tried several combination. Does anyone encountered the same problem?

I appreciate if someone can help me with my puzzle. Thanks a lot!

Parents
  • 0
    •  Analog Employees 
    on Feb 2, 2021 2:34 PM

    Hi,

    You appear to be using the HMC739 out of it's specification. This VCO is only specified to operate between 23.8GHz to 26.8GHz. Can you send attached your ADISimPLL file if you have it?

  • Hi, sorry. I made a mistake. The desired frequency is 24 GHz (22.4 GHz is operated by HMC738 in previous design). Actually, I have tried so many configurations. I fixed the phase margin to around 45 degrees, and varied the loop bandwidth. The loop filter seems to malfunction, in which the output frequency is with a 3dB-bandwidth around ±10 MHz centered at 24GHz. The attached is one of the configuration I used.

    Output frequency := 1.5 GHz*16 = 24 GHz

    PDF := 5 MHz

    Reference freq := 25 MHz

    Loop BW := 20.3 kHz

    Phase Margin := 45.1 degrees

    Thank you for your kind help. I am sorry to forget to save the spectrum analyzer's result. But the phenomenon is that the output is far from -93dBc/Hz claimed in the datasheet. I guess the key is the design of loop filter.

    ADF4150HV_HMC739.zip

  • Sir, I feel so grateful for your patience on my question.

    Do you mean that the cutoff frequency is 10MHz?

    Yes, sir. The spectrum is not a single frequency but with a range. The bandwidth of the RF output is around 20 MHz, centered at my desired frequency which is 24 GHz.

    This shows -102.75 dBc/Hz with this configuration. 

    The simulation results in ADIsimPLL are always good enough. But when I applied the schematic calculated by ADIsimPLL on my real board, the phase noise is not as good as it predicts, e.g., a 20 MHz bandwidth RFout can be detected instead of a single frequency. So, I wonder if I erroneously used ADIsimPLL, or I need to check my board.

  • I am sorry I don't have the 24 GHz output result by HMC739+ADF4150HV on hand, but the result is very similar to the spectrum in the following figure. A nearly single frequency RF output can be hardly achieved with the schematic calculated by ADIsimPLL.

    My desired output is 22.4GHz, so here is the spectrum result.

    I also tried to fine-tune the resistors/capacitors around the recommended resistance/capacitance, but the improvement is subtle.

  • 0
    •  Analog Employees 
    on Feb 4, 2021 10:55 AM in reply to realmadridman

    Hi, are you using our eval boards or is this a custom board? Could you send your list of writes how you're programming ADF4150HV.

    Are you using 5.1k Rset on your board?

  • Thank you, sir.

    I am using a custom board. So far, the RF frequency can be stabilized, but the output bandwidth is too broad that can be up to 20 MHz.

    And, yes, the Rset is 5.1k ohm, and the registers value is given in the enclosed Excel file. 

    ADF4150HV_registers.zip


    I am not sure if the table is comprehensible.

    Thank you for sparing your precious time answering my questions.

  • +1
    •  Analog Employees 
    on Mar 31, 2021 10:01 AM in reply to realmadridman

    Hi,

    I've had a look through the settings that you have sent. From what I can see, R counter has been set incorrectly. You have set Reference Divider = 1 (register 2 bit 24) and R counter = 10 in your configuration. This would result in Fpfd = 1.25MHz for a Reference = 25MHz. This would then mean that the loop filter components selected would not be valid for this Fpfd. 

    To ensure you have the correct values set for your configuration, you can download the applications software from the ADF4150HV product page. This will allow you to obtain the calculated values to program. Link to software here.

Reply
  • +1
    •  Analog Employees 
    on Mar 31, 2021 10:01 AM in reply to realmadridman

    Hi,

    I've had a look through the settings that you have sent. From what I can see, R counter has been set incorrectly. You have set Reference Divider = 1 (register 2 bit 24) and R counter = 10 in your configuration. This would result in Fpfd = 1.25MHz for a Reference = 25MHz. This would then mean that the loop filter components selected would not be valid for this Fpfd. 

    To ensure you have the correct values set for your configuration, you can download the applications software from the ADF4150HV product page. This will allow you to obtain the calculated values to program. Link to software here.

Children
No Data