AD9851 Ringing?

I am using an AD9851 with +5V supply in an FM signal generator application. The frequency control word is updated serially each 11.77 microseconds to provide the needed FM deviation. The udfcw pulse produces ringing on both rise and fall transitions. The signal is single ended and piped through the on-board 70MHz low pass filter. I have a cloned board but the filter appears to be identical to the filter on the 'clock' evaluation board. The output is terminated with 100 ohms. The scope probe is x10, 6.5pFd. Is this filter not suitable for this application or is there reason to suspect something else? Thanks.

  • I apologize if my first question seems silly. I should have furnished more background details. I have tried a number of measures to stop the ringing, which occurs at about 70 MHz, including: improving dc power bypassing, separating digital and analog lines, shielding and adding an EMI filter to the ac side of the dc power supply. All is to no avail. The only thing affecting the ringing to even a small degree is rearranging the grounding system including testing equipment.

    I modeled the filter with PSPICE and it shows ringing like I'm measuring is possible, but it requires a fairly strong dc pulse on the RF input to rise to the high level I observe. It seems like the ringing may be caused by induction or by circulating ground currents. Could it be that this filter designed for the clock application is not suitable for my application? Any suggestions on how to proceed to determine the cause/fix? I would appreciate your expertise. Thanks.

  • 0
    •  Analog Employees 
    on Dec 26, 2015 4:46 PM over 5 years ago

    Hi Jerry,

    I apologize for the late response. It might be that the filter you are using isn't good enough for the intended application or the filter's characteristic was changed when implemented into the board (refer to AN-837). For the former, I would suggest to run a simulation for your filter design. I know that you have already used PSPICE for your simulation but try a dedicated filter design and simulation tool as a reference. I'm aware that our boards have a built in filter in it but it was optimized for image rejection. There is a trade-off between frequency and time-domain responses.  In other words, depending on your application you have to compromise either one of the filter's performance. As basic, filter simulation tool have this step response calculation which is good for your application. A more advanced filter tool can do simulations for sinusoidal inputs of changing frequency.

    Best Regards


  • Hi Louijie,

    Thanks for your reply and the information on the filter application notes. I will give the filter design tool a go in the next few days to see what falls out.

    Best Regards 

  • Hi Louijie.

    I am not certain if I found the correct filter design tool. I found the website a bit tricky to navigate. I used a tool that designs active low pass filters. Nice tool, but only somewhat useful for my project.

    I am now seeing a slowly progressing failure of some component(s) of the evaluation board. The RF output is starting to roll-off around 20 MHz instead of 70 MHz and is down about 34 dB at 70 MHz. This is observed on the unfiltered output of the AD9851 as well, but not as severe - down about 25 dB. I am replacing the evaluation board and then I shall continue. Thanks for your assistance.

    Best Regards


  • 0
    •  Analog Employees 
    on Aug 2, 2018 3:17 PM over 2 years ago
    This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.

    Thank you,
    EZ Admin