• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 46 subscribers
  • Views 509 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Original question:

DAC38J84: spurs issues

DAC39RF10EVM: spurs on specific frequencies

Yewei Li
Prodigy 20 points
Part Number: DAC39RF10EVM
Other Parts Discussed in Thread: LMX1204

Tool/software:

Hi TI,

    I am using DAC39RF10EVM and HSDC Pro to generate single-tone signals (DEM/Dither are both enabled), the sampling rate fs is 1GHz, the input frequency fin can range from 0 to fs/2.

    But I found some unexpected spurs on specific frequencies, they are usually equal-spaced in the output spectrum, like when I set fin = 60MHz and 70MHz, the output spectrum is:

    But when I set other input frequencies like fin = 110MHz, 160MHz, there are no such spurs :

    I want to know what is the cause of these spurs, and is there any suggestion on how to solve them? I also found similar questions in the forum, but it didn't say where the spurs came from: DAC38J84: spurs issues - Data converters forum - Data converters - TI E2E support forums.

    Below is my DAC39RF10EVM GUI configuration:

 

    Thanks, looking forward to your reply.

  • 0
    TI__Expert 3580 points

    Hello, 

    Are you feeding full scale data into the DAC? Its best to back off by 1dbFS or some over saturation in the DES2X interpolation filter can occur. As these spurs are very low it takes a very small amount of gain to cause them to show up.

    HSDC pro has a bug that does not properly digitally back off the data it generates. If you have a way to generate a CW tone I would recommend using that with 1dbFS back off. I have attached a 60Mhz and 70MHz CW normalized to 1GSPS backed off by 1dBFS for you to try. sine_60MHz_1GSPS_real_m1dbFS.csvsine_70MHz_1GSPS_real_m1dbFS.csv

    Regards, 

    Matt

  • 0 in reply to Matthew Kramer
    Prodigy 20 points

    Hi,

    Thanks, this did solve my problem, when I change the output mode to NRZ, these spurs also disappear. Besides this, I have another question about spurs.

    Now I set the sampling rate fs = 10.24GHz as the manual indicated, the output mode is NRZ, and there are 4 symmetrical spurs in the output spectrum, like when fin = 1GHz, 1.1GHz, the output spectrum is:

    Spur 4 and 7, 5 and 6 are symmetrical to fin, and they always have equal distance to it. When fin changes, the increment of spur 7 is 1.90464GHz - 1.80224GHz = 0.1024GHz = fs / 100. Some literature says that the spurs can be caused by the mixture of fin/harmonics and sampling clock, but I can't figure out how they are generated through mathematical expressions. Do you have any ideas on the cause of these spurs? Below is my GUI configuration:

    Thanks.

  • 0 in reply to Yewei Li
    TI__Expert 3580 points

    Hey Yewei, 

    I wonder if this is being generated from the LMX1204 on the board. This device generates the FPGA clock and can inject some spurs into the DACCLK that can appear on the output. 

    I put the device in DDS mode, (No FPGA required) and I do not see these spurs at all. Could you try resetting the DAC board (via reset button) and loading this configuration file? Just go to "Console" in the GUI and select "Load File". This will program the DAC to generate a 1G Sine wave in DDS mode (no FPGA or FPGA clocks required). If we can confirm the spurs are not there the next step will be modifying the DAC EVM to use an external reference clock. 

    Regards, 

    Matt


    EVM_write_log.cfg