• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 46 subscribers
  • Views 631 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.

ACPR decrease at DAC5688

Johnny Kim
Intellectual 330 points

Other Parts Discussed in Thread: DAC5688, TRF370333

Hi.

 

I got ACPR decreasing problem using DAC5688.

Fdata=122.88MHz

NCO=15.36MHz/138.32MHz

FDAC=491.52MHz

PLL off

X4 interpolation

All data is measured data at moduler TRF370333 ouput

As captured figure in attached file, when NCO set as 138.32MHz, noise level is increasing more than 15.36MHz abnormally.

I can find just 5dBc decreasing at 138.32MHz compare with 15.36MHz.

pls, advise me to figure out this problem.

6472.DAC5688.zip

  • 0
    TI__Guru* 80395 points

    Hi Jonny,

    You may see a some decrease in ACPR performance as the output frequency becomes higher. As you have mentioned, the ACPR performance should not differ by 17dB with 184.32MHz.

    I noticed the output power of the 184.32MHz is higher than 15.36MHz. Was the digital gain between the two frequencies different? 

    Typically, the output power would decrease with frequency since the DAC output has the sinx/x response that rolls off with respect to frequency. You may want to double check the gain setting or even the inverse sinx/x response filter to see if it is enabled. If the signal is saturated digitally, then the ACPR performance will degrade.

    Please also send me the registers files or schematic (if you can) so I may double check your settings. 

    -KH

  • 0 in reply to Kang Hsia
    Intellectual 330 points

    Hi Kang.

    Thanks for your supporting.

     


     

    Attached Schematic & register file.

    Enabled QMC & SINC filter.

    QMC gian =1 / offset = 0

     

    Fullscreen 7220.DAC5688 Setting.txt Download 
      addr data
 ------------
  0x01 0x1E 
  0x02 0xEF 
  0x03 0x00 
  0x04 0x00 
  0x05 0x82 
  0x06 0x00 
  0x07 0x00 
  0x08 0x00 
  0x09 0x00 
  0x0A 0x00 
  0x0B 0x60 
  0x0C 0x00 
  0x0D 0x00 
  0x0E 0x00 
  0x0F 0x24 
  0x10 0x00 
  0x11 0x00 
  0x12 0x00 
  0x13 0x00 
  0x14 0x00 
  0x15 0x00 
  0x16 0x55 
  0x17 0x08 
  0x18 0x80 
  0x19 0x00 
  0x1A 0x8E 
  0x1B 0xFF 
  0x1C 0x00 
  0x1D 0x00 
  0x1E 0x80

    3386.DAC5688_Schematic.zip

     


     

  • 0 in reply to Johnny Kim
    TI__Guru* 80395 points

    Johnny,

    The DAC register settings and schematic look correct to me. The decrease in ACPR performance with increasing in IF is most like due to the inverse sinc filter being enabled and the data was not backed-off correctly. The inverse sinc filter at 180MHz IF can have 2dB of digital gain, and your data must be backed-off by at least 2dB in order to not have digital saturation.

    Also, the NCO gain is set to enabled. The I/Q signal may also be saturated with full-scale input. For detail, please refer to the DAC5688 datasheet fine-mixer section. 

    -KH

  • 0 in reply to Kang Hsia
    Intellectual 330 points

    Hi Kang.

    I have tested ACPR again with disable SINC filter and mixer gain, then ouput signal and noise level are reduced , but ACPR performance is same, not increasing, because decreasing level of signal & noise are same, so CDMA ACPR spec. doesn't meet.

    refer to the attached captured spectrum with disabled SINC filter & mixer gain, and another spectrum is with 4dB increading digital at source.

    Have any other solution?

    Additionaly I wonder why noise level increasing at 180MHz IF, I can see good performance at 15.36MHz IF. is it NCO phase noise performance ?

     

    2526.ACPR_Capture.zip

  • 0 in reply to Johnny Kim
    TI__Guru* 80395 points

    Johnny,

     

    Apologies for the late reply. For some reason the initial post did not go through the E2E system.

     

    The first ACPR capture that you had sent shows the power at 184.32MHz IF is higher than the 15.36MHz IF. This could happen if you are sending a full-scale digital CMDA waveform to the DAC and enabled the internal inverse sinc filter and mixer gain. In this case, you must back-off the data by the amount of total digital gain inside the DAC to avoid digital saturation. For 491.52MHz Fdac, 184.32MHz IF, you will need to back-off at least 2dB to avoid digital saturation. Could you please confirm if your system had digital saturation before? I need to understand this before further debugging.

     

    For the later ACPR captures that you had sent, is the blue line 15.36MHz IF output?

     

    To proceed our troubleshooting:

    1. use the same digital output signal level from your source, only adjust the NCO to 15.36MHz and 184.32MHz to see the ACPR performance. Do *not* enable inverse sinc filters and NCO mixer gain. Compare the result between the two cases. The decrease in ACPR at higher output frequency is expected. For example, Figure 10 of the DAC5688 datasheet shows about 6dB decrease in ACPR from 15.36MHz to 184.32MHz.
    2. Since you are using Agilent spectrum analyzer, it is possible that you can measure ACPR with noise correction feature enabled. The noise correction can calibrate out the internal analyzer noise and display the true ACPR of your system output. I believe you can go to “measurement” > “ACP” > noise correction
    3. You may want to check your analog low pass filter to see if the gain flatness is the right design. I have checked your AC response in TINA simulation and it doesn’t seem to have much ripple at 184.32MHz. You may want to double check by sweeping the power (using fixed digital gain) across the IF to check for ripple. Experiment #1 result may help as well. I have also attached TI TINA file for the filter network in case you need it. You can use it to simulate the AC response. 
    -KH
    Current_Sink_DAC_to_3p3VMod_DC_5Vsupply.TSC