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.

The problem with DAC5682Z @1GSPS, single channel output

Other Parts Discussed in Thread: DAC5682Z, CDCE62005, DAC5681Z, DAC5681, TPS54620

Now, I am using DAC5682Z as the follow figure show:

I want to generate signal between 175MHz to 425MHz, so I write a test program which generate a 425MHz signal to test the DAC5682Z.

 

I use CDCE62005 to generate 1GHz and 500MHz clock for DAC5682Z and FPGA.

In FPGA, I use 4 bits OSERDES to generate DDR data for the DAC5682Z.

 

And, I configure the DAC5682Z’s registers through SPI interface as follow:

 

1, Configure the CDCE62005 to generate the clocks for DAC5682Z and FPGA.

/*Provide stable clocks for DAC5682Z and FPGA */

 

2, Reset DAC5682Z through the RESETB pin

/* Set he DAC5682Z to initial status */

 

3, CONFIG8 0x07 = 0x04;

/* Set DLL restart */

 

4, CONFIG2 0x02 = 0x80

/*Set the DAC5682Z to single DAC mode*/

 

5, CONFIG1 0x01 = 0x00

/* Disable the filters */

 

6, CONFIG5 0x05 = 0x82

/*Set the serial interface in 4 pin mode

 Bypass the PLL */

 

7, CONFIG10 0x0a = 0xc0

/*when DCLKP/N is between 325MHz to 500MHz, the register should be 0xc0 through datasheet.*/

 

8, CONFIG8 0x08 = 0x00;

 

9, Read the register 0 to check if the DLL is locked.

 

 

I want to generate a 425MHz signal, but the generate signal is actual 375MHz? why?  

  • Hi Xiang,

    Your register settings seem to be okay. Does the spectrum for 375MHz look clean or are there other signals in the spectrum? What is the level of the 375 MHz signal?

    If the spectrum is clean, then it seems that you're sending the wrong data to the DAC. It will only output what you send it.

    If the spectrum has other spurs in it (other than DAC images) than there may be something else at fault. My first instinct is that you're not syncing the DAC correctly. Can you try using the SIF_SYNC instead? After verifying that the DLL is locked write 0x71 to CONFIG3 to choose the SIF_SYNC option, then write 0x73 to CONFIG3 to perform the SIF_SYNC.

    Regards,
    Matt Guibord 

  • Thank you for your help.

    Now, I use the software synchronization, and the DAC output the right signal. When I use DAC5682Z in two channels mode, I just always set the pin SYNCP/SYNCN “1” and the DAC5682Z can output the right signal. Why in single channel mode the software synchronization should be used?

    In my application, the output signal is between 175MHz and 425MHz, and the SFDR should be greater than 60dBm.

    There is no detailed introduction about IF output and SFDR in DAC5682Z’s datasheet, but I find it in DAC5681Z’s datasheet and DAC5681’s datasheet.

    If it is said in Figure 3 that, when the output signal is 175MHz, the SFDR is almost 58dBm, when the output signal is 425MHz, the SFDR is almost 50dBm.

    And, the actual test result by spectrum analyzer is showed as follow figures.

    The output signal is 425MHz.

    The output signal is 425MHz.

    The output signal is 300MHz.

     

    The output signal is 300MHz. 

    Does the DAC5682z work right?

    Thanks.

  • Hi Xiang,

    It's not that the software SYNC must be used in single-DAC mode, but rather that you are not properly syncing the DAC. For instance, you are always sending ones from the FPGA, but the DAC needs to see a rising edge on the SYNC pin. You should send all zeros to the DAC SYNC input until you are ready to synchronize and begin transmission. At that point, all ones should be sent, but only after the DAC has been programmed.

    The spurs around the fundamental are a little weird, and could possibly be coming from a switching power supply. Do you have a supply switching at 500 kHz?

    In the 300 MHz plot, the spur you have indicated is the third harmonic. Increase the attenuation on your sprectrum analyzer to verify that you are not over driving the input. You'll know if it's over driving if the second and third harmonics drop in power. You should continue to increase the attenuation until the harmonics no longer drop. It's a little hard to tell, but you are also likely limited by the third harmonic in the 425 MHz case.

    Regards,
    Matt Guibord 

  • Hi Matt Guibord,

    1. I think I have made a mistake about sync before. Either hardware sync or software sync should generate a rising edge.

    2. I have checked my board, the 3.3V power is generated by TPS54620, and the register between RT and GND is 100K, so the switching frequency is 500KHz. Now I change the device to TPS78601 (LDO), and test the output signal (300MHz) by spectrum analyzer as follow figure show.

                                                            The output signal is 300MHz

    The spur has improved, is the output signal is right?

    Thanks.

    Xiang song.

  • Hi Xiang,

    It seems like those spurs are coming directly from the power supply. Have you completely removed the TPS54620 from the system or is it still running? Do you have other switching power supplies?

    Regards,
    Matt Guibord 

  • Hi, Matt Guiboard

    Now, my power supply has change to as this:

    5V->TPS54620->3V3->TPS78601->1.8V

    5V->TPS54620->3V3->TPS78601->3.3V

    Thanks.

  • Hi Xiang,

    Was this the configuration in the last spectrum analyzer picture or is this what you're trying now?

    Both switchers could show up on the DAC output,

    Regards,
    Matt Guibord 

  • Hi Matt Guibord,

    Yes, this configuration is used for last spectrum analyzer picture.

    The spur @ 500KHz has just changed smaller.

    Does IT has some application document for design clock source and power supply for high speed DAC and ADC?

    Thanks,

    xiang song.