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.

DAC5682z behavior query

Other Parts Discussed in Thread: DAC5682Z

Hi All,

I am trying to use DAC5682z EVM board with a Stratix II GX eval board using HSMC connectors to interface them. I am generating a simple saw-tooth pattern from the FPGA. The FPGA clock (DCLK half rate clock) and hence the data rate to the DAC board is set to 400 MHz. I am using a low frequency signal generator to provide the reference clock to the PLL in the DAC board to generate a (full rate) sampling clock at 800 MHz. I use the following code to generate the saw-tooth wave.

always @ (posedge DCLK)

DAC_Data_Reg <= DAC_Data_Reg + step_size;

Where DAC_Data_Reg and step_size are both 16 bit wide.

The problem:

The DAC works fine for small values of step_size (upto say 16'h0080). If I increase the step_size from (16'h0001) to say (16'h0200), the recovered output is highly distorted, the lower edge (starting point) of the saw-tooth is rounded up instead of being a sharp corner. Moreover, the transition from 0xFFFF to 0x0000 is not smooth even at small step sizes. The output shows distortion as the code changes from 0xFFFF to 0x0000.

Q1. I want to ask if there is some restriction on the maximum step change at the DAC digital inputs e.g. is there a limit that the output be changed less than (say) 10 mV  per data update for proper operation?

Q2. Since changing the step_size also changes the frequency of the applied SAW tooth pattern so is there a limit on the frequency of the applied pattern?

Q2. Is there a setup and hold requirement for the DAC when using the DLL to adjust the data to clock skew? If yes, then please mention the specification in reply.

Regards,

Abdul Rehman Javed.

  • Abdul Rehman

    Are you looking at transformer coupled output on the DAC EVM? In that case, the saw tooth might be getting distorted by the transformer. Can you try with single sinusoid first to verify if everything in your setup is working as expected. The only limitation that the DAC imposes on the input data rate is that it has to be less than 0.4*fDACCLK/INTERPOLATION if any interpolation is being used (i.e. maximum frequency component in the input data has to be less than 0.4*fDACCLK/INTERPOLATION). If no interpolation is being used, then input data rate is limited by Nyquist only.

    From your description, my understanding is that you are running the DAC at 800 MSPS. DCLK from FPGA to DAC is 400 MSPS and you are not using any interpolation. You have got to be using the DAC in single channel mode to make the DAC accept your ramp data for a single channel only. Can you send me your DAC and CDC (clocking chip on the DAC EVM) settings so that I can verify?

    I am assuming you used the PLL in the CDC chip to generate the DAC CLK. Have you tried feeding in a clock of 800 MHz directly from the signal source to the DAC EVM? Or use the PLL in the DAC to multiply the slower clock to 800 MHz?

    DAC5682z expects the DATA and DCLK to be edge aligned and the datasheet specifies the maximum tolerable limits on skew between DATA and DCLK on page 9. For DCLK of 400 MHz, the maximum tolerable skew between DCLK and any of DATA bits is [-600,300] ps.

    Regards