• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 47 subscribers
  • Views 1041 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.

DAC 8831 Noise Floor.

Umar Khan
Intellectual 320 points
Other Parts Discussed in Thread: DAC8831, OPA735

hello,

            I am using the DAC 8831 and observed that the output is noisy. I searched around and learned that DACs have a noise floor. I was expecting that the noise would be above a certain frequency and I would just be able to filter that out, but it seems to be present at all frequencies. 

So I did an experiment to try to quantify this noise. I took the example code from Tom in SLA246, and set the two lines where the data is sent out as follows

...

MCBSP_write(hMcbsp0, 0x8000); // 5-6K McBSPb Header

...

MCBSP_write(hMcbsp1, 0x8000);// 5-6K McBSPa Header

so the output should be zero. Then I recorded the DAC output using a National Instruments DAQ at 1M samples per second. I got the following output and corresponding FFT

So the DAC seems to add a noise signal that is between +-10mV. Since I want to be able to generate signals in the 20kHz range and since the magnitude of variation in these signals is small, this noise is a problem.  

Is there any way around this? Low pass filtering can be tried but that would restrict the max frequency that can be generated by the DAC.

I am using this DAC with C6713 DSK + 5-9K interface board.

Best regards,

                           Umar Khan.

  • 0
    TI__Guru**** 180931 points

    Hi Umar,

    Which NI DAQ are you using here?  Can you also let us know if you modified the SLAA246 code or the DAC8831 EVM at all?

  • 0 in reply to Tom Hendrick
    Intellectual 320 points

    hi,

    I am using the NI USB 6356. I made no   modification to the EVM. It runs in bipolar mode. The only modification in the SLA246 code is the one that I mentioned,

    i.e, instead of  sending the sine and cosine tables I send 0x8000 so that the DAC output should become zero. 

    Umar

  • 0 in reply to Umar Khan
    TI__Mastermind 43915 points

    Hi Umar,

    A couple questions/comments -

    - Are you continuously writing to the device while looking at the output from the DAC? If so, you need to be concerned with digital feed-through causing, what looks like, noise on the output. You will want to halt communication when doing noise tests to avoid feed-through glitches.

    - Ref3025 may not be the cleanest reference when trying to reproduce data sheet noise specs. REF5025 with a 1uF on the trim pin, 10uF low ESR tant capacitor on the output with a RC filter would be a better fit.

    - The DAC8831 is an unbuffered DAC meaning that it consists of basically a variable resistor string. The noise you will be measuring from the DAC will be the resistor thermal noise which is code dependent. The OPA735 on the output is going to dominate the 1/f noise so keep that in mind as you are taking measurements. If you want the true DAC noise, you would need to look at the output of the DAC prior to the buffer amplifier.

    Regards,

    Tony CAlabria