Forum Discussion

Altera_Forum's avatar
Altera_Forum
Icon for Honored Contributor rankHonored Contributor
10 years ago

How does dither value affect the sine wave output of Altera NCO IP Core

Good day!

I tried plotting a sine wave in Microsoft Excel by using the following equation in page 3-1 of NCO IP Core User Guide v14.1:

s(nT) = Asin[2π(fO+ fFM)nT+ ϕPM+ ϕDITH)]

I observed that varying the dither value (ϕDITH) and holding other parameters constant, shifts the sine wave right or left.

However I think this is NOT the correct effect of the dither value based on the following:

1. Description of Phase Dithering in page 3-6 of NCO IP Core User Guide v14.1

2. Actual observation in Altera NCO time and frequency domain graphs

Is my understanding correct?

What is the correct equation(with Phase Dithering) for plotting the sine wave generated by Altera NCO IP Core?

Thanks!

6 Replies

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    10 years ago

    did add a constant for dither, that will change phase by a constant value. you should add random noise.

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    10 years ago

    Hi kaz, thanks for the reply.

    "did add a constant for dither, that will change phase by a constant value."

    -> Yes, when I tried adding a constant for dither, the phase was changed by a constant value.

    But is this the expected effect of dithering?

    Based on page 3-6 of NCO IP Core User Guide v14.1, phase dithering provides improvement in SFDR by reducing noise localization.

    "you should add random noise"

    -> To where should I add noise?
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    10 years ago

    if you look at equation:

    s(nT) = Asin[2π(fO+ fFM)nT+ ϕPM+ ϕDITH)]

    then adding constant for dith makes it no different to the equation from PM. The idea of dith is random values(+/- some suitable range) added in above equation in place of dith.

    You actually don't need to do that to prove the concept, it is already proven but for own exercise dith is dith not a constant.
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    10 years ago

    Hi kaz, Thanks!

    I understand now that ϕDITH in the equation is noise of random value.

    What I don't understand is that in Altera NCO IP Core, when Dither Level is adjusted, the time domain graph is not changed but the frequency domain graph is affected. Can somebody explain how this happens?

    Thanks!
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    10 years ago

    both change. It depends on display graphics and normally frequency domain gives more visibility of what is in time domain