# THD of sinewave

Discussion in 'Electronic Design' started by Jamie Morken, Feb 26, 2008.

2. ### John O'FlahertyGuest

With all that noise, it's hard to see how much actual harmonic
distortion there is. It looks as if it could be 7% THD plus noise,
though.

3. ### Guest

John is right. Subtraction of the fundamental results in distortion
plus noise. To measure THD, you must extract (via sharp bandpass
filters) each harmonic component. You then take the square root of
the sum of squares of the harmonics to get the THD. Audio analyzers
automate this process, but they are very expensiv
Regards,
Jon

4. ### Martin GriffithGuest

On Tue, 26 Feb 2008 08:23:49 -0800 (PST), in sci.electronics.design
This might help
http://audio.rightmark.org/products/rmaa.shtml

havent used it in years, but it might prove useful

martin

6. ### EeyoreGuest

There a number of free programs that will calculate audio THD using a sound
card.

Graham

7. ### EeyoreGuest

Remove the noise and I doubt it'll be over a couple of percent.

Graham

9. ### john jardineGuest

Yep. Same thought here.

10. ### WimpieGuest

Hello Jamie,

Assuming that the frequency spectrum of the noisy deviation from a
sine wave is in the band of interest, this one can have (N+D)/S = 0.07
(7%), but I expected less then 7%, because it looks like a reasonable
sine wave. Did you square the components before adding, as THD is
defined as a power ratio.

If it was a pure sinewave with additive noise (within band of
interest), I expected N/S to be better (< 0.004, power ratio).

Best regards,

Wim
PA3DJS
www.tetech.nl (Dutch)

11. ### WimpieGuest

Hello Jamie,

Assuming that the frequency spectrum of the noise deviation from a
sine wave is in the band of interest, this one can have (N+D)/S = 0.07
(7%), but I expected less then 7%, because it looks like a reasonable
sine wave. Did you square the components before adding, as THD is
defined as a power ratio.

If it was a pure sine wave with additive noise (within band of
interest), I expected N/S to be better (< 0.004, power ratio).

Best regards,

Wim
PA3DJS
www.tetech.nl (Dutch)

12. ### Tom BruhnsGuest

Given that you have captured the waveform as a .jpg, and that you've
subtracted the fundamental, I assume you have the data in digital
form, or it would be reasonably easy to get it into that form. Then
just run an FFT on it, and see where the energy is spectrally (versus
frequency). It would be best if the block you FFT contains an
integral number of cycles of the fundamental; else be careful to
window it properly. All the tools you need to do the windowing, FFT,
converting to dB, and displaying graphically, are contained in the