Why is the Labview THD measurement higher than our Matlab THD measurement?

[TCPlomp]

Hi Solitude,

could you show the code you used (including LV version) and some data you have.

Have you looked at the help for the functions both MATLAB and LV?
Have you tested with 'real' signals or generated signals?
If you have tested with real signals could you post (a bit) of them here?

Ton

[solitude]

Hi Ton,

The waveform is a sweep frequency.  We are using a standard formula to compute the total harmonic distortion in our Matlab application.  We have some flexibility in the extracting the bins (energy) for computing the THD,  whereas in the Labview application we depend on its formulas.  The THD is approximately 0.3 percent in Matlab, and 5.4 percent in Labview for the same sweep frequency.  Does the Labview require additional VIs to overcome the errors in computing a THD for a sweep frequency?  Is the THD computation in Labview strictly designed for a single frequency waveform?

Solitude

[TCPlomp]

It's been a while that I used that VI, but I believe you open it to see what code is executed. And I believe it is targeted at a single frequency!
Have you tried using Mathscript in LV (8 and up)?

Ton

[solitude]

Hi Ton,

We have had some discussion on the THD.  Our application is built from Labview 6, and we have looked into the VIs, but some of those VIs are password protected.  I believe the THD does require a single tone frequency and we are trying to implement it using a sweep frequency.  The Matlab is our own formulas and so we can work with the math and produce better THD results.  I have attached a document from long ago that describes the THD in Labview if the frequency varies.

Wayne

[johnsold]

Wayne,

Post the formulas you are using in Matlab. Someone can probably suggest a way to implement them in LV. If the same formulas are used both places the results should be the same.

Lynn

[solitude]

Lynn,

Thanks for your reply.

Another engineer has taken a look at the formulas for both the Labview implementation and the Matlab.  There is an understanding that the THD is difficult to measure for a frequency sweep.  It turns out that in the Matlab formula is a bit more custom and so produces different results than the Labview.

Wayne