Different results in harmonic analisys

[GoceJ]

I am analysing data in the same time with the:
1.Harmonic Analyser VI and
2.RealFFT VI-(for getting the 1_sided FFT, than graphing that data to see the harmonics and afterwards I am using Peak Detector VI).
The frequency of the main harmonics and the peaks for them from the other VI are pretty different, and so are the peaks and amplitudes. Which method is more precise. Thank you.

[<Mike Ross> mross@smpcorp.com]

My guess is they both do precisely what they are supposed to. But your question was, "Which method is more precise?"

Can you generate a calibrated wavform? If you had a function generator you could mess around with that and compare the results.

There are lots of tools for comparing a couple of datasets. But you need to have exactly the same inputs to compare the results.

You can simulate an input with an array of data. I just used the Basic Harmonic Analyzer Example VI to create an array of data for a sinusoidal signal. You could do likewise and write that out as a spreadsheet file and use it to test and compare other VI's.

I have a VI that compares many of the Filter VI's. I save a dataset and pass it through my favor
ite ones and view them overlayed on top of the raw data. This tells me how much they are distorting the signal. There are various coefficients that can be tweaked to get different results. I guess I can attach that so you can look at it.

I couldn't find any VI's in the examples for LV6.1 with the names you gave. Where exactly did your VI's come from?

Mike

[GoceJ]

Hi, thank you for your repply. I have LabVIEW 5.1 the VI's are located in Functions/Signal Processing/Measurement. As you can see from the atachment I analyse the same data twice: once directly with the Harmonic Analyser(THD) and the second time with Real FFT and afterwards with the Peak Detector.
They both give me different results, maybe the best way is to use device for analysing the harmonics and see which of the results are better. Thank you. Bye.

[<Mike Ross> mross@smpcorp.com]

Hi,

With regard to harmonics, I see close correlation between the FFT graph and the Harmonics Analyzer array output. Do you wish for closer correlation?

I came to this conclusion by using the zoom functions and changing the y-scale of the graph. If you set the y-scale to a maximum of 1.5 then you will see the are small peaks in the FFT at very nearly the same freqencies as shown in the array returned by the Harmonics Analyzer VI.

The Harmonics Analyzer documentation is pretty good. The question to ask is: Do you like what it does? Those results seem to make sense. They are confirmed by the FFT (as they are the result of FFT). They are easier to read than the FFT.

The results of the Peak Detector are not so good
it seems. You are operating on the output of the FFT before passing the data to the Peak Detector. Are you sure you have that part right?

I should stop trying to advise you at this point. It is clearly outside my areas of expertise. I barely understand FFT, and work with it not at all. I don't have or use LV5.1 and I am missing some components to run the VIs.

Bon chance,

Mike

[GoceJ]

Thank you for your time,
best regards, Goce

[Primson]

We use TCI filters. They have been very reliable & the software works like a charm. I think they were called "Trans-Coil" before.

http://www.clrwtr.com/TCI-Trans-Coil.html