Incorrect thd measurements using harmonicanalyser

[alun]

We are using a PCI4451 to generate and analyse sinewaves.
As a test and for calibration we have looped a channel input to channel output.

The output is also displayed on a scope and spectrum analyser.

A sinewave is generated with a specified amount of THD added.

This displays and measures correctly at the outpu of the card.

The input channel is sampled. The AC amplitude and frequency are correctly measured. The THD and noise measurements are not.

See zip file for program details.

Thanks
Alun

[Bryan F]

Hello Alun --

Thank you for your interest in DSA hardware and analysis! My name is Bryan Fry -- I am writing from the DSA group in Engineering at NI. I am fairly certain that the issue you are experiencing relates to "spectral leakage" if the tone you are generating is not precisely on-bin for the FFT analysis. This possibility is something that is inherent in the THD method used in the CVI analysis library.

The THD functions in LabVIEW (version 6.0 and higher) use a more advanced peak detection algorithm that is insensitive to whether or not the tones are on-bin. I'd suggest that you send me an email at bryan.fry@ni.com and we can discuss the best workaround for this issue.

Thanks!
Bryan

[rorOttawa]

If this is the case than why does the help for this function not tell you of this problem and the limitations of the function.

In fact if the function does not deliver
what it says it does, why advertise that it does
give you THD and THD+noise

Why has CVI not been upgraded to have the labview
function.