Harmonic Content, FFT, DFT, and Fourier Analysis in EXCEL

Have you tried 'stretching' the input data to 64 points, using interpolation to spread the 36 values?

Nope, but I will try to strech the data.

I hope you mean that I create additional fictive points that are interpollated between the original measuring points ?

Isn't there a Discrete Fourrier Transform in Excel ?

"I hope you mean that I create additional fictive points that are interpollated between the original measuring points ?"

Yes. As a trivial example, here's a data set of 5 points:

... and the plot:

Here it is stretched to 8 points, using linear interpolation:

... and the plot:

Sorry, can't help with the discrete FT.

Have you tried adding half your zeros at the beginning of the signal train, and half at the end? The other option is to clip your data to 32 points, then run the FFT four more times, shifting the data window up one data point each time. Fortunately, the Excel FFT runs pretty fast, especially with the small amount of data you are dealing with. If you have good data, you should not see too much change in the transform as you shift the window...

Thirty six data points is a pretty short data record...I'm not sure your results are going to be valid with only 36 data points...

I haven't got a clue what you're talking about here: so I'm sort of probing.

"The measuring points are all values from a fundamental and its higher harmonics."

Do you mean the measurements were taken in the frequency domain rather than the time domain?

The purpose behind all this is to examine the shape (the form, the way it looks) of a global magnetic field.

The global magnetic field is created by a number of coils. Each coil can increase or decrease the magnetic field at the local point where this coil is situated. So, all the coils together can reform the shape of the global magnetic field.

The coils are mounted in slots I have exactly 36 slots, this gives me the 36 measuring points.

Each measuring point represents a local magnetic field. It looks like a bar graph where the length of each bar a representation is of the local magnetic field.

The total bar graph gives me a representation of the global magnetic field.

With Fourrier, I want to know the harmonic content of this bar graph.

So each measuring point represent a magnetic field at a particular distance.

Can you provide us with the 36 data points?

The FFT is a math function that operates on a time series, but what you have is a "distance series". Also, the samples need to be taken from a single trigger. I am not concluding that your attempt to use the FFT is irrelevant, only that it is outside of the strict definition of an FFT.

If I were setting up an experiement like you described, I would want to sample each sensor with an A/D taking more points (say 512), and all triggered at the same time (perhaps with an RF transmitted trigger?). Then I could take the FFT of all 36 locations, and easily compare the phase of any harmonic from one location to any other. There would be lots of other ways to analyze these data as well (Cepstrum, for instance).

Add many repeated sequences in a row to form a larger sample of say 1024. The chopping effect (window) will become less significant as you add points. Then use the FFT utility to find the harmonics spectrum.

This could be the solution for my problem.