Microchip Technology AC164120 User Manual
Signal Analysis PICtail™ Daughter Board User’s Guide
DS51476A-page 18
2004 Microchip Technology Inc.
2.6
FFT (FAST FOURIER TRANSFORM)
The FFT displays the Fast Fourier Transformation of the data in the frequency domain.
This display is typically used to analyze AC signals or to help locate the source of AC
noise.
This display is typically used to analyze AC signals or to help locate the source of AC
noise.
FIGURE 2-11:
FFT DISPLAY
The size of the data set of the FFT analysis determines the resolution and accuracy of
the FFT. The accuracy is equal to:
the FFT. The accuracy is equal to:
Where:
n = number of converted bits
N = number of data points
N = number of data points
In instances where the input signal is not periodic with respect to the sampling fre-
quency of the converter (as is the case with this tool), windowing algorithms are advan-
tageous when looking at FFT results. The windows that are available with the Signal
Analysis Application are the Blackman, Hamming, Hanning and Rectangular. Basically,
an FFT window is multiplied times the measured signal data set taken during the spec-
ified length of time of the conversion. Typically this multiple reduces the magnitude of
the beginning and end of the sample and consequently minimizing discontinuities seen
with the FFT calculation.
quency of the converter (as is the case with this tool), windowing algorithms are advan-
tageous when looking at FFT results. The windows that are available with the Signal
Analysis Application are the Blackman, Hamming, Hanning and Rectangular. Basically,
an FFT window is multiplied times the measured signal data set taken during the spec-
ified length of time of the conversion. Typically this multiple reduces the magnitude of
the beginning and end of the sample and consequently minimizing discontinuities seen
with the FFT calculation.
Blackman – Window has a bell-shaped characteristic similar to the Blackman-Harris
window. The peak resolution of this window is not as fine as the Hanning, but the
responses flares out less and the rejection of the sidelobes is better.
window. The peak resolution of this window is not as fine as the Hanning, but the
responses flares out less and the rejection of the sidelobes is better.
Hamming – Window has a bell-shaped characteristic. The initial samples from the
conversion data are multiplied by a small number as are the last samples. With this
window, the side lobes adjacent to the main lobes are lower than the results from the
Hanning Window.
conversion data are multiplied by a small number as are the last samples. With this
window, the side lobes adjacent to the main lobes are lower than the results from the
Hanning Window.
Hanning – Window has a bell-shaped characteristic. The initial samples are multiplied
by zero as well a the latest samples. The samples between the beginning and end are
multiplied with the Hanning bell-shape curve. The side lobes of this window are farther
from main lobe as compared to the Hanning Windows. This window is typically used
for harmonic analysis of continuous time signals as well as random noise.
by zero as well a the latest samples. The samples between the beginning and end are
multiplied with the Hanning bell-shape curve. The side lobes of this window are farther
from main lobe as compared to the Hanning Windows. This window is typically used
for harmonic analysis of continuous time signals as well as random noise.
FFT Window Selection
FFT Accuracy = (4/(n
√N)dB)