Table of Contents
Advertisement
The amplitude and frequency uncertainty of the FFT display depends on the choice of the window, and the
analyzer sweeptime. Amplitude uncertainty is maximum when the spectral component falls midway between the
filter shapes. Passbands that are flatter in shape, like the FLATTOP filter, contribute less amplitude uncertainty, but
frequency resolution and sensitivity are compromised (see TWNDOW)
Of the three algorithms, the FLATTOP has the least amplitude uncertainty and greatest frequency uncertainty.
Worst-case accuracy is - 0.1 dB. Use this passband when transforming periodic signals.
The UNIFORM algorithm has the least frequency uncertainty and greatest amplitude uncertainty. Worst-case
amplitude uncertainty is 3.9 dB and its 3 dB resolution bandwidth is 60% of the HANNING bandwidth. The
UNIFORM algorithm contains no time domain window weighting. Use it for transforming noise signals or tran-
sients that fully decay within one sweeptime period.
The HANNING algorithm is a traditional passband window found in most real time analyzers. It offers a compro-
mise between the FLATTOP and UNIFORM shapes. Its amplitude uncertainty is - 1.5 dB, and its 3 dB band-
width is 40% of the FLATTOP bandwidth.
The FFT results are displayed on the spectrum analyzer in logarithmic scale. For the X dimension, the frequency at
the left side of the graph is 0 Hz, and at the right side is Fmax. Fmax can be calculated using a few simple equations
and the sweeptime of the analyzer.
The sweeptime divided by the number of trace array elements containing amplitude information (in this case,
1000) is equal to the sampling period. The inverse of the sampling period is the sampling rate. The sampling rate
divided by two yields Fmax. For example, let the sweeptime of the analyzer be 20 msec. 20 msec divided by 1000
equals 20 psec, the value of the sampling period. The sampling rate is l/20 psec. Fmax equals l/20 pet divided
by 2, or 25 kHz.
The fourier transforms of the window functions are shown in the following figure. Use these graphs to estimate
resolution and amplitude uncertainty of a fourier transform display. Each horizontal division of the graphs equals
elements.
(Continued) FFT
Programming 1 1 1
Advertisement
Table of Contents
