Tektronix TDS5000B Series, Online Help Manual

Page: 293 / 420

293
transitions between the start and end
points cause discontinuities in the
waveform that introduce high-
frequency transients. These transients
add false frequency information to
the frequency domain record.
Applying a window function to the
waveform record changes the
waveform so that the start and stop
values are close to each other,
reducing the discontinuities. This
results in an FFT measurement that
more accurately reflects the actual
signal frequency components. The
shape of the window determines how
well it resolves frequency or
magnitude information.
FFT Window Characteristics
Each FFT window is a trade-off between frequency resolution and magnitude accuracy. What you
want to measure and your source signal characteristics help determine which window to use. Use
the following guidelines to select the best window.
FFT Window Characteristics Best for measuring
Rectangular Best frequency, worst
magnitude resolution. This
is essentially the same as no
window. Transients or bursts where the signal
levels before and after the event are
nearly equal.
Equal-amplitude sine waves with
frequencies that are very close.
Broadband random noise with a relatively
slow varying spectrum.
Hamming,
Hanning Better frequency, poorer
magnitude accuracy than
Rectangular. Hamming has
slightly better frequency
resolution than Hanning. Sine, periodic, and narrow-band random
noise.
Transients or bursts where the signal
levels before and after the event are
significantly different.
Blackman-Harris Best magnitude, worst
frequency resolution.
Predominantly single frequency signals
to look for higher order harmonics.
You can also determine the best window empirically by first selecting the Rectangular window,
and then selecting (in the following order) the Hamming, Hanning, and Blackman-Harris windows
until the frequency components merge. Use the window just prior to where the frequencies merge
for the best compromise between resolution and amplitude accuracy.