Measurement Computing WBK Options - Manual (Page 124)

Maximizing Alias Protection
What is Aliasing?
Aliasing is a phenomenon of sampled data systems wherein a high frequency signal is misrepresented as a
low frequency signal when the A/D converter sampling rate being used is too slow. This misrepresentation
can result in severe data corruption and incorrect FFT results. Aliasing is a well-documented data
acquisition effect, and interested users are encouraged to research detailed information that is available on-
line from companies such as Analog Devices and Texas Instruments. This text aims to not supplant those
resources, but to provide most users with sufficient knowledge to avoid most alias problems through
proper filter and sampling rate configuration.
For a given sampling rate, F
S
, input signals of frequency up to F
S
/2 will be processed correctly. However,
input signals above F
S
/2 are subject to aliasing. For example, a sampling rate of 100 kHz can process
signals up to 50 kHz without aliasing. An input signal of 90 kHz, however, will be aliased. Specifically, it
will appear in the sampled data as a signal of frequency F
S
-F
IN
, which in this case is 100 kHz-90 kHz = 10
kHz.
Aliasing, and its prevention, should be a consideration in all sampled data systems. This is especially
important in mechanical vibration measurements, because most mechanical systems exhibit a resonance
apart from their fundamental frequency. That is, there may be signal energy present that has the potential
to be aliased that is unknown to the user. And the worst part of aliasing is that its effects are
indistinguishable from real input signals. That is, in the given example, it is not apparent to the user
whether the 10 kHz energy is real or an alias.
Aliasing Protection using the WBK18
F
IN
/F
C
Gain (dB)
<0.1 -0.15
0.1 -0.15
0.2 -0.18
0.3 -0.22
0.4 -0.27
0.5 -0.33
0.6 -0.39
0.7 -0.47
0.8 -0.68
0.9 -1.4
1.0 -3.6
1.25 -16
1.5 -28
1.75 -39
2.0 -48
2.5 -63
2.8 -70
3.0 -75
3.3 -80
3.5 -86
The WBK18 provides alias rejection via its 8-pole filter. This filter has an
extremely steep roll-off characteristic, very closely achieving an ideal “brick-
wall” response. It consequently passes frequencies of interest without
significant attenuation, but significantly attenuates frequencies just above. Its
attenuation is so high that most alias frequency energy is reduced to a level
below the noise floor of the measurement system. However, it must be
configured correctly to achieve these results. In general, the cutoff frequency
(F
C
) of the filter should be set as close to, but above, the highest input
frequency of interest. This will maximize the alias rejection it provides.
The 8-pole filter provides excellent filter response. However, no filter is
perfect, meaning that some signal attenuation occurs for frequencies just
below F
C
and maximum attenuation is not exhibited for frequencies just
above F
C
.
For reference, the typical response of the 8-pole filter is provided in the table
at the right and the graphs that follow.
4.0 -95
WBK18, Dynamic Signal Input Module 926896 WBK18, pg. 9