Somat eDAQlite, Руководство пользователя

Страница: 213 / 228

I2773-4.
4
en 213
Somat eDAQ
lite
15 Digital Filtering
The following sections provide information on the user-configurable digital filtering
available for ELHLS and ELBRG channels. For information on the analog filtering for
each layer, refer to layer data sheet.
15.1 Signal Aliasing
In the process of converting analog input signals to digital data representations, signal
aliasing can occur if the digital sample rate is too low compared to the frequency
content of the analog signal. This is often referred to as under sampling the analog
signal. This section discusses how the eDAQ
lite
handles aliasing.
The eDAQ lite has analog guard filters that work in combination with user selectable
digital filters to provide low pass filters that limit the frequency content of the digitized
signal. Furthermore, if an anti-aliasing filter is selected, this filtering guarantees that
the digitized signal is not aliased by higher frequency content components of the
analog signal. In other words, the digitized signal accurately represents all frequency
content of the analog signal below the nominal low pass filter cut-off value). Note that
the analog and digital filters do not have infinitely sharp low pass characteristics. The
amount of filter attenuation as a function of frequency is highly dependent on the
type(s) of digital filters used.
Use of these filters is most critical for low-level signal conditioning (e.g., for low level
Strain SMART Module channels). Low-level signal conditioners are more susceptible
to both eDAQ
lite
external and eDAQ
lite
internal electronic noise aliasing, since the
strain signals are typically in the millivolt range. The signal conditioner gain amplifiers
amplify the noise components as well as the actual strain signal components. In harsh
electromagnetic interference (EMI) environments, the noise contributions can even be
larger in magnitude than the actual strain signal contributions. Fortunately, these EMI
contributions often have much higher frequency content than the actual strain
contributions and can therefore be eliminated with the use of the appropriate analog
and digital filters.
Note that there are situations where the EMI frequency content is in the same range
as the actual signal frequency content. One classic example of this is 60 Hz AC power
line noise induction. The analog and digital filters cannot eliminate this type of signal
corruption, which is not aliasing in the strict definition. In this scenario, the 60 Hz noise
must be eliminated before it enters the eDAQ
lite
signal conditioner.
All of the above discussion on the use of anti-aliasing filters to ensure that the digital
data acquired accurately represents the input signal refers to accurately representing
the frequency content of the input signal. It by no means ensures that the digital data
will represent the input signal in terms of providing accurate peak-valley data that is
critical to time domain analyses (such as fatigue analysis). In fact, to get peak-valley
data that is guaranteed to provide 1% amplitude accuracy, the sample rate must be
over 20 times the maximum frequency content of the input signal.
15.2 Digital Filter Characteristics
TCE provides two types of digital filters for analog input channels. One type emulates
an eight-pole, analog Butterworth filter. The second type is an equiripple, linear-phase,
finite impulse response (FIR) digital filter.
HBM: public