Miscellaneous - IMC C SERIES User Manual

5.2 Miscellaneous

5.2.1 Filter settings
5.2.1.1 Theoretical background
The filter setting is especially important in a signal-sampling measurement system: the theory of digital
signal processing and especially the sampling theorem (Shannon, Nyquist) state that for such a system,
the signal must be restricted to a limited frequency band to ensure that the signal has only negligible
frequency components beyond one-half of the sampling frequency ("Nyquist-frequency"). Otherwise,
"aliasing" can result – distortions which cannot be removed even by subsequent filtering.
A C-SERIES device is a sampling system in which the sampling frequency, which must be set in the
configuration menu, is subject to this constraint. The low-pass filter frequency selected thus hinges on
how band-limited the signal to be sampled at that rate is.
The control AAF for the filter setting stands for "Automatic Anti-aliasing Filter", and automatically selects
the filter frequency in adaptation to the sampling rate selected. The rule this is based on is given by:
AAF-Filter frequency (-80 dB) = sampling frequency * 0.6 = Nyquist frequency * 1.2
AAF-Filter frequency (-0.1 dB) = sampling frequency * 0.4 = Nyquist frequency * 0.8
5.2.1.2 General filter concept
imc C-SERIES system architecture is actually a two-step system in which the analog signals are sampled at
a fixed "primary" sampling rate (analog-digital conversion with Sigma-Delta ADCs). Therefore a fixed-
frequency analog low-pass filter prevents aliasing errors to this primary rate. The value of this primary
rate is not visible from the outside, depends on the channel type and is generally greater than or equal to
the sampling rate which is selected in the settings interface. The filter to be set is realized as a digital
filter, which offers the advantage of an exact magnitude and phase shift. This is especially important for
the sake of matching of channels which are jointly subjected to math operations.
If slow data rates (f_sample) are set in the system configuration, then digital anti-aliasing filters (low-pass
filters) ensure compliance with the conditions for the Sampling Theorem. One distinguishes among three
cases.
5.2.1.3 Implemented filters
Filter-setting "Filter-Type: without":
Only the (analog) anti-aliasing filter, matched to the primary data rate is in effect, along with digital
frequency response correction downstream, which provides a steep frequency response.
This setting can be useful if maximum bandwidth reserves are to be used and there are theoretical
limitations on the measured signal's spectral distribution, which justify not performing total filtering.
Filter-setting "Filter-Type: AAF":
The (digital) anti-aliasing filters are elliptical Cauer filters. Their "tight" characteristic curve in the
frequency range makes it possible to have the cutoff frequencies approach the sampling and Nyquist
frequencies much closer without having to make a compromise between the bandwidth and freedom
from aliasing.
© 2018 imc Test & Measurement GmbH
Hardware configuration of all devices
imc C-SERIES - Manual, Version 4 R 3 - 2018-10-19
75