LeCroy DDA 7 Zi series, Руководство оператора

Страница: 369 / 544

Operator's Manual
WP700Zi-OM-E-RevA
368
On the instrument, when the Custom filter kind is selected, a file selection box will appear. Select the file saved
from Mathcad; the coefficients will be used in the filter.
Multirate Filters
Description
In many of today's development environments, digital filter design has become most challenging. Specifications
typically require higher order filters, implying increased storage capacity for filter coefficients and higher
processing power. Moreover, high-order filters can be difficult, if not impossible, to design. In applications such as
3G wireless systems, for example, at the receiver end data must be filtered very tightly in large magnitude in order
to be processed.
Although the LeCroy DFP option provides many filter types, the correlation between edge frequencies and sample
rate may be a limiting factor: edge frequencies are limited from 1% to 49.5% of the sample rate, while the
minimum transition width region is 1% of the sample rate.
Multirate, multistage filters are a practical solution for the design and implementation of FIR filters with narrow
spectral constraints. Multirate filters change the input data rate at one or more intermediate points within the filter
itself, while maintaining an output rate that is identical to the input rate. This approach provides a solution with
greatly reduced filter lengths, as compared to standard single-rate filters.
This can be achieved in two or more simple steps. First, a filter (with a relatively limited edge frequency) is applied
and the results are decimated. Then, a second filter is applied to the decimated waveform, substantially reducing
the lower edge frequency limit.
Example
A sine wave with a frequency of 3 MHz has a higher frequency noise component. A low-pass filter is required to
remove the noise component. The sample rate of the scope is 2 GS/s. The minimum edge frequency of the low-
pass filter for this sample rate is 20 MHz. While this filter is sufficient for removing part of the noise, it cannot
remove the high frequency component completely. In such a case, the problem can be solved in two stages.
Figure 5-21. 1. A noisy sine wave with a frequency of
3 MHz.
Figure 5-22. 2. The first low-pass filter with 20 MHz
edge frequency and 30 MHz transition region is
applied.
Figure 5-23. 3. A sparsed version of trace A.
Figure 5-24. 4. A second low-pass filter with an edge
frequency of 5 MHz and a transition region width of 6
MHz is applied to the sparsed trace in 3.
The last trace shows the zoomed signal, which was filtered by a multistage filtering method. Notice that all high
frequency noise components were removed.
Specifications
x
The pass-band gain of all filters (except custom) is normalized to 1.
x
FIR Coefficients: 2001 max.
x
IIR Stages: 29 max.
x
Filter Kinds: high pass, low pass, band pass, band stop, raised cosine, raised-root cosine, Gaussian,
custom
x
IIR Filter Types: Butterworth, Chebyshev, Inverse Chebyshev, Bessel