Frequency and filtering – Agilent Technologies VXI E1439 User Manual
Page 40
30
Using the Agilent E1439
Frequency and filtering
Frequency and filtering
The Agilent E1439’s center frequency is normally set at zero (baseband path) and 70 MHz for the
IF signal path. However, you may set the center frequency to a non-zero value in order to examine
a narrower span away from baseband (zoom measurement).The frequency band of interest,
represented by digitized time data samples from the ADC, is mixed with the E1439 digital LO, a
complex exponential, at the desired center frequency. As a result, the frequency band of interest in
the input signal is shifted to a complex signal centered around dc.
multi-module systems” on page 39
for special considerations with respect to changing the center
frequency in multi-module systems.
The default filter for E1439 measurements is an analog anti-alias filter. However, you may further
isolate the frequency band of interest for more detailed analysis by using digital filtering. A
decimating digital filter simultaneously decreases the bandwidth of the signal and decreases the
sample rate. The built-in digital filters conform to the Nyquist sampling criterion, which
guarantees that the output sample rate may be reduced by the same factor as the signal bandwidth
reduction while still maintaining a complete representation of the underlying bandlimited signal.
For each octave step in bandwidth reduction (except for the first octave), the E1439 digital filters
automatically reduce the data rate by discarding alternate output samples. This process, called
decimation, results in an output sample rate that is nominally four times the signal bandwidth
whenever sigBw>0. This is still double the theoretical rate necessary to fully characterize the band
limited signal. However, because the digital filters do not have a perfectly abrupt cutoff, the
sample rate cannot be reduced to the theoretical limit without some aliasing of signals in the
transition frequency band of the filters. In many applications, this limited aliasing potential is not
important. For this reason you may optionally choose to apply a final factor-of-two decimation.
See the Technical Specifications for detailed information on the digital filter shapes.
The decimation process used to reduce the output sample rate is driven from a "decimation
counter" that keeps track of which samples to save and which ones to discard for each of the
octave bandwidth reduction filter stages. In multi-module systems where synchronous sampling is
required, the decimation counters in all the modules must be synchronous with each other.