Rohde & Schwarz R&S FSP Series, Operating Manual

Page: 70 / 796

R&S FSP Getting Started
Measurements on Modulated Signals
2.46 Operating Manual 1164.4556.12 - 06
4. Bandwidth selection
If the amplitude distribution is measured, the resolution bandwidth must be set in
a way that the complete spectrum of the signal to be measured falls within the
bandwidth. This is the only way of ensuring that all the amplitudes will pass
through the IF filter without being distorted. If the selected resolution bandwidth is
too small for a digitally modulated signal, the amplitude distribution at the output
of the IF filter becomes a Gaussian distribution according to the central limit
theorem and so corresponds to a white noise signal. The true amplitude
distribution of the signal therefore cannot be determined.
A video bandwidth which is large in comparison to the resolution bandwidth (
≥
3 x
RBW) must be selected. This ensures that the amplitude peaks of the signal are
not smoothed by the lowpass effect of the video filter. The video bandwidth is set
automatically during statistics measurements.
Since the video bandwidth of the R&S FSP is limited to 10 MHz, lowpass filtering
occurs during measurements with a resolution bandwidth of 10 MHz. Additional
band-limiting occurs at a resolution bandwidth of 10 MHz due to the lowpass
filtering at the output of the log amplifier. The latter limits the video signal to a
bandwidth of 8 MHz in order to obtain sufficient suppression of the 20.4 MHz IF.
The level range of the signal amplitudes, e.g. during APD white-noise
measurements, is smaller. For broadband-modulated signals such as W-CDMA
signals, the effect depends on the bandwidth occupied by the signal. At a signal
bandwidth of 4 MHz, the amplitude distribution can be measured correctly with the
effective video bandwidth.
5. Selecting the number of samples
For statistics measurements with the R&S FSP, the number of samples N
Samples
is entered for statistical evaluation instead of the sweep time. Since only
statistically independent samples contribute to statistics, the measurement or
sweep time is calculated automatically. It is indicated on the R&S FSP display.
The samples are statistically independent if the time difference is at least 1/RBW.
The sweep time SWT is, therefore, expressed as follows:
SWT = N
Samples
/RBW