Agilent Technologies Agilent 33500 Series, Руководство пользователя

Страница: 247 / 278

246
Chapter 5 Tutorial
Quasi-Gaussian Noise
5
Quasi-Gaussian Noise
The Noise waveform is generated to be optimized for both quantitative
and qualitative statistical properties. Though not truly random, it does
not repeat for more than 50 years of continuous operation.
The Noise waveform’s voltage distribution is approximately gaussian.
Unlike the ideal distribution, though, there is zero probability of getting
a voltage outside the range established by the instrument’s Vpp setting.
The distribution’s crest factor (peak voltage divided by RMS voltage) is
approximately 4.6.
You can vary the Noise bandwidth continuously from 1 mHz to 30 MHz.
The energy in the noise signal is concentrated in a band from DC to the
selected bandwidth, so the signal has a higher spectral density in the
band of interest when the bandwidth setting is lower. In audio work, for
example, you might wish to set the bandwidth to 30 kHz, making the
signal strength in the audio band 30 dB higher than if the bandwidth
were set to 30 MHz.
PRBS
A Pseudo-Random Bit Sequence (PRBS) waveform has two states (high
and low) like a Pulse or Square wave, but it switches between them in a
manner that is difficult to predict without knowing the sequence
generation algorithm. A PRBS is generated by a linear-feedback shift
register (LFSR), an example of which is shown below.