Stanford Research Systems SR785, Руководство по эксплуатации и программированию

Страница: 104 / 546

2-24 FFT Averaging
SR785 Dynamic Signal Analyzer
FFT Averaging
Averaging successive measurements together improves accuracy and repeatability. For
measurements computed from multiple spectra, averaging is performed on measurement
results or individual spectra depending upon the measurement definition. Time records
are not averaged.
Just as the SR785 computes all measurements in a given measurement group
simultaneously, all averaging variants of any measurement are also computed
simultaneously. For instance if a frequency response measurement is paused or finished,
you can look at the rms averaged, vector averaged, peak hold averaged, or instantaneous
(not averaged) frequency response by adjusting the <Display Avg> softkey, there's no
need to acquire new data to look at a different averaging type.
You can control whether the SR785 calculates averaged quantities for measurements
with the <Compute Avgs> Softkey. If <Compute Avgs> is set to "Yes," the SR785 will
compute averages for all measurements. If <Compute Avgs> is "No" the SR785 will not
compute any averages and all quantities diplayed will be instantaneous values.
RMS Averaging
RMS averaging involves averaging the result of multiplying a complex quantity by the
complex conjugate of another complex quantity. For instance the RMS averaged FFT is
defined as:
RMSAvg( FFT1 ) =
√
(<FFT1* • FFT1>)
The precise definition of what "RMS Averaging" means for each measurement is given
in the description of each measurement. Baiscally, since RMS averaging always involves
averaging the "square" of a quantity, RMS averaging reduces fluctuations in the data but
does not reduce the actual noise floor (squared values never cancel). With a sufficient
number of averages, a very good approximation of the actual noise can be obtained.
Note that the definition given above always yields a real quantity whose phase is
zero.This is not true for all RMS averaged quantities computed by the SR785 however.
Both the rms averaged frequency response and the rms averaged cross spectrum are
complex quantities whose phase is not necessarily zero.
Vector Averaging
Vector averaging computes the average of the real part (X) and imaginary part (Y) of a
measurement according to
VecAvg (X,Y) = (<X> ,<Y>)
ie. the Vector average of a complex quantity is the complex quantity formed by the
average of its real and imaginary parts independantly
Linear averaging computes the equally weighted mean of X and Y over N measurements.
Exponential averaging weights new data more than old data and yields a continuous
moving average.