Atek HP 8753D, Руководство пользователя, Страница: 271 / 677

While only a single ow path is shown, two identical paths are available , corresponding to
channel 1 and channel 2. When the channels are uncoupled, each channel is processed and
controlled independently .
Data point denition: A \data point" or \point" is a single piece of data representing a
measurement at a single stimulus value . Most data processing operations are performed
point-by-point; some involve more than one point.
Sweep denition: A \sweep" is a series of consecutive data point measurements , taken over
a sequence of stimulus values . A few data processing operations require that a full sweep
of data is available . The number of points per sweep can be dened by the user . The units
of the stimulus values (such as power , frequency , and time) can change , depending on the
sweep mode , although this does not generally aect the data processing path.
Processing Details
The ADC
The ADC (analog-to-digital converter) converts the R, A, and B inputs (already down-converted
to a xed low frequency IF) into digital words . (The A UX INPUT connector on the rear panel is
a fourth input.) The ADC switches rapidly between these inputs , so they are converted nearly
simultaneously .
IF Detection
This detection occurs in the digital lter , which performs the discrete F ourier transform
(DFT) on the digital words . The samples are converted into complex number pairs (real plus
imaginary , R+jX). The complex numbers represent both the magnitude and phase of the IF
signal. If the A UX INPUT is selected, the imaginary part of the pair is set to zero . The DFT
lter shape can be altered by changing the IF bandwidth, which is a highly eective technique
for noise reduction.
Ratio Calculations
These calculations are performed if the selected measurement is a ratio of two inputs (for
example , A/R or B/R). This is a complex divide operation. If the selected measurement is
absolute (such as A or B), no calculations are performed. The R, A, and B values are also split
into channel data at this point.
Sampler/IF Correction
The next digital processing technique used is sampler/IF correction. This process digitally
corrects for frequency response errors (both magnitude and phase , primarily sampler rollo) in
the analog down-conversion path.
Sweep-T o-Sweep A veraging
A veraging is another noise reduction technique . This calculation involves taking the complex
exponential average of several consecutive sweeps . This technique cannot be used with
single-input measurements .
Application and Operation Concepts 6-5