Answers to RF Cafe Quiz #66

All RF Cafe quizzes would make perfect fodder for employment interviews for technicians or engineers - particularly those who are fresh out of school or are relatively new to the work world. Come to think of it, they would make equally excellent study material for the same persons who are going to be interviewed for a job.

**Click here for the complete list of *** RF Cafe
Quizzes*.

Note: Many answers contain passages quoted in whole or in part from the text.

**Return to RF Cafe Quiz #66**

This
quiz is based on the information presented in *
Spectrum and Network Measurements*, 2nd Edition, by Robert A. Witte.
"This comprehensive treatment of frequency domain measurements successfully consolidates all the pertinent theory into one
text. It covers the theory and practice of spectrum and network measurements in electronic systems. It also provides thorough
coverage of Fourier analysis, transmission lines, intermodulation distortion, signal-to-noise ratio and S-parameters."
The book was graciously provided by Keysight Technologies.

Note: Some of these books are available as prizes in the monthly RF Cafe Giveaway.

**1.** Which type of FFT (Fast Fourier Transform) window should be used when measuring a fast transient
signal?

a) Exponential

A decaying exponential response leaves the beginning portion of the measured waveform undisturbed while forcing the end of the time record to zero. (p 60)

**2.** Which statement regarding the DFT (Discrete Fourier Transform) is true?

a) A DFT is only an approximation of a Fourier transform

The DFT differs from the Fourier transform in many ways, including an assumption that the waveform begin measured is periodic, which can cause significant 'leakage' artifacts on the display, and the limited dataset can limit the dynamic range and noise performance. (p 38)

**3. ** When might a non-zero coherence (between input and output signals) measurement be false?

b) When a third signal affects both the input and output signals

A significant third signal that gets into both the input and output signal can cause the instrument to include its contribution to the result. (p 70)

**4. ** Where is the video filter of a swept spectrum analyzer located in the signal chain?

d) Between the detector and the display

The video filter smooths out the data displayed on the monitor and therefore sits between the detector and the display. (p 90)

**5.** Where, if at all, on the spectrum analyzer display does LO (local oscillator) feedthrough show
up?

a) At 0 Hz (DC)

When the LO frequency is the same as or very near the IF (intermediate frequency), the difference frequency of the mixing process generates signal content at or near 0 Hz. (p 95)

**6.** What is the purpose of the 'Zero Span' function on a spectrum analyzer?

c) It shows the amplitude characteristic of a specific frequency over time

In the Zero Span mode, the output of the detector is swept over time at the center frequency of the spectrum analyzer (with a bandwidth that includes its sidebands), showing how the amplitude of signal changes over time. (page 114)

**7.** What is the term given for a long-term frequency shift?

a) Drift

Long-term changes in frequency are considered drift due to temperature and/or mechanical changes to the circuitry. (p 162)

**8.** What is the term given for a short-term frequency shift?

b) Phase noise

Short-term changes in frequency are considered phase noise due to perturbations such as jitter, white noise, unintentional modulation, or any other short-lived phenomenon. (p162)

**9.** Spectrum analyzer predetection filtering reduces which aspect of noise?

d) Noise present in the input signal

Predetection and filtering works on the raw input signal before it is affected by noise components introduced by the analyzer circuitry. (p 177)

**10.** What is the rule-of-thumb swept spectrum analyzer correction factor used for noise power measurements?

c) 2.5 dB

Because of the effective noise bandwidth of the filter (usually Gaussian) and the log video amplifier in the display circuitry, many - if not most - spectrum analyzers display a noise level around 2.5 dB lower than the actual noise present. (p 158)