What Is a Spectrum Analyzer? A Beginner's Guide

A spectrum analyzer shows the frequencies present in an electronic signal and measures their levels. While an oscilloscope displays how a signal changes over time, a spectrum analyzer helps you identify frequency components such as harmonics and unwanted signals. Engineers use it for radio frequency (RF) testing and diagnosing electromagnetic interference (EMI).

Many modern signal analyzers combine traditional spectrum measurements with more advanced capabilities, such as analyzing complex or modulated signals, depending on their installed software and options. This guide uses "spectrum analyzer" to refer to both.

A spectrum analyzer typically measures signals without supplying its own test stimulus. Some models include an optional tracking generator for basic magnitude-only response measurements. To measure a component's response, including both magnitude and phase, a vector network (VNA) is generally the better choice. The used Keysight E5071C offers frequency coverage up to 20 GHz, while Keysight recommends the E5080B for new purchases.

A swept-tuned spectrum analyzer scans a range of frequencies and measures the signal level at each one, creating a display of amplitude versus frequency. Keysight's guide How Does a Spectrum Analyzer Work? explains the mixer and detector stages. The main stages are:

• Input attenuator: Reduces the signal level before it reaches the input mixer.
• Mixer and local oscillator (LO): The local oscillator produces a signal that mixes with the incoming signal, converting selected frequency components to an intermediate frequency (IF). As the LO sweeps, different frequency components pass through the IF stage.
• IF filter: Passes a narrow band of frequencies. The filter's width determines the resolution bandwidth (RBW).
• Detector and video filter: The detector measures the level of the filtered signal, while the video filter can smooth the displayed trace.
• Display: Shows the measured signal level against frequency.

The right type depends on the signals you need to see. Frequency range varies by model, so check the data sheet for the analyzer you're considering.

Swept-tuned analyzers are the traditional design. The local oscillator (LO) sweeps so the analyzer tunes across the selected frequency span, while the intermediate frequency (IF) filter determines the resolution bandwidth. Newer swept analyzers use digital signal processing, so check the data sheet for how a given model works.

• Simple signal path: The block diagram above shows the main stages from input to display.
• One narrow band at a time: At any instant the analyzer measures only the narrow band it is tuned to, so a short event elsewhere in the span can be missed.
• Resolution depends on the filter: Closely spaced signals need a narrow RBW, which slows the sweep.

Digital spectrum analyzers convert incoming signals into digital data for processing. A common method is the Fast Fourier Transform (FFT), which calculates the spectrum of a captured time-domain signal.

• Whole band at once: An FFT analyzer can compute many frequencies from one capture instead of stepping through them one at a time.
• Vector signal analyzers: With vector analysis software, a signal analyzer can examine complex signals, such as modulated carriers.

A swept-tuned analyzer measures one narrow band at a time, so it can miss what happens elsewhere in the span. A real-time spectrum analyzer processes every sample within a stated bandwidth without gaps. Keysight's 89600B help describes a swept-tuned analyzer as displaying results sequentially and a real-time analyzer as displaying the energy at all frequency components simultaneously.

• Short events: Useful for intermittent interference and signal bursts that can be shorter than a sweep.
• Limits: The analysis bandwidth and the shortest event the analyzer can show depend on the model, so read the specification for the one you plan to use.

These 7 steps walk you through setting up a spectrum analyzer and interpreting the results. Start with the 4 settings that have the greatest impact on your measurements.

• Use a cable rated for the range: Connect the signal source to the RF input using a coaxial cable rated for the frequencies you are measuring.
• Respect the input limit: Never exceed the maximum input level on the data sheet.

• Know the expected frequency: This guides the start and stop frequencies, or a center frequency and span. The span is the difference between start and stop.
• Start wide, then narrow: A wide span helps you locate signals across the frequency range, while a narrower span lets you examine individual peaks in more detail.

• Understand RBW: Resolution bandwidth is the width of the IF filter. A narrower RBW separates closely spaced signals and lowers the displayed noise floor, but slows the sweep.
• Match the RBW to the task: Use a narrow RBW for close signals and a wide one for a quick overview. Thermal noise power is proportional to bandwidth, so doubling the RBW generally raises the displayed noise floor by approximately 3 dB. A University of British Columbia course note explains how an RBW that is too wide can prevent the analyzer from distinguishing closely spaced signals and increase the noise included in the measurement.

• Reference level: Sets the level at the top of the display. Adjust it so the largest expected signal fits on the screen.
• Attenuation: Reducing the signal amplitude can prevent distortion and saturation. Set it by hand or use the automatic setting if your analyzer has one. A 2020 Keysight blog describes the input mixer level as a trade-off between distortion performance and noise sensitivity.

• Read the peaks: Start with the expected signal. Peaks indicate energy at particular frequencies.
• Find the noise floor: Look at the baseline between signal peaks to estimate the displayed noise floor. Record the RBW with your measurement.
• Use markers: Place a marker on a peak to read its frequency and amplitude.
• Compare with what you expect: An unexpected peak or level can point to distortion, interference, or a setup error.

• Harmonics and spurious signals: Harmonics appear at integer multiples of the fundamental. Keysight's EMI testing page describes using a signal analyzer or EMI receiver to troubleshoot and identify suspected disturbances.
• Noise measurement: Examine a frequency range without expected signals. Record the RBW so the noise reading can be interpreted correctly.
• Demodulation: Some analyzers can demodulate modulated signals to analyze their content.

• Capture the display: Save the trace as an image or data file for documentation or review.
• Annotate your results: Label key features so they are easy to recall.
• Review the data: Check whether the results match your expectations and note any measurements that need further investigation.

Accuracy depends on the instrument model and its calibration, so read the data sheet for the figures that matter to your measurement.