Talk 3: Spectrum Analyzer Detection Modes

This talks explains how to select the best type of detection for any given type of received signal.
Have you ever wondered how a spectrum analyzer works, how to properly adjust all of the analyzer's parameters, or why a stair-step pattern initially appears on a spectrum analyzer screen when you turn it on? Do you know how to precisely calculate the analyzer's sensitivity in your head, merely by glancing at the screen display without any signal present? Are you uncertain about how much gain, and how low a noise figure, you ought to specify when you are ordering a low-noise amplifier (LNA) for a radio receiver? Do you want to know the difference between noise figure and noise factor? Do you wonder how to diagnose and solve radio interference problems?
If you have questions about how to make good radio spectrum measurements or how to diagnose interference problems, you will find the answers in the NTIA Seminar Series on Spectrum Measurement Theory and Techniques. In this series of talks, an NTIA engineer at the Institute for Telecommunication Sciences (ITS) laboratory in Boulder, CO, discusses the fundamentals of radio spectrum measurements. The speaker, Frank Sanders, who has nearly thirty years of experience in this field, recognizes that even for many engineers who routinely use spectrum analyzers, the fundamentals of how they work and how to use them may be a bit murky; even in university lab classes the instructors do not always understand these machines very well themselves.
Most of the talks, which are 80-100 minutes long, are divided into two parts. In the first portion of each video, Sanders explores a particular aspect of radio spectrum measurement technique or theory with a whiteboard lecture. In the second part, the lessons of the whiteboard discussion are implemented with actual measurement hardware and radio signals. A few of the talks, which for example involve large numbers of photographs of radar systems, are videos of his Microsoft Powerpoint presentations.
In this series, Sanders explains spectrum analyzer functionality in terms of convolution bandwidth and shows how, when convolution is understood along with the mechanics of analyzer design, spectrum analyzer operations and outputs become easy to understand and use. Other topics include (1) what you need to know to use spectrum analyzers to examine all types of radio signals, including mobile radios, radars, and digital data links; (2) the use of low noise amplifiers and how to specify the right gain and noise figure for your receiver and measurement applications; (3) how radar systems work, and how to understand and interpret the signals that you see coming from radars; (4) the ways that radio interference can occur; (5) a methodical approach for diagnosing and solving radio interference problems; (6) the math needed to convert spectrum analyzer measurements into field strengths of radio signals; and (7) the proper conversions for radiation hazard calculations.

[@shih-haowang4788]

This is so instructive for learning how spectral analyzer works and how to start to use it. Also interesting demonstration with microwave oven and radar. Thanks for uploading the great contents

[@drwagner14]

Super useful, thank you!

[@Dickwooten]

Much appreciated!

[@James-wb1iq]

Thanks! This is great.

[@rodericksibelius8472]

I wonder what model of this Agilent 'Spectrum Analyzer' used in these demonstrations.

[@donaldviszneki8251]

I am skeptical that all analog spectrum analyzers connect detector output to an ADC. Don't old analog spectrum analyzers with CRT displays amplify the analog detector output and use it to control electron beam deflection in the CRT?
Thanks for this amazing resource.

[@zga8]

Hi not sure I am understand ...at 10:50 , bins plot X axis is the time domain ,or frequency ?

[@badgereclipse]

Time Domain

[@lunarjournal]

23:27 @NTIAGov I think there is a mistake when the instructor said that the values for Xi represent power?. In most spectrum analyzer user manuals i've come across the RMS detector mode uses the signal voltage for Xi. Thus effectively calculating RMS voltage which can then be converted into average power given a reference impedance. If Xi represented power, you would be calculating RMS power not average power. Someone correct me if I am wrong.

[@institutefortelecommunicat1216]

Hello. This is the reply from Frank Sanders, via the Publications Office at the Institute for Telecommunication Sciences (ITS).
The machines ordinarily measure power (voltage squared). They do not do ordinarily a direct voltage measurement. (The exception is a power meter with a bolometer head.) Having measured the input power through a log amp (unless the user has manually selected a linear-amp option-which is rare), the machine divides by 10 and antilogs to get to the voltage. Then they compute the average from multiple samples, and then log the result and multiply by 10 to get back to power.
Older analog machines could not do this; they did not offer RMS power measurement options. The way they got close to it in the oldedd days was to use a narrow video bandwidth. The newer, digital machines offer the RMS feature described above. But again, it’s not a direct measurement of RMS.

[@karielf8947]

The notion of bins is not clear ar all for someone like me. Is it in the basé of time when you talk about bin. And what is the relation ship with frequency step, and rbw???