Real-time spectrum analyzer helps troubleshoot radio links

The number of wireless transmitters around the world has increased dramatically over the past decade. The Internet of Things (IoT) has driven the need for low-cost, easy-to-implement chipsets to enable wireless connectivity. Wireless technologies such as Bluetooth, ZigBee, and WiFi (802.11) have made connectivity ubiquitous and widely used in homes, vehicles, and businesses. These technologies have been widely adopted by engineers because they have constantly updated and improved standards, do not require wireless spectrum licenses, and have reference designs that have been approved by regulators.

Needless to say, the spectrum in 2.4 GHz is the most popular operating area for these unlicensed and economical applications. To facilitate spectrum sharing, wireless standards must adopt advanced hardware and software features, including time domain multiple access, clear channel assessment, adaptive frequency control, etc. Engineers find that even if they use "certified" solutions, they still face many problems in establishing radio links and maintaining communications.

As IoT technology rapidly develops, the problem will only get worse, with more and more signals competing for spectrum space. Some people even renamed IoT to "Interference of Things", which is not an abuse. Although IoT can share spectrum without a license in practice, the challenges it brings are indeed not to be underestimated. IoT adds very complex RF control chips to a large number of electronic devices, such as smart systems, cars and various home appliances. Each device may generate more noise and interference. Although designers can add additional shielding and other noise reduction features to products to reduce the amount of noise, this will increase the cost to consumers, and the relatively small number of channels makes it difficult to cope with the endless number of electronic devices.

Another challenge in solving these problems is that there is no perfect system to detect noise. Trying to find the problem in the city is particularly difficult compared to the suburbs or the countryside because there are so many electronic devices in use. In addition, noise varies with various issues, such as location and weather, which further complicates the problem. Physical interference is already a challenge, and it is bound to get worse over time.

Characterizing wireless links

Debugging and characterizing wireless connections requires some basic knowledge of the type of radio being analyzed, including:

The frequency or channel you are operating on so you know which bands of the spectrum to look in.

The type of wireless connection (Bluetooth, WiFi, Zigbee, NFC) to determine what kind of spectrum signature is expected. This will also indicate special modes such as TDMA, frequency hopping, FDMA operation, etc.

Transmitter power levels, which gives an idea of ​​the expected interference levels.

Minimum receiver sensitivity, which specifies how sensitive the receiver is to interfering signals.

With this information, we can use a spectrum analyzer to understand the wireless link to some extent and characterize the RF environment. A spectrum analyzer is an essential tool when making measurements in the wireless spectrum. Figure 1 is a simplified block diagram of a traditional swept spectrum analyzer for reference.

Figure 1: This is a simplified block diagram of a traditional swept spectrum analyzer.

Superheterodyne spectrum analyzers (SAs) have been used for many years. The challenge of using this type of instrument is the "swept" nature of their operation. What is measured on the spectrum display is not coherent in time and may not accurately represent the spectral information (especially for TDMA signals). Even the fastest swept spectrum analyzers are limited in their ability to view transmitters that use frequency hopping techniques. In addition to the basic frequency vs. amplitude display, some manufacturers provide three-dimensional spectrogram information. In a swept spectrum analyzer, this information is derived from multiple sweeps, so the timing information is essentially only an approximation of what might occur in a pulsed or frequency agile transmitter.

A real-time spectrum analyzer (RTSA) provides essentially the same functionality as a traditional spectrum analyzer, with some key features added. Figure 2 is a block diagram of a basic real-time spectrum analyzer for reference.

Figure 2: This is a block diagram of a basic real-time spectrum analyzer.

One of the main differences between an RTSA and a basic signal analyzer is the RTSA bandwidth specification. For any bandwidth below the maximum real-time bandwidth, the RTSA does not have to sweep the frequency, but can continuously capture spectral information. The RTSA is also not limited to one display at a time. It can analyze spectrum, spectrogram, and modulation information simultaneously, and because this information comes from continuous acquisition, it is time-correlated.

RTSA is particularly well suited for analyzing systems that use TDMA protocols such as WiFi, Bluetooth, ASK/FSK. One of the biggest issues for devices using unlicensed bands is managing the effects of multiple transceivers sharing the same spectrum. Almost all regulations require that devices operating in unlicensed bands must not cause interference and must accept any interference that exists. RTSA is particularly well suited for quantifying the effects of interference because it continuously captures the spectrum information.

Important RTSA features include fast spectrum rates (10,000 to 3,000,000 acquisitions per second), the ability to continuously record spectrum data and the entire RF environment as it changes over time. Other key features include time, frequency, and amplitude triggering and correlation of time-domain, frequency-domain, and modulation measurements.

Figure 3 shows a real-time spectrum analyzer’s digital phosphor spectrum display. As with a typical spectrum analyzer, the display shows frequency versus amplitude information. In addition, the pixels in the display are colored to tell us how often RF energy is measured at that pixel (pixel occupancy). With digital phosphor spectrum measurements, the user can specify a fade function that provides a phosphor effect that simulates the display effects used in CRT-based oscilloscopes. This adds a periodic dimension to the display, showing how often the signal is actually being measured in the frequency band of interest.

Figure 3: The RTSA's phosphor effect simulates the effect of a CRT-based oscilloscope, allowing you to see how often a signal is measured within the frequency band being measured.

This form of real-time spectrum display allows us to "see" what the receiver "sees" and gain greater insight into what is happening in the frequency band of interest. However, it does not provide sufficient information about the potential effects of interfering signals. Due to this feature, the spectrum display does not show the time interleaving of signals. Using "zero span" measurements can provide sufficient detail about pulse amplitude and duration, but no frequency information.

The spectrogram measurement is designed to solve exactly this type of problem. Like the spectrum display, it shows low frequencies on the left and high frequencies on the right. Unlike the basic spectrum display, it uses color to represent amplitude, and all of this information is plotted relative to time on the Y-axis. The spectrogram is essentially a barcode recorder that shows spectral activity over time.

In a swept spectrum analyzer, this spectrogram is not temporally coherent because the instrument is sweeping the frequencies. The analyzer sweeps the frequencies, meaning that the trace points on the left side of the span occur earlier than the trace points on the right side. Therefore, there is no timing relationship in the spectrogram captured by a swept spectrum analyzer.

However, the spectrogram created by the RTSA is made up of continuously recorded spectrum data rather than a sweep. Another benefit of the RTSA is full domain correlation, so the information in the spectrogram can be directly correlated with other measurements, such as modulation, power, and CCDF.

Figure 4 is an example of a digital phosphor display and a spectrogram. The digital phosphor display in this example shows a great deal of detail about the signals that are present. In the center of the frame is a weaker, broadband signal with a large crest factor. This signal has high channel occupancy (near continuous) given the bright or “warm” color. Also visible in the frame is a Wi-Fi signal that appears to be operating at 2.437 GHz (WiFi channel 6). There are a dozen other signals in the frame at varying frequencies and powers. Given the shape of the spectrum and the frequencies used, these signals are likely coming from Bluetooth devices.

Figure 4: The RTSA enables full-domain correlation, and the information in the spectrogram on the left can be directly correlated with other measurements.

Although many different services use the spectrum measured above, these signals are time interleaved so there is little or no loss in link quality when using active spectrum sharing techniques. Routine spectrum analysis increasingly requires real-time spectrum analyzer technology to verify that the link is operating as expected. Historically, RTSAs have been limited to niche applications, but modern wireless designs clearly require the processing power and flexibility of real-time spectrum analysis to debug system-level issues and characterize operating modes.