Just as with DC power-rail signals, probing of RF signals should be done with as little noise and as non-invasively as possible and with the best possible signal fidelity.
Many IoT devices use wireless methods to communicate with other devices or with host systems. Just as with DC power-rail signals, probing of RF signals should be done with as little noise and as non-invasively as possible and with the best possible signal fidelity. Effective probing opens the door to Wi-Fi and Bluetooth signal analysis with RF demodulation, vector signal analysis, and spectrum analysis.
Acquiring wireless signals typically isn’t difficult with IoT devices. In many cases, wireless antennas are connected to the IoT’s printed-circuit board (PCB) with U.FL connectors, which are ultra-small, surface-mount coaxial connectors. A U.FL connector on the IoT’s PCB enables a clean, direct connection to the device’s RF output.
Teledyne LeCroy’s RP4030 active voltage-rail probe is a good option for RF applications due to its low noise, high offset range, and 4-GHz bandwidth. Figure 1 depicts the RP4030 connected to the U.FL connector on an IoT device’s PCB. In this case, 20-dB attenuators have been added to the front end of the RP4030 Probus interface to bring the signal within the probe’s 1.6-V pk-pk dynamic range. The waveform can then be rescaled on the oscilloscope to reflect the true measurement of the acquired signal.
Figure 1: Many IoT devices accept wireless antennas using U.FL connectors.
A typical IoT wireless IC supports various network standards, including IEEE 802.11 and Bluetooth, and will operate in the 2.4- and 5-GHz bands. There also are various modulation modes such as CCK and orthogonal frequency division multiplexing (OFDM) with PBSK, QPSK, 16 QAM, 64 QAM, and others for the engineer to validate.
Modern oscilloscopes can bring the following tools to bear on the RF validation challenge:
Let’s look at Bluetooth Low Energy (BLE) advertising packets and how they might be analyzed using RF-burst amplitude and frequency demodulation. Referring to Figure 2, a BLE slave advertises on channels 37, 38, and 39, which are assigned to 2.402, 2.426, and 2.480 GHz, respectively. BLE channels are space 2 MHz apart. A BLE master listens for the advertising packets and sends a Connect Request to a slave it wants to connect with. One metric we want to look at is how centered each of these packets are around its channel’s mean frequency, and how much the spectral content is spread around that mean frequency.
Figure 2: Bluetooth Low Energy slaves advertise on channels 37-39; BLE masters listen for these advertising packets and send a Connect Request to slaves.
Referring to Figure 3, the top-left grid shows the BLE advertising bursts. Using the oscilloscope’s amplitude-demodulation math function, we can create a burst envelope waveform with which to measure the bursts’s width, period, and duty cycle. These can be seen as the light-green trace bordering the burst envelope waveform.
The Internet of Things (IoT) has come a long way from just being an industry hype to becoming one of the main drivers for the semiconductor industry. This month’s In Focus looks at the latest developments happening in the IoT space and new innovations that are being enabled by it.
The next grid down at left is a zoom trace of the Channel 37 burst. Below that in magenta is a fast-Fourier transform of that signal to measure its spectral content. Spectral analysis reveals that the burst is centered around 2.402 GHz and that most of its energy is contained within the desired 2-MHz bandwidth as expected.
Figure 3: An amplitude-demodulated BLE advertising burst (shown in top-left grid above) reveals details regarding period, burst width, duty cycle, and more.
The frequency demodulator math function can also be used on the GFSK-modulated burst to demodulate it into its serial-data bit stream (Figure 3, top right grid). The data can then be decoded to read the packet’s preamble, axis, address, the advertising channel’s header, and the packet’s payload data directly on the oscilloscope.
Teledyne LeCroy’s serial-data analysis software provides some extremely flexible capabilities, such as the length gate feature that enables performance of independent parallel analysis on each of the BLE advertising bursts simultaneously. The screen capture of Figure 4 shows Channel 37 on lane one, Channel 38 on lane two, and Channel 39 on lane three. We can view the spectrum of each RF burst and render eye diagrams of the frequency-demodulated data to analyze the signal quality of the underlying bit streams.
In the measurement table at the bottom of Figure 4, we can find reports of the eye height, width, and peak-to-peak and RMS jitter, as well as other serial-data metrics that provide the user with deep insight into what’s happening with their serial-data bit stream.
Figure 4: The length gate feature of Teledyne LeCroy’s serial-data analysis software permits parallel analysis of each BLE advertising bursts simultaneously.
Meanwhile, Figure 5 depicts an example of how an oscilloscope can frequency-demodulate an RF burst and then Manchester-decode the underlying bit stream. In this case, the application is a vehicular key fob; typically, RF signals from key fobs are ASK- or FSK-modulated. Analysis requires only a few simple steps: The user selects the modulation type, enters the carrier frequency, and launches the Manchester decoder.
Figure 5: This screen capture depicts frequency demodulation and subsequent Manchester decoding of the bit stream.
Fundamentally, oscilloscopes are time-domain instruments, but they often provide powerful capabilities for debugging of frequency-related signal problems. Many Teledyne LeCroy oscilloscopes can be provided with a spectrum-analysis application to simplify setting up and using the instrument for frequency-domain analysis.
Figure 6 shows a capture of an 802.11 RF burst from an IoT device. The burst waveform, which appears in magenta at top right, was captured using Teledyne LeCroy’s RP4030 active voltage-rail probe along with a 20-dB attenuator. The grid has been rescaled to accurately report the units of the signal under test. Shown at top center is a 3D map of the spectral content. The table at top left identifies significant frequency peaks along with their frequencies and amplitudes.
Figure 6: Examining an 802.11 RF burst in the frequency domain details significant frequency peaks.
In another illustration of spectrum analysis, Figure 7 shows an FFT of the spectral content of BLE advertising bursts. The full acquisition is at left; on the right is a zoom trace of Channel 38. The spectrum analysis shows that most of the signal’s energy is contained within a 2-MHz window as the BLE specification stipulates.
Figure 7: A fast-Fourier transform of BLE advertising bursts shows that the signal complies with the 2-MHz bandwidth stipulation in the BLE specification.
As mentioned above, setup for spectrum analysis is made easier by the application software. At the bottom of Figure 8 is its setup dialog. If you’re familiar with using a dedicated spectrum analyzer, the oscilloscope application has similar functionality. The user has control over the center frequency and span, resolution bandwidth, and the operation mode of the trace, as well as the window type and output units. At the top of Figure 4 is the 3D spectrogram view, which provides a history of the waveform’s frequency modulation.
Figure 8: The setup dialog for the oscilloscope’s spectrum-analysis application looks much the same as those found on dedicated spectrum analyzers.
Orthogonal frequency-division multiplexing (OFDM) packets can be extensively analyzed using the oscilloscope’s vector signal analysis application. This tool offers comprehensive OFDM analysis and visualization capabilities (Figure 9).
Figure 9: The vector signal analysis tool provides comprehensive analysis of OFDM packets.
The pilot error-vector magnitude (EVM), data EVM, and average EVM can be plotted by subcarrier and by symbol. The tool also can report measurement EVM percentage values as well as magnitude error and pilot signals.
The tool supports the 802.11a protocol, but other user-defined protocols can be tested by loading a metadata file. Figure 5 also shows the ability to render constellation diagrams, which provide visualization of the phase angle and amplitude of the resulting modulated signal. The VSA software contains a full list of selectable measurement parameters, including such items as Q factor, data and pilot EVM, SNR, and phase error. It allows for analysis of up to eight data streams in parallel. User-defined Matlab blocks can be inserted at any point in the processing chain for complete customization of the DSP process.