Analyzing a device’s current consumption over time provides many beneficial insights into its behavior. Not only is this important for battery powered devices to predict battery lifetimes, but for medical devices it is essential to prevent possible injury or loss of life. However, capturing and then analyzing a device’s current draw over many hours or days to identify abnormal events presents a near unsolvable challenge if performed manually.
There are many “traditional” ways to do this analysis; possible choices include: a digital multimeter, an oscilloscope with a current probe, or a power supply with current monitoring capabilities. All these solutions have their strengths and weaknesses, and all of them involve making tradeoffs between bandwidth/sampling rate, measurement resolution, and minimum measurable current. In addition, none of these solutions offer any ability to automatically analyze captured data and identify the locations of anomalous behavior. To answer this unmet need, Keysight developed the CX3300A Device Current Waveform Analyzer.
Despite its tongue-twisting name, you can view the CX3300A as a specialized oscilloscope for measuring current (although it can of course also measure voltage). Typical of high-performance current measurement, the CX3300A’s current sensors require a break in the current path (using either one of the CX330A’s low-burden voltage sensors or a shunt resistor that is already existent on the DUT board). The following picture shows the CX3324A with some of the different probes and sensor heads that it supports.
The CX3300A series supports the following sensor options that cover a broad current and voltage measurement range. You can choose the appropriate sensor combinations according to your requirements of sensitivity, bandwidth, and connectivity. The following lists show the available sensors along with their application.
Current measurement:
- CX1101A Single-channel current sensor
- CX1102A Dual-channel current sensor
- CX1103A Low-side current sensor
- CX1104A Selectable shunt current sensor
Voltage and current measurement:
- CX1105A Ultra-low noise differential sensor
Voltage measurement:
- CX1151A Passive probe interface adapter
The 4-channel CX3324A also includes an 8-channel digital input (connector on the rear) to enable you to track and record digital signals such as the controller’s I/O or data bus. This is obviously helpful when you need to correlate aberrant device behavior with specific machine actions.
The unique capability of the CX3300A is its optional data logger mode and associated waveform analytics capability. In data logger mode the CX3300A can capture waveform data at up to 10 MSa/s for up to 100 hours. However, it should be obvious that this capability alone without some means to automatically analyze the captured data would not be very useful. Therefore, this option also includes a waveform analytics capability that uses AI techniques to automatically place captured waveform data into groups based on their similarity. Because the software performs the grouping as it collects data, you can begin analyzing data immediately even if the database contains hundreds of thousands of waveforms.
The following picture shows a snapshot of the waveform analytics software, illustrating how easy it is to locate an anomalous waveform amongst thousands of similar waveforms. Notice that you can also pinpoint exactly where in time each waveform occurs, which makes it easy to correlate waveform activity with other operations occurring on the device under test.
Of course, an alternative solution to the CX3300A is to use a Keysight N2820A/N2821A high sensitivity, high dynamic range current probe with one of our S-Series Infinium oscilloscopes. However, this solution can only measure currents down to 100 nA, whereas the CX3300A can measure currents down to 150 pA. Perhaps more importantly, the oscilloscope based solution does not provide any means to automatically analyze the waveforms it captures.
If you want to do a deeper dive into the CX3300A’s waveform analytics software, check out the related content link “Using Anomalous Waveform Analytics for IoT Device Analysis”. This modular training includes a software demo.