Oscilloscopes detect ECU disturbances from EMI
When you perform EMC tests, you often think of emissions measurements made with a spectrum analyzer. But, there are EMC applications for oscilloscopes. A relatively underutilized use for oscilloscopes in EMC testing is for real-time functional performance evaluation, including deviation detection, of a device under test (DUT) during exposure to a disturbance. An oscilloscope can help you document how EMC affects your product’s operation. We often use oscilloscopes but need to electrically isolate them from the EUT that’s inside a chamber.
The term "deviation" refers to an EUT’s response to a disturbance where one or more functions exceed allowable tolerances. These functions and tolerances are defined in an EMC test plan document, uniquely developed for the specific device, and approved by all concerned parties before testing commences.
Standard practice in the automotive industry has been to perform component-level tests to determine a device’s immunity to disturbances such as ESD (electrostatic discharge), transients on power and I/O lines, conducted RF, and radiated magnetic and electric fields. These tests are conducted prior to full vehicle immunity testing. Acceptance criteria for immunity, such as RF field strength levels the DUT must endure, are defined in an OEM’s engineering specifications while the procedures are typically performed to international standards.
The test setup that’s common to most component-level immunity tests consists of a wire harness and a load simulator, which contains actual and/or electrically equivalent loads that represent the DUT’s interface with the vehicle. The DUT is exercised in one or more modes of operation, defined in the test plan, and exposed to a disturbance. During exposure to the disturbance, the DUT functions are monitored for a response exceeding an allowable tolerance. Typical to RF immunity tests, detection of a deviation requires determination of the device’s immunity threshold, a process where the magnitude of the disturbance is reduced significantly and increased in fine increments until the deviation recurs.
If the DUT has a CAN communication bus, then some information concerning its functional state can be sent over the bus. Unfortunately, other monitored functions details can’t transfer over the bus. Examples include the analog signals of a sensor or a PWM (pulse-width modulation) output to drive an actuator. We must measure these functions with an appropriate instrument.
RF immunity tests are typically performed in shielded chambers to prevent exposure of laboratory personnel to hazardous fields and to prevent malfunction of sensitive equipment. The conducted RF immunity test described in ISO 11452-4 utilizes a clamp-on current injection probe to induce RF current into the EUT harness at frequencies from 1 MHz to 400 MHz at levels ranging from tens to hundreds of milliamps. Those currents create fields near the test bench at levels high enough to effect operation of unshielded equipment. The radiated RF immunity test described in ISO/IEC 61000-4-21 utilizes a reverberant chamber containing a mechanical mode tuner which, when a sufficient number of tuner positions have been obtained at a given test frequency, produces a statistically uniform field within the useable volume of the chamber. The test frequency range is 300 MHz to 3 GHz with field strengths as high as 200 V/m (CW and AM) and 600 V/m (radar pulses).
Maintaining the integrity of the shielded chamber prohibits directly connecting measurement instrumentation to the test setup over conductive cabling. Inside the shielded chamber, RF fields couple to the cable, which then acts as a radiating antenna outside the chamber. To avert that problem, we use isolated connections using RF hardened fiber-optic transmitter and receiver sets. The converted signals exit the chamber though non-conductive fiber-optic cables routed through waveguides having a lower cutoff frequency above the frequency range of the test. The optical signals are converted back electrical form by the receiver, which is connected to the measurement instrumentation.
In Figure 1, the test setup (not shown) and RF hardened fiber optic transmitters are placed within the useable volume of the reverb chamber on a foam bench having a relative permittivity less than 1.4.
Figure 1. Reverb chamber equipped with a mode tuner (right). Transmit and receive antennas not pictured.
Once available outside the chamber, the signals are typically routed to data-acquisition system, which often requires custom software to analyze and compare the signal information to allowable tolerances and decide if the if EUT meets the specified requirements. Unlike many sensors, ECUs (electronic control units) may have several signals to monitor and evaluate measurements to acceptance limits and the software needed can come at a high development cost. Instead, we use an array of oscilloscopes in place of a complex, custom data acquisition system. Because oscilloscopes are already equipped with mask testing and parameter limit test abilities, they can address many, if not all, of the test requirements directly, without any significant amount of software development time needed.
Figure 2 shows the open doorway to the reverberation chamber, which is to the right of the test bench. On the left side, fiber optic cables, receiver and an array of oscilloscopes for performing real-time analysis.
Figure 2. An array of oscilloscopes is used for real-time analysis of the DUT response to radiated electric fields.
We use waveform masks in the oscilloscope to compare the waveform shapes during exposure to a disturbance relative to the shapes with no disturbance present. The dimensions of the mask depend on the acceptance criteria defined in the test plan.
Figures 3, 4, and 5, show the output of a simulated ECU. For confidentiality reasons, simulated data is used which closely approximates what signals may be monitored with a typical ECU. Channels 1 and 2 show simulated PWM signals which control an output driver actuator signal. The simulated actuator signal is captured on Channel 3, and a CAN split voltage is displayed on Channel 4.
Figure 3 shows the acquisition with mask testing turned off, the wave shape of each signal is observed. The oscilloscope is Edge-triggered on Channel 2, and all four waveforms are captured synchronously.
Figure 3. Simulated ECU output signals include PWM signals on Channels 1 and 2, an Actuator Driver Output on Channel 3, and a CAN Split voltage on Channel 4.
Figure 4 shows mask testing. The mask shape verifies that the signal’s high level, low level, frequency, duty cycle, and other criteria fit within tolerance limits described in the test plan. The mask thickness forms the specified tolerance band around a defined nominal value, which verifies that each acquired waveform doesn’t deviate by more than a specified percentage beyond the defined nominal value. In this example, all waveforms meet all of the specified test criteria. Note that the oscilloscope, set for edge triggering, continuously monitors for deviations using the predefined mask criteria. The oscilloscope triggers on an edge occurring on Channel 2, and the scope is configured to identify and document each of the deviations as they occur.
Figure 4. Simulated ECU output signals show that the PWM signals on Channels 1 and 2, the Actuator Driver Output on Channel 3, and the CAN Split voltage on Channel 4 all fit within the defined tolerance masks, resulting in passing mask test criteria.
In Figure 5, the simulated ECU exhibits an out-of-tolerance response during exposure to an electric field, 1 kHz amplitude modulated. Amplitude of the PWM signals is reduced and their duty cycles increase. Additionally, the modulation frequency superimposes onto the signal during their high state. The Driver Output waveform suggests an indirect effect from the disturbance as it merely responds to the PWM input signals. Unlike the other three signals, the CAN Split signal is not affected by the EMI and continues to produce a compliant result. This type of mask testing allows for multiple criteria to be rapidly tested in real time.
Figure 5. When subject to EMI, the simulated ECU PWM and Actuator Driver Output signals each exceed the tolerance mask test criteria and the scope notifies the operator that a deviation has occurred.
In addition to waveform mask testing, pass/fail limits are also applied to the parametric data to ensure that the numerical measurement results also comply with specified limits. Note that on the screen image in Figure 5, the scope has indicated the three deviations which have occurred denoted with the red "Fail" message on the screen under the test criteria. In the event of a mask failure or parameter limit failure, an oscilloscope can also automatically execute actions, such as saving the waveform data to be used for direct comparison and documentation, saving a screen image to be used for documentation and evaluation, generating a pulse out of the oscilloscope to assist with test automation, and sounding an alarm to inform the test operator when a deviation occurs.
Although oscilloscopes are well equipped to perform the rapid parametric measurements required for detecting deviations in EMC immunity testing, they have been often overlooked, mainly due to lack of awareness and lack of sufficient oscilloscope channel count. Using an array of oscilloscopes is potentially the most efficient and cost effective method to qualify ECU signal and actuator outputs during immunity tests, since most of the functionality using pass/fail mask and parameter limit testing has already been implemented, saving design engineers significant cost and time for functional testing compared with the costly software development time needed to implement custom data acquisition systems to perform the same rigorous test requirements of EMC deviation detection.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
