Detect and analyze EFT events

, , & -December 21, 2016

Energy bursts from electrical fast transient (EFT) pulses can couple into nearby electrical paths, risking digital signal corruption of electronic systems, leading to potential malfunctions such as unintended latching or resets. You can use an oscilloscope to capture, display, and analyze EFT events. Many EMC engineers are, however, unaware of this capability and the benefits it provides. By using features found on many oscilloscopes, you can trigger on EFT events even when a long time might pass between them.

You can also detect "runt" pulses caused by EFTs. Finally, you can calculate the energy in an EFT pulse. With that information, you can make design changes to improve EFT immunity. EFT events occur when current flow is instantaneously interrupted, resulting in arcing between contacts that can disrupt circuits and systems. The arcs create EM fields that can then couple into circuit paths through cables, traces, and connectors. Common causes for EFT include relay-contact bounce, opening and closing of circuit breakers, switching of inductive loads, and powering down equipment. Breakdown of the air gap between electrical contacts often triggers a rapid burst of EFT pulses.

Sequential capture

A sequential acquisition is ideal for capturing many fast pulses in rapid succession (such as EFT pulses) or when you need to capture a narrow slice of events separated by long time periods (such as EFT bursts). In sequential-capture mode, an oscilloscope can display a complete waveform that consists of a number of fixed-size segments. The oscilloscope utilizes the sequence acquisition by setting the desired number of segments, maximum segment length, and total available memory, which determines the actual number of events it can acquire.

The use of a sequential timebase mode has a double benefit for EFT analysis because it allows the capture of both complex sequences of events over large time intervals in fine detail, while also ignoring the uninteresting periods (long gap times) between events. You can make timing measurements between events on selected segments using the full precision of the acquisition timebase. Figure 1 shows a diagram of a sequential timebase capture operation.

Figure 1. Sequential timebase acquisition operation removes "dead" time between events of interest.

A sequential timebase capture automatically acquires time stamps for each segment, which lets you qualify the frequency of EFT burst events. In addition, utilizing a sequential timebase mode allows for the inclusion of advanced triggering to isolate a rare event, which you can use detect erroneous EFT pulse shapes.

Figure 2 depicts a series of EFT bursts acquired as segments in a sequential capture mode. Note that the process of sequential capture removed the long gap time between bursts, leaving only the desired burst waveform within the acquisition. Time stamp information is displayed for each burst, showing the date and time of acquisition, the time since the start of sequencing, and the intersegment time corresponding to the EFT interburst time.

Figure 2. Sequential-capture mode lets an oscilloscope display EFT bursts acquired as time-stamped segments.

In contrast, Figure 3 shows EFT pulses, rather than bursts, acquired as segments.

Figure 3. The oscilloscope captures EFT pulses, time-stamping them as segments.

The oscilloscope could potentially capture tens of thousands of individual pulses. Note that the time scaling for sequential burst capture here is 2 msec/division (corresponding to a 20 msec time capture window), while the time scaling for sequential pulse capture is 100 nsec/division (corresponding to a 1 µsec time capture window).

The intersegment time stamps in EFT burst capture mode shows an interburst timing of approximately 100 msec between bursts, while the intersegment time stamps in EFT pulse capture mode shows an interpulse timing of approximately 100 µsec between pulses. The time scaling between the two captures differs by a factor of 1000, highlighting the contrast between the characterization of either individual EFT pulses or EFT bursts.

Locate EFT Anomalies
Qualifying a unit to meet EFT immunity requires the use of a generator (Figure 4) to simulate rapid EFT pulses. To ensure that the EFT simulator generates correct pulse shapes and burst timings, you must test it with an oscilloscope. To identify potential problems, a sequenced acquisition can be obtained, and EFT anomalies can be detected using sequenced acquisition display modes (Figure 5). In this case, an EFT pulse with about half of the normal pulse amplitude is acquired in the sequence.

Figure 4. An EFT generator creates the test pulses.

Figure 5. A signal anomaly caused by EFT (left) also appears in the waterfall display (right).

The runt EFT pulse in Fig. 5 is quite visible compared with the normal-height EFT pulses using both the overlay display (left plot) and the waterfall display (right plot). A potential root cause of EFT pulse anomalies such as the one captured in Fig. 5 could include a bad contact within the transient simulator causing the pulse to not reach its full amplitude.

Figure 6 shows an acquisition of EFT bursts, captured on Channel 2 (pink, top left grid). Seven consecutive zooms (Z2 through Z8) show the big picture and the details of the EFT bursts, as each zoom shows increasingly more detail than the previous zoom. These zooms range from a view of a single burst to the view of a single pulse within a burst. The use of a multi-grid display format lets you see each trace within its own display grid area.

Figure 5. Parameter measurements calculate phenomena such as burst gap time and pulse energy.

Numerical Qualification of EFT Events
Thus far, we've focused on visualization techniques for identifying EFT pulse quality issues. Equally powerful for the characterization and debug of EFT events is the use of numerical parameter measurements.

Fig. 5 incorporates the use of eight measurement parameters, each of which provide further insight into the acquired EFT burst and pulses. By restricting the measurement region to a specific gated area, measurement parameter P1 characterizes the frequency of pulses within a single burst. Other measurement parameters can be independently gated over different areas of interest. In the measurement table (Figure 5, bottom left), parameter P1 reports an EFT frequency of approximately 10 kHz.

Because you can view the pulse edges between the end of one burst and the start of the next as a negative-going pulse, you can use a width-measurement parameter to automate measuring the idle time between bursts. Parameter P2 uses a negative-polarity pulse width measurement to determine this burst gap time.

Note that another measurement parameter of pulse width is used in P4, but has a positive pulse polarity selected, and so is measuring pulse widths rather than burst gap times. For this reason, the pulse width measurement result of P4 (EFT pulse width) is on the order of 150 nsec, while the pulse width measurement result of P2 (EFT gap time) is about 90 msec, a significant difference in magnitude.

The math operator, F2 (pink, upper right grid), is the square of channel C2, allowing the units to be output in V². Energy in joules is computed using the chain of math and measurements operators F2, P6, P7, and P8.

Starting with P6, area under the curve is determined by applying the area parameter to the squared function F2. Next, a measurement constant of 50 Ω is set in parameter P7. Finally, by dividing P6/P7 using the ratio operator in P8, energy in microjoules can be computed.


The acquisition of sequenced waveforms can both simplify the capture EFT events and provide assistance in visualizing anomalies and variation among pulses. Familiar measurement parameters, when applied in a new way (for example negative pulse width to quantify EFT burst time), and the chaining together of math operators and measurement parameters (for example to compute energy) can provide powerful and novel insight into electrical fast transient pulse characteristics.

Also see

Loading comments...

Write a Comment

To comment please Log In