EMC summit gives engineers tips and fundamentals

An IEEE EMC Society chapter provided training on FCC Part 15, shielding effectiveness, and EMC art & science.

Boxborough, Mass — Interference among electronic devices never goes away. Engineers who aren’t EMC experts seem to need constant introductions and reminders of how to reduce interference because every design is different. EMI/EMC issues keep popping up, which keeps EMC engineers employed putting out fires.

On September 27, engineers gathered at test lab NTS (now Element) for the 2023 EMI/EMC Summit, organized by the IEEE EMC Society Boston Chapter. The summit consisted of four one-hour presentations and a tabletop show featuring EMC-related products.

FCC Part 15
The sessions opened with a presentation from Element’s Duy Ho on FCC compliance. Ho discussed the process of compliance as it related to EMC and specific absorption rate (SAR), a measure of RF exposure in humans (Figure 1). The most common radio compliance rule for EMI is Part 15 from the Code of Federal Regulations (CFR) Title 47. Part 15 covers covers emissions. Unlike in many other countries, FCC regulations don’t require EMC susceptibility tests.

Figure 1. EMC testing for FCC compliance consisted of radiated emissions and SAR measurements.

In describing the compliance procedures, Ho noted that ANSI standards specify the testing details. “Standards are open to interpretation,” noted Ho. “Labs and companies might have differences that need to be worked out.”

A company that plans to get a product tested for FCC compliance must develop a knowledge database (KBD) that describes the product’s design and construction details. A company may perform pre-compliance tests and present results to the test lab, whose engineers know the regulations. A test lab might provide feedback. “The tradeoff is test versus cost,” said Ho.

Ho then described the kind of test facilities that a lab might provide: test chambers, antennas, signal generators, amplifiers, and spectrum analyzers. Test chambers cover 9 kHz to 300 GHz. The size of equipment under test (EUT) determines the size of the test chamber because the distance from the EUT to the test antenna greatly affects emissions measurements.

Ho concluded his talk by noting that 5G is still new. Companies building 5G equipment can go to a Telecommunication Certification Body (TCB) for design reviews and testing. The TCB may have to submit the device to the FCC for final certification.

EMC engineer Scott Carlson followed Ho with a discussion on shielding effectiveness. In his talk, Carlson used a connector as an example. He also covered how such measurements are performed on shielding panels used in test chambers.

After covering relevant standards, Carlson explained that some standards require testing in a GTEM cell while others use a reverb chamber. For connectors, ANSI-EIA-364-66A-2000 specifies a reverb chamber, while SCTE 48-1 specifies a GTEM cell. The chamber, whose insides are very reflective, uses a tuner to steer the energy inside the chamber. Carlson explained that both methods produce consistent test results.

Shielding effectiveness
Carlson then explained how to measure the shielding effectiveness of a panel for materials you might use in a shielded enclosure for pre-compliance EMC testing. Figure 2 shows the test setup. A transmit antenna inside a chamber illuminates the panel under test with RF energy of known frequencies and power. The receive antenna and EMI receiver or spectrum analyzer outside the chamber measure the energy that leaks through the panel. The ratio of received power to transmitted power represents shielding effectiveness.

Figure 2. Copper pipes shield test cables that must enter EMC test chambers.

As with any EMI test chamber, cables need to pass through its walls. Carlson showed an example of using a pipe that shields a cable and holds it in place during a test (Figure 3). The construction may not be pretty, but it works.

Figure 3. A test setup for a panel’s shielding effectiveness uses a transmit antenna in a chamber and a receiver outside the chamber.

EMC art and science
The afternoon’s two sessions featured longtime “Cheerfully mostly retired” EMC engineer Colin Brench, who discussed the art and science of EMC design (Figure 4).

Figure 4. According to Colin Brench, EMC design and test is a combination of art and science.

“Anyone can do DC,” Brench told the engineers. That’s when he displayed the diagram in Figure 5. In the DC world, you close the switch, and current flows through the load — end of story. That eventually happens, but first, a mechanical switch bounces, creating high-frequency electric fields that can radiate from the wires, which act as antennas. That’s nonlinear behavior. Sometime later, you get a DC current, but only with a stable, resistive load. Today’s loads have impedance that can change rapidly.

Figure 5. A simple DC circuit has AC characteristics when the switch closes.

As Figure 4 shows, models make up the science of EMC. We use them to model circuit and field behavior. Brench asked, “What are your tools’ capabilities, and how much detail do you need to accurately model a circuit?” That depends on what circuit behavior is important. If all you care about in the Figure 5 circuit is DC, then all you need is Ohm’s Law to calculate current.

If you’ve been around EMC engineers, you may have heard that “ground is for carrots and potatoes,” to quote Bruce Archambault. EMC and RF engineers think in terms of return paths and their impedances. Those impedances, combined with changing currents, create EMI. “Current is what matters,” said Brench. “Voltage doesn’t exist,” he continued. Voltage is derived from the integral of electric fields. “Maxwell’s equations refer to fields and current.” After all, the base unit for electricity is current, not voltage.

When it comes to modeling, Brench noted that “you can do things perfectly well and be perfectly wrong.” By that, he referred to doing measurements to back up your models. “Are you a modeler or a doer? You need to have some idea of what’s right, and that means measurements.” He cited the difference between older and younger engineers by saying that young engineers model while old engineers build and take measurements. Brench recommended that you take any measurements that you can before bringing your product to a test lab.

Brench then turned his sights to electrostatic discharge (ESD). Having done ESD work on 224 Gb/sec electrical data links, Brench noted what ESD can do to the system running at those data rates. The signals operate at 56 GHz using PAM4 modulation where data is clocked on rising and falling edges, two bits per symbol, thus 224 Gb/sec.

You may think of ESD as a high-frequency event. It was, but not compared to today’s signals. As Brench noted. Today’s high-speed serial data links can have 10 to 20 transitions in the rise time of an ESD pulse. By today’s standards, ESD takes a long time. Thus, system firmware and software must take that into account when encoding data.

In between the technical sessions, attendees had a chance to meet with local representatives of EMC test equipment and suppression product manufacturers (Figure 6). While most tables displayed data sheets, Absolute EMC’s table included a demonstration of how AC mains filters can reduce emissions radiating from power cords.