Effective Ways of Measuring “Noisy Ground” Currents

Introduction

If your job involves electronics design, system development, electromagnetic interference (EMI) testing and troubleshooting, you might encounter the terms "noisy ground" and "clean ground." Technically, there is no such thing as "noisy or clean ground." However, these terms are commonly used in the engineering world despite their inaccuracies. They often give engineers the impression that the earth/ground is an infinite sink of electrical noise, disregarding the physics that dictate all current must flow in loops.

Despite our reluctance to accept these terms, their widespread use in engineering cannot be denied. The reality is that certain structures—such as the ground plane on a PCB, chassis in a vehicle, or the green/yellow earthing wire in a large system—can carry unwanted or uncontrolled return currents. The crucial point here is the uncontrolled nature of this current.

When a dedicated return path is provided for power/signal transmission, only a small amount of "noisy current" will be present on other conductors. However, errors in PCB design, product assembly, or system installations can lead to a lack of a dedicated ground, causing return current to appear on other conductors. This phenomenon is known as common mode current, and the affected conductors become the "noisy ground." It's important to note that common mode current on a conductor can directly contribute to radiated emissions.

Suppose you encounter an EMI-related issue, whether it's related to passing a regulatory EMC test or ensuring the signal and power integrity of your product. Your next step is to diagnose the issue, akin to doctors treating a patient. This involves identifying the noise and locating the source of the "noisy ground".

Because the noise is radio-frequency (RF) by nature, with frequency contents ranging from a few kHz to the GHz range, measuring the noise presents a challenge, especially for those unfamiliar with RF engineering. This article presents four useful techniques that engineers can apply:

Technique 1: Using a magnetic field loop with a wire[1]

One method of locating noisy areas of a circuit ground is to attach a wire acting as an antenna and observe if it carries substantial common-mode current. Figure 1 demonstrates this concept. A square magnetic loop attached to a wire checks a ground pin of a connector for its ability to drive common-mode current on a cable. The magnetic loop is a homemade unshielded type (a commercial shielded type will work as well), and the loop is attached to the wire by paper masking tape. The wire length is about 30 cm long, a length useful in the 150 to 200 MHz range (as the wire behaves as a quarter-wave length antenna). The magnetic loop size, such as the one demonstrated in the picture, has a frequency response ranging from about 100 MHz to just above 1 GHz, so it is sufficient for this kind of work. A spectrum analyser with a 50 Ohm input impedance is usually used for frequency analysis. As can be seen from the picture, the green trace shows the ground noise measured.

It should be noted that there are two drawbacks to this technique. One is that the loop and the small wire themselves could couple noise from the board (via near-field coupling), making this method impractical in a system where many noisy components surround each other. The other problem is that it could potentially load the circuit under investigation (see details later).

Figure 1 Demonstration of measuring ground noise on a PCB ground connector pin

Technique 2: Using an RF current probe with a ground wire

Based on the same principle, this method is perhaps much better suited for troubleshooting ground noise issues in large systems (such as large-sized cabinets or vehicles). One can connect a one-metre-long wire to the ground that could potentially be noisy, place an RF current probe around this ground lead, and check the common-mode noise. If a one-metre-long wire is used, then the effective frequency range will be up to 100 MHz. Using a shorter wire means the frequency range can be extended to a few hundred MHz.

When using this method, one might need to slide the current probe along the wire to different locations. Depending on the frequency contents you are trying to measure, there could be a standing wave on the cable. Therefore, it is always worth checking a few points along the wire.

Again, a spectrum analyser can be used, but an oscilloscope with a 50-ohm input impedance can also suffice. This is useful where engineers may carry a battery-powered oscilloscope, but the bandwidth of the scope needs to be at least 200 MHz in this case. Figure 2 demonstrates this concept, where the RF current probe is connected to an oscilloscope. In this example, the heatsink of the device is not grounded to the 0V of the PCB, but rather, the heatsink is the chassis. As a result, a large amount of RF current was observed in the system ground.

Figure 2 Using an RF current probe to measure ground noise

Technique 3: Using a homemade resistive current probe [2]

In this example, during an electric vehicle's charging test, significantly high levels of noise caused conducted emission failure. We measured the surface current on the chassis bracket in the vehicle, where the earth wire was connected, using a homemade resistive current probe. This type of probe was first introduced in [3], and its construction is shown in Figure 3. Essentially, the probe is a resistor network built with a coaxial cable, it works to about 300 MHz (just the region where cable radiation is usually a problem). For readers who are interested in how this probe works and how to construct one yourself, please refer to reference [3] for details. It was observed that when the onboard charger was operational, the low-frequency noise on the chassis was significantly high, as shown in Figure 3(b).

Figure 3 Using a homemade resistive probe to measure the surface current on the chassis

Technique 4: Using a surface current probe:

Both Techniques 1 and 2 have potential issues. Physical contact could "electrically load" the board/system, potentially affecting its operation. Therefore, a non-contact measurement technique is preferred. A surface current probe, essentially a half-donut-shaped RF current probe, has proven effective (it works by magnetic field coupling between the surface current on the metal structure and the magnetic pickup coil, therefore, there is no hard wire connection).

In Figure 4, we demonstrate how a surface current probe can effectively pick up high-frequency RF current on a metal surface. In this case, a shielded cable is compromised by a bad connection on the connector side, causing surface current to flow on the enclosure. A surface current probe picked up the noise on the surface.

Figure 4 A surface current probe is used to measure noise on the enclosure case

Conclusion

When common mode current begins to flow through uncontrolled structures—be they referred to as earth, ground, or 0V—they can result in emission issues. Identifying and locating the source and coupling path is essential for ultimately resolving these noise issues. Hopefully, the four ground noise measurement techniques discussed in this article serve their intended purpose.

Reference

[1] D Smith, A Measurement Technique for Locating EMI "Hot" Areas on Boards or Systems, [Link] https://www.emcesd.com/tt2006/tt080106.htm

[2] M Zhang, Troubleshooting Low-Frequency Common Mode Emissions, Sigal Integrity Journal, [Link] https://www.signalintegrityjournal.com/articles/2965-troubleshooting-low-frequency-common-mode-emissions

[3] D Smith, A Resistive Current Probe, [Link] https://emcesd.com/tt070100.htm