MiDEN Technical Insights

Near-Field Probe EMI Measurement Guide

Introduction

Near-field probe EMI measurement is a fast way to find which part of a switching converter radiates, and an easy way to fool yourself. Move the probe a few millimetres, rotate it, or let its cable brush the bench, and the spectrum can change by more than it does when you alter the circuit. The probe is not, by itself, a calibrated field meter; it is a small antenna whose output depends on distance, orientation, the field it couples to, and the current on its own shield. This guide explains what the probe actually senses, how those variables distort a reading, and how to set up a measurement you can trust when comparing a layout change or a filter modification.

Technical Analysis: What the Probe Actually Senses

Most near-field probes fall into two families. A magnetic loop probe responds to the rate of change of magnetic flux through its loop, so its output rises with frequency and with the area of the current loop it is looking at. An electric-field probe responds capacitively to the voltage on a nearby conductor; it is most useful on high dv/dt nodes, where the displacement current through the probe is largest. Both are directional: a loop reads maximum when its axis is aligned with the field, and an E-field probe reads maximum when its tip points at the source.

The near field is not the far field. Close to the board, the electric and magnetic components are not in a fixed ratio, and the probe reads whichever component it couples to most strongly. That is useful for localization and misleading for prediction: a strong near-field reading at 80 MHz does not prove the conducted emission at the mains terminal will fail, and a quiet reading does not prove the board is clean.

Treat the probe as a relative instrument. Unless it has a calibrated transfer factor and the geometry is controlled, the reading is a comparative number, not a field strength in V/m or A/m. That is enough to answer "which area is noisy" and "did the change help", provided the geometry is fixed. The general diagnostic sequence follows the Rohde & Schwarz discussion published by Power Electronics News in October 2026: correlate the spectrum with the switching event, then trace the path. The probe behaviour described below is standard near-field measurement practice, and the exact response depends on the probe and fixture.

Design Challenges: Height, Orientation and Cable Current

Four variables usually dominate the reading before the circuit does.

• Height. The near field falls quickly with distance. A probe held by hand at an unknown gap is not a repeatable instrument. Fix the gap with a spacer or a simple stand, record it, and use the same gap for every comparison.
• Orientation. A loop probe is directional. Rotate it to the maximum for a source, then record the angle and keep it. Comparing a vertical reading with a horizontal one is comparing two different measurements.
• Cable common-mode current. The probe cable shield can carry current and act as part of the antenna. Keep the cable route fixed and do not touch the cable during a sweep. A ferrite sleeve on the probe cable is a candidate control measure, not a universal requirement: compare a sweep with and without it and keep the configuration that gives a repeatable result.
• Receiver settings. Resolution bandwidth, span, reference level and preamp gain all change the displayed amplitude. Save the setup and recall it for every A/B comparison.

Two further effects are easy to miss. A probe close to a high-impedance node can load it capacitively and shift the frequency of the very ringing you are measuring. And an unshielded probe can respond to the electric field even when you are trying to measure a current loop. A useful control check is to move your hand toward the probe; if the reading changes, the setup is coupling to your body and the measurement is not yet under control.

Application Scenarios: Localizing a Source and Comparing a Filter

Two workflows justify the effort.

• Source localization on a SiC or GaN board. Scan at a fixed height over the switching loop, the heatsink, the transformer and the input cable. Note the frequencies where the probe peaks, rotate for maximum at each peak, and compare the peak frequency with the switching edge and ringing seen on the oscilloscope. The goal is a short list of physical areas, not a compliance number.
• A/B comparison of a filter change. Fix the probe relative to the same point, keep load, input voltage, cable length and receiver settings constant, and change one component at a time. A common-mode choke change, for example, should be judged at the input cable and the choke body with the probe in the same position before and after. If the probe moves between runs, the result cannot be attributed to the component.

Use both probe types when the coupling mode is unclear. A strong E-field response at an isolated heatsink is a diagnostic clue for capacitive coupling from a switching node; a strong loop response around a cable or inductor is a clue for a current loop. Treat these as clues, not conclusions: confirm the path with a controlled change, such as moving the probe, adding a temporary shield or changing one component, before deciding whether a shield, layout change, common-mode choke or differential inductor is appropriate.

Product Connection: Where MiDEN Magnetics Fit

Near-field probing does not replace the filter, but it tells you which filter position is worth changing.

• The common mode choke series addresses current that flows out and back through the cable and chassis. Probe the input cable and the choke body before and after a change to see whether the common-mode path moved.
• The power inductor series addresses the differential-mode path. Probe the switching loop and the inductor body when the differential component is the suspected path.
• Custom magnetic components are the route when the winding geometry, shielding or mounting has to change to move the field away from a sensitive area.

The broader suppression workflow is covered in the article on EMI suppression techniques in power supplies, and the server input-filter case in common mode choke for server power supply. The wider component set is listed under magnetic component applications.

Frequently Asked Questions

How high should a near-field probe be?
There is no universal number. Closer gives more signal but more loading and less repeatability. Choose a gap that gives a clear peak, fix it with a spacer, and record it so every comparison uses the same geometry.

Can I compare near-field readings between different boards?
Only if the probe, gap, angle, cable route and receiver settings are identical. Otherwise treat each reading as relative to its own setup, not as an absolute field measurement.

Why does the reading change when I touch the probe cable?
The cable shield can carry common-mode current and become part of the antenna. Fix the cable route and keep hands away during the sweep. A ferrite sleeve is one control to test, but it is not required on every probe; verify whether it changes the reading on your setup.