by Kenneth Wyatt and Steve Sandler

Published by Signal Integrity Journal, July 28, 2022

The question on whether the ground return plane should be cut away under the switch node or inductor of DC-DC converters has been an ongoing debate. 

The argument for an adjacent solid plane under all converter circuitry has been to contain the electromagnetic fields in the dielectric space between the circuit traces and return plane; that this would prevent the spread of EMI around the board.

Others argue the capacitive coupling between circuit traces and return plane should be cut away to minimize the capacitive coupling due to large dV/dt swings of the converter switch node with corresponding contamination of the return plane with EMI. In addition, there is concern for induced eddy currents on any solid plane below the inductor. Although, any eddy currents generated by the inductor tend to self-cancel and reduce the overall field level from the inductor (depending on the type and mounting orientation).

My colleague, Steve Sandler, decided to investigate the issue by designing four identical buck converter circuits (Figure 1) with component layout (Figure 2), built on a two-layer stack-up (Figure 3). The experiment was designed to compare (1) a solid return plane with, (2) a cutout at the switch node pad, (3) a cutout under the switch inductor, and (4) a cutout under both the switch node and inductor. See Figure 4 for the four return plane designs.

Converter Design

Figure 2. A graphic diagram of the component layout.

Figure 3. The stack-up design used. The bottom layer was solid, except for the voids added.

The four boards were labeled as (see Figure 4):

1-1(solid return plane)

0-1(cutout under SW node)

1-0(cutout under inductor, L1)

0-0(cutout under both SW node and L1)

EMC Testing

The two biggest questions in my mind were how the performance of the four boards compared for radiated and conducted emissions.

Click here to read entire article.