Don't Route Over Ground Plane Splits — Part 4: Poor Man's 3D Field Patterns

Terrain maps

The 2D heatmaps in Part 3 work, but flat color can be hard to read. This time the field amplitude is mapped to height. Strong fields form mountains, weak fields form valleys. Color follows the same mapping. The result is a topographic map of the H-field directly above the board.

All three boards are normalized to the same scale. The tallest peak across all boards and frequencies sets the maximum. This means the heights are directly comparable: if the intact board shows a small ridge and the split board shows a mountain at the same frequency, that's a real difference, not a scaling artifact.

The green rectangle is the PCB. Drag to rotate, scroll to zoom, use the slider to step through frequencies.

Scan setup

Same scan data as Part 3, just rendered as topographic maps instead of flat heatmaps.

Good board

The intact board is quiet. The field concentrates in a narrow ridge above the signal trace. Step through frequencies and you'll see the same pattern: a tight line that drops off steeply on either side. The return current flows directly under the trace, keeping the loop area small and the radiation contained.

Bad board

Very different picture. The split forces the return current to detour around the slot, creating a large loop that acts as an antenna. The terrain shows broad mountains covering most of the board. At many frequencies the whole board lights up. Compare this to the intact board at the same frequency and the difference is obvious.

Bad board with stitching capacitors

The stitching caps bridge the slot with a low-impedance path for the return current. The terrain sits somewhere between the other two boards. At most frequencies the caps pull the field down compared to the bare split. But at certain frequencies a ridge appears right where the caps sit. The red arrow marks the cap location when this happens. The caps aren't a magic fix. They introduce their own resonances.

What the terrain shows

• Intact board: low ridges along the trace. Tight, contained fields.
• Split board: tall, broad mountains. The field spreads across the whole board.
• Caps board: reduced peaks compared to the split. The caps help, but don't fully recover the intact board's performance.

The same scale across all three makes the comparison direct. Pick a frequency on one board, then open the other two at the same frequency. The difference in terrain height shows the difference in corrected near-field probe response at that scan height.