Engineer uses GPT-6 to script Altium + openEMS simulations that auto-optimize GND voids on high-speed PCBs

BLUECOW009 · x · 2026-09-07

An engineer shares a workflow where GPT-6 goes deep into PCB electromagnetic optimization: it reads stackup and pad info via the Altium PCB API, generates and runs openEMS simulations, and computes the optimal GND void size for large pads that cause impedance dips on high-speed lines, applying it automatically to the design.

Ground-plane voids under large pads minimize local impedance dips on high-speed designs. All pads currently receive the same expansion value; the author plans to port the setup to the palace EM solver to solve per-pad expansion values automatically.

The takeaway: AI is no longer just drawing PCBs — it's running simulations and optimizing design parameters on its own.

Original post →

More from coding & agent

coding & agent channel →