The 1968 GE Patent That Founded Solid-State Transformer Technology
- US Patent 3,517,300 filed by William McMurray at GE Schenectady in 1968
- McMurray named it 'electronic transformer'—first documented SST concept
- Design replaced iron core with a high-frequency converter and compact HF transformer
- Patent predated practical GTO and IGBT switching devices by 15-20 years
- Concept remained theoretical for four decades until semiconductor costs fell enough to validate it
William McMurray, a research engineer at General Electric’s Schenectady campus, filed United States patent 3,517,300 in 1968. The patent proposed an “electronic transformer” that replaced the iron core of a conventional transformer with a high-frequency converter stage and a compact magnetic link tuned for kilohertz operation. Fifty-seven years later, that patent is the foundation of every commercial solid-state transformer (SST) shipping in 2026.
The motivation behind the McMurray patent was specific: European railways operated at 16.66 Hz, which forced locomotive transformers to be several times larger and heavier than the 50 Hz or 60 Hz utility units of the same kVA rating. A high-frequency conversion stage promised a transformer at a fraction of the weight and volume.
The architecture in the patent is recognizable in modern SST designs. Thyristor switches modulate the input at a chosen high frequency. A medium-frequency transformer (MFT) provides galvanic isolation and step-down at the chosen frequency. An output stage reconstructs the low-frequency or DC waveform required by the load. Today’s commercial topologies (modular multilevel, dual-active-bridge, AC-DC-AC two-stage) all trace to this skeleton.
The name “electronic transformer” did not survive. In 1980, Brooks coined the term “solid-state transformer.” In 1995, the Electric Power Research Institute (EPRI) launched its Intelligent Universal Transformer (IUT) program, the first attempt to commercialize an SST as a distribution-grid building block rather than a traction-system component.
The McMurray architecture languished commercially for almost 60 years because iron-core transformers, at 99 percent efficiency and zero moving parts, were impossible to beat on cost-per-kVA. Three converging loads in 2024 and 2025 finally flipped the equation: gigawatt-scale AI data centers, megawatt-class EV charging, and DC-native renewable interconnects, all of which value the SST’s active power-quality control and DC-bus output over the iron transformer’s raw cost advantage.
ABB, Hitachi Energy, GE Vernova (the direct corporate descendant of McMurray’s employer), Eaton, and a wave of US startups including DG Matrix, Amperesand, and Heron Power are now shipping or developing products derived directly from the architecture McMurray sketched on a patent drawing in 1968.
Why It Matters
Utility procurement officers and grid planners benchmarking SST proposals should treat the McMurray lineage as engineering provenance, not as evidence of operational maturity. The relevant question is whether a given vendor’s 2026 product has accumulated enough megawatt-hours of field service to support a 25-year asset-life decision, not whether the underlying topology was patented six decades ago.
Critical Perspective
The McMurray patent is rightly celebrated as the conceptual origin of the solid-state transformer, but several caveats apply when reading commercial SST coverage that traces a clean line from 1968 to 2026. The patent itself was a niche solution for railway traction, not a general-purpose distribution transformer; commercial railway adoption took another four decades and never displaced the iron-core unit in mainline service. The semiconductor switching speeds, voltage ratings, and thermal limits available to McMurray in 1968 ruled out medium-voltage utility deployment, and the same is partly still true: every shipping medium-voltage SST in 2026 still relies on series-stacked silicon carbide modules and modular multi-level architectures rather than a single high-voltage device, exactly the workaround McMurray would have needed. Coverage that frames the 57-year gap as validation obscures that most of the gap was waiting for materials science (SiC, GaN), not waiting for engineers to take McMurray seriously. Vendors with a commercial SST to sell have a clear incentive to invoke the McMurray pedigree; readers should weight that incentive when evaluating shipping-product claims against laboratory results.
Related Coverage
Concept remained theoretical for four decades until semiconductor costs fell enough to validate it
