The Circuit That Remembers: How Memristors Could Slash Power Converter Energy by 92%

When you charge your phone or laptop, a tiny device inside the charger called a power converter is working hard to change wall electricity into the right format for your battery. These converters normally use computer chips called DSPs (short for digital signal processors) to control how electricity flows. Now, scientists have built a new kind of circuit that could make those chargers use far less energy.

The new design uses components called memristors, which are special because they can remember how much electrical charge has passed through them — sort of like how your brain stores memories. Because of this ability, memristors use much less power than traditional computer chips. Researchers found that their memristor-based circuit uses 92% less energy than a standard DSP-based system.

The team, led by Fanfu Wu, designed a circuit that operates at around 200,000 cycles per second — a range called kilohertz. More specifically, the circuit can be programmed to switch between 144.7 thousand cycles and 204.2 thousand cycles per second, giving engineers flexibility depending on what device they're building it for.

Power converters are everywhere: phone chargers, laptop adapters, electric car charging stations, and solar panel systems all use them. If even a fraction of these devices used the new memristor design, the energy savings could be enormous. The paper, published online as a preprint on arXiv, was submitted to the journal npj Power Electronics for peer review.

The researchers achieved their results using commercially available memristors — not custom-made laboratory components — which means companies could potentially scale up production. The design bridges a gap between pure digital control (flexible but power-hungry) and pure analog control (energy-efficient but harder to program). By combining the best of both approaches, the memristor circuit offers programmability without the energy cost.

The researchers noted that memristors can achieve up to 1,000 times less energy consumption in certain computing applications like neuromorphic computing, which mimics how the human brain processes information. The new design brings those energy-saving benefits to the practical world of power conversion — technology that billions of people interact with every day.