Active Current Balancing for the Current-Sensorless Interleaved Boost Converter Under Severe Branch Asymmetries

In the world of modern electronics, power is often managed by devices called converters, which act like efficient translators, taking electricity from a source and adjusting it to the exact voltage a device needs to run. One popular design, known as an interleaved boost converter, works by splitting the heavy lifting of this job between two parallel pathways, or branches. Imagine a team of two porters carrying a heavy load; if they share the weight equally, they can move faster and with less strain on their bodies. In these electrical circuits, this equal sharing reduces heat and wear on the components. However, in the real world, no two branches are perfectly identical. Tiny differences in the materials used to build the inductors and resistors within each branch can cause one side to carry far more current than the other. This imbalance is like one porter straining under most of the weight while the other barely lifts a finger, leading to overheating and potential failure. Traditionally, engineers have solved this by installing sensors on each branch to measure the current directly, but these sensors add cost, complexity, and points of failure to the system.

The challenge, then, is to keep the two branches working in harmony without the need for these extra sensors. A researcher at Northern Illinois University has proposed a new way to achieve this balance by looking only at the final output voltage of the converter. The core idea relies on a subtle detail: even when the system is running smoothly, the voltage at the output is not perfectly steady. It ripples, rising and falling in a rapid, high-frequency pattern. Because the two branches switch on and off at slightly different times, these ripples create tiny, distinct peaks in the voltage waveform. The researcher discovered that if the two branches are carrying unequal currents, the height of these peaks will differ. By carefully measuring the output voltage and using a mathematical filter to reconstruct the internal behavior of the circuit, the system can detect these peak differences. A simple control loop then makes a tiny adjustment: it slightly increases the "on" time for the weaker branch and decreases it for the stronger one. This subtle nudge forces the currents to equalize without ever needing to measure the current itself.

To test this idea, the researcher ran detailed computer simulations of a two-phase converter operating under difficult conditions. The setup was designed to mimic a worst-case industrial scenario where the components in the two branches were significantly mismatched, with differences as large as twenty percent in their electrical properties. The simulations covered a wide range of operating conditions, testing how the system performed when the duty cycle—the ratio of time the switch is on versus off—was varied from low to high values. In every scenario where the branches were unbalanced, the system without the new balancing method showed a severe mismatch, with one branch carrying nearly forty percent more current than the other. This level of imbalance would typically cause significant inefficiency and heat buildup.

When the new active balancing method was engaged, the results changed dramatically. For most of the tested operating ranges, specifically when the duty cycle was between 0.45 and 0.60, the system successfully reduced the current mismatch to less than one percent. In these cases, the two branches shared the load almost perfectly, despite the underlying hardware differences. The system also proved robust against changes in the output capacitor's properties, maintaining balance even when the capacitor's internal resistance varied. Crucially, the method achieved this balance while keeping the output voltage steady, ensuring that the device being powered received a consistent supply of electricity.

However, the simulations also revealed a limit to how far this technique could be pushed. When the duty cycle dropped below 0.40, the effectiveness of the method began to fade. In these lower ranges, the voltage ripples became so small that they were overwhelmed by other electrical noise and the inherent resistance of the components. The difference in peak heights no longer provided a clear signal about the current imbalance, and the balancing error rose back up to around twelve to twenty-three percent. This suggests that while the method is highly effective for a broad range of standard operations, it is not a universal solution for every possible setting. The findings indicate that for this specific type of converter, there is a practical lower boundary around a duty cycle of 0.40, below which the relationship between the voltage peaks and the current balance breaks down.

The study concludes that this output-voltage-based approach offers a viable path toward sensorless current balancing, provided the system operates within its identified boundaries. The research establishes that it is possible to correct severe hardware asymmetries using only the information available at the output terminal. While the results are currently based on computer simulations, they provide a strong foundation for future work. The next logical step would be to build a physical prototype to see how the method holds up against real-world noise and measurement errors, and to explore whether the technique can be extended to converters with more than two parallel branches. For now, the work demonstrates a clear and elegant way to keep electrical systems balanced without the need for extra hardware, turning a complex control problem into a matter of listening closely to the voltage itself.