Optimizing power efficiency and density in renewable energy and industrial systems with real-time microcontrollers

Power electronics designers work to improve power efficiency and power density in industrial and automotive systems, including multi-axis drives, solar energy, energy storage, electric vehicle charging stations and electric vehicle on-board chargers.

One of the main design challenges for these systems is to achieve better real-time control performance while reducing system cost. A common approach to address this challenge is to use highly integrated microcontrollers (MCUs) with ultra-low latency control loop processing for analog and control peripherals.

Control performance in real time: latency is key

Before we dive into application examples, let's take a brief look at "latency". In multi-axis drives, robots, photovoltaic inverters with energy storage systems, EV charging stations and electric vehicles, control performance is directly related to how quickly the MCU can sample, process and control signals. Figure 1 illustrates the relationship between a real-time signal chain and signal latency, which is the time from when an analog-to-digital converter (ADC) measures a signal, to when the CPU processes the information, and when a pulse-width modulator (PWM) controls power. This time needs to be as small as possible to achieve ultra-low latency control loop processing.

Figure 1 : Concepts of real-time performance and latency

For digital power supplies, achieving higher power density means increasing the DC/DC switching frequency from 50kHz to 100kHz, 500kHz or higher. If you are using an MCU running at 100MHz and the regulation loop is synchronized to the PWM frequency, at 10kHz the number of available CPU cycles between PWM interrupts is 10,000, which drops to 1,000 at 100kHz. As frequency increases, the time available to detect process control decreases, so you need to optimize the MCU architecture to save as much time as possible on each cycle in the real-time signal chain.

Enabling next-generation power supplies in photovoltaic inverters and energy storage systems

As shown in Figure 2, the photovoltaic inverter market continues to develop, and hybrid inverters with integrated energy storage systems have emerged, bringing the challenge of controlling bidirectional energy conversion. The single-chip architecture requires the use of an MCU with many high-resolution PWM channels and additional high-bandwidth ADC inputs, such as the TMS320F28P650DK C2000™ 32-bit MCU.

Figure 2 : Photovoltaic inverter architecture with integrated energy storage system

To meet the growing demand for renewable energy in many applications, photovoltaic inverters require higher power efficiency and better total harmonic distortion performance. One approach is to use newer multi-phase multi-level inverter power architectures. This type of architecture is typically implemented with a complex set of power algorithms and additional external logic (such as complex programmable logic devices or field-programmable gate arrays) to safely turn the power switches on and off using the correct sequence. This approach increases board space and system cost.

MCUs that support on-board customization, minimal dead-band, and illegal combinational logic (MCU features used to prevent destructive power-up/power-down sequences) in different PWM modules allow designers to reduce or even move Remove external logic, further simplifying the design.

In addition, it is important to tightly couple the PWM unit and the integrated analog window comparator to provide overcurrent and overvoltage protection for the power converter. Depending on the power topology, the MCU you choose may need to be equipped with a PWM unit capable of achieving peak and valley current mode control of the resonant mode converter.

Easier, faster integration in EV on-board chargers

As the number of electric vehicles grows worldwide, designers need to find new solutions to further integrate on-board chargers and reduce their costs. A typical implementation is two MCUs isolated from each other, one for vehicle charger power factor correction and the other for vehicle charger DC/DC.

Although using a single MCU increases the isolation components required to send signals back to the MCU, the increased cost is offset by savings from reduced component count, including fewer CAN transceivers, voltage regulators, power management ICs, op amps, and Implement the isolation required for communication back to the host MCU.

Figure 3 shows a single MCU controlling a three-phase onboard charger power stage topology up to 22kW. The PFC stage is a two-phase interleaved totem pole, while the DC/DC stage is a dual capacitor-inductor-inductor-inductor-capacitor (CLLLC), which reduces transformer size and field-effect transistor current levels.

Figure 3 : Three-phase electric vehicle on-board charger ( PFC vs. DC/DC ) controlled by a single MCU

After determining the minimum MCU hardware resources required (PWM, ADC, comparators), you may also want to achieve more software integration while reducing CPU overhead. Because integration enables sampling of more signals on a single device, selecting an MCU such as an ADC that includes built-in hardware-based oversampling and offset calibration simplifies software design, resulting in a more cycle-efficient MCU and Ability to run control loops faster.

Another challenge is software integration of multiple tasks with different real-time constraints: PFC, DC/DC as well as auxiliary control and safety need to coexist, making software development even more complex.

Moving from a single-core MCU to a multi-core MCU architecture and allocating memory, PWM and analog resources among the MCU cores can help allocate different control loop frequencies to multiple cores , for example, one core controls the PFC and another Run two CLLLC. Each core runs the control loop at a different independent frequency: the totem pole is usually a fixed frequency, but the onboard charger’s DC/DC power conversion stage (Figure 3) is constantly changing. Using a multi-core architecture also enables more reliable and sophisticated overcurrent and overvoltage protection (because each control loop can be optimized for each core), eliminating the need for external monitoring components and reducing costs.

Electric cars will be fully charged in minutes, every home will use photovoltaic and energy storage systems, factories will use more efficient robots and automate with a smaller energy footprint... Innovations in real-time control MCUs will enable cleaner , paving the way for a safer and more efficient world.