6.2 Control Loop Methodology

The following topic is thoroughly discussed in the Microchip University class “Charging Batteries From Solar”.

The solar charger control system design uses an unconventional-looking control loop compared to a normal voltage-mode power supply. Typically, the inner high-speed control loop focuses on the output of a converter as the key feedback parameter. This value is usually the output voltage, but in battery charging circuits, it can also be the output current. If the output voltage drops slightly, the duty cycle of the converter is increased to deliver more power to the load. The input power source impedance is expected to be low enough to provide whatever power is needed by the converter.

When charging from solar, the input impedance dynamically changes constantly with the variations in sun angle, temperature, position and most dynamically, the chance disturbance in the illumination due to a passing object. Something as simple as the shadow from a passing airplane, bird or even a falling leaf can increase the output impedance of a solar panel 100x for a moment. The result of the increase in source impedance on the converter is the immediate loss of input current and the subsequent voltage crash on the input capacitor. If the control loop is regulating an output parameter (voltage/current), then an input voltage feed-forward method must be incorporated at very high bandwidth to prevent the input voltage from crashing. This crash would cause the converter to shut off, or even worse, to run backwards as the input voltage drops below the output voltage. It is possible to increase the input capacitor to slow down the input voltage drop, but the capacitor size required is not physically practical or cost-effective.

To prevent these issues, this design uses the panel input voltage as the feedback parameter to the inner control loop. This keeps the panel voltage fixed until the converter makes a change in the operating point with the DAC-5 set point. If the illumination drops on the panel, the converter will respond by reducing the duty cycle. This is exactly the opposite of what a typical converter would do. This input voltage-based inner-loop control method results in high bandwidth responses to light perturbations without software intervention. The outer control loops, consisting of MPPT tracking and battery current/voltage monitoring, will adjust to the new illumination level within seconds. Outside the inner loop, parameter changes are low-frequency slow shifts in voltages, currents and operating points. The simple 8-bit microcontroller used can easily provide 10 Hz or better sampling and system updates.

Note that both output overvoltage and battery temperature monitoring are handled with dedicated hardware comparators (COMP-4 and COMP-6), which have direct access to the ‘Auto Shutdown’ input of the COG output that controls the MOSFETs. Even if the on-board firmware is inactive or in Sleep mode, these comparators will shut down the converter if required.