The Nomura MPPTs evaluated by the team have a fairly unique design compared to most other solar car MPPTs. Instead of having each MPPT designed to output the full battery pack voltage, Nomura MPPTs are designed to only output up to 37 V and must be stacked in series on their output sides in order to charge the battery. This configuration is advantageous because it enables each MPPT to have a smaller string of PV cells on its input side, making the entire array significantly more robust against uneven lighting. Moreover, due to the large surface area of the SunPower cells being used on MSXII, the total stackable cell voltage for the entire car is quite low and would only support around 2 strings for traditional MPPTs.
However, this also posed some interesting and potentially confusing questions about how MPPTs in a "stack" coordinate together to output the correct pack voltage. After some testing with 3 Nomura MPPTs, this article has been written to document our best understanding of how they work to hopefully reduce confusion about these devices moving forward.
Considerations for MPPT stacking
When stacking MPPTs in series, it is crucial that the total voltage across the whole stack remains around nominal. This becomes a challenge in this configuration because the series configuration forces all MPPT output currents to be equal. In order to maximize total power utilization from each MPPT's input string in situations where the strings are producing different amounts of power, each MPPT is designed to be able to adjust its output voltage depending on its input power such that the total stack's nominal voltage is distributed by the ratios of each MPPT's input power, not just equally.
This begs the question: when one MPPT lowers its output voltage due to a weaker input string, how do other MPPTs in the stack "know" to increase their output voltage, while also ensuring that under normal conditions the stack voltage never exceeds the pack voltage?
Modes of Operation
To answer this question, it is important to first understand the three operating regions of the Nomura MPPT. First, consider a simplified design of the MPPT that does not need to take into account pack overvoltage, and can operate at an unlimited boost ratio (output/input ratio). Furthermore, assume there is only one MPPT in the series stack (the MPPT directly connects to the battery). For a given input power, the MPPT should try to output as close to this input power as possible. The output voltage is constrained by the state of charge of the battery.
From basic power equations, the MPPT's output current should be a 1/x function of its required output voltage. Physically, the boost converter on the MPPT will try its hardest to fill the output capacitor until the pack draws the same amount of current that the boost converter is able to provide (itself being limited by the input side PV cells), at which point equilibrium is reached. We assume the MPPT is never capable of supplying enough current to cause overcurrent.
This, of course, results in issues when the battery pack is nearing full charge and is drawing a low current from the MPPT. If the MPPT were to continue to try to maintain the same output power, its output voltage would increase without bound and eventually cause a battery overvoltage. The Nomura MPPT thus implements a voltage limiting function, tuned via potentiometers, which prevents the boost converter from continuing to charge up the output capacitor once it has reached a preset limit.
Additionally, because the MPPT is only capable of boosting, not stepping down the voltage of the input side PV string, there exists a feed-through mode where the MPPT will directly connect the input string to the output in situations where the output voltage is less than the minimum boost voltage, 1.1 x Vin. This bypasses the boost circuitry entirely and results in no power point tracking, but elegantly handles a short-circuit and can still deliver some power in certain cases.
Outside of the overvoltage-limit and feed-through modes, the MPPT is in constant power mode, where its output current is directly determined by the output voltage and input power. All three regions are shown in the drawing below. Note that the IV curve can be idealized as a doubly-truncated 1/x curve, where the untruncated region corresponds to the constant-power region.
