Purpose

A maximum power point tracker is used to extract the maximum amount of power from our solar panels. As conditions change, the optimal load characteristic to maximize power transfer changes. Since we cannot constantly change our load impedance (the battery), a DCDC converter is used to transform the impedance of the load. An algorithm adjusts the duty cycle of the DCDC converter to keep the panel at it’s optimal operating point.

We wish to stop giving MPPT manufacturers a lot of money, so we are trying to make our own. These should be interoperable with the current Nomura MPPTs since the rest of our system is designed for their use already.

Requirements

• At least 7A, 10V-30V input

• At least 40V output

• Analog control (FW doesn’t have bandwidth)

• Series operation and interoperability with SPV1020 (Nomura)

MPPT Control

Fundamentals

The relationship between the output current and voltage of a PV panel is described by an IV curve (red). The point along the IV curve at which the panel operates is called the operating point, which will change depending on the load. There is a specific operating point at Vmp which maximizes the product of voltage and current, power. As the irradiance on the panel changes, so will the IV curve and thus Vmp. The goal of an MPPT is to find and maintain the panel at Vmp in changing conditions.

MPPTs adjust the operating point of a panel using a DC-DC converter situated between the load and the panel. If the power consumed by the load is voltage dependent, such as the case of a resistor or battery load, the DCDC conversion ratio can be adjusted to change the input impedance of the converter, setting the operating point of the panel. For example, a 10V panel connected to a 10 Ohm resistor will result in 10W of power supplied. However, if this is below the 22.5W MPP of the panel, a boost converter can boost the panel voltage to 15V, allowing the load to consume 22.5W of power and consequently allowing the panel to supply 22.5W.

Incremental Conductance Method

The operating point of the panel along the PV curve can be determined by calculating dP/dV, where dP/dV > 0 means it is below the MPP, dP/dV = 0 means it is at the MPP, and dP/dV < 0 means it is above the MPP.

An analog multiplier can be used to obtain power. A high pass filter is used to obtain dP/dt and dV/dt, and a comparator compares the sign of the 2 time derivatives to obtain the sign of dP/dV. The output of the comparator indicates whether it is positive or negative, which is then used to adjust the operating point of the panel in the appropriate direction.

Operating Point Regulation

The DCDC converter employed uses voltage regulation to maintain it’s operating point. Typical DCDC converters use output voltage regulation, that is, they seek to maintain a set output voltage. Since the load power is related to the voltage applied to the load, if the panel is operating below MPP the output voltage of the DCDC can be raised to increase the power drawn from the panel and shift it’s operating point right towards the MPP, or vice versa.

Simulations with output voltage regulation show that it is much harder to maintain stability for a battery load. Due to the binary nature of the analog control system, the operating point of the panel is constantly changing. Thus, the output voltage oscillates around the ideal voltage which keeps the panel at its MPP. However, for a battery load, a small change in applied voltage results in a large change in received power, thus the input voltage of the PV panel oscillates a much larger amount ( >1V).

When the DCDC converter is set to regulate it’s input voltage, the only oscillations present are the input ripple of the converter and the small oscillations caused by the control system, greatly enhancing stability.

Note that to use a conventional DCDC controller to regulate the input voltage rather than the output, the feedback signal needs to be reversed. This is because on the output side, if the voltage is too low the converter needs to increase the duty cycle, thus increasing the supplied current to raise the output voltage. However, if the input voltage is too low, the duty cycle needs to be decreased to reduce the current drawn and shift left on the IV curve.

Next Steps

1. Using input regulation alone means that no output voltage regulation is required, but it is still needed as the maximum output voltage should not exceed the maximum charge voltage of the battery. Thus, a secondary feedback loop needs take over from the MPPT loop to limit the output voltage after it surpass set point by forcing it to operate below MPP. This control loop should not saturate the MPPT loop, otherwise it may fail to find the MPP when it is needed. Current regulation is built in to the DCDC controller so it is not something we need to add.

2. Investigate methods to prevent saturation of MPP controller 9409801.pdf (hindawi.com)

3. Investigate load current sensing to eliminate the need for a multiplier. Since a battery keeps the output voltage at nearly a constant voltage (in the timescales relevant to the control loop), the output power can be sensed by measuring the output current, which can be used to find dP/dV instead of multiplying panel input voltage and current.

This is the simulation we have so far, that you can add to for these next steps:

Power Stage

Function

Array configuration

Seems boost only MPPTs are sufficient for MS15, as even with the lowest battery pack voltage the boost ratio is at the minimum of 1.1. It is possible for Vmpp to be slightly higher if the irradiance is above 1000W/m2, but I don’t think that’s too much of a concern.

Since pack voltage can’t go any higher (limited by motors), a buck-boost topology will be needed for MS16 since array area is increasing to 6m2, or the array will need to be split into 3 strings.

Topology

Since we only need boost mode for now, we will stick with that as it is the simplest option. The same controller can be used in a SEPIC configuration if step-up/down is desired in the future.

Primary FETs

GSFN6982 Good-Ark Semiconductor | Discrete Semiconductor Products | DigiKey

Gate charge sensitivity vs Rdson at 50kHz assuming 4A input and 5.5V Vgs is ~30nC/mR, so basically gate charge is not a concern lower Rdson = better. The sensitivity can be scaled linearly to different parameters.

 Gate Charge Loss

Other things we need:

• Inductor: AGP4233-473ME (100khz) or -683 (50khz), larger

  • Lower frequency → higher efficiency (less switching loss), more mass (larger coils)

• Other simpler components in spreadsheet, seem feasible.

• ~4 mOhm at 7.5V

Current Progress:

MPPT with output regulation:

updated preadsheet with boost converter specs:

To investigate: not using a multiplier

• Potentially via output current sensing?

  • Battery should keep (in a short timeframe) stable voltage, so sensing the output current would give us a proxy for power

Next steps:

Create schematic