How a Solar Inverter Works: DC to AC and Grid Sync

A solar inverter sits between a photovoltaic array, which produces direct current, and a building’s alternating-current wiring. Its job is not simply to change one voltage into another. It has to extract the most power the array can give at any instant, shape that power into a sine wave that matches the grid in frequency, phase and amplitude, and stop feeding the grid the moment the grid is no longer present. Those three jobs are separate mechanisms, and most inverter behaviour that looks strange on a monitoring graph comes from one of them.

The switching stage: building a sine wave from DC

The array delivers DC at a voltage that varies with temperature and irradiance. The inverter’s power stage turns that into AC by switching semiconductor devices on and off thousands of times per second. A simple square wave would be useless; the switching is controlled so that the average output follows a sine wave, and a filter on the output removes the high-frequency switching components, leaving a smooth 50 or 60 Hz waveform.

Two topologies are common. A transformerless inverter avoids a bulky iron-core transformer and tends to be lighter and slightly more efficient, but it must manage the DC path to the grid carefully. An inverter with a high-frequency transformer isolates the array from the grid electrically, which some installations require. Neither is universally better; the choice is a design decision about isolation, weight and local code.

The switching devices are the parts that carry the current and dissipate heat, and they are the reason an inverter has a temperature derating curve. On a hot afternoon the inverter may reduce output not because the array is weak but because its own electronics are near their thermal limit.

Maximum power point tracking

A photovoltaic cell has a current-voltage curve. At any given irradiance and cell temperature there is one operating point on that curve where the product of current and voltage, and therefore power, is highest. That point moves: as the cell heats up, the voltage at the maximum power point falls; as light increases, the current rises. The maximum power point tracker is a control loop inside the inverter that continually adjusts the operating voltage to sit on that peak.

A common method is perturb-and-observe: the tracker nudges the voltage slightly, measures whether power went up or down, and moves in the direction that increased it. Under uniform light this converges quickly. Under partly shaded conditions the power-voltage curve can have more than one local peak, and a simple tracker can settle on a lower one. That is why shade behaviour differs between inverter types and why module-level electronics exist.

The tracker’s effect is visible in string sizing. A string of modules must stay within the inverter’s input voltage window across the range of temperatures the site will see. On a cold, clear morning the open-circuit voltage of a silicon string rises; on a hot roof at midday the maximum-power voltage falls. The inverter cannot track a point outside its window.

Grid synchronisation and anti-islanding

A grid-tie inverter does not create its own frequency. It measures the grid’s voltage waveform and injects current in phase with it. Before connecting, it checks that the grid voltage and frequency are within limits; after connecting, it continuously adjusts its output to stay synchronised. If the grid drifts, the inverter follows until the drift exceeds the protection settings, at which point it disconnects.

Anti-islanding is the requirement that the inverter stop energising a section of wiring when the grid is down. If it kept running, a line worker could encounter live conductors that are supposed to be dead. The inverter detects loss of grid by watching for a sudden change in voltage or frequency, or by actively injecting a small disturbance and checking whether the grid absorbs it. Either way, the response is a disconnection within a defined time.

This is why a grid-tie inverter cannot run during a blackout, even if the sun is shining. It is a safety function, not a limitation of the panels. Hybrid and off-grid inverters add a transfer mechanism and a battery so that selected circuits can run isolated from the grid, but that is a different topology with its own switching and grounding requirements.

Why the inverter often fails first

The array has no moving parts and degrades slowly. The inverter contains electrolytic capacitors, cooling fans, relays and power semiconductors, all of which age. It runs at elevated temperature in a enclosure that is often mounted in sun or in a garage, and it processes the full array power continuously. A typical inverter’s useful life is shorter than a module’s, and replacement is a normal maintenance event rather than a defect.

The failure modes are characteristic. Capacitor wear shows up as a gradual loss of output or a shutdown on hot days. Fan failure leads to overtemperature trips. Relay wear affects the grid-disconnect function. Monitoring that shows a inverter repeatedly restarting at midday, or derating before the array is hot, points at the inverter rather than the array.

Topologies in brief

A string inverter takes the DC from one or more series strings and converts it centrally. It is the simplest arrangement and the one where module-level mismatch and shade affect the whole string. A microinverter mounts at each module and performs the DC-to-AC conversion there, so each module has its own tracker; the trade is more electronics on the roof and a different maintenance model. A hybrid inverter includes a battery interface and a transfer mechanism, allowing stored energy to supply selected loads when the grid is absent. Each topology changes where the maximum power point is tracked and where the grid interface sits, not the underlying physics.

What people get wrong

The most common misdiagnosis is to blame the panels for a production shortfall that belongs to the inverter. A graph that looks wrong often shows the inverter doing its job: clipping at midday when the array is oversized, derating when the heatsink is hot, or disconnecting when grid voltage rises above the protection threshold. The mistake is natural because the panels are the visible part of the system and the inverter is a box with a display. The array’s behaviour is slow and predictable; the inverter’s is fast and conditional, so its interventions look like faults. When that shortfall leads to a financing dispute, the Borrower’s Guide explains your consumer-credit rights and how to challenge a lender’s claim.

The second misdiagnosis is expecting a grid-tie inverter to run during an outage. The inverter is designed to stop, and it stops. That is not a failure to be reset; it is the anti-islanding function working.

Reading inverter data

Most inverters report through a manufacturer app or monitoring portal, and many also expose a local data interface such as Modbus. What can be read, and at which register, is defined in the manufacturer’s documentation and differs between models, so no addresses are given here. The useful readings are the same everywhere: AC output power, DC input voltage and current for each MPPT input, internal temperature, and the event log. Repeated grid overvoltage or frequency events point at the grid side or the inverter’s settings rather than the array; one MPPT input reading well below its neighbours under the same sky points at that string. Reading this data through the portal involves no contact with the DC side at all.

Any work on the DC side of a PV system, including opening a combiner or disconnecting a connector, is licensed electrical work in most jurisdictions. The array produces dangerous voltage whenever light falls on it and cannot be switched off at the panel.

Net metering, interconnection rules, export limits and anti-islanding settings are set by state and utility and change. The reader should check the local interconnection agreement and the utility’s requirements rather than assuming a rule from another jurisdiction applies.