How solar panels work, from photon to breaker panel

A rooftop solar installation looks simple from the street: a rectangular array of dark panels, a box on the wall, a cable running down to the meter. Underneath that simplicity is a chain of energy conversions that spans nanometres to metres — from electron transitions inside a silicon crystal to the 50 Hz alternating current that a washing machine expects. Understanding each link in that chain is useful not because homeowners need to design cells, but because every failure mode, every efficiency loss, and every wiring decision on a real installation traces back to one of these conversion steps.

This article follows the energy from photon absorption through to the point of interconnection with the grid. It is deliberately end-to-end: cell physics, module assembly, string design, inverter behaviour, and the safety equipment that sits between the array and the building's consumer unit. For a narrower treatment of the cell-level physics alone, see how a solar cell turns light into current.

The goal is to give enough grounding that component datasheets, commissioning reports and troubleshooting symptoms make sense in context, rather than being memorised as isolated facts.

The photovoltaic effect in plain terms

A solar cell is, at its core, a large-area diode. It consists of a wafer of crystalline silicon (or a thin film) that has been deliberately doped to create two regions with different electrical character: an n-type layer with a surplus of free electrons and a p-type layer with a surplus of "holes" (electron vacancies). Where these two layers meet — the p-n junction — an internal electric field forms.

When a photon with sufficient energy strikes the silicon lattice, it can knock an electron out of its bond, creating an electron-hole pair. Silicon's bandgap is about 1.12 eV, which corresponds to light with a wavelength shorter than roughly 1100 nm — most of the visible spectrum and part of the near infrared. Photons with less energy than the bandgap pass through or are absorbed as heat; photons with much more energy than the bandgap lose the excess as heat as well. This is the fundamental reason no single-junction silicon cell can exceed the Shockley-Queisser limit of about 33.7% efficiency under standard test conditions.

The internal field at the junction sweeps the newly freed electron toward the n-side and the hole toward the p-side before they can recombine. Metal contacts — the fine silver grid lines visible on the front of a cell and a full-area or fine grid on the rear — collect these separated charges. Connect an external circuit between front and back contacts and you have a current source: the photovoltaic effect.

From cell to module

A single cell produces roughly 0.5–0.7 V at its maximum power point, which is far too low to be useful. Manufacturers connect 60, 66, 72 or 120/144 half-cut cells in series to build a module with a usable operating voltage, typically 30–50 V for residential panels. The cells are laminated between a front glass sheet, encapsulant (usually EVA or POE), the cell matrix, a backsheet or rear glass, and an aluminium frame.

Two design choices matter disproportionately to real-world performance:

• Half-cut cells. Cutting cells in half halves the current per cell, which cuts resistive (I²R) losses in the cell's own conductors roughly by four, and reduces the impact of partial shading because the module is typically wired so a shadow on one half does not force the whole string to a bypass condition simultaneously.
• Bypass diodes. Each module contains typically three bypass diodes, one per sub-string of cells. If a sub-string is shaded or damaged, the diode allows current from the rest of the string to route around it rather than forcing the shaded cells into reverse bias, where they would dissipate power as heat (a hot spot). This is covered in more depth in the article on bifacial modules, where rear-side shading patterns are more complex.

Module electrical ratings

Datasheets specify performance at Standard Test Conditions (STC: 1000 W/m² irradiance, 25°C cell temperature, AM1.5 spectrum) and increasingly at NOCT (Nominal Operating Cell Temperature, closer to real rooftop conditions).

String design and the inverter's job

Modules are wired in series into "strings" to build up voltage, and strings may be wired in parallel to build up current. The inverter's DC input window must accommodate the string voltage across the full expected temperature range: Voc rises as cell temperature falls, so the coldest expected morning sets the maximum string length, and the hottest expected operating temperature sets the practical minimum for maximum power point tracking (MPPT) to stay within range. Use /calculators/inverter-sizing to check a given module and inverter combination against these limits rather than relying on rules of thumb.

The inverter's central job is MPPT: continuously adjusting the operating voltage of each string, many times per second, to find the point on the current-voltage (I-V) curve where the product of voltage and current is maximised. This point shifts constantly with irradiance, temperature and shading. A string inverter typically has one to four independent MPPT channels; a system with roof faces in different orientations, or partial shading from a chimney, benefits from putting each affected string on its own MPPT channel, or from using module-level power electronics (power optimisers or microinverters) that perform MPPT at the individual module level.

Worked example: sizing a string

Take a module with Voc = 41.2 V at STC and a temperature coefficient of Voc of -0.27%/°C, installed in a location where the lowest expected ambient temperature is -10°C (cell temperature at low irradiance tracks close to ambient). The inverter's maximum DC input voltage is 600 V.

Temperature difference from STC: 25°C − (−10°C) = 35°C. Voltage rise: 41.2 V × 0.27% × 35 = 41.2 × 0.0027 × 35 = 3.89 V. Adjusted Voc per module: 41.2 + 3.89 = 45.09 V.

Maximum string length: 600 V ÷ 45.09 V = 13.3, rounded down to 13 modules per string. A designer would typically apply a further safety margin (often 5–10%) recommended by the inverter manufacturer, which might bring the practical maximum to 12 modules — this is exactly the calculation that inverter and module compatibility tools automate, and it is worth re-checking whenever either component in an installation changes.

Converting DC to AC

The inverter's second job is the DC-to-AC conversion itself. Modern residential inverters use high-frequency switching (IGBT or MOSFET-based H-bridge topologies) combined with pulse-width modulation to synthesise a sine wave, followed by an output filter to strip out switching harmonics. Grid-tied inverters must also synchronise phase, frequency and voltage with the grid before closing their output relay, and must disconnect within defined time limits if grid parameters go out of range — a function called anti-islanding protection, required so the inverter does not continue to energise a section of grid that utility workers believe is dead.

Efficiency of this DC-AC stage in a good modern inverter is 97–98.5% at typical loading, but efficiency curves are not flat: most inverters are least efficient at very low output (below ~10% of rated power) and near their peak, with a broad efficient plateau in between. Oversizing the DC array relative to the inverter's AC rating (a DC/AC ratio of 1.1–1.3 is common practice) pushes more of the year's operation into that efficient plateau and only clips power at the very peak of clear summer days.

From inverter to consumer unit

Downstream of the inverter, the installation must meet the same safety requirements as any other electrical circuit, plus PV-specific requirements. An AC isolator allows the inverter to be disconnected for maintenance. Overcurrent protection (a circuit breaker sized to the inverter's maximum AC output current) protects the cable run into the building's distribution board. In many jurisdictions a residual current device (RCD) — potentially a type specified for PV inverters with DC fault current components — provides earth fault protection. Metering, whether a simple generation meter or a smart meter with import/export registers, sits at the boundary with the grid.

This is also where export limitation, if required by the local distribution network operator, is enforced — either by the inverter's own settings or by an external device that curtails output based on a current transformer reading at the grid connection point. The regulatory and contractual side of this boundary is covered in grid connection and export rules explained.

Losses along the chain

No stage in this chain is loss-free, and system design is largely the discipline of managing the accumulation of small losses.

These figures are why a system's realised annual yield (commonly expressed in kWh/kWp) is always meaningfully below what a naive multiplication of STC power by annual sunshine hours would suggest, and why performance ratio (actual yield divided by theoretical yield at STC-equivalent irradiance) is the standard metric for judging whether an installed system is behaving as designed.

Why this matters for system economics

None of this is academic. Every percentage point of loss recovered — through better string design, a well-chosen DC/AC ratio, or simply keeping panels clean — compounds over a 25-year module warranty period. The financial consequence of these engineering choices is explored in the honest payback calculation, and it is worth running any proposed system through /calculators/payback using the realistic performance ratio a local installer can substantiate, rather than a headline efficiency number from a datasheet.

Standards and references

• IEC 61215 — Design qualification and type approval for terrestrial PV modules; the mechanical and environmental stress tests a module design must pass.
• IEC 61730 — PV module safety qualification, covering electrical shock, fire and mechanical hazard protection.
• IEC 62109 — Safety of power converters for use in PV power systems, applicable to inverters.
• IEC 62548 — Design requirements for PV arrays, including string sizing, DC cabling and protection.
• IEC 60364-7-712 — Low-voltage electrical installations, particular requirements for PV power supply systems.
• IEEE 1547 — Standard for interconnecting distributed energy resources with electric power systems, widely referenced for anti-islanding and grid support functions.

Frequently asked questions

Why do solar panels lose efficiency as they get hotter?

Higher cell temperature increases the thermal energy of charge carriers, which raises the reverse saturation current of the p-n junction and reduces the open-circuit voltage more than it increases short-circuit current. The net effect is lower power output, typically -0.30% to -0.45% of rated power per degree Celsius above 25°C for crystalline silicon. This is a semiconductor property, not a manufacturing defect, and is quoted on every datasheet as the temperature coefficient of Pmax.

What is the difference between Voc and the operating voltage?

Voc (open-circuit voltage) is measured with no load connected and represents the maximum possible voltage the module can produce. Under normal operation feeding an MPPT inverter, the module runs at its maximum power point voltage (Vmp), which is typically 80–85% of Voc. Designers use Voc, not Vmp, to check against an inverter's maximum input voltage because Voc occurs whenever the inverter is off or disconnected, such as during a fault.

Do bypass diodes fully solve the shading problem?

No. Bypass diodes prevent the extreme case of hot-spot heating and limit power loss to roughly the shaded fraction of the module rather than the whole string, but a shaded sub-string still contributes zero power while its diode conducts. Partial shading remains a meaningful loss, which is why module-level optimisers or microinverters are often specified for roofs with persistent partial shading from chimneys, vents or trees.

Can a solar system work without an inverter?

Off-grid DC systems that run DC appliances directly, or that charge batteries, can operate without a grid-tied inverter, though they typically still use a charge controller performing MPPT. Any system intended to power standard AC appliances or to export to the public grid requires an inverter to produce compliant AC, synchronised in frequency, voltage and phase with the grid it connects to.

How is a system's real-world output different from its rated power?

Rated power is measured at Standard Test Conditions, an artificial benchmark (1000 W/m², 25°C cell temperature) rarely sustained in the field. Real installations run at a performance ratio, typically 75–90%, that accounts for temperature derating, soiling, inverter conversion losses, DC cabling resistance and shading. Annual energy yield in kWh is calculated from rated power, expected irradiation for the site and orientation, and this performance ratio, not from rated power alone.

Sources

• NREL — Solar Photovoltaic Technology Basics
• PVEducation — Solar Cell Operation
• Fraunhofer ISE — Photovoltaics Report
• IEC — TC 82 Solar photovoltaic energy systems
• IEA PVPS — Reports