Skip to content
Product

Module mismatch and cell variation

Unlisted

A module’s output is limited by the interaction of its series-connected cells. When cells differ, the strings they form cannot all operate at their individual maximum-power points, and the module produces less power than the sum of its cells. This shortfall is the mismatch loss, reported on the JV output tab.

This page explains how cell-to-cell variation becomes a distribution of module output when modifiers are run as a module population, and how two modelling choices, the device configuration and the way file values are referenced to the SunSolve circuit, change what that distribution represents.

A population run produces a spread of module outputs. That spread has two distinct origins:

  • Cell variation: genuine cell-to-cell differences, supplied either by parameterised distributions or by an imported file of measured cells.
  • Stochastic noise: the run-to-run variation of the Monte Carlo ray tracer, described in stochastic error.

Both broaden the module-output distribution, so a population result only isolates the effect of cell variation when the stochastic contribution is small relative to it.

Isolating cell variation with a unit-cell module

Section link: Isolating cell variation with a unit-cell module

The device configuration sets which of the two sources is present in the optics.

  • A unit-cell module solves the optics of a single representative cell and applies that result to every cell position. The optical result carries no cell-to-cell stochastic noise, so a population run varies only through the modifiers. The resulting distribution is due to cell variation alone.
  • A full module ray-traces every cell position. Each cell then carries stochastic noise from the finite ray count, which adds to the cell variation. With too few rays this noise can dominate the distribution; the ray count must be raised until the mismatch loss converges before the cell-variation spread can be read reliably.

Referencing cells to the median cell

Section link: Referencing cells to the median cell

When cells come from a measured-cell file, each imported cell is expressed relative to the file’s median cell, ranked by Isc. A cell with 2% more J01 than the median cell takes 2% more J01 than the SunSolve circuit. The measured variation is therefore applied as ratios on top of the simulated circuit, so the module’s own optical and resistive behaviour is retained and the file contributes only the relative spread.

Because each cell is scaled about the median cell, the circuit must describe that median cell. The calibration step in module population from measured cells lines the two up. On a unit-cell module the optics are identical at every position, so the output spread reflects the measured cell variation alone.

Cell position and white space

Section link: Cell position and white space

In a full module the light reaching a cell depends on its position. Cells beside the module perimeter or a cell-group gap couple additional light reflected from the white space around them, so at high ray counts the per-cell current settles into a position-dependent pattern rather than a uniform value. This position dependence is a systematic contribution to mismatch that is present in a full module but absent from a unit-cell module.