A string inverter and a microinverter both turn direct current from photovoltaic modules into alternating current that follows the grid. The difference is where the electronics sit and how many modules share one power-point tracker. A string inverter takes a series string, sometimes several strings, and converts once. A microinverter converts at the module, so each module has its own tracker. Neither one, by itself, is a battery or a backup system. The choice is about shade, layout, and what happens when one module is weak.
What a string inverter assumes
Modules in series share one current. The string’s current is limited by the weakest module in that series. A hard shadow, a cracked cell, or a module pointed a different way pulls the others down. Bypass diodes reduce the damage from a fully shaded substring. They do not turn a shaded string into an unshaded one. Shade losses are nonlinear, which is why a small shadow can cost more energy than its area suggests. The mechanism is in how shade affects PV output.
A string inverter is a good fit when the modules on one input see the same sun: one roof plane, no chimney, no tree that covers two modules after 2 p.m. Multiple MPPT inputs let you put two orientations on one inverter without mixing them on the same tracker. Mixing east and west on one input is still a mistake. Central inverters are the plant-scale cousin of this idea, with a dated comparison in central versus string solar inverters. A house is almost never that central machine.
The DC voltage of a string is the reason the design has to respect cold open-circuit voltage. More modules in series means more voltage. The inverter’s maximum input is a hard ceiling. Microinverters avoid that long series voltage because each one sees a single module, typically a low DC voltage. They do not avoid the rule that equipment must be rated for the voltage it actually sees.
What a microinverter changes
Each module works at its own power point. A shaded module loses its own output and does not set the current for a series string, because there is no series string of modules on the DC side. NREL’s 2016 update on module-level power electronics found that the products evaluated recovered about 25 to 35 percent of annual shading losses. The larger difference between the trials was how many strings were in parallel, not the brand name. Read that sentence carefully. The recovery is a share of the loss shade already caused. It is not a 35 percent energy bonus on a clear, unshaded roof. If your roof has no shade and one orientation, the microinverter’s shade advantage is a solution to a problem you do not have.
You also get module-level data, if the monitoring is installed and used. A dead module shows up as one module, not as a slightly low string that you notice in October. Monitoring in general is in what a solar monitoring system is. More electronics on a hot roof means more devices that can fail, each usually smaller to replace than a whole string inverter, and each in a harsher place than a garage wall. There is no universal life comparison in this article. Warranty years are whatever each contract says. PNNL’s note that inverter warranties are often shorter than module performance warranties applies to both architectures as a planning remark, not as a winner.
AC voltage still has to match the service. A microinverter listed for 220–240 volts is for that service. It is not a 120/240 volt split-phase device unless the listing says so. One US utility’s houses are 120/240. Many other countries are near 230 volts. Phase and voltage are in single-phase versus three-phase solar inverters. Grid-tied behavior is the same family either way: follow the utility, and cease to energize an unintentional island within 2 seconds under IEEE 1547-2018.
How to choose without a brand chart
| Roof | Architecture that matches the problem |
|---|---|
| One clear plane, one orientation | A string inverter is a straightforward match. Module-level electronics are optional, not required by the shade. |
| Chimneys, trees, or several roof faces | Microinverters, or several MPPT inputs kept separate, so one weak module does not set the current for the rest. |
| Backup power is the actual goal | Neither architecture provides it by default. You need a hybrid or a battery system that can island. See hybrid versus grid-tied inverters. |
Power optimizers are a third layout: electronics at the module, conversion still in a string inverter. They are not microinverters. This page does not assign them the 25 to 35 percent figure unless the product was in that evaluation. Ask whether the optimizer you are shown was the kind of module-level electronics NREL was talking about, and still remember that the figure is recovery of a shading loss.
Failure and weather still have to be sorted the same way. A low day can be cloud, heat, or dirt. Replacing a string inverter because one shaded module pulled the string down does not fix the shade. Sorting those afternoons is in solar inverter faults and the weather.
Frequently asked questions
Are microinverters more efficient than string inverters?
Not as a blanket rule. On a shaded, mixed roof they can recover a share of the shading loss, about 25 to 35 percent of that loss for the module-level products NREL evaluated. On a clear single-plane roof the conversion efficiency is whatever each datasheet prints at the load you actually run. Compare those sheets. Do not add 35 percent to a clear roof.
Can I mix a microinverter and a string inverter on one house?
Yes, if each one is listed for the service, the interconnection allows it, and you are not trying to parallel their outputs in a way neither manual allows. People do this when one roof face is shaded and another is not. It is two systems, not one hybrid.
Do microinverters work in a blackout?
A standard grid-tied microinverter stops when the grid stops, same as a string inverter. Backup requires equipment that can disconnect from the utility and form an island. The small box under the module does not do that by itself.
Sources and references
Recovery of about 25 to 35 percent of shading losses for the module-level electronics evaluated, with string count mattering more than brand: NREL, 2016. Cease-to-energize within 2 seconds: IEEE 1547-2018 primer. One utility’s residential voltage: SWEPCO handbook, 2024. Typical inverter warranty lengths as contracts, not as a life ranking of the two architectures: Pacific Northwest National Laboratory.
