String inverters vs microinverters vs optimisers, compared

Choosing an inverter architecture is one of the few decisions in a residential or small commercial PV design that cannot easily be reversed later. String inverters, microinverters, and DC power optimisers all convert the same DC energy into usable AC power, but they differ fundamentally in where power conversion and maximum power point tracking happen, and that difference cascades into shading tolerance, monitoring resolution, upfront cost, and how the system fails when something goes wrong.

This article compares the three architectures on the dimensions that actually matter for a design decision, rather than treating it as a simple efficiency contest. None of the three is universally correct; the right choice depends on roof geometry, shading, budget, and how much visibility into performance the owner wants.

Readers unfamiliar with the underlying tracking behaviour should first see MPPT explained, since much of what follows depends on where MPP tracking is physically located in each architecture.

Architecture overview

String inverters sit at one central point, typically near the switchboard, and each MPPT input on the inverter tracks one series string of modules as a single unit. The DC-to-AC conversion happens once, centrally.

Microinverters are installed at each individual module, converting that module's DC output to grid-synchronised AC on the roof. Each module effectively becomes its own independent AC generator, wired together on an AC branch circuit.

DC power optimisers sit at each module like microinverters but perform only DC-to-DC conversion and MPPT at the module level; the optimised DC output from all modules is then combined into a string and sent to a central inverter for the actual DC-to-AC conversion.

The distinguishing question for any of the three is: at what point does independent MPP tracking stop, and where does DC get converted to AC?

Shading behaviour

Shading is where the architectures diverge most sharply. In a string inverter system, the whole string's current is set by its weakest module — a single partially shaded panel forces bypass diodes into play and can create a multi-peak IV curve for the entire string, discussed in detail in MPPT explained. Losses from moderate shading on a fraction of one string can easily reach double digits as a percentage of that string's output, and depending on inverter tracking sophistication may even affect adjacent unshaded modules sharing the same MPPT channel.

Microinverters and optimisers each track (and in the optimiser's case, at least partially decouple) the shaded module's output independently, so a shadow on one panel affects only that panel's contribution rather than dragging down its neighbours. This is the single strongest argument for module-level power electronics (MLPE) on roofs with chimneys, dormers, trees, or complex multi-plane geometry.

On a genuinely unobstructed, single-orientation roof, this advantage is largely theoretical — there is little mismatch to correct in the first place, and the cost premium of MLPE buys little.

Monitoring granularity

String inverters typically report at the string or inverter level: total DC input per MPPT channel, AC output, and basic fault codes. Diagnosing an underperforming individual module requires manual IV curve tracing or clamp-meter work on site.

Microinverters and optimisers report per-module data — voltage, current, and often temperature — through the monitoring portal, making it straightforward to spot a single failing or soiled module from a dashboard without a site visit. This granularity has real value for larger commercial arrays or for owners who want early warning of degradation, but it also multiplies the number of communicating devices that must stay online for full monitoring coverage.

Cost and lifecycle

String inverters generally carry the lowest upfront hardware cost per watt and the fewest total components, since much of the electronics are concentrated in one accessible unit. Warranties are typically 5–12 years as standard, extendable to 20–25 years at extra cost, and replacement — when eventually needed — is a single, ground-level swap.

Microinverters and optimisers add a per-module hardware cost, usually recovered partially through improved shaded-roof yield and, for microinverters, avoidance of a separate central inverter (though a communication gateway/hub is still required). Their standard warranties tend to run longer, commonly 20–25 years for microinverters, aligning with module warranties, since replacement of a roof-mounted unit is more disruptive than swapping a ground-level inverter.

Over a 25-year system life, the total cost comparison depends heavily on assumed failure rates and the labour cost of accessing roof-mounted components — a return visit to replace one of forty roof-mounted microinverters costs more in labour proportionally than replacing one central string inverter that failed once.

Worked example: shading loss comparison

Consider a small residential array of 12 modules in one string, each rated 400 W, with one module in partial morning shade for roughly two hours a day, during which it produces only 30% of its unshaded output.

String inverter case: for the two shaded hours, the string current is limited by the shaded module (bypass diode activation aside, assume the tracker settles on a compromise operating point delivering roughly 55% of unshaded string power during that window, a typical mid-range outcome for single-module partial shading on a 12-module string).

Unshaded string power ≈ 12 × 400 W = 4,800 W Shaded-period power ≈ 4,800 W × 0.55 ≈ 2,640 W Loss during shaded window ≈ 4,800 − 2,640 = 2,160 W, i.e. 45% of that period's potential output

Microinverter/optimiser case: only the shaded module's own output drops.

Shaded module unshaded output ≈ 400 W; at 30% ≈ 120 W Remaining 11 modules unaffected ≈ 11 × 400 = 4,400 W Total during shaded window ≈ 4,400 + 120 = 4,520 W Loss ≈ 4,800 − 4,520 = 280 W, i.e. about 6% of that period's potential output

Over the two-hour shaded window each day, the string inverter architecture loses roughly 2.16 kWh vs 0.28 kWh for MLPE, a difference of about 1.9 kWh/day in this specific scenario — illustrating why shading severity, not efficiency specs, usually drives the architecture decision.

Failure modes

A central string inverter failure is more disruptive per incident (the whole array or string stops producing) but is rare and easy to diagnose and fix. A roof-mounted microinverter or optimiser failure affects only one module but is statistically more likely to occur simply because there are far more units deployed, and each replacement requires roof access.

Comparison matrix

Design and safety implications

Module-level AC conversion (microinverters) reduces the DC voltage present on the roof to each module's own operating voltage, which lowers DC arc-fault risk compared with a string inverter or optimiser system where several hundred volts DC exists in the string wiring across the whole array. This is one reason some jurisdictions favour or mandate rapid-shutdown provisions for rooftop DC — a requirement optimisers satisfy by dropping their output to a safe voltage on command, and microinverters satisfy inherently. See grid connection and export rules explained for how these safety requirements interact with grid connection approval.

String sizing constraints differ too: a string inverter design must respect the inverter's MPPT voltage window across temperature extremes for the whole string (see the worked example in MPPT explained), whereas optimiser and microinverter designs decouple module count per string from voltage-window arithmetic, since each module already sits at its own regulated operating point before reaching the central inverter or grid tie.

Standards and references

• IEC 62109-1/-2: safety standard for power converters in PV systems, covering both central inverters and module-level electronics.
• IEC 60364-7-712: low-voltage electrical installation requirements specific to PV supply, including DC wiring and isolation considerations relevant to string vs module-level architectures.
• IEEE 1547: interconnection requirements for distributed energy resources, applicable to both microinverters and central inverters interfacing with the grid.
• IEC 62548: PV array design requirements, including guidance on rapid shutdown and DC arc-fault mitigation relevant to architecture choice.

Frequently asked questions

Are microinverters always the best choice for a shaded roof?

They are usually the strongest technical choice for meaningfully shaded or multi-orientation roofs, but "best" also depends on budget and long-term service access. DC optimisers paired with a string inverter achieve similar per-module shading tolerance at a somewhat different cost and monitoring profile, and are a reasonable alternative where a central inverter is preferred.

Do optimisers eliminate all string-level constraints?

No. Optimisers remove per-module MPPT mismatch, but the string as a whole still connects to a central inverter with its own DC voltage window and current limits, so string length and configuration still require calculation, just with more flexibility than a plain string inverter design.

Is a mixed system possible — string inverter on one roof face, microinverters on another?

Not directly through one inverter, since string inverters and microinverters are different product categories, but a site can have a string inverter serving an unshaded south roof and a separate microinverter branch circuit serving a shaded or complex section, provided both connect appropriately at the switchboard.

How does architecture choice affect future system expansion?

Microinverter and optimiser systems are generally easier to expand incrementally, since additional modules with their own module-level electronics can often be added to an existing AC or DC bus without recalculating a whole string's voltage window, subject to the central inverter or combiner capacity. Expanding a string inverter system usually requires adding a new MPPT input or a second inverter.

Do microinverters or optimisers reduce total system efficiency compared with string inverters?

The DC-DC or DC-AC conversion stage at the module level introduces its own small conversion loss, typically in the range of 1–3%, similar in order of magnitude to gains recovered from better shading tolerance and mismatch reduction. On unshaded roofs this conversion loss can leave MLPE slightly behind a well-matched string inverter system in raw efficiency; on shaded roofs the mismatch recovery usually more than compensates.

Sources

• NREL: PV System Performance and Reliability — comparative data on inverter architecture performance.
• Fraunhofer ISE Photovoltaics Report — market and technology overview including inverter trends.
• IEA PVPS — international reports on PV system technology and reliability.
• PVEducation: Shading effects — technical background on partial shading and mismatch.