Solar Inverter Efficiency: Peak vs European Explained

Two efficiency numbers, and only one of them is useful

Every inverter datasheet carries at least two efficiency figures, and marketing material almost always quotes the larger one.

Peak efficiency is the single best measurement the unit ever produces: one specific load level, one specific DC input voltage, one controlled temperature. It is a real number, but it describes a moment the inverter spends almost none of its life in.

European efficiency — stored here as euro_efficiency_pct — is a weighted average across the load levels an inverter actually experiences over a year in a temperate climate. It is the honest figure, and the gap between the two tells you a great deal about how a machine behaves away from its sweet spot.

The SolarEdge SE6000H makes the point better than any argument. It is rated at 99.2% peak — among the highest figures in this database — but 97.5% European. That is a gap of 1.7 points. The Fronius Verto 33.3, by contrast, is rated 97.47% peak and 97.36% European: a gap of 0.11 points, a machine that barely notices what load it is running at. Judge only by the headline number and you would rank these two in the wrong order for a real installation.

What the weighting actually represents

The European efficiency formula is not an average of the curve. It is a weighted sum of six measurements taken at 5%, 10%, 20%, 30%, 50% and 100% of rated power, with the weights chosen to approximate how a rooftop array in central Europe distributes its energy over a year: roughly 3%, 6%, 13%, 10%, 48% and 20% respectively.

Read that list and the design pressure becomes obvious. Nearly half the weight sits at 50% load, and only a fifth at full power. An inverter spends most of its productive year at part load — mornings, evenings, cloud, winter, and every hour of a mildly oversized array. The weighting deliberately punishes designs that are only good when saturated.

The low end matters more than its weight suggests, too. Below about 10% load the inverter's own standby and switching losses are a large fraction of what is passing through it, which is where cheap designs bleed. The 5% and 10% points carry 9% of the weight between them, and they are where the peak-to-European gap usually opens up.

A parallel figure, CEC efficiency, uses the same method with weights fitted to Californian irradiance. It reads slightly differently but answers the same question.

What one point of efficiency is worth

Efficiency differences look trivially small until you multiply them by a year.

Take a 6 kWp array in a location yielding around 1,400 kWh per kWp per year — a reasonable southern-European figure. That is about 8,400 kWh of DC energy arriving at the inverter annually. Each full percentage point of European efficiency is therefore worth roughly 84 kWh a year, and each tenth of a point about 8 kWh.

Apply that to real hardware. Between the Huawei SUN2000-3-6KTL-L1 at 97.8% European and the Deye SUN3/3.6/5/6K-SG04LP1 at 96.5% there are 1.3 points, or about 110 kWh a year on that array. Over a 20-year service life that is roughly 2,200 kWh — real, but modest, and easily swamped by a poor string layout or an MPPT window that leaves a channel idle every morning.

That is the honest framing. Efficiency is worth optimising once the sizing and tracking decisions are right, not before them. A machine two points better on paper that forces both roof pitches onto one tracker will lose more than it gains.

Why transformerless designs win

Older grid-tied inverters used a mains-frequency transformer to provide galvanic isolation between the array and the grid. That transformer is heavy, and its core and copper losses are present whenever the machine is energised — including at the low load levels the European weighting cares most about.

Transformerless (also called non-isolated) designs remove it and handle isolation requirements electronically instead, with residual-current monitoring on the DC side. The efficiency gain shows up exactly where it matters, and the field bears it out: the Fronius Tauro ECO 100-3-P and the Huawei SUN2000-50KTL-M3 are both recorded as transformerless, at 98.2% and 98.0% European respectively, while the Deye SUN120–136K-G01P3 records its topology as non-isolated and reaches 98.2%.

Nearly every string inverter sold for grid connection today is transformerless. Where isolation is still mandated — certain off-grid and battery configurations, or specific national grid codes — the transformer returns and a point or so of efficiency goes with it.

THD: the quality of what you export

Efficiency counts how much energy reaches the grid. Total harmonic distortion — thd_pct — describes what shape it arrives in.

An inverter synthesises a sine wave by switching a DC bus at high frequency. The result is never perfectly sinusoidal, and THD quantifies the residue: the fraction of the output current that sits at multiples of 50 or 60 Hz rather than at the fundamental. Harmonic current does no useful work. It heats transformers and cabling, it interferes with sensitive equipment on the same circuit, and grid codes generally cap it at 5% at rated output.

Good hardware clears that limit with room to spare. The Solis S5-GR3P records 2%, and the Huawei SUN2000 units and the whole Deye grid-tied range record 3%. Note the qualifier that always applies: THD is specified at rated output. At 10% load, when the switching residue is a much larger fraction of a much smaller current, distortion is invariably worse — which is another argument against oversizing an inverter relative to its array.

Real models ranked by European efficiency

Two patterns are worth naming. First, the large three-phase machines cluster tightly at the top — above about 20 kW, 98% European is the expected standard rather than a distinguishing feature. Second, several units in this database publish only a peak figure: the Growatt MAX 320K-X at 99.03%, the Growatt MID 20-30KTL3-X2 at 98.75% and the GoodWe MS G4 at 98.1% have no European efficiency recorded. Treat a missing weighted figure as missing information, not as a good one.

Reading the fields on a listing

On every model page here, efficiency_pct is the peak value, euro_efficiency_pct is the weighted value, and thd_pct is the distortion at rated output. Compare units on the middle field, use the first only to gauge the gap, and check the third against your grid code. Then go back to how to choose a solar inverter for everything the efficiency figures do not cover.