Put 1,300 kW of modules behind a 1,000 kW inverter and you have a DC:AC ratio of 1.3. Somebody always asks whether that means throwing away 300 kW.
It does not, most of the time, and understanding why is the difference between a system that earns well and one that either wastes inverter capacity or clips away real revenue.
Module nameplate is measured at Standard Test Conditions: 1,000 W/m² irradiance, 25°C cell temperature, AM1.5 spectrum. Those conditions are a laboratory construct. On a real roof they essentially never coincide.
When irradiance hits 1,000 W/m² on a clear summer day, your modules are not at 25°C – they are at 55-65°C. With a temperature coefficient around -0.34%/°C, that alone strips 10-13% off output. Add soiling, wiring losses and mismatch, and a 1,300 kW array delivers maybe 1,050-1,120 kW at its genuine annual peak.
So a 1.3 ratio does not clip 300 kW. It clips for a small number of hours a year, by a small margin.
Clipping is the inverter capping output when DC input exceeds what it can convert. The energy is not converted; it is simply not harvested.
For a well-chosen ratio in a temperate climate, annual clipping losses typically run 0.2-1.5% of potential generation. In exchange you get materially better inverter utilisation across the whole year – more hours near rated output, better conversion efficiency at part load, and lower cost per kWh delivered because the inverter is the expensive box.
The trade is: give up a sliver of peak-hour energy to gain far more shoulder-hour energy per dollar of inverter.
| Situation | Typical DC:AC | Reasoning |
|---|---|---|
| Northern Europe, UK, Nordics | 1.25 - 1.40 | Peak irradiance rare; heavy overload nearly free |
| Central Europe, northern US | 1.20 - 1.30 | Balanced; the common default |
| Southern Europe, US Southwest | 1.10 - 1.25 | Frequent high irradiance; clipping accumulates |
| East or west facing arrays | 1.30 - 1.50 | Flatter production curve, low peak |
| Hard export limit | Often 1.35+ | You are capped anyway; maximise shoulder hours |
| Paired with battery storage | 1.30 - 1.45 | Clipped energy can charge the battery instead |
Export-limited connections. If your grid agreement caps export at 500 kW, the inverter is capped there regardless. Loading more DC behind it fills more of the day at that cap, which is exactly what you want. This is common on constrained European networks and it changes the arithmetic completely.
East-west layouts. A dual-orientation array has two modest humps rather than one sharp midday peak. Total peak power is inherently lower, so heavier DC loading barely clips at all while flattening the delivery profile – which usually matches commercial consumption better anyway.
DC-coupled storage. If a battery sits on the DC bus, energy that would have clipped goes into the battery instead. Clipping stops being a loss and becomes free charging.
The mistake I see most often is applying a ratio someone read in a webinar without checking it against local irradiance data. A 1.35 ratio that is nearly lossless in Manchester will clip meaningfully in Seville, and the difference is worth real money over twenty-five years.
The second mistake is ignoring the inverter's own limits. Every inverter has a maximum DC input power and a maximum DC input current, and they are separate constraints. Exceeding the current limit can void warranty even when power looks acceptable. Check both against the manufacturer's stated maximums, at the coldest expected temperature, when Voc is highest.
Run the simulation at three ratios – say 1.15, 1.25 and 1.35 – with your actual site's irradiance dataset, and compare annual yield against inverter cost. The curve is usually flat near the top, which is good news: there is a broad band of acceptable answers, and precision matters less than avoiding the extremes.
If a designer cannot show you that comparison, they picked the number from habit. It might still be a fine number. But you should know which it is.