In 50 words: In four years of hourly PVGIS data for 22 cities, a DC/AC ratio of 1.2 clips 0.5% or less of a solar array's energy everywhere. The highest ratio under 1% loss runs from about 1.26 in Denver to above 1.5 in Miami and Chennai. Cold, clear sites clip most.
Solar arrays are usually sized larger than their inverter. A 7.2 kW array on a 6 kW inverter has a DC/AC ratio of 1.2. On the sunniest hours the panels could deliver more than the inverter's rating, and the inverter "clips" the excess. Most guides quote a single example of how much that costs. The answer depends heavily on climate, so we computed it for 22 cities.
We took PVGIS 5.3's hourly output for 2020 to 2023 for 1 kW of panels at each city's best fixed angle with 14% system losses, and measured how much energy sat above the inverter limit at each ratio. Try your own sizes in the DC/AC ratio and clipping calculator.
Clipping by ratio, city by city
Share of the year's energy lost to clipping. "Peak hour" is the highest hourly output per kW of panels over the four years.
| City | Peak hour (W per kW) | 1.2 | 1.3 | 1.4 | 1.5 | Highest ratio under 1% | Under 2% |
|---|---|---|---|---|---|---|---|
| Denver (US) | 1,000 | 0.49% | 1.57% | 3.44% | 5.89% | 1.26 | 1.32 |
| Montréal (Canada) | 1,003 | 0.36% | 1.03% | 2.15% | 3.74% | 1.30 | 1.39 |
| Chicago (US) | 977 | 0.27% | 0.95% | 2.32% | 4.32% | 1.30 | 1.38 |
| Calgary (Canada) | 1,035 | 0.24% | 0.87% | 2.14% | 4.01% | 1.31 | 1.39 |
| Toronto (Canada) | 979 | 0.22% | 0.80% | 1.92% | 3.54% | 1.32 | 1.40 |
| New York (US) | 975 | 0.20% | 0.78% | 1.91% | 3.55% | 1.32 | 1.41 |
| Phoenix (US) | 936 | 0.03% | 0.37% | 1.45% | 3.43% | 1.36 | 1.43 |
| Munich (Germany) | 937 | 0.05% | 0.35% | 1.16% | 2.55% | 1.38 | 1.46 |
| Los Angeles (US) | 877 | 0.00% | 0.18% | 1.14% | 3.15% | 1.39 | 1.44 |
| Berlin (Germany) | 902 | 0.03% | 0.32% | 1.05% | 2.24% | 1.39 | 1.48 |
| London (UK) | 886 | 0.03% | 0.30% | 1.02% | 2.09% | 1.40 | 1.49 |
| Manchester (UK) | 907 | 0.03% | 0.28% | 0.96% | 1.95% | 1.40 | 1.50 |
| Houston (US) | 914 | 0.02% | 0.24% | 0.87% | 1.91% | 1.41 | 1.51 |
| Vancouver (Canada) | 922 | 0.03% | 0.20% | 0.83% | 2.00% | 1.41 | 1.50 |
| Delhi (India) | 848 | 0.00% | 0.09% | 0.68% | 2.02% | 1.42 | 1.50 |
| Jaipur (India) | 829 | 0.00% | 0.07% | 0.67% | 2.11% | 1.42 | 1.49 |
| Edinburgh (UK) | 899 | 0.02% | 0.21% | 0.77% | 1.69% | 1.42 | 1.53 |
| Seattle (US) | 901 | 0.01% | 0.13% | 0.70% | 1.83% | 1.43 | 1.51 |
| Bengaluru (India) | 831 | 0.00% | 0.10% | 0.72% | 1.80% | 1.43 | 1.51 |
| Mumbai (India) | 808 | 0.00% | 0.01% | 0.38% | 1.47% | 1.46 | 1.53 |
| Miami (US) | 871 | 0.00% | 0.04% | 0.24% | 0.79% | 1.52 | above 1.6 |
| Chennai (India) | 770 | 0.00% | 0.00% | 0.09% | 0.61% | 1.54 | above 1.6 |
What the numbers show
1.2 is safe almost everywhere
At a ratio of 1.2, the worst city in our set, Denver, clips 0.49% of the year's energy, and most clip under 0.1%. For a 7 kW home array that is at most about 65 kWh a year, and far less in most cities.
Cold and clear clips most; hot and humid least
The cities that clip most (Denver, Montréal, Chicago, Calgary, Toronto, New York) all reach close to 1,000 W per kW of panels on their best hours: cool air keeps the panels efficient, so they come closest to their rating. Panels lose output as they heat up, so hot cities rarely see panels near their rating: the peak hour in Chennai, Mumbai, Jaipur, Bengaluru and Delhi was 770-850 W per kW. Chennai clips nothing until a ratio of 1.4.
Above 1.3 the loss climbs quickly
Clipping grows faster than the ratio. In New York it goes from 0.78% at 1.3 to 1.91% at 1.4 and 3.55% at 1.5. Whether that is worth it is a cost question: if extra panels cost less per kWh than the energy they add, a higher ratio still pays.
The gain from oversizing
Clipping is only half the picture. On the same inverter, more panels mean more energy in every hour below the limit, which is most hours. In New York, an array at 1.3 has 30% more panels than one at 1.0 and delivers about 29% more energy, because only 0.78% is clipped. At 1.5 it has 50% more panels and delivers about 45% more.
Choosing a ratio
- Start at 1.2-1.3 for a south-facing home array in the northern US, Canada, the UK or Germany; the loss stays near or under 1%.
- Go to 1.4 or higher in hot climates such as India, Miami or Houston, where clipping stays under about 1% to 1.4.
- East-west and shaded arrays can go higher. Their output peaks lower and spreads across the day, so they clip less than these best-angle figures.
- Check the inverter's DC input limits, not just its AC rating: maximum DC power, voltage and current all cap how many panels it accepts. The string sizing calculator checks voltage and current.
A note on hourly data
Clipping only removes output above a limit, so averaging over an hour can only hide peaks, never add them. Real clipping is therefore at least the figures here. Treat them as a lower bound and leave margin at higher ratios.
Related
Sources
Researched and drafted by the Earth Energy Log Editorial Desk with AI assistance. Reviewed by Pruthvi A. before publication.
How we work: Editorial standards · AI disclosure · Report an error


