Solar Inverter Clipping: DC/AC Ratio and Yield Loss

Last updated: November 2026

Inverter clipping happens when an array produces more DC power than the inverter can convert. The inverter holds its AC output at its rating and the surplus is lost. It is the price of DC oversizing, and the right amount of it is a design decision, not a defect.

This guide explains how the inverter loading ratio (ILR, also called the DC/AC ratio) drives clipping, what NREL researchers measured on real systems, how to pick a starting ratio for residential, commercial and utility-scale projects, and where inverter warranty limits cap the answer. It also covers bifacial gain, storage, per-MPPT loading and a worked example with assumed inputs.

In this guide:

• What clipping is, with the funnel analogy
• Why DC oversizing can earn more energy than it loses
• What NREL measured about clipping and hourly-model bias
• Starting DC/AC ratios for residential, commercial and utility-scale systems
• Climate corrections, bifacial gain and per-MPPT loading
• Inverter DC input limits and warranty traps
• A worked 25-year example with assumed inputs

What Inverter Clipping Actually Is

Inverter clipping is the slice of DC power the inverter throws away because its AC output is already maxed out. Think of a kitchen funnel. Pour water through it slowly and every drop passes. Pour it faster than the spout can drain and water piles up at the rim, then spills over the edge. The spout is the AC nameplate. The water that spills is clipped energy.

A solar array’s DC output rises and falls with sunlight. On a clear day at solar noon, output peaks. If the array nameplate is 6,000 W and the inverter AC nameplate is 5,000 W, anything above 5,000 W of useful DC cannot leave the inverter as AC. The inverter moves the array operating point off its maximum power point (MPP) through maximum power point tracking (MPPT), and the lost DC is dissipated as heat.

Three things matter here. First, clipping is normal and intentional in any system with an inverter loading ratio (ILR) above 1.0. Second, clipping shows up in inverter monitoring logs as a flat AC plateau on clear-sky days. Third, a monitoring portal can label an AC limit setting as clipping, so check the setting before you blame the array.

For definitions, see the inverter clipping glossary entry and the inverter loading ratio page.

Why DC Oversizing Can Earn More Than It Loses

DC oversizing works because panels rarely produce their nameplate. Nameplate is the value at Standard Test Conditions (STC): 1,000 W/m² irradiance, 25 °C cell temperature, AM1.5 spectrum. Real conditions are different. Cells run hotter than 25 °C, and soiling, partial shade, wiring losses and module degradation take more. In most hours a 6 kW STC array produces well under 6 kW, so pairing it with a 5 kW inverter does not waste much. It keeps the inverter near full output for longer.

Here is the logic in plain terms. A 1.0 ILR design runs the inverter below its rated AC capacity for most of the year. A higher ILR pushes the inverter to full AC output across more hours, especially mornings, evenings, winter and cloudy days. The cost is some clipping at noon on clear days. The gain is a fatter generation profile in every other hour. Energy is the area under the curve, not its peak.

How Much Energy Does Clipping Cost? What NREL Measured

Annual clipping loss rises slowly at low ILRs and faster at high ones, and hourly yield models tend to understate it. That is the finding of Anderson et al. (NREL, Southern Company and DNV), whose 2022 preprint The Effect of Inverter Loading Ratio on Energy Estimate Bias compares hourly model results with measured one-minute power data from real plants.

The mechanism is simple. Within an hour, irradiance spikes push array power above the clipping point for a few minutes even when the hourly average sits below it. An hourly model sees no clipping, so it overstates energy. The authors express the error as a percent of annual production. The table below shows their results for the NIST test system.

Source: Table II of Anderson et al. (2022), NIST site, percent of annual production.

Three points follow from the paper:

1. The bias grows with ILR at every site studied. Across the six systems, the PVsyst bias at ILR 1.5 ranged from 1.9% to 4.5%.
2. It is not limited to humid climates. The authors report 2 to 3% bias at ILR 1.5 even for the semi-arid NREL system.
3. The two smallest arrays showed the largest bias, 4.4% and 4.5% at ILR 1.5, which the authors link to less spatial averaging of irradiance.

This is model error, not the whole clipping loss. The total energy clipped at a given ILR is larger and depends on climate and mounting. The authors also note that the effect is not a recoverable loss from poor operation. It is energy a coarse model assumed would be produced and never was.

NREL’s 2024 Annual Technology Baseline for utility-scale PV assumes an ILR of 1.34 for its base year and future projections. The NREL paper describes ILRs of 1.2 to 1.5 as typical of new systems today.

Trackers and east-west layouts flatten and widen the daily output curve, so more hours sit near the AC limit. Run fixed-tilt and tracker cases separately instead of reusing one clipping number.

DC/AC Ratio Definitions That Actually Matter

The terms get used loosely. Three definitions sit underneath the math:

• DC nameplate (kWp). The sum of module rated power at STC, in kilowatt-peak.
• AC nameplate (kWac). The inverter’s rated continuous AC output, in kilowatts.
• DC/AC ratio = ILR = kWp ÷ kWac. A 7.5 kWp array on a 6 kWac inverter gives 7.5 ÷ 6 = 1.25.

A second number, the inverter’s maximum DC input, is the warranty-relevant figure. Manufacturers publish it differently: some give a maximum DC power, some a maximum array power or a percentage of AC rating, and nearly all give a maximum DC current per MPPT. The limit varies by model, so read the datasheet of the exact inverter you plan to use.

Two other terms get confused with clipping and should not be:

• Curtailment. A grid-side limit imposed by the utility or market operator. It is not inverter behavior.
• Operating-window limits. When string voltage falls outside the inverter’s MPPT window, for example when very hot modules push the maximum power voltage below the MPPT minimum, the inverter cannot track the maximum power point. That is a string design issue, not an ILR issue.

Typical DC/AC Ratios for Residential Systems

Residential designs usually start between about 1.15 and 1.35, and the number moves with climate, export rules and whether a battery is planned. The ranges below are design rules of thumb, not published standards, and no table replaces a simulation of the actual roof.

Local rules also cap the ratio. Some programs set a maximum array-to-inverter ratio for incentive eligibility, and network operators set export limits. Check the current installer guidelines of the program, utility or regulator before you fix the ratio.

Where nothing caps the design, 1.25 to 1.30 is a sensible residential starting point. Below about 1.20 you are likely leaving energy on the roof, and above about 1.40 clipping piles up on the clear summer days when production should peak.

Commercial Typical Ratios

Commercial rooftop sits between residential and utility-scale practice. These are starting ranges used in design work, not a standard:

The storage row is the most consequential change of recent years. With a DC-coupled battery on the same inverter, surplus DC can charge the pack instead of being thrown away, so a higher ratio loses less. Read AC coupled vs DC coupled battery solar for how the two layouts differ.

When you pitch commercial buyers, explain how the ratio conversation changes once a battery is in scope. Without storage, the aim is the lowest cost per kWh. With storage, the aim shifts to shifting energy into the hours that pay most.

Utility-Scale DC/AC Ratios

Utility-scale plants run higher ratios than rooftops because extra DC is cheap relative to the inverter, grid connection and land already paid for. NREL’s 2024 ATB assumes 1.34 for utility-scale PV, and the NREL paper describes 1.2 to 1.5 as typical of new systems.

What pulls utility ratios upward:

• Module cost. Module prices have fallen sharply since 2020, so extra DC capacity costs less than the energy it adds.
• PPA structures. Many newer PPAs pay less at midday and more in shoulder hours. A higher ILR shifts production into those hours.
• Degradation planning. Modules lose output each year, so the effective ratio drifts down over the plant’s life. Some EPCs start higher for that reason.
• Tracker maturity. Single-axis trackers extend the daily output curve, so a higher ILR captures more energy on a tracker plant.

What pushes back:

• Grid export caps. Where AC export is capped, extra DC is wasted unless it goes into storage.
• Curtailment risk. In oversupplied markets, midday curtailment compounds clipping.
• Inverter limits. The inverter’s maximum DC input and current limits cap the ratio regardless of economics.

Climate-Specific Recommendations

Climate does most of the work in setting the right ILR. Two variables drive it: how often the system reaches clear-sky peak DC, and how hot the panels run. The target ranges below are rules of thumb.

Clear, sunny climates (Spain, Arizona, MENA, India). Hot modules lose output against STC, so clipping risk is moderate at moderate ratios. But clear-sky hours are frequent, so the clipped kWh still add up. Target ILR: 1.15 to 1.30.

Mixed-sky markets (US East Coast, central Europe, southern Australia). A mix of clear and partly cloudy days with moderate cell temperatures. Brief cloud-edge enhancement can push DC above STC for short periods. Target ILR: 1.25 to 1.35.

Cloudy, cool markets (UK, Germany, Netherlands, Pacific Northwest). Few clear-sky peak hours per year, and cool cells keep DC near STC when the sun does break through. The array spends most hours well below nameplate. Target ILR: 1.30 to 1.45.

Extreme cold markets (Canada, Nordic Europe). Cold modules can produce above STC, and cold-temperature Voc compliance matters more than ILR. Keep the ratio modest, 1.20 to 1.30.

This is the counterintuitive part. Cloudy climates can, and usually should, run higher ILRs than sunny ones. Clipping is a peak-hours problem, and cloudy markets have fewer peak hours, so you can add more DC without paying for it in clipped energy.

Bifacial System Oversizing Math

Bifacial panels add rear-side energy that depends on ground albedo, mounting height and module bifaciality. That rear gain is real DC power, and it reaches the inverter like front-side DC.

This matters for ILR because many designers still use the front-side STC nameplate to calculate it. Suppose a 600 W bifacial module delivers a 10 percent rear gain in the field. It then behaves like a 660 W module. Across a 1 MW array, 1.10 MWp behaves like 1.21 MWp, and a design that targeted 1.30 using STC nameplate is really running about 1.43.

The correction is simple but often skipped:

• Effective ILR for bifacial = (STC nameplate × (1 + rear gain)) ÷ AC nameplate.

With an expected 10 percent rear gain, a true target of 1.30 to 1.35 means an STC-nameplate ILR of about 1.20 to 1.25. Take the rear gain from a simulation that models rear irradiance row by row, because the gain is not uniform. Front rows of a ground-mount block usually see different albedo effects than interior rows. See bifacial solar panel design for the rear-irradiance breakdown.

Inverter DC Input Limits

The warranty-relevant number is the inverter’s published maximum DC input. This guide does not quote per-model figures, because they differ by manufacturer and model and the manufacturer’s own datasheet is the only reliable source. Read these items for the exact inverter you plan to use:

If you plan to go beyond a published limit, get written confirmation from the manufacturer first. Check both the power limit and the per-MPPT current limit, because the current limit can bind first.

Warranty Traps to Avoid

Three oversizing mistakes cause most warranty disputes:

1. Per-MPPT current exceeded. Total DC power is inside the limit, but two parallel strings on one MPPT push short-circuit current above the rated input current. The input stage can fail, and the manufacturer may decline the claim.
2. Bifacial gain ignored. The designer sized on STC nameplate, and rear gain pushes the effective DC past the maximum input. Some manufacturers ask for a margin above STC nameplate.
3. Temperature derating not applied. Where a datasheet gives a lower allowed input at high ambient temperature, a site that is hot for many hours a year can be outside the limit while looking compliant on paper.

The fix is to read the datasheet and warranty carefully, apply any derating the manufacturer requires, include bifacial and tracker gains in the DC figure, and size strings against the MPPT voltage window at both temperature extremes.

Worked Example: 25-Year Value of a Higher DC/AC Ratio

This example uses assumed inputs to show how to compare ratios. It is not project data. Replace every assumption with your own simulation results.

Assumed inputs:

• Inverter capacity: 400 kWac
• Specific yield before clipping: 950 kWh per kWp per year
• Degradation: 0.4% per year, so 25 years of output equals 23.84 times the year-one figure
• Retail value of energy: 0.22 per kWh (any currency)
• Clipping loss at each ratio: assumed values of 0.2%, 1.4%, 3.1% and 7.2%. These come from the exercise, not from a source, and your own minute-resolution simulation should replace them.

The pattern matters more than the totals. Each added kWp earns less than the one before it, because more of its output lands in the hours that clip. In this example the second step adds about 711 kWh per added kWp per year, and the third adds about 518. If an added kWp costs less than its 25-year value, the step pays. If it costs more, stop at the previous ratio.

That is the logic people miss when they argue about clipping. Clipping is a known cost and the extra energy is a known revenue. The best ratio maximizes revenue minus cost, not the ratio with zero clipping. Ratios above the inverter’s maximum DC input also fail on warranty grounds before they fail on economics.

For the financial side of the same math, see the generation and financial tool.

What Most Clipping Guides Miss

Five patterns come up again and again in clipping discussions:

1. Monitoring portals can overstate clipping when AC limits are misconfigured. An AC limit set in firmware below the inverter’s nameplate, often for export compliance, produces a flat plateau that the portal logs as clipping. The cause is the setting, not DC excess. Read the AC limit from the inverter, not the dashboard.

2. Hourly modeling underestimates real clipping. The NREL study found hourly-model energy bias of about 0 to 4 percent at ILR 1.2 to 1.5, rising with the ratio. If your model reports low clipping at a high ratio, test it at a finer time step before you rely on it.

3. Bifacial gain is a common warranty trap. Designers anchor on STC nameplate when they check maximum DC input. With bifacial modules, that undercounts real DC.

4. Trackers shift the clipping curve up. A tracker holds high output across more of the day than a fixed-tilt array, so the same ILR clips more. Guides that quote one clipping number for every mounting type mislead.

5. Curtailment and clipping do not simply add. If grid curtailment already forces output below the inverter’s AC rating during the same hours that DC excess would clip, the inverter has no clipping to do. Treat curtailment as the binding limit while it is active, then clipping in the remaining hours. Adding the two as separate losses double-counts.

Sources and Further Technical Reading

These references are worth opening for any design you will have to defend to a lender or auditor:

• Anderson et al., NREL (2022): the preprint on ILR-dependent energy estimate bias used in this guide.
• Prilliman, Keith and Hobbs, NREL (2024): a side-by-side comparison of sub-hourly clipping models.
• Sandia PV Performance Modeling Collaborative: reference material on PV and inverter performance modeling.
• IEA PVPS Task 13: reports on PV system performance and reliability.

Three Design Scenarios, Three Ratios

The clearest way to see why one ratio is not the answer everywhere is to compare three typical situations. These are scenarios built to show the reasoning, not reports of real projects.

Scenario A: cloudy-climate home with a battery planned later. The array spends most hours well below nameplate, so a ratio in the upper part of the residential range is low risk. If the hybrid inverter’s datasheet allows it, leaving headroom lets the owner add modules later without swapping the inverter. Check the datasheet limit before you promise that.

Scenario B: sunny tracker ground-mount with a shaped PPA. Clipping looks high on paper at a high ratio. It can still be acceptable when the PPA pays little at midday, because most clipped hours were low-revenue hours anyway. Weigh the hourly revenue, not the clipping percentage alone.

Scenario C: split-orientation commercial roof on a multi-MPPT inverter. The east and west strings never peak at the same time as the south strings. An aggressive overall ratio would clip the south side without helping the others, so set the ratio per MPPT and check how the blended figure behaves.

The lesson across the three is that the ratio is not one number. Different MPPTs can carry different ratios, different climates need different ratios, and different revenue structures pull the optimum in different directions.

Per-MPPT Ratio Optimization

A multi-MPPT string inverter behaves like several smaller inverters sharing one AC nameplate. The total AC output cannot exceed the inverter’s rating, but each MPPT can be loaded differently. This is a high-impact optimization that many designers skip.

The rule: load the MPPTs that never peak at the same time more heavily than the MPPTs that do. The values below are examples, and every MPPT must stay inside its own input current and power limits.

• South-facing strings (northern hemisphere) reach peak DC at solar noon. Load them at or below the project target ILR, for example 1.25.
• East-facing strings peak in the morning, before the south strings. They can carry more, for example 1.35.
• West-facing strings peak in the afternoon, after the south strings, and can also carry more.
• Shaded or poorly oriented strings rarely reach full nameplate. They can carry the most, because they will not drive the inverter into saturation.

The combined ILR on the inverter nameplate might be 1.30, but the inverter rarely sees a coincident peak across all MPPTs. The result can be more annual energy at lower modeled clipping than a uniform design. The size of the gain depends on the roof layout, so simulate it.

When to Pair DC Oversizing with Storage

DC-coupled storage on the same inverter changes the ILR math. Surplus DC that would have clipped charges the battery, then dispatches later in the day. Net energy loss can fall sharply at high ratios.

The decision tree:

• Net metering at parity (1:1 export value). Storage adds little for clipping recovery alone. Stay at a moderate ILR (1.20 to 1.30).
• Net billing or a feed-in tariff below retail. Self-consumption is worth more than export, so storage justifies a higher ILR (1.40 to 1.55).
• Time-of-use tariff with peak evening rates. Storage shifts midday clipping into evening dispatch, which can justify 1.45 to 1.60.
• PPA with shaped pricing. If midday is low-paid, storage moves clipped midday energy into higher-priced hours. A ratio of 1.40 to 1.50 can work.
• Grid export cap. Without storage, anything above the cap is wasted. With DC-coupled storage, the system can run an effective ILR of 1.50 or higher with little loss. See the grid export limitation rules by country for what caps apply where.

All ranges in this list are rules of thumb. The economics depend on the storage price, the tariff structure and the inverter’s DC input tolerance.

Tradeoffs in Plain Language

You are choosing between objectives, and they pull in different directions:

A single ratio cannot serve every objective. Pick yours before you size, not after.

Putting It Together: Design Checklist

The sizing workflow, in the order an experienced designer runs it:

1. Read the inverter datasheet. Note AC nameplate, maximum DC input power, maximum DC input current per MPPT, MPPT voltage window and any temperature derating clauses.
2. Pick the target ILR based on climate, market segment and economic objective. Use the tables above as starting points.
3. Calculate DC nameplate from the target ILR. Round to a whole number of modules per string.
4. Apply a bifacial correction if applicable. Reduce the STC-nameplate ILR by the expected rear gain.
5. Check string voltage at both temperature extremes. Cold Voc must stay below the inverter’s maximum DC voltage, and hot Vmp must stay above the MPPT minimum.
6. Verify per-MPPT current at Isc is below the maximum DC input current.
7. Run a clipping simulation at minute resolution if your tool offers it. Confirm modeled clipping is below the threshold you agreed with the client.
8. Compare cost of energy at the chosen ratio against one step lower and one step higher. The lowest cost wins.
9. Document the design margin for warranty. Keep the DC figure at or below the manufacturer’s maximum DC input.

Steps 7 and 8 need a simulation tool that reports clipping at a fine time step. Check which resolution your tool uses before you rely on its clipping number. Steps 1 to 6 you can do by hand with the datasheet.

Conclusion: Three Actions to Take This Week

1. Audit one recent design. Pull the ILR and modeled clipping. If the ratio is below 1.20, check whether the extra modules would pay. If it is above 1.45, check the warranty limit and the bifacial gain assumption.
2. Pull the datasheets of the inverters you use most. Note the maximum DC input power, maximum DC input current and any temperature derating clause. If your design template does not have these, you are guessing.
3. Run a side-by-side comparison at three ratios on your next quote. Try 1.20, 1.30 and 1.40, and pick the ratio that wins on cost of energy, not the one with the lowest clipping percentage.

The DC/AC ratio is one of the highest-impact choices in solar design, and it is easy to leave at a default. To compare ratios on your own projects, book a SurgePV demo or read how the solar design software handles array layout and yield modeling.

Frequently Asked Questions

What is solar inverter clipping?

Solar inverter clipping is the energy lost when the array’s DC output exceeds what the inverter can convert to AC. The inverter holds its output at its AC rating and the surplus is not exported. It shows up as a flat AC plateau on clear days, and clipping grows as the DC/AC ratio rises.

What is the ideal DC to AC ratio for solar?

There is no single ideal ratio. NREL researchers describe 1.2 to 1.5 as the typical range for new systems, and NREL’s 2024 ATB assumes 1.34 for utility-scale PV. The right value depends on climate, tariff, storage and the inverter’s published maximum DC input.

How much energy does inverter clipping waste?

It depends on the ratio, climate and mounting, and it rises faster at high ratios. In a 2022 NREL study, hourly models overstated annual energy by 0.6% at a ratio of 1.2 and 2.0% at 1.5 for one test site, because they missed sub-hourly clipping. Simulate your own site at minute resolution.

Is DC oversizing worth it for solar systems?

Often yes, when extra modules cost less than a larger inverter and the added output in morning, evening and cloudy hours outweighs clipping at noon. It stops paying when export caps, warranty limits or high clipping remove the gain. Compare the energy value of each added kWp with its installed cost.

Does DC oversizing void the inverter warranty?

It can if you exceed the manufacturer’s published maximum DC input power or current. Limits differ by model, and bifacial gain or high ambient temperature can push real DC above the datasheet figure. Read the datasheet and warranty for the exact model, and get written approval before exceeding it.

What is the inverter loading ratio (ILR)?

The inverter loading ratio (ILR) is the DC nameplate of the array divided by the AC nameplate of the inverter, and it is the same thing as the DC/AC ratio. A 6 kW array on a 5 kW inverter has an ILR of 1.20. NREL describes 1.2 to 1.5 as typical for new systems.

How is clipping different from curtailment?

Clipping is an inverter-side limit, while curtailment is a grid-side limit. The AC rating caps clipped output, but a utility or market operator orders curtailment, usually for network or price reasons. The two can overlap in export-capped systems, so model them as one combined limit.