Solar Calc

CEC vs Peak Efficiency: Reading Inverter Ratings Honestly

ByIndependent solar research and calculators

CEC vs Peak Efficiency: Reading Inverter Ratings Honestly

Look at almost any inverter spec sheet and you’ll find two efficiency numbers that don’t match. One might read “peak efficiency 98.5%,” the other “CEC weighted efficiency 97.5%.” A single percentage point apart doesn’t sound like much, and a salesperson will happily quote you the larger one, but the gap between those two figures is the difference between a best-case laboratory moment and the efficiency your inverter will actually average across a full year of real, uneven sunlight. Knowing which number to trust — and, more usefully, why they differ at all — is what separates reading a spec sheet from being marketed to by one.

The inverter’s job is to convert the DC electricity your panels produce into the AC electricity your home and the grid can use, and no conversion is perfectly lossless. A small fraction of the energy passing through becomes heat inside the inverter’s electronics rather than reaching your loads. Efficiency is simply the share that survives the conversion intact. The catch, and the reason two numbers exist in the first place, is that this share isn’t a single fixed value stamped into the hardware. It changes constantly, minute by minute, depending on how hard the inverter is working at that instant — and any honest attempt to describe the inverter with one number has to grapple with a value that’s moving all day long.

Why a single efficiency number can’t describe an inverter

An inverter’s efficiency depends heavily on its load — how much power is flowing through it at a given moment relative to its maximum rated capacity. Plot efficiency against load and you don’t get a flat line; you get a curve with a distinct shape, and that shape is the whole reason a lone number is misleading. At very light load, in the dim minutes just after sunrise or under heavy cloud, efficiency is genuinely poor. The inverter has its own baseline power draw just to keep its electronics awake and its grid-monitoring active, and when only a trickle of energy is passing through, that fixed overhead eats a large share of it. As production climbs through the morning, efficiency rises quickly and then flattens into a broad plateau, usually somewhere between about 30% and 70% of rated load. That plateau is where a grid-tied inverter spends the overwhelming majority of a normal day, cruising along at high efficiency while the sun does its work. Near full load, efficiency often dips a touch again, because the internal components run hottest and resistive losses grow when the inverter is pushing near its ceiling.

The peak efficiency figure is drawn from the very top of that plateau — the single highest point on the entire curve, achieved at one specific load and one specific input voltage under favorable conditions the manufacturer selected. It’s a real, measured value. It is not a lie. But quoting only that peak is like describing a car by its top speed: technically accurate, occasionally reached, and nearly useless for predicting the miles you actually drive. Your inverter will touch its peak efficiency for brief stretches on good days and spend the rest of its life at every other point on the curve, including the unflattering low-load edges of every single morning and evening. A number that captures only the best instant tells you almost nothing about the aggregate energy the inverter will actually deliver over a year, which is the only quantity that shows up in your production and your savings. This is precisely why a second, more representative number had to be invented — because the peak figure, left on its own, systematically overstates real-world performance in a way that happens to favor whoever is selling the inverter.

It’s worth understanding where the lost energy actually goes, because that’s what gives the curve its shape. An inverter loses power in a few distinct ways at once. There’s a more or less fixed overhead — the energy its own control electronics, cooling fans, and grid-monitoring circuits consume simply to be switched on and standing by — and that fixed draw is brutal at low load, where it’s a large fraction of a tiny throughput, but nearly invisible at high load, where it’s a rounding error against a flood of power. Then there are losses that scale with the power flowing through: switching losses in the transistors that chop DC into AC thousands of times a second, and resistive losses in the transformer, the magnetics, and the wiring, all of which grow as current rises. The interplay of a fixed overhead that dominates at the bottom and current-dependent losses that swell toward the top is exactly what produces the characteristic hump — poor at the edges, best somewhere in the broad middle. Temperature layers onto all of it, since a hot inverter runs slightly less efficiently and, past a point, actively derates its output to protect itself, which is why ventilation and a shaded, cool mounting spot quietly matter to lifetime yield. Architecture enters here, too: a string inverter handles the whole array’s power through one box and rides its curve as a unit, while microinverters each convert one or two panels’ worth of power on their own small curves, so partial shade or a single mismatched panel drags down only its own little converter rather than the entire string. The shapes differ, but the same physics of fixed overhead plus load-dependent loss governs every one of them. Manufacturers know all of this intimately, which is precisely why they can select the single operating point — a specific load, a specific input voltage, a controlled temperature — where their inverter looks its absolute best, and print that number in large type. The peak figure isn’t fabricated; it’s cherry-picked, and the difference matters when the picking is done by the party trying to sell you the box.

What CEC weighting does, and why it’s the honest number

The CEC weighted efficiency figure — named for the California Energy Commission testing protocol that defined it — exists to fix exactly the problem the peak number creates. Instead of reporting the single best point on the curve, it measures efficiency at several load levels spread across that curve and blends them into one weighted average, with the weights deliberately chosen to reflect how much time a real grid-tied inverter actually spends at each load. In practice the protocol samples efficiency at load points such as 10%, 20%, 30%, 50%, 75%, and 100% of rated output, then weights the mid-range points most heavily, because that mid-range is where the inverter genuinely operates most of the time. The result is a number that intentionally folds in the less flattering parts of the curve: the light-load inefficiency at the edges of every day and the slight droop near full power. That’s why the CEC weighted efficiency always comes in at or below the peak figure — typically by a few tenths of a percent up to about a full percent. A European equivalent called “Euro efficiency” does the same thing with weightings tuned to cloudier northern climates that spend more time at lower loads, so on some spec sheets you’ll see that number instead, or listed right alongside the CEC one.

For estimating what an inverter will do on your roof over a year, the CEC weighted figure is the honest predictor, and it should be the number you compare and the number you feed into any production estimate. It already accounts for the reality that your system operates across the full span of loads as the sun climbs and sinks, not just at the flattering peak. If you’re comparing two inverters, lining up their CEC weighted efficiencies is a genuine apples-to-apples comparison; lining up their peak figures can quietly mislead you, because two inverters with identical peaks can have differently shaped curves and therefore different weighted averages — one might hold its efficiency well at low load while the other sags, and the peak number hides that entirely. The practical size of the difference is modest but real and worth putting in dollars. A one-percent efficiency gap on a system producing, say, 12,000 kWh a year is roughly 120 kWh annually. That’s not enormous in any single year, but it compounds across the two-decade-plus life of the system and costs you nothing extra to get right at the point of choosing, which makes it a free improvement rather than a trade-off.

There’s a further subtlety worth knowing, because it reinforces why the honest number is a weighted, condition-aware one rather than a lone best case. Efficiency also depends on the DC input voltage the panels feed the inverter, so some manufacturers report CEC weighted efficiency at three separate voltages — the low, nominal, and high ends of the inverter’s operating window — and the figures differ slightly between them. If your array’s string voltage tends to sit toward one end of that window, as it will in very cold weather when panel voltage rises or very hot weather when it falls, then the efficiency measured near that voltage is the more relevant one for your actual conditions. It’s a small effect, rarely decisive on its own, but it’s one more reminder that a single confident percentage rarely tells the whole truth about a device whose performance genuinely varies with load, temperature, and voltage all at once. Weighted efficiency is also just one input among several that determine how much energy the inverter harvests: its behavior under partial shade, the quality of its maximum power point tracking that squeezes energy from panels operating at different points, and how it’s sized relative to the array all matter too. On sizing in particular, the DC-to-AC ratio interacts directly with the efficiency curve, because an inverter loaded closer to its sweet spot spends more of its hours in the high-efficiency plateau and less at the inefficient low-load edges.

For a homeowner in a genuinely cloudy region, it’s worth noting which weighting a spec sheet used, because the Euro-efficiency protocol leans harder on the low-load points that a gray climate spends more of its time at, and a number tuned to sunny-state conditions can read a touch optimistic where the sun is scarce. The practical instruction stays simple regardless of climate: find the weighted figure, confirm which protocol produced it, and compare like against like — CEC against CEC, or Euro against Euro — rather than pitting one product’s flattering peak against another’s honest weighted average and imagining you’ve learned something useful. When the numbers are matched properly, the differences that remain are small, real, and yours to keep essentially for free by choosing the better-shaped curve.

Reading it on a real spec sheet

When a spec sheet lists both numbers, the simplest working rule is to treat the peak figure as marketing and the CEC weighted figure as engineering. Use the weighted number to compare products against one another and to feed any production estimate you’re building. And if a listing shows only a peak efficiency and conspicuously omits the weighted one, treat that omission as a small signal in itself — the honest, informative, harder-to-flatter number is precisely the one that’s missing, and its absence usually isn’t an accident on a document written to sell.

Picture two spec sheets laid side by side. Inverter A boasts a 98.6% peak and lists a 97.4% CEC weighted figure; inverter B shows a more modest 98.2% peak but a 97.9% CEC weighted number. Read only the peaks and A wins. Read the weighted figures — the ones that reflect a real, uneven year — and B is actually the more efficient machine across the loads it will spend its life at, because its curve holds up better away from that single flattering peak. That reversal is the whole reason the distinction earns your attention: the number that sells the inverter and the number that predicts its output can rank two products in opposite orders, and only one of those rankings ever shows up on your electric bill. While you’re comparing the efficiency lines, a few neighboring specs on the same sheet deserve a glance, because they shape real-world yield as much as a fraction of a percent does. The MPPT voltage range tells you the window within which the inverter can actually track your panels’ output, and an array whose voltage drifts outside that window in extreme heat or cold loses production the efficiency rating never captures. The maximum input current and power ratings govern how much array you can pair with the unit and how much clipping to expect on bright days. And the warranty term, often the quietest line on the page, may matter more to your twenty-five-year outcome than any efficiency digit, because an inverter that fails at year twelve and needs replacing costs far more than the sliver of energy a slightly better curve would ever have saved. None of this requires an engineering degree, only the discipline to read past the largest, boldest number to the ones the manufacturer would rather you skimmed. The bold number is chosen to impress; the smaller, weighted, condition-aware numbers are the ones written to describe how the device behaves on an ordinary Tuesday in April — which is the only behavior that ever reaches your meter.

There’s a reason to keep the whole efficiency question in proportion even as you take the trouble to get it right. The spread between a good inverter and a mediocre one, measured in weighted efficiency, is usually under a single percentage point — real money over decades, but small next to the losses a poorly sited array, heavy shade, or an undersized service can quietly impose. So use the CEC number to break ties and to reject any product that hides it, but don’t let a chase for the last tenth of a percent pull you toward an inverter that’s less reliable, worse supported, or a poor architectural fit for your particular roof. The efficiency figure earns its place as one honest input among several, valuable precisely because it’s cheap to read correctly and easy to be misled on, not because it settles the decision by itself. Get it right, weigh it alongside reliability and fit, and then stop optimizing it — the returns past that point are too thin to chase.

None of this should drive the entire inverter decision, and it’s worth keeping the fraction of a percent in perspective. Reliability, the length and terms of the warranty, the quality of the monitoring, how gracefully the unit handles shade, and how the architecture — string, microinverter, or optimizer — fits your particular roof usually matter more to your long-run satisfaction than a sliver of conversion efficiency. But once you’ve narrowed the field to two solid, comparable candidates and you’re trying to break the tie, the CEC weighted figure is the one that reflects real days full of changing light rather than a single ideal instant staged in a lab. To see how inverter efficiency folds into the overall production picture for your specific system size — alongside sun hours, orientation, and the other system losses that stack up between the panels and your meter — a solar panel calculator puts this one piece in its proper context rather than letting a single spec-sheet number stand in for the whole system.

Related reading