Is My Solar System Underperforming? A Step-by-Step Diagnosis

Quick answer: Before you call anyone, establish what the system should produce. Compare a full year of monitoring data against an independent model for your address, tilt and orientation. If the gap is under about ten percent, that is usually normal. If it is larger, work through the causes in order: a fault you can see in the app, shading that has grown, soiling, a failed component, then finally the panels themselves.

What this guide covers

• Step 1: establish the baseline
• Step 2: read the production curve
• The causes, ranked by likelihood
• Step 3: the checks you can do yourself
• What is normal and not worth chasing
• When to call the installer, and what to send
• Common mistakes
• FAQ

Step 1: establish the baseline

You cannot judge a shortfall without a target. Most owners skip this and end up arguing from a feeling.

Your installer’s original proposal should include a modelled annual production figure in kilowatt-hours. Find it. That is the number the system was sold on.

Then build an independent check. Public modelling tools from national laboratories let you enter your address, system size, tilt, azimuth and an assumed loss figure, and return an expected annual output. Run it with honest inputs, including a realistic shading allowance.

Now compare three numbers. The proposal figure, the independent model, and your actual production over a full twelve months. Twelve months matters, because a summer month tells you nothing about a system that fails in October.

If actual output sits within roughly ten percent of the model, the system is behaving normally. Weather varies year to year, and a cloudy year alone can account for several percent.

Step 2: read the production curve

Open the monitoring app and look at three views. A single clear day, a month, and the full year.

The clear day. A healthy day is a smooth arch. It rises through the morning, peaks near solar noon and falls away. Notches or steps in that arch usually mean shading crossing the array.

A flat top at exactly the inverter’s rated output is not a fault. That is clipping, and it is normally a deliberate design choice, explained in inverter clipping.

The month. Look for zero days. A system that produced nothing on a sunny day had a fault that day. Note the dates.

The year. Expect a strong seasonal shape, with summer several times winter output in northern latitudes. A year that steps down and never recovers points to a component failure on a specific date.

If your system has panel-level monitoring, sort panels by output. One or two panels well below their neighbours is a clear lead.

The causes, ranked by likelihood

Notice the order. The dramatic causes are usually simple, and the exotic ones are usually not the answer.

Step 3: the checks you can do yourself

These are safe and take under an hour. Do not open enclosures or touch wiring.

Look at the inverter. Note any fault code and the status light. Photograph the display. Codes are grouped by type, and the manual explains them.

Check the breakers. A solar breaker in the main panel or a separate disconnect can trip. If one has, reset it once and watch whether it holds.

Confirm the data is real. If the app shows nothing, check whether the inverter itself shows production. A dead gateway or a changed wifi password stops the data, not the power.

Walk the array’s shadow path. Go outside mid-morning, midday and late afternoon. Look for shade on any panel. Trees grow, and neighbours build.

Look at the glass. Heavy soiling, bird mess, sap or lichen all cost output. See panel cleaning for when cleaning actually pays.

Check for visible damage. Cracked glass, a discoloured cell, a scorched connector or a hanging cable all warrant a professional visit.

Rule out the bill. A higher bill does not always mean lower production. A tariff change, a new appliance or an electric vehicle can raise consumption while the array performs normally.

What is normal and not worth chasing

Some perfectly healthy behaviour looks alarming.

Winter output is a small fraction of summer output at higher latitudes. That is geometry, not a fault.

Hot afternoons produce less than cool mornings at the same brightness. Panels lose efficiency as they heat up, and roof cells run far hotter than the air.

Output declines slowly every year. Modern panels typically lose well under one percent annually, so a decade old system might produce a few percent less than when new. Read degradation rates for realistic figures.

Light cloud can briefly increase output above the clear-sky value through edge-of-cloud reflection. Short spikes are not an error.

And a year with more cloud than average simply produces less. Compare against a long-run model, not against last year.

When to call the installer, and what to send

Call when production stopped entirely, when a fault code persists after one reset, when a panel or string is clearly dead, or when the annual gap against the model exceeds roughly fifteen percent with no obvious explanation.

Send a package rather than a complaint. Include the twelve month production export, a screenshot of one clear-day curve showing the problem, the inverter fault code and photographs of the display, and the modelled annual figure you are comparing against.

Ask three specific questions. What is the diagnosis? Is the part under warranty, and does that warranty cover labour? What is the expected repair timeline?

Labour is the trap. Many warranties supply a replacement part but do not pay the technician. That distinction is covered in the inverter warranty guide.

If your installer has closed, a different company can usually service the system, provided you hold the plan set and monitoring credentials. Read how to vet a solar company before choosing one.

Common mistakes

Judging on one month. Always compare a full year.

Comparing to a neighbour. Different roof, tilt, shading and system size make it meaningless.

Assuming a higher bill means broken panels. Consumption changes too.

Repeatedly resetting a tripping breaker. A breaker that trips again is reporting a fault.

Cleaning panels aggressively. Abrasives and high pressure damage the surface and can void warranties.

Waiting until the warranty expires. Report faults while cover still applies.

FAQ

How much production loss is normal?
Within about ten percent of a good model is normal, given weather variation. Beyond roughly fifteen percent, investigate.

Why did my solar production drop suddenly?
A sudden step down usually means a component failed or a breaker tripped. Check the inverter status and the monitoring date the drop began.

Do dirty panels really matter?
Usually a few percent, more in dusty or pollen heavy areas. Rain clears most of it in temperate climates.

Can one bad panel affect the whole system?
On a series string, yes. With microinverters or optimizers the effect is contained to that panel.

How do I know if it is the panels or the inverter?
If everything stopped at once, suspect the inverter or a breaker. If one panel or one string is low, suspect that component.

Is lower winter output a fault?
No. Short days and a low sun angle reduce output substantially, especially at higher latitudes.

Should I expect the panels to fail after ten years?
No. Panels usually outlive their inverter. The inverter is normally the first significant replacement.

Where to go next

Read inverter troubleshooting for fault codes, tree shading for the most common creeping loss, and monitoring apps for getting better data.

Independent performance modelling is published by the National Renewable Energy Laboratory, with system guidance from the U.S. Department of Energy.