A solar system underperforming diagnosis checklist identifies whether low photovoltaic production comes from weather, soiling, shading, equipment failure, grid limits, design assumptions, or battery control. Compare actual AC energy with weather-adjusted expectations first, then use monitoring data and safe visual checks before qualified technicians test live DC circuits.
Key Facts at a Glance
A production deficit is meaningful only after comparing the same period with irradiance, temperature, snow, and system availability.
A single low module usually indicates a localized optimizer, microinverter, connector, shade, or module problem.
A complete string with low current but normal voltage commonly points to shading, soiling, a fuse, or a conductor fault.
Module temperature commonly reduces silicon PV power by about 0.35%-0.45% for each degree Celsius above 25°C.
A sustained weather-adjusted deficit above 15% deserves investigation, but the warranty threshold may be different.
Never disconnect DC connectors, open combiner boxes, or megger a live array as a homeowner.
What Counts as Underperformance?
Solar system underperformance means actual AC energy is materially below a defensible expected value after accounting for sunlight, temperature, snow, shading, outages, inverter clipping, and export restrictions. A system that produces less energy on a cloudy week is not necessarily defective.
The energy path is straightforward:
Sunlight → PV modules produce DC → inverter produces AC → home, battery, or utility receives energy.
Expected production normally comes from the original PVsyst, HelioScope, Aurora Solar, or PVWatts model. The model may use a typical meteorological year rather than the actual weather at the property, so a month-by-month comparison can exaggerate or hide a fault.
A useful first calculation is:
Loss percentage = (1 – actual AC kWh ÷ weather-adjusted expected AC kWh) × 100
Use AC meter or inverter output, not module wattage. Module nameplate capacity describes rated DC power under Standard Test Conditions, while the homeowner receives AC energy after inverter conversion and other losses.
Which measurements matter most?
The most useful first measurements are AC kWh, plane-of-array irradiance, module temperature, operating hours, inverter availability, and curtailed energy. Commercial operators should also track performance ratio, specific yield, clipping, availability, and export limitation.
| Measurement | Meaning | Typical use | Main limitation |
|---|---|---|---|
| AC energy | Delivered electricity in kWh | Home bills and monthly trends | Includes clipping and curtailment |
| DC nameplate | Module capacity in kWdc | System sizing and inverter ratio | Not expected daily output |
| Specific yield | kWh per kWdc | Comparing periods or sites | Needs reliable system size |
| Performance ratio | Actual output relative to irradiance | Technical fleet comparison | Sensor quality affects result |
| Availability | Percent of time online | Inverter and outage review | Online status does not prove full power |
| Production ratio | Actual divided by modeled output | Warranty and design review | Model assumptions may be wrong |
Step 1: Establish the Baseline
Establish the baseline by exporting at least 30 days of AC production, retrieving the original design estimate, and comparing matching calendar periods with local irradiance data. A one-day comparison is too noisy for diagnosis because clouds, temperature, snow, and grid events can change output sharply.
Collect these records before changing anything:
- Inverter or monitoring-portal production for the last 12-24 months.
- Original monthly kWh estimates and system size in kWdc.
- Utility interval data, if available.
- Local irradiance, cloud cover, snow, and extreme-temperature records.
- Inverter fault logs, availability reports, and installer service history.
- Battery charge, state-of-charge, and export-limit records where applicable.
Calculate specific yield by dividing AC kWh by installed kWdc. For example, 1,050 kWh from a 7 kWdc array equals 150 kWh/kWdc for that period. Compare the result with the design model and nearby systems, rather than with a national average.
NREL’s PVWatts documentation describes the tool as one that “estimates the energy production of grid-connected photovoltaic systems.” That estimate is a planning baseline, not a guarantee that every month will match the modeled value.
What should trigger an investigation?
A sustained 10%-15% weather-adjusted deficit is a practical investigation threshold, while a 20% or larger deficit usually merits prompt professional review. These are screening rules, not universal warranty standards; contracts may specify annual energy, availability, or performance-ratio guarantees instead.
| Observed deficit | Likely interpretation | Recommended response | Review period |
|---|---|---|---|
| 0%-10% | Weather, model variation, normal losses | Continue monitoring | 30-90 days |
| 10%-15% | Possible soiling, shade, sensor, or availability issue | Perform remote and visual checks | 14-30 days |
| 15%-25% | Probable unresolved system issue | Arrange structured diagnosis | 7-14 days |
| More than 25% | Major outage, inverter, string, or export problem | Escalate promptly | Same week |
| 100% loss | Shutdown, grid outage, breaker, or monitoring failure | Check status and safety conditions | Same day |
Step 2: Normalize for Weather
Normalize production against actual irradiance before blaming the equipment. A region receiving 20% less sunlight than the model period can produce roughly 20% less energy, although temperature, cloud structure, snow, and inverter behavior prevent a perfect one-to-one relationship.
Use plane-of-array irradiance when available because a south-facing tilted array does not receive the same sunlight as a horizontal weather station. A commercial pyranometer or calibrated reference cell near the array is more useful than a distant airport sunshine-hour total.
Temperature has two opposing effects. Hot cells reduce instantaneous voltage and power, while cold clear days can produce strong output. Silicon modules commonly have a maximum-power temperature coefficient near -0.35% to -0.45% per °C, but the precise value belongs to the module datasheet.
Snow requires separate treatment. A thin, bright snow cover can block nearly all production even when surrounding ground appears sunny, while edge melt can create partial-string behavior. Do not climb onto a roof to remove snow unless the array manufacturer and a qualified professional approve the method.
How does the time pattern reveal the cause?
The production shape often identifies the fault before a technician visits. A midday-only collapse suggests thermal derating, grid overvoltage, or export control; a missing morning or afternoon segment suggests shade, tracker behavior, or an orientation-specific issue.
| Time-series pattern | Strongest candidates | Confirming evidence | First action |
|---|---|---|---|
| Flat zero all day | AC breaker, inverter shutdown, grid outage | Inverter status and utility power | Check portal and electrical panel |
| Normal morning, zero after noon | Overheating, grid voltage, export limit | Fault timestamps and inverter temperature | Review midday alerts |
| Lower output every morning | New tree or building shade | Shade profile by season | Compare clear-day curves |
| One abrupt permanent step down | Failed inverter, optimizer, fuse | Event date and component data | Open warranty case |
| Uniform gradual decline | Soiling, sensor drift, degradation | Visual inspection and irradiance comparison | Clean or calibrate |
| Sawtooth midday clipping | DC oversizing or control limit | Repeated flat power ceiling | Compare inverter rating |
Step 3: Read Monitoring Data
Read the monitoring portal at 5-minute, 15-minute, or hourly resolution and compare component-level output on the same clear day. Enphase Enlighten, SolarEdge Monitoring, SMA Sunny Portal, Fronius Solar.web, and many utility portals expose different levels of data, so missing panel data does not prove panel failure.
Start with system status, then work downward:
- Record inverter uptime and every fault code.
- Check whether the portal shows estimated data instead of measured data.
- Compare each string, optimizer, or microinverter with electrically similar neighbors.
- Look for repeated faults at the same clock time.
- Compare DC power, AC power, voltage, current, and temperature.
- Check clipping, curtailment, and grid-export limits.
- Verify the monitoring gateway has not lost internet communication.
A single low microinverter channel suggests a localized component, connector, shade, or module issue. A complete low string points toward shading, a blown fuse, open conductor, connector failure, or inverter input-channel fault. Similar low output across every component points more strongly toward weather, soiling, grid control, sensor error, or an incorrect baseline.
A monitoring gap is not the same as a production gap. Utility interval data and the inverter’s internal lifetime counter can distinguish a communications failure from an energy failure.
Step 4: Inspect the Array Safely
Inspect modules from ground level or with binoculars for soiling, bird droppings, leaves, snow, cracked glass, delamination, browning, and new shade. Roof access, module removal, DC connector work, and electrical enclosure opening require appropriate fall protection and qualified personnel.
Check the array perimeter as well as the module faces. Tree growth, a new satellite dish, chimney shadow, construction scaffolding, overhead wires, and seasonal vegetation can create losses that were absent during commissioning.
Soiling usually produces a broad, gradual decline, but one large bird dropping can create a sharply localized hot cell. Agricultural dust, pollen, marine salt, and ash may remain after light rain. Rain cleans some dust but can leave mud rings and mineral deposits.
Cleaning is not automatically beneficial. The U.S. Department of Energy notes that rain often provides adequate cleaning for many residential systems, whereas dry climates, low panel tilt, agricultural exposure, and bird activity increase the case for scheduled washing.
When should panels be cleaned?
Clean modules when measured soiling loss exceeds the cleaning cost, or when visible deposits cover active cell areas. Use early morning or evening conditions, manufacturer-approved water and a soft nonabrasive tool, and avoid pressure washers, abrasive brushes, detergents that leave films, and cold water on heat-stressed glass.
| Site condition | Typical cleaning interval | Likely soiling concern | Practical decision |
|---|---|---|---|
| Rainy, 30° or steeper roof | 12-24 months | 2%-5% annual loss | Inspect before scheduling |
| Dry suburban site | 6-12 months | 5%-10% seasonal loss | Compare clear-day curves |
| Agricultural or construction site | 1-3 months | 10%-30% without cleaning | Measure before and after |
| Coastal location | 3-6 months | Salt film and corrosion | Use demineralized rinse |
| Heavy bird activity | As needed | Local hotspots | Add deterrence or guards |
| Snow-prone roof | Weather-dependent | Near-total temporary loss | Use safe professional removal |
Step 5: Perform Electrical Tests
Electrical tests can confirm a solar fault, but live PV strings can produce lethal voltage and dangerous DC arcs even when the AC disconnect is off. Only qualified solar electricians should isolate strings, measure open-circuit voltage, measure current, perform insulation-resistance tests, or open combiner equipment.
A technician generally compares string open-circuit voltage, operating voltage, and current with the module count, datasheet temperature coefficient, and neighboring strings. A low voltage can indicate missing modules, an open connector, bypass-diode behavior, or wiring damage. Normal voltage with low current more often indicates shade, soiling, a fuse issue, or a high-resistance connection.
Insulation-resistance testing, often called a megger test, looks for leakage from conductors to ground caused by pinched cable, water ingress, damaged insulation, or connector contamination. Sensitive electronics must be isolated according to manufacturer instructions before testing because an incorrect test can damage optimizers or inverters.
| Test | What it measures | Fault signature | Safety requirement |
|---|---|---|---|
| String Voc | Open-circuit DC voltage | Low or absent voltage | Qualified technician, isolated circuit |
| String operating voltage | Voltage under load | Mismatch between strings | Meter rated for PV DC |
| String current | Current under sunlight | Low current with normal voltage | Calibrated clamp meter |
| Insulation resistance | Leakage to ground | Low resistance or unstable value | Electronics isolated correctly |
| IV-curve trace | Current versus voltage | Fill-factor and mismatch defects | Specialist PV test equipment |
| Polarity check | Positive and negative orientation | Reversed conductors | De-energized, rated equipment |
Do not reset a fault repeatedly when an inverter reports insulation failure, arc fault, ground fault, or overtemperature. Repeated resets can conceal a worsening hazard and may complicate warranty evidence.
Step 6: Confirm With Thermal and IV Tests
Thermal imaging is useful when performed under strong, reasonably uniform sunlight and interpreted by a trained technician. A hotspot, diode anomaly, connector heating, or mismatched module can appear as a temperature pattern, but reflections, wind, cloud edges, and different mounting conditions can create false positives.
A useful infrared scan normally occurs near peak irradiance with the array producing meaningful current. The technician should record irradiance, ambient temperature, wind, inverter state, and image location. A thermal image without operating context is evidence for follow-up, not proof of a failed module.
An IV-curve test gives more diagnostic detail than a simple voltage reading. It can reveal reduced short-circuit current from shade or soiling, reduced fill factor from resistance, and abnormal knee shape from cell or bypass-diode problems.
Electroluminescence imaging may reveal microcracks that thermal imaging misses, but it generally requires specialized equipment and controlled conditions. It is most useful for disputed module defects, manufacturing claims, and large commercial arrays.
Which Cause Matches the Symptom?
The best diagnosis comes from combining symptom, location, time pattern, and electrical behavior rather than selecting the most familiar fault. Inverter faults affect blocks or the whole array, while shade and module defects usually create repeatable component-level signatures.
| Symptom | Probable cause | Confirmation | Corrective path |
|---|---|---|---|
| Entire system is zero | AC breaker, grid outage, inverter failure | Status screen and utility check | Qualified reset or service call |
| One microinverter is zero | Microinverter, connector, or module | Neighbor comparison | Component warranty claim |
| One string has low current | Shade, soiling, fuse, conductor | String current comparison | Inspect, then electrical test |
| One string has low voltage | Open circuit or missing module | Voc and continuity testing | Technician repair |
| Midday shutdowns | Grid overvoltage or heat | Fault timestamps and voltage | Utility or inverter investigation |
| Repeated flat power ceiling | Inverter clipping or export limit | DC and AC power traces | Review design and utility settings |
| Uniform seasonal decline | Weather, snow, shade, soiling | Irradiance and visual records | Normalize, clean, or trim |
| Rapid unexplained decline | Hardware, water ingress, sensor error | Event log and IV or thermal test | Escalate promptly |
| Battery reaches full early | Charge controller or full battery | State-of-charge history | Check loads and charge settings |
| Utility credits fall only | Rate or billing change | Metered kWh versus bill | Contact utility, not installer |
Grid overvoltage deserves special attention. A distribution circuit with high daytime voltage can cause an inverter to reduce output or disconnect under its configured interconnection rules. Replacing the inverter without measuring voltage at the inverter terminals may leave the real cause untouched.
Which Checks Are Safe to Do Yourself?
Homeowners can safely perform remote data review, utility-bill comparison, ground-level observation, and monitoring-alert setup. Homeowners should not open DC disconnects under load, unplug MC4-style connectors, enter a roof array, remove modules, or test conductors with a multimeter.
A safe homeowner checklist includes:
- Confirm the utility has power and the inverter portal has recent data.
- Photograph the inverter display without opening its enclosure.
- Check whether the dedicated AC solar breaker is visibly tripped.
- Record the exact fault code and timestamp.
- Compare a clear day with a similar clear day from the prior year.
- Look for shade, snow, leaves, bird droppings, and visible damage from ground level.
- Check whether the monitoring portal reports estimated or missing data.
- Contact the installer when faults persist, production is zero, or any DC warning appears.
A breaker reset may be appropriate only when the manufacturer’s instructions specifically permit it and no burning smell, water, exposed cable, repeated trip, or arc-fault warning exists. Never keep resetting a breaker that trips again.
How Much Does Diagnosis Cost?
Remote diagnosis commonly costs $0 when the portal is accessible, while a professional site visit typically costs $200-$450 in the United States. Thermal imaging, IV-curve testing, roof access, travel distance, commercial reporting, and after-hours work can increase the total.
These are typical market ranges, not guaranteed prices. Warranty coverage may reduce the homeowner’s cost when the failed component and labor qualify under the installer or manufacturer terms.
| Service | Typical price | Typical duration | Best use |
|---|---|---|---|
| Remote monitoring review | $0-$150 | 1-2 hours | Fault codes and production trends |
| Visual and electrical site visit | $200-$450 | 2-4 hours | Strings, inverter, and wiring |
| Thermal drone or roof scan | $300-$600 | 1-2 hours onsite | Large arrays and hotspots |
| Professional module cleaning | $150-$350 | 1-3 hours | Soiling-loss confirmation |
| Inverter replacement | $1,500-$3,500 | 3-6 hours | Out-of-warranty failure |
| Commercial IV-curve survey | $500-$2,000+ | 2-8 hours | String and module ranking |
Ask the technician whether the quote includes a written report, inverter-log extraction, string measurements, thermal images, travel, roof access, and a warranty submission. A low headline price can exclude the test that actually identifies the fault.
DIY or professional diagnosis?
| Approach | Cost | Detectable issues | Main limitation |
|---|---|---|---|
| Monitoring review | $0-$150 | Outages, clipping, component gaps | Cannot prove hidden DC faults |
| Ground inspection | $0 | Shade, snow, dirt, visible damage | Cannot inspect backsheets or wiring |
| Electrician testing | $200-$450 | Voltage, current, insulation faults | Requires scheduling |
| Thermal survey | $300-$600 | Hotspots, connectors, bypass diodes | Needs suitable sunlight and skill |
| Full commercial survey | $500-$2,000+ | Fleet ranking and root-cause evidence | Poor value for a small array |
How Should Different Systems Be Diagnosed?
Residential systems need portal alerts and clear-day comparisons, commercial systems need irradiance-normalized fleet analytics, and off-grid systems need battery and charge-controller analysis. The same low AC output can represent a fault in one architecture and normal energy management in another.
Residential arrays: Set email or push alerts for inverter faults, inspect nearby shade twice yearly, and retain annual production reports. A microinverter system makes one-panel comparisons easier, while a central string inverter requires string-level interpretation.
Commercial arrays: Track performance ratio, availability, irradiance, clipping, tracker position, inverter temperature, and export limits. A SCADA system should distinguish communication loss, curtailment, and equipment downtime, because all three can appear as missing production.
Off-grid systems: Check battery state of charge, charge-controller current, battery temperature, generator settings, and load dispatch. A full battery can intentionally reduce PV production because the charge controller has nowhere safe to send additional energy.
Systems with optimizers: Compare optimizer-level voltage and power, but remember that portal estimates can persist after communications loss. Confirm suspected optimizer faults with installer diagnostics rather than replacing modules based on one anomalous data point.
Mistakes That Produce False Diagnoses
False diagnoses usually come from using the wrong comparison period, confusing data availability with energy production, or treating a model estimate as a guaranteed output. Correcting the measurement method often resolves the apparent underperformance without replacing hardware.
- Comparing unrelated days: Compare similar irradiance and temperature periods, preferably using 15-minute or hourly data.
- Using utility credits as production data: Rate changes, net-metering rules, taxes, and billing-cycle dates can reduce credits while metered kWh remains unchanged.
- Ignoring inverter clipping: A DC array larger than the inverter can produce a repeated midday ceiling by design.
- Cleaning hot modules with cold water: Thermal shock and slipping hazards make midday washing unsafe.
- Resetting faults without records: Photograph codes and export logs before resetting so intermittent events remain documented.
- Blaming annual degradation for a sudden loss: Typical crystalline-silicon degradation is often around 0.5%-0.8% per year, while a sudden 20% step change suggests another cause.
- Trusting one thermal image: Reflections and wind can mimic hotspots; correlate infrared patterns with electrical measurements.
- Ignoring export control: A utility-imposed zero-export or grid-support setting can limit AC output even when the array and inverter are healthy.
One practitioner rule is especially valuable: compare components only with electrically equivalent components under the same sun. A west-facing module should not be judged against a south-facing module at 10 a.m.
Repair, Warranty, and Escalation Path
Document the deficit before authorizing repairs, then escalate from data evidence to inspection, testing, and warranty action. A strong service request includes the affected dates, expected and actual kWh, weather source, screenshots, fault codes, component identifiers, and a clear request for root-cause findings.
Use this sequence:
- Export production and fault data.
- Confirm the baseline and weather adjustment.
- Check utility status, AC breaker condition, and monitoring communication.
- Record visible shade, soiling, snow, and damage.
- Ask the installer for a diagnostic visit if the deficit persists.
- Request string, inverter, insulation, thermal, or IV testing as appropriate.
- Obtain a written repair scope and warranty determination.
- Verify production after repair during comparable weather.
The installer should not close a case merely because the inverter reconnects. Verify the affected string or component against neighboring output and confirm that fault recurrence has stopped.
What evidence helps a warranty claim?
Warranty evidence is strongest when it links a specific component to a repeatable production signature. Include portal screenshots, serial numbers, installation date, weather records, technician measurements, and post-repair output.
Manufacturer warranties differ substantially. Product warranties may cover modules or inverters for 10-25 years, while workmanship coverage may be shorter and labor reimbursement may require approved procedures. Read the actual agreement rather than relying on a sales summary.
FAQ
Can a solar system underperform in its first year?
Yes. First-year output can fall below the design estimate because the model uses typical weather, while the actual year may be cloudier, hotter, snowier, or more shaded. A first-year assessment should use irradiance-normalized data and verify commissioning settings before treating a modest shortfall as an equipment defect.
How much does shade reduce solar panel output?
Shade loss ranges from minor to severe, depending on timing, area, module wiring, bypass-diode behavior, and system architecture. A narrow shadow across one cell row can reduce a module disproportionately, while recurring shade across several series-connected modules can reduce an entire string during the affected hours.
Should I replace the inverter or panels first?
Replace neither until monitoring and electrical testing identify the failed component. A failed inverter can affect the complete array, but grid overvoltage, export limiting, a tripped breaker, or a wiring fault can create the same symptom and make replacement ineffective.
Does rain clean solar panels enough?
Rain often removes loose dust from adequately tilted residential modules, but it may not remove bird droppings, agricultural residue, salt, ash, or mineral film. Compare output before and after professional cleaning when the array has visible deposits and the suspected loss exceeds the cleaning cost.
Why does my battery system stop producing solar at noon?
A battery system may curtail PV when the battery is full, household loads are low, export is restricted, or the charge controller reaches a configured limit. Check state-of-charge, battery temperature, load demand, export settings, and inverter alerts before diagnosing a module or inverter failure.
How long should a solar diagnostic visit take?
A basic site visit usually takes 2-4 hours, while a remote review can take 1-2 hours. Large commercial systems, intermittent faults, roof access, thermal surveys, and IV-curve testing may require additional visits or a full operating-day observation.
The Bottom Line
A solar system underperforming diagnosis checklist should begin with weather-adjusted AC production, not immediate panel replacement. Establish the baseline, inspect monitoring data, check safe external conditions, and use qualified electrical or thermal testing when the evidence points to strings, wiring, insulation, grid voltage, or internal module faults. A sustained deficit above 15% is a practical escalation point, while zero production or repeated safety faults require faster action.