On a sunny Saturday in Phoenix, Maria watches her inverter light blink. Her array was meant to cover the afternoon A/C, yet output is flat. Two weeks later in Portland, a small commercial site trips its main breaker after export protection failed.

Key Takeaways

  • Three priority checks prevent most faults: vet the installer, verify commissioning, confirm protection settings.
  • Neglected commissioning can turn a well-sized system into recurring downtime. Fixes cost less than repeated service calls.
  • Simple steps catch over 70% of common issues. That includes one 10-minute stability test and a protection readout.
  • Use your retail rate to value lost production. For example, 18¢/kWh today, with wide state variation.
  • Turn downtime into dollars: Lost kWh = system kW × hours × availability factor. Cost = lost kWh × rate.
  • Start early. Strong vetting and contract language decide what gets proven at handover.

Start the work before the first panel is mounted. Good vetting and contract language set what you can test and expect at handover.

How to vet installers and contracts before signing

Strong vetting reduces callbacks and hidden risks. The goal is clear scope, testable deliverables, and proof of competence.

Check the installer’s record. Request evidence of at least 50 completed self-consumption systems (onsite solar use). Ask for three recent local references with addresses and contact names. Confirm licensing and insurer details with current policy dates. Compare at least three detailed quotes to see how workmanship terms shift risk onto the owner.

Define scope in writing. Require a 2-year workmanship warranty that sits alongside equipment warranties. List acceptance criteria for performance and safety tests. Include pass or fail thresholds, retest rules, and punch-list closure steps. If the installer resists measurable outcomes, expect loose ends after handover.

Ask for specific pre-installation documents. These include a single-line diagram (one-page electrical path) and the exact inverter model. Request the firmware version and protection-setting templates. Ask for a proposed test plan. The plan should name instruments, data intervals, and environmental notes.

Set numeric thresholds upfront. Require inverter string mismatch limits at ≤5%. Cap the maximum allowed annual loss-of-production at, for example, 24 hours before remedies apply. Define how remedies work. A clear rule is faster than a dispute.

Request a sample commissioning report from a recent job. Look for time-stamped logs, minute-level data, and photos of labels. Verify that performance was compared to a model adjusted for irradiance and temperature. Owners who read even one past report usually spot gaps fast.

Watch for red flags. Vague phrases like “commissioning performed” without logs are not acceptable. Refusal to set protection values to grid-friendly thresholds is another warning sign. If the team cannot export a settings file on request, their process is weak.

A field observation shows the value. On August 19 at 2:14 p.m., a monitoring screenshot showed 0 kW from one string under clear sun. A loose connector was found and fixed the same day. That quick catch prevented weeks of intermittent faults.

Commissioning and on-site testing: what to verify

Commissioning (first energy-up checks) is your quality gate. The aim is stable operation, safe behavior, and usable data.

Run a sustained generation stability test near peak sun. Use minute-level resolution. Log AC power, voltage, and frequency over a defined window. Set acceptance thresholds. Keep voltage deviation within ≤2% during the test with minimal drift. Record the ambient temperature and irradiance (sunlight intensity) to explain expected output.

Verify maximum power point tracking, or MPPT (inverter optimization of panel output). The inverter should track within 1–3% of expected power under stable light. If clouds pass, note time windows that still qualify for evaluation. Owners who stand beside the tech during this test gain confidence and context.

Capture string-level I‑V readings (current–voltage sweep). Compare strings for mismatch and shade effects. Investigate any string more than 5% off its peers under the same conditions. Photograph torque markings and labels for later reference.

A spring review found one string averaging 5.2 kW while its pair made 4.7 kW. That was an about 9% mismatch traced to micro-shading. Re‑routing the conduit cleared the obstruction and restored balance.

At midday on a clear day, a stability run averaged 3.8 kW on a 5.5 kW array. The model predicted 4.3 kW under those conditions. A quick wash restored the expected range.

Overnight logs from 11 p.m. to 4 a.m. showed three inverter restarts totaling 19 minutes. The reboots did not affect daytime yield. They did delay a scheduled utility event report. That detail helped the owner explain a one-time alert.

Check modeled versus measured output. Example calculation: test at midday under steady light. If a 7.2 kW array averages 5.6 kW, that is about 78% of nameplate. This is acceptable if the model predicted 75–85% for that temperature and irradiance.

Log data access. Ensure the system records minute-level data and that the owner receives credentials. Request cloud access or a CSV export for the first 12 months. Establish who maintains the logger clock and how daylight saving shifts are handled. Without logs, fault reviews become guesswork.

Perform transient checks where safe and permitted. Follow manufacturer procedures only. Confirm the inverter rides through brief dips without nuisance trips. Record firmware versions and settings. If a parameter deviates from the approved design, document it and get written signoff.

One field observation shows why detail matters. On a clear Tuesday at 1:15 p.m., logs showed a 2.3% voltage drift on one phase. MPPT wobbled for two minutes, then stabilized. Tightening a neutral connection stopped the drift and ended random resets.

Deliver a complete package at handover. Include a signed commissioning report. Attach time-stamped production logs for the test window, and a signed list of resolved punch-list items. Keep copies in cloud storage and on paper.

Once stable operation is proven, shift focus to protection settings. Anti-islanding (automatic shutdown when the grid is off) must work as designed.

Protection settings, anti-islanding and grid-safety checks

Correct protection settings keep your system safe and neighbor-friendly. Leave handover with clear numbers and a plain-English summary.

One-page protection summary to record on-site:

  • Set under-frequency trip at approximately 59.3 Hz.
  • Program over-frequency trip around 60.5 Hz.
  • Reconnection deadband: follow values specified in the design.
  • Apply time delays, for example 2 seconds before trip, unless the utility requires otherwise.
  • Limit grid export to the amount approved in the signed scope.

Confirm all setpoints match the approved design. That includes under and over frequency, voltage trips, anti-islanding, and export limits. Request an exported readout of programmed setpoints. Take screenshots and a text file if possible. Owners who have lived through a windy-evening trip know the value of exact thresholds.

Test the trip function using manufacturer-recommended procedures. Do not inject faults into the grid without coordination. Use simulated signals or test ports when provided. Record actual trip times and recovery behavior. Note the reconnection time and whether ramp rates are limited.

Document firmware versions and whether autonomous ride-through is enabled. Keep any deviation from the approved model in writing. Store the settings file with the commissioning report. If a later update changes limits, save both versions.

A short field observation underscores this. During a spring storm at 3:22 p.m., an inverter tripped at 59.1 Hz. It stayed off for five minutes. The delay was set far above the intended value. Restoring the correct delay eliminated nightly nuisance events.

Summary and Recommendation

Follow a focused due diligence path. Do that early. First, vet installers for experience and proof: at least 50 similar installs and three local references. Second, insist on documented commissioning with a defined stability test and minute-level logs. Third, secure explicit protection-setting readouts that match your approved design.

Quantify the risk with your retail rate. For example, 18¢/kWh is a reasonable planning value today. In one example calculation, 48 hours of downtime at 3 kW average equals 144 kWh per year. That translates to roughly $26 in avoided self-consumption value each year.

Strengthen contracts with clear levers. Require acceptance testing with named thresholds and signed logs. Hold a modest retention until 30 days after commissioning, subject to issue closure. Tie the workmanship warranty to measured performance against the agreed criteria.

Owners who apply this checklist usually see fewer recurring faults. Aim for a 70%+ reduction in handover-related issues within the first year. The result is safer operation, steadier output, and fewer surprises when the weather turns.