The cruel thing about a microinverter failure is how quiet it is. When one panel out of twenty stops producing, your system output drops about 5% — small enough to hide inside normal weather variation for months. Your bill creeps up a little, the app still shows a healthy-looking production curve, and nobody notices until the annual numbers look off. The fix is simple: learn to read the monitoring data you already have, and give it five minutes a week.
We check our own system’s per-string output weekly, and the habit has paid for itself — not because failures are common, but because the only way to catch a single-panel problem early is to look below the whole-system number.
Know What Your Monitoring Platform Can Actually Show You
Not all solar monitoring is equal, and the differences matter enormously for failure detection:
- Enphase Enlighten (microinverter systems): shows per-panel production. Each microinverter reports individually, so you can see exactly which panel is underperforming, when it started, and by how much. This is the gold standard for spotting a single failing unit.
- Tesla app and most string-inverter platforms: show system-level or string-level output only. You can see that something is down, but not which panel. Detection here means pattern-reading — comparing today’s curve against a known-good baseline and against sister strings.
- SolarEdge (power optimizers): per-panel visibility similar to Enphase, through the SolarEdge monitoring portal.
If you have microinverters, make sure per-panel view is enabled — some installers leave homeowner accounts on the summary view. If you have a string system, your unit of detection is the string, and a single weak panel is genuinely harder to isolate; that’s not a monitoring failure, it’s the architecture.
Establish a Baseline Before You Need One
Anomaly detection only works if you know what normal looks like. Spend your first few weeks of monitoring building three reference points:
- The clear-day curve. On a cloudless day, production traces a smooth arc from sunrise to sunset. Screenshot one for each season. Dips in that arc that repeat at the same time daily are shading (a chimney, a tree, a vent pipe) — normal for your roof, and important to know so you don’t mistake them for faults later.
- Panel-to-panel spread. On a per-panel platform, identically oriented panels on a clear day should produce within a few percent of each other. Note which panels always read slightly low and why (partial shade, different tilt). That’s your fingerprint.
- Monthly totals. Any long-run drift shows up here first on string-level systems.
Symptom-to-Cause: What the Data Is Telling You
Most monitoring anomalies map to a short list of causes. This table covers the patterns worth memorizing:
| What you see | Most likely cause | What to do |
|---|---|---|
| One panel at zero midday, neighbors normal | Failed microinverter, or its DC connection | Note the serial/position, contact your installer — likely a warranty swap |
| One panel missing from reporting entirely (no data, not zero) | Communications dropout (microinverters report over powerline or wireless) | Often self-resolves or fixes with a gateway reboot; the panel may still be producing |
| One panel 10–25% low at the same hours every day | Shading or soiling on that panel | Check the roof sightlines; clean or trim, no service call needed |
| One panel slowly declining over weeks relative to neighbors | Degrading microinverter or panel, failing connector | Watch it, document it, flag to installer if the gap keeps widening |
| An entire string or branch at zero | Tripped breaker, failed branch connection, or string inverter fault | Check your electrical panel first, then call the installer |
| Whole system down, sunny day | Inverter fault, gateway offline, or utility-side issue | Check inverter status lights and your internet before assuming the worst |
| Whole system reading low on a bright day | Clouds you didn’t notice, haze/smoke, high heat (panels lose efficiency when hot), or seasonal sun angle | Usually nothing — verify against weather before reacting |
The single most diagnostic distinction in that table: zero output versus missing data. A microinverter producing nothing is a hardware problem. A microinverter reporting nothing is usually a communications problem, and the panel is often working fine. Platforms display these differently — learn how yours does.
The Five-Minute Weekly Check
Daily monitoring is overkill and breeds false alarms. Weekly is the sweet spot — frequent enough that a failure costs you days of one panel’s output rather than months. Here’s a routine that takes five minutes:
- Open the per-panel (or per-string) view for the best-weather day of the past week. Comparing on a clear day removes cloud noise.
- Scan for outliers. Any panel more than ~10% below its identically oriented neighbors that wasn’t low in your baseline fingerprint gets a second look.
- Check reporting completeness. Same number of devices reporting as last week?
- Glance at the weekly total against the same week’s total from your records. Trending within seasonal norms? Done.
On our own roof the weekly per-string glance is exactly this — thirty seconds of “do the strings still match each other,” and a closer look only when they don’t. String-level checking can’t name the guilty panel, but it reliably tells you when to bring in someone who can.
When to Call — and What to Have Ready
Call your installer (or the manufacturer’s support line) when you have a confirmed hardware symptom: a panel at zero on a clear day, a widening gap between twins, or a whole branch down that isn’t a tripped breaker. Before you call, gather:
- The affected device’s serial number or map position from the monitoring platform
- Screenshots showing the anomaly and the date it started
- A comparison screenshot of a normal day
That evidence usually turns a “we’ll schedule a diagnostic visit” into a “we’ll bring a replacement unit.” Microinverters generally carry long warranties — Enphase’s current IQ series is warrantied for 25 years — so on most systems the part is free and you’re only ever negotiating labor, which some installers cover for a period after installation. Check your paperwork before agreeing to charges.
Don’t Chase Ghosts
A final calibration point: most dips are weather. Clouds, haze, wildfire smoke, heat, snow cover, and the shrinking arc of winter sun all cut production, sometimes dramatically, and none of them are faults. The signal that separates a real failure from noise is asymmetry — one panel behaving differently from its identical neighbors, or one string diverging from its twin. Weather hits everything; failures hit one thing. Keep that rule in mind and a five-minute weekly habit will catch the real problems early without turning you into someone who panics at every cloudy Tuesday.