Producing less than it used to? Cleaning vs. a real repair.
Short answer: A solar production drop is one of two things: soiling (dirt, pollen, ash, fixed by a warranty-safe clean) or a fault (a dead microinverter, optimizer, connector, or string, fixed by a licensed C-46 repair). They look different in your monitoring app. Soiling is gradual and even, and it recovers after rain or a clean. A fault is sudden, or shows one panel or string flat-lining while the rest are fine, and cleaning does nothing. Reading the difference before you spend money saves you buying a $600 diagnostic for a $200 clean's job, or worse, cleaning a system whose real problem is a failed inverter.
What soiling looks like
Dirt, pollen, and ash build up slowly and land on the whole array roughly evenly. In your monitoring app, soiling shows up as:
- A gradual, even decline, every panel dims a little; the whole system's peak output drifts down over weeks
- Recovery after rain, a good rain restores some production, which is the tell that the loss was on the glass
- A seasonal pattern, worst in late summer (dust) and after wildfire smoke; better in the wet months
- No error codes, the inverter is healthy and reporting; it's just working with less light
Typical soiling loss in the Sacramento Valley runs 6 to 12% on panels not cleaned in 6+ months, and higher in the dust-and-fire zones of Yolo and Solano. This is the case a warranty-safe clean fixes, soft-brush + softened water, never pressure washing.
What a fault looks like
An electrical failure behaves nothing like dirt. In the app it shows up as:
- A sudden step down, production drops on a specific day and never comes back, rain or not
- One panel or one string at zero, on Enphase or SolarEdge (which report per panel), a single module or a whole string reads flat while its neighbors are normal
- An error code, ground fault, arc fault, DC voltage fault, or a rapid-shutdown that won't clear
- A total blackout, no data reporting at all, which usually means the inverter or the monitoring gateway itself
- Heat-linked dropouts, production that cuts out in the afternoon heat and returns when it cools, the signature of a failing connector
None of this is fixed by cleaning. It needs a licensed C-46 repair, the fault isolated from your monitoring data, the failed component tested and replaced, and the system re-commissioned.
The cheap test that settles it
When it's genuinely unclear, the honest and cheapest first move is a warranty-safe clean with a before/after monitoring reading. It costs a fraction of a repair diagnostic, and it answers the question definitively:
- Production comes back after the clean, it was soiling. You're done, and you have the kWh recovery documented.
- Production doesn't move, it's a fault, and now you have data that proves it. The repair specialist starts a step ahead instead of running a diagnostic from scratch.
A reputable directory won't sell you a repair when a clean would do, or a clean when the inverter is clearly dead. The monitoring data tells the truth; the job is to read it before quoting.
When in doubt, look at the shape of the loss, not the size. A 10% drop that's even across every panel and eased after the last rain is dirt, book a clean. A 10% drop that appeared overnight, or that traces to one dark panel in the app, is a fault, book a repair. The annual health check is the middle path: it reviews the monitoring, thermal-scans for microcracks, and tells you which bucket you're in before anything breaks.