Your inverter isn't producing. What to check before calling service

Zero is a different problem from "somehow less". When a system produces less, you're looking for a gap against expectation and the answer is usually boring — weather, orientation, shade. When it produces nothing, something is interrupted, and it costs money on every sunny day.

The good news: in a fair share of cases it's only the app that stopped. The panels and the inverter are working, the data just isn't getting anywhere. That's why the order below matters — steps 1 and 2 rule out the most common reason people call service for nothing.

(Looking for "it works, but produces too little"? That case is covered in your solar is producing less than expected.)

1. It's a time when zero is correct

It sounds trivial, but it's the most common false alarm. An inverter doesn't only switch off at night — it needs a minimum input voltage to start, so on a heavily overcast winter day it can hover around zero all day. Morning start-up and evening shutdown are gradual, not a step.

The inverter knows about this state and reports it. In SolaX's table it's code 100, "waiting for operation" — a normal state, not a fault. Our diagnostics therefore never raises it as a finding.

When it's worth going further: the day is clear, the sun is on the panels, and output stays at zero.

2. Is the system down, or only the reporting?

Data from your system travels from the inverter to a communication module (a Wi-Fi or LAN "dongle"), from there to the manufacturer's server, and only then to the app. Any link in that chain can break without anything touching the panels.

Signs that only the reporting is broken:

• The inverter display shows output while the app shows zero.
• The meter isn't spinning into import at midday — the house is drawing its electricity from somewhere other than the grid.
• Zero started in a step, at one specific moment, typically after a power cut, a Wi-Fi password change, or a new router.

The usual culprit is a dongle that lost Wi-Fi. The fix is normally a restart or reconnecting it to the network per the manufacturer's instructions — and it isn't a fault anyone needs to fix on the roof.

3. The inverter is reporting a fault itself

This is the most valuable piece of information available, and it's routinely overlooked. An inverter usually knows about its own fault and says so with a status code — you don't have to infer anything from production data.

The crucial distinction is between two classes:

• A recoverable fault (code 103 on SolaX) — a reaction to circumstances, typically voltage or frequency fluctuation on the grid. The inverter disconnects itself, waits, and tries again. Once in a while this is normal operation.
• A permanent fault (code 104) — the inverter won't resolve this one on its own. While it lasts, the system probably isn't producing, and it's a case for service.

The difference has its own article: inverter status codes.

4. A tripped breaker or an open isolator

A solar system can be switched off in two places: the AC breaker in the consumer unit and the DC isolator at the inverter. Both get switched off from time to time — after an inspection, after electrical work, or after a storm with a voltage surge.

You can tell by looking: the breaker in the off position, the rotary isolator on "0" instead of "1". Don't do anything beyond looking — if a breaker tripped on its own, something tripped it, and switching it back without knowing what is a bad idea. What you found is information for service, not an instruction to act.

5. A grid problem, not a system problem

An inverter is required by regulation to disconnect when grid voltage or frequency leaves the permitted range. In practice this mostly affects the ends of long rural feeders and areas dense with solar: at midday, when everyone is exporting, grid voltage rises and inverters disconnect one after another.

The signature is distinctive: zero through the best midday hours, normal output in the morning and afternoon. It isn't a fault of your system and service can't fix it alone — it's resolved with the distribution network operator, possibly by a settings change only a qualified person may make.

6. Covered panels

Snow can cut output to practically nothing and, at our latitudes, sit there for days. A thick layer of leaves does the same, as does — in the extreme — a tarp left after work on the roof. It isn't a defect, but it's worth knowing that the cause may be visible from the garden.

How long has this been going on?

Almost nobody asks this early enough. An outage is easy to recognise: a month in which the system produced practically nothing where there should have been output. In our diagnostics that's a rule triggering below 2 % of the expected monthly production, and it's reported straight away as a serious problem.

The problem isn't recognising it — it's that nobody is looking. An outage found a quarter late costs a quarter of production, and it's harder to prove when it started if you ever claim. Which is the argument for monitoring that notices by itself, rather than opening an app regularly and hoping you spot it.

What to tell service

The call will be much shorter if you have:

1. When zero started — the day, ideally the time, and whether it was a step or gradual.
2. What the inverter display shows compared with the app.
3. The status or error code, if one is showing.
4. Whether it repeats — constantly, or only at certain times of day.
5. What you've already ruled out — breaker on, panels clear of snow, app connected.

Where Slunora helps

The quick check spots a missing month — enter monthly output and an outage appears as a finding with the month named, not as an impression.

If you have a SolaX, you can connect the inverter and the watching runs by itself: we read several times a day and take the inverter's status code and battery state exactly as the device reports them about itself. That's the one kind of finding where we infer nothing from production data — which also makes it the most certain finding we can give.

And the other way round: if the data doesn't point at an outage, we say so. Zero in the app and zero in production are not the same thing, and telling them apart is worth one look at the data.