How do I know if my solar panels are working?

You opened the app, saw a green dot and a graph with some bars on it, and you still can't tell whether your solar panels are working. You're not missing something obvious. You're asking a harder question than the one the app was built to answer.

There are really two questions hiding in this one, and it helps to pull them apart. Are the panels on and reporting? You can settle that yourself in about ten minutes, and the first half of this page walks you through it. Are they producing what they should be producing? That's the second question, and almost nothing you already own will answer it.

If you read nothing else, read this. Your system is on if the gateway is talking to the internet, today's production climbs through the daylight hours, and every panel is reporting. Your system is correct only if that production matches what an array your size, pointed your direction, should have made under the weather that actually happened over your roof. Those are two different tests. The second one is where the money is.

Question one: are they on?

Start here, because it's quick and because a real fault often shows up in the first two minutes. Pick a clear day and do these in order.

1. Look at the hardware itself

Your system has a communications box, usually in the garage, the basement, or next to the main panel. On an Enphase system that's the gateway. On a string system it's the inverter's own display or its cellular stick. Every one of them has status lights and a legend printed in the manual.

Steady green almost always means normal operation. Amber, red, or a blinking pattern means the equipment wants attention, and the manual will tell you which pattern maps to which fault. A box with no lights at all usually means it lost power, and it's worth checking whether a breaker tripped before you call anyone.

2. Watch today's curve in the portal

Open your monitoring app and look at today, not the month. A working array draws a hill. Production starts near sunrise, rises through the morning, tops out somewhere near the middle of the day, and tapers to zero at sunset.

What you're looking for is a shape that quits early. A hill that flattens at 1pm on a cloudless day, or a day that's simply flat at zero while the sun is out, is a real signal and worth a call. Our guide to reading your solar monitoring data goes through the graphs screen by screen if yours is hard to interpret.

3. Check that every panel is reporting

If you have microinverters or optimizers, your portal has a roof-layout view with one tile per panel. Every tile should be reporting a number during daylight. A dark or greyed tile means that panel's electronics aren't phoning home, and on most systems that means those panels aren't producing either.

This is the check people skip, and it's the one that catches quiet failures. A single dead microinverter on a thirty-panel roof costs you a slice of production and never triggers anything you'd notice on the summary screen. If you're deciding whether per-panel data matters, we wrote up when panel-level monitoring earns its keep.

4. Cross-check against something outside the app

Your utility meter and your bill are an independent witness. If the app claims you generated a healthy number of kilowatt-hours last month and your bill shows no offset at all, one of the two is wrong, and it's worth finding out which before you go further.

This also catches the case where the panels are fine but the export side isn't, which is a different problem with a different phone number attached to it.

Green means reporting. It doesn't mean right.

Most people stop here, and the useful part sits just past it. A monitoring platform is a communications tool. It answers one thing: did the equipment phone home, and what number did it send? It has no opinion about whether that number should have been higher.

Your dot is green when a microinverter reports 180 watts at noon in June. Your dot is also green when the same microinverter reports 240 watts under the same sky. Both are "working" as far as the portal is concerned. Only one of them is right.

The two comparisons homeowners reach for both fall apart on contact:

• Comparing to last year. Weather differs year to year. A cloudier summer produces less from a perfectly healthy array, and a sunnier one can hide a real fault behind a bigger number.
• Comparing to the installer's original estimate. That estimate was built on a typical weather year, not on the weather you actually got. It was never meant to be a monthly scorecard, and using it as one produces false alarms and false comfort in roughly equal measure.

We pulled that apart in more detail in expected versus actual solar production, including the losses that are normal and the ones that aren't.

Question two: are they producing what they should?

This is the question the whole page is really about, and once you've seen it you can't unsee it. "On" is a yes-or-no answer about equipment. "Correct" is a number, and it only exists once somebody computes what your specific roof should have produced.

Doing that honestly takes four ingredients:

• Your array as it actually is. How many panels, at what wattage, facing which direction, at what angle, with what shading and at which hours of the day. A generic estimate for "a 9 kW system" tells you nothing about your roof.
• The weather that actually happened. Not a typical year and not a regional average. The sunlight and the temperatures over your address, day by day, for the period you're checking.
• Normal losses separated from abnormal ones. Heat, wiring, dust, and inverter limits all take a predictable bite. Subtracting the expected bite is what leaves a gap that means something.
• A day-by-day comparison. Compare at the annual level and a cloudy fortnight and a failed panel blur into the same number. Compare daily and they separate.

None of that is exotic, and the underlying model is public. If you want the mechanics, we published how OwlWatt calculates your system's expected production, including which inputs go in and why an installer can audit every one of them.

The money nobody writes to remind you about

People spend evenings on state unclaimed-property sites, hunting old escrow refunds and forgotten accounts. The pull isn't the amount. It's the shape of the thing: money somebody owes you, that nobody's job it is to remind you about.

A production guarantee has that shape. If your contract includes one, an installer put a kilowatt-hour number in writing and agreed to make you whole if the system falls short. Nobody at that company has a calendar reminder to check it on your behalf. The difference from unclaimed property is important and I'd rather say it up front: there's no fund sitting in a database with your name on it. The gap between what you were promised and what you got only exists once someone measures it.

So measure it. Use your own numbers, not mine. Take the annual production your system was supposed to make, and take the rate off your own electric bill.

Say the number in your paperwork is 10,000 kWh a year and your rate is $0.30 per kWh. Those are stated examples. Swap in yours before you draw any conclusion.

Illustrative arithmetic on a stated 10,000 kWh year at a stated $0.30/kWh. Use your own expected production and your own rate.

Two things make those rows bigger than they look. Most guarantees are assessed per contract year, so a gap that's real repeats annually for as long as the guarantee runs, often 10 to 25 years. And the smallest row on that table, $150 a year, already costs more than watching for it: OwlWatt is $9.99 a month, or $87.89 a year. At the 10% row it's roughly three and a half times the subscription.

Now the honest half. OwlWatt measures the gap and documents it. It doesn't recover the money. Whether a claim actually gets paid depends on what your contract says, and past a certain point it depends on a licensed attorney in your state who can use the documentation as evidence. I won't promise you a dollar outcome and you should be wary of anyone who does.

The other honest half: the answer may be that there's no meaningful gap. Plenty of systems produce what they were supposed to produce. Knowing that is worth something too. It ends the nagging, and it means the next time your bill jumps you can look somewhere other than the roof.

What to do next

In order, and you can stop at any rung that answers your question.

1. Run the four "is it on" checks above. Lights, today's curve, per-panel reporting, and the bill. If something's plainly broken, that's your call to make this week, while warranty windows are still open.
2. Find your promised number. Dig out the contract and the proposal. You're looking for a guaranteed or estimated annual kilowatt-hour figure, and for any clause about how and when it gets assessed. Our walkthrough of whether your performance guarantee is being met shows where those clauses usually hide.
3. Get an independent expected-output number. Not the installer's, and not last year's. You can start with our rough free solar underproduction estimator to see whether you're in the neighborhood of a problem at all.
4. If there's a gap, document it before you argue about it. A dated, weather-adjusted record of actual versus expected is a different conversation than "it feels low." A production guarantee is only as good as the evidence you bring to it.

One note on who's telling you this. OwlWatt sells no panels, installs nothing, and takes no referral fee from any installer. We're the independent party doing the measurement, which is the only position from which the number is worth anything when you're disagreeing with the company that sold you the system.