HOW-TO GUIDE
How to Detect a Faulty Solar Optimizer When One Panel Produces Less Power
By the OhmSavvy Research Team ·
Rooftop PV wiring is live whenever there is daylight
- Solar modules generate voltage as soon as light hits them. Switching off the inverter or the AC breaker does not make the panel wiring dead, and a string can reach several hundred volts DC.
- Never unplug PV connectors, open a combiner box, open the inverter, or touch rapid-shutdown or optimizer hardware. A DC arc does not extinguish itself the way an AC arc tends to, and disconnecting under load can start one.
- If you see scorched connectors, melted cable, smoke from the array or a burning smell, keep everyone away, shut the system down using the procedure on the inverter label if it is safe to reach, call the installer, and call 911 for any fire.
What an optimizer does, and how it fails
An optimizer is a small DC-to-DC converter mounted behind each panel. It tracks the maximum power point of its own module and adjusts its output so the series string can carry a common current without the weakest panel dictating it. In the US, rapid-shutdown requirements in recent National Electrical Code editions are one reason module-level electronics are so common; in the EU and UK, optimizers are used mostly to deal with partial shade on complex roofs. Most optimizers also report per-panel data to the inverter, which is what makes your app's panel layout view possible.
When one fails, it tends to fail in a recognizable way. It may pass no power at all, so the panel reads zero while its neighbors are busy. It may limit output to a fraction of normal and stay there. It may lose its data link, so the app shows a gray or missing tile while the string current looks unaffected. Or it may cut in and out, which shows up as a jagged line where the others are smooth. A damaged module, a bad connector between module and optimizer, or a cracked cell string can produce some of the same pictures, which is why the hardware verdict needs an installer.
Symptoms and likely causes
| Symptom | Likely cause | What to check | Who should fix it |
|---|---|---|---|
| Panel low early or late, normal at midday | Shading from a tree, chimney or vent | Same panel across a full clear day | Owner (trim trees); installer for layout |
| Panel a few percent low every day, better after rain | Soiling: dust, pollen, bird droppings | Look from the ground; compare before and after rain | Owner via a cleaning service |
| Small, constant gap that never changes | Normal module mismatch or ageing | Gap stays within a few percent | No action needed |
| Zero all day while neighbors produce | Failed optimizer, module or connector | Midday reading on a clear day | Installer (warranty claim) |
| Output capped at a flat line while others curve | Optimizer limiting or derating | Shape of the daily curve | Installer |
| Tile gray or missing, string output normal | Optimizer communication fault | Inverter and gateway status | Installer or manufacturer support |
| Whole string drops, inverter shows an error | Open string, ground or arc fault | Error code and time | Installer only |
The column that matters most is the second-to-last. Every shading or dirt explanation leaves the panel close to its neighbors at solar noon on a cloudless day, because that is when shadows are shortest and light is strongest. A hardware fault is indifferent to the time of day.
1Read the daily output pattern
Open the panel-level view for a clear day, and pick the panels that share the suspect's roof face, tilt and orientation. Panels on a different roof plane will legitimately peak at a different time, so comparing an east-facing module with a south-facing one tells you nothing. Then look at three things: the daily energy total for each panel, the shape of each panel's curve, and the suspect's value at around solar noon compared with its row.
As a rough guide, panels of the same model on the same plane commonly land within a few percent of one another. A spread of that size is ordinary manufacturing tolerance and ageing, not a fault. A panel that is consistently 10 to 20% behind deserves a closer look for dirt or partial shade. A panel that delivers a small fraction of its neighbors, or nothing, on a bright day is the signature of a failed optimizer, module or connection. Repeat the comparison over several clear days before drawing a conclusion, since one hazy afternoon or a passing cloud can scramble a single day's numbers.
Two more tells help. If the suspect's curve has the same smooth hump as the others but is shorter, something is reducing the light or the panel's capacity. If it rises normally, then snaps to a flat ceiling and stays there, the electronics are limiting it. Many apps let you export the per-panel data as a spreadsheet; a week of exports makes the installer's job much quicker.
2Checks you can make from the ground
- Walk the shadows. At the times the suspect panel is low, go outside and look at where shadows fall. New tree growth, a satellite dish, a vent stack or a neighbor's extension can shade one module for a few hours without anyone noticing.
- Look, don't climb. From the ground or a window, binoculars will show leaves, bird droppings, a cracked glass sheet or a browned patch. Leave roof access to the installer or a qualified cleaning service.
- Read the inverter. Note any error or warning code with its date and time, and photograph the inverter display. Don't reset it repeatedly hoping the fault clears.
- Check the data link. If several tiles went gray at once, the problem is more likely the monitoring gateway, router or the inverter's internet connection than a group of optimizers failing together.
- Check the layout map. Installers occasionally map the optimizer serial numbers to the wrong positions in the app. If the "weak" tile sits in a spot that never gets shade, the physical panel it represents may be somewhere else.
3How an installer tests the optimizer
You may hear the installer describe tests that sound like ordinary meter work, so it helps to know why they are not. With optimizers, string voltage behaves differently from a plain string: many systems hold each optimizer at a low safety output until the inverter signals it to start, so an open-circuit reading on the string does not look like the sum of the panels' ratings. Interpreting it depends on the specific product's documentation.
A qualified technician will typically shut the system down by the manufacturer's procedure, confirm the rapid-shutdown state, and then work at the suspect position. Checks can include the panel's own open-circuit voltage at its leads with the optimizer bypassed, the optimizer's output in its safety state, a thermal camera scan for hot cells or connectors, and an IV-curve tracer, a dedicated PV tester that sweeps the module from short circuit to open circuit and shows whether its curve matches the datasheet. A DC clamp meter rated for PV voltages confirms string current under production without breaking the circuit. These are tools rated CAT III 1000 V or higher for DC, used with insulated gloves and eye protection, and they are why a homeowner's general multimeter has no part in this. The panel-level background is in testing a single solar panel with a multimeter, which applies to an unconnected module on the ground, not a live roof array.
If the module tests healthy and the optimizer doesn't, the optimizer is replaced and the new serial number is registered in the monitoring system. If the module itself is damaged, the warranty question shifts to the panel maker. Either way, ask for the before-and-after readings in writing.
When to call the installer
Meters installers use on PV strings
CAT III 1500V, 1500 V DC, DC clamp to 999.9 A, DC power and an audio polarity alarm. Built for PV work.
Budget AC/DC clamp to 1000 A with 1000 V DC, rated CAT IV 600V / CAT III 1000V. No PV-specific functions.
True RMS multimeter to 1000 V AC/DC, rated CAT IV 600V / CAT III 1000V, for checks on a panel off the roof.
As an Amazon Associate I earn from qualifying purchases. Prices and availability are shown on Amazon.
Frequently asked questions
Will one bad optimizer drag down the whole string?
Usually not by much. Optimizers exist so that one weak module doesn't set the current for the rest of the string, so the healthy panels keep producing near their normal output. You lose roughly that one panel's share. If the optimizer fails in a way that opens the series circuit, though, the whole string can drop out, and the inverter will usually raise an error.
Can I swap the suspect optimizer with a neighbor's to test it?
Not as a homeowner. It means working on a roof, disconnecting DC connectors on modules that are live whenever light falls on them, and re-registering hardware with the monitoring system. Installers do sometimes swap parts to confirm a diagnosis, under their own safety procedure and with the system shut down correctly.
My panel reads zero only in the morning. Is the optimizer dead?
Probably not. A panel that sits at zero or near zero for the first hours and then catches up with the others is almost always in shade early in the day, from a chimney, a tree, a neighboring roof or another row of panels. A failed optimizer tends to stay at zero, or stay capped, all day including at noon on a clear day.
Does this work the same with microinverters?
The detective work does. Microinverters also report per panel, and the same comparison against neighbors, at midday, on a clear day, separates shading from hardware faults. The difference is that a microinverter converts to AC on the roof, so a failure takes out that one panel's output and nothing else.
Can I replace a solar optimizer myself?
It is not a homeowner job. The optimizer sits under the panel on the roof, in a DC string that can carry several hundred volts in daylight, and the replacement has to be a compatible model that is registered or paired in the monitoring system. Most makers also route warranty replacements through an installer, so a DIY swap can complicate the claim. Send the panel data to the installer instead.
