SunPower Panels “Not Working”? Dirt Is the Problem More Often Than Defects

If your SunPower system’s production numbers have been sliding for weeks, don’t assume the panels are failing. Check the dirt first. On the eleven commercial systems I oversee, soiling-related losses produced more “not working” tickets than actual SunPower panel faults — and one cleaning recovered 18.7% of daily output on a system that had been flagged for a warranty claim. I’ve been signing cleaning and maintenance invoices for six years, and this is where I’ve landed: solar PV panel cleaning is one of the highest-ROI maintenance items we approve, but only on the right sites, at the right frequency. On some sites, the smartest maintenance decision is no clean at all.

Quick background so you know who’s talking. I’m the procurement manager at a small solar O&M contractor. I manage a $1.3M annual maintenance budget, I’ve negotiated with more than 30 cleaning and service vendors, and every number below is tied to an invoice or a production log in our cost tracking system.

Why dirty panels look like dead panels

Solar panels rarely fail the way a light bulb does. Real panel faults usually show up as sharp, step-shaped drops — a string goes to zero, an inverter throws a fault code, a microinverter stops reporting. Soiling doesn’t do that. Soiling is a slow creep: output drifts down 1% one week, 0.8% the next, flattens after rain, then keeps sliding during the next dry stretch.

If your weather-normalized production curve looks like a gentle slide and your site is dusty, agricultural, or near a construction zone, the cause is almost certainly dirt — not degradation.

Here’s the counterintuitive part: a 5% seasonal soiling loss costs you more energy in one quarter than ten years of SunPower’s warranty-grade degradation. SunPower panels are warrantied to degrade at roughly 0.25% per year, which works out to about 92% of original output after 25 years. That’s excellent, and it doesn’t matter much if a film of dust is eating 5% of your output every quarter. You can pay a premium for low-degradation hardware and lose the entire advantage to a dirty glass surface. The monitoring dashboard can’t tell the two apart, because both look identical on a line chart.

This isn’t a niche concern. NREL’s soiling research has put average global PV losses at 3–4% of total energy, with far higher numbers in dry regions. Failing to budget for cleaning is failing to budget for reality.

The cleaning line item I cut — and the $7,000 lesson

In my first year, I thought panel cleaning was a low-value upsell. It wasn’t in our maintenance template. We had a formal process for inverter service, torque checks, and thermal scanning, but nothing for soiling. That process gap cost us about $7,000 on one mid-sized commercial rooftop in 2022.

Here’s what happened. The client said their SunPower solar panels were not working. Normalized output had been sliding for eight weeks, from 4.2 kWh/kW per day down to 3.4. The inverter showed no faults. We sent a technician out twice; both times the system checked out as nominal. The client was understandably frustrated — the system was plainly earning less than it had the year before.

On the third visit, an experienced tech took thermal images and found uneven heating across the array. The culprit wasn’t failed cells. It was a thin layer of agricultural dust, pollen, and bird droppings, barely visible from ground level. We finally scheduled a deionized-water clean: $1,340. Output recovered by nearly 20% within a week.

Now the real cost. Two unnecessary service dispatches added up to about $1,100. Eight weeks of suppressed production during peak pricing cost roughly $4,200. Then the clean itself. Total: around $6,600, and the damage to the client’s confidence was worse than the invoice.

That was the first time. After the third false “not working” report in the same region, I made the fix permanent: before dispatching for a module fault, someone checks the 30-day weather-normalized trend and the site’s soiling history. It should not have taken three incidents for me to build that step.

What solar PV panel cleaning actually costs

Here’s where our invoices have landed in the last four years:

• Ground-mount arrays with good access: $0.45–$0.70 per panel. At this price, cleaning almost always pays when soiling is visible.
• Commercial rooftop arrays: $1.20–$2.10 per panel, depending on height, tilt, and water access. A typical 400 kW rooftop clean runs us $1,400–$1,900.
• Systems that need rope access or a lift: $2.50+ per panel. This is where the ROI math flips.

The biggest cost driver isn’t the washing. It’s logistics — water supply, containment, safety certification, insurance. Contractors who show up with a pressure washer and tap water are cheaper for a reason: hard water leaves mineral deposits that can make next season’s soiling worse. We only approve deionized water on commercial sites.

Here’s the framework I use before approving a cleaning. Take a 400 kW array with 6% annual soiling loss. That’s roughly 24,000 kWh per year. At a commercial rate of $0.21/kWh, the loss is worth about $5,040. Two full cleans at $1,600 each net you around $1,840 — before counting the softer benefits like lower thermal stress and cleaner production data.

But change the inputs and the answer flips. At an $0.08 export rate and $2.50-per-panel access cost, that same array needs a soiling loss above 25% before washing earns its keep. The cleaning decision is a math problem, not a rule of thumb.

When I recommend doing nothing

Not every site needs a cleaning contract. I manage a cluster of SunPower installations in Sydney, several with battery storage, and regular rainfall there does most of the work for free. We schedule one clean per year on those systems, usually after the dry season, and that’s it. Some sites in genuinely wet climates never need a paid wash.

If you have a battery, there’s an extra trap. A SunPower battery installation shifts your consumption and your utility bill, so comparing energy bills month to month tells you almost nothing about whether the array is underperforming. Use the PV production meter, not the bill. We’ve seen Sydney clients panic about “failing panels” when the real story was just a dusty array plus battery time-shifting.

I also talk people out of cleaning when the decline is sudden. A 20% drop in under 48 hours is usually a hardware or grid issue, not soiling. That deserves a service call, not a wash. And if you’re on a low-value export tariff, clean once, measure the recovery, and let that number decide the next step.

Test the math with one clean

You don’t need to sign a year-long contract to find out if cleaning makes sense. Clean one array or one inverter string, then compare weather-normalized output for the five days before and after. If the recovered kWh times your tariff clears the cost of the clean, you’ve got your answer. Do it again next season and track the result. That logged data is what turns cleaning from an opinion into a budget line item.

So before you file a SunPower warranty claim or order a replacement inverter, ask a simpler question: when was this glass last washed? If nobody can give you a date, your first move shouldn’t be new hardware. It should be a wash, a measurement, and a little patience. Most of the time, the panels weren’t broken. They were just dirty.