10 Reasons Solar Systems Fail Early (and How Proper Engineering Prevents Them)

A well-built solar array should still be producing power in 2050. Modules carry 25-year performance warranties, and long-term field studies put median degradation near 0.5% of output per year. So when a system dies at eighteen months, the sun did not fail and the technology did not fail. Somebody's engineering did.

If you have heard the horror stories, you are not being paranoid. Across Nigeria, Egypt and Sudan we keep getting called to sites where a business paid full price for solar and got two good years, or two good months. The owner is back on diesel and suspicious of the next quotation.

This article explains why solar systems fail early, using the ten causes we see most often on site. For each one you get what goes wrong, the warning signs you can spot before you sign, and what proper engineering does differently. At the end there is a due-diligence checklist to take into any vendor meeting.

Why Solar Systems Fail: It Is Almost Never the Panels

Here is the uncomfortable truth about most solar system problems. The hardware in a failed installation is usually still capable of working. What killed it was a decision made before anyone climbed onto the roof.

Solar looks simple from outside. Panels, an inverter, some batteries, a few cables. That apparent simplicity is why the market fills with installers who have no electrical engineering behind them.

A designed system has calculations behind every component: load profile, voltage window, conductor sizing, thermal margin, protection coordination. An assembled system has a parts list. Both look identical on day one.

Not sure which camp your quotes fall into? The ten failures below work as a test. Our engineering and renewable energy services exist because too much of the equipment in this region gets sold rather than designed.

1. The Load Audit Nobody Did

What goes wrong. The installer counts your air conditioners, does arithmetic in his head, and quotes. Nobody measures consumption over a full working cycle, separates continuous load from surge load, or asks what runs at 2am. The system is sized to a guess, and the guess runs low because a smaller number wins the job.

In early 2024, a bakery owner in Ikeja we will call Adebayo paid just over 9 million naira for a 15 kW system sold as "enough for the whole bakery." The vendor never metered anything. Four months in, the deck ovens and dough mixer were tripping the inverter every morning at 6am, because the starting current of the mixer motor and chiller compressor together was roughly triple what the inverter could deliver. The generator came back on, and his promised three-year payback is now closer to never.

Warning signs. No clamp meter or data logger appears on site. Nobody asks for twelve months of bills or diesel records. The proposal quotes kilowatts but never kilowatt-hours per day.

What proper engineering does differently. A real load audit logs your supply for at least a week, captures the daily kWh profile hour by hour, and identifies every motor with high inrush. Sizing then follows measured demand plus a growth allowance. Our guide on how to size a solar system for a business sets out the method.

2. Counterfeit Panels and No-Name Batteries

What goes wrong. The module says 550W and delivers 380W on a clear day. The battery is rated 5 kWh and holds 3.1 kWh from the first cycle. Counterfeit and relabelled equipment moves through informal import channels in volume, priced to be irresistible.

Fake modules fail in a recognisable pattern. Output is plausible at first, then hotspots develop where cell quality is inconsistent, backsheets crack under UV, and junction boxes delaminate. These are the same solar panel failure causes documented in international field surveys, arriving years sooner. Batteries are worse: B-grade cells with no real management system drift out of balance within months, and one weak cell can take the whole pack into thermal runaway.

Warning signs. The brand appears on no tier-one manufacturer list. The datasheet is a low-resolution image rather than a PDF from the maker's own domain. No verifiable serials, no IEC 61215 or 61730 certification for modules, no IEC 62619 for batteries.

What proper engineering does differently. Components are specified before they are sourced, with traceability written into the contract: named manufacturers, verifiable serials, factory certificates matched to the delivered batch. We publish the equipment we stand behind in our equipment shop rather than substituting whatever landed at the port that month.

The ten engineering and procurement decisions behind most early solar system failures; none is a technology problem. Source: MIMAH field engineering analysis across 150+ projects in 12+ countries.

3. Inverter and Battery Voltage That Do Not Match

What goes wrong. An inverter has a battery voltage window, a charging profile and a communication protocol. So does the battery. When these are mismatched the system still switches on, which is why the fault survives commissioning and surfaces later.

The common version is a 48V lithium bank paired with an inverter whose charge algorithm was written for lead-acid. It absorbs at the wrong voltage and the battery management system starts cutting off to protect itself. The other version is string voltage exceeding the inverter's DC input limit on a cold morning, because open-circuit voltage rises as temperature falls.

Warning signs. Battery and inverter come from unrelated brands with no published compatibility list. Nobody mentions closed-loop CAN or RS485 communication. String voltage sits near the inverter ceiling with no cold-temperature correction.

What proper engineering does differently. Voltage windows are checked at both temperature extremes for your actual site, and pairings come from tested compatibility matrices with communication configured so battery and inverter agree on charge limits. If you are still choosing hardware, our breakdown of how to choose a solar inverter in Africa covers the questions that matter.

4. Cable Sizing That Causes Voltage Drop and Fire Risk

What goes wrong. Cable is the easiest place to shave cost, because nobody inspects the cross-section of a conductor inside a conduit. Drop from 25mm² to 16mm² on a battery run and you save real money on copper. You also add resistance, and resistance becomes heat and lost voltage.

Excessive drop makes the inverter read a lower voltage than the bank holds, so it cuts off early and cycles the pack harder than it should. Owners blame the batteries. The batteries are fine. Fire risk sits on the same side: direct current does not cross zero the way AC does, so an arc at an undersized termination sustains itself.

Warning signs. Cable quoted as a lump sum with no cross-section in mm². No voltage drop calculation in the document. Mixed-brand DC connectors on site, or household cable used outdoors instead of UV-rated solar cable. No torque tool during termination.

What proper engineering does differently. Conductors are sized to hold voltage drop within roughly 1% on the DC battery circuit and 2% to 3% on the array, derated for ambient temperature and conduit grouping. DC connectors are matched pairs from one manufacturer, terminations are torqued to specification and recorded, and fuses and isolators are correctly rated throughout.

5. No Surge or Lightning Protection

What goes wrong. A rooftop array is a large conductive structure in an exposed position, wired into every sensitive device you own. In regions with severe storms, and much of West Africa qualifies, an indirect strike hundreds of metres away can induce a surge on the DC string big enough to destroy an inverter instantly.

The strike does not need to hit you. Induced surges travel down the array wiring into the MPPT stage, then into anything on the AC output. Grid-side transients do the same damage from the other direction where utility supply is unstable.

Warning signs. No Type 2 DC surge protection in the combiner box. No AC protection at the distribution board. No earth resistance test result in the handover pack. Array frames bonded with thin wire rather than a rated earthing conductor.

What proper engineering does differently. Surge protection is coordinated across the system: devices at the array and inverter input, AC protection at the point of connection, all referenced to a tested earthing system with a measured resistance value. On exposed sites a lightning risk assessment comes first, because an air termination system costs a fraction of one destroyed inverter.

Want to see what engineered installations look like in the field? Our project portfolio documents commercial and industrial systems across 12 countries, with the protection and commissioning detail most vendors omit.

6. Mounting That Fails in Storms or Cooks the Array

What goes wrong. Mounting looks like the least technical part of the job and is quietly one of the most consequential. It has to survive wind uplift for 25 years, keep water out of your roof, avoid galvanic corrosion between dissimilar metals, and let air move behind the modules.

Wind failures happen because nobody calculated uplift. Panels at roof edges and corners see far higher suction than those mid-array, and one gust finds every under-specified fixing at once. The quieter failure is thermal: modules flush to a roof deck run far hotter than modules with airflow behind them, and output falls roughly 0.3% to 0.5% per degree above 25°C.

Expected service life by component in a well-built system; failures well inside these windows point to design or installation, not technology. Source: MIMAH analysis based on Photovoltaic Degradation Rates: An Analytical Review (NREL, 2012) and manufacturer warranty terms.

Warning signs. No wind load calculation for your location and roof height. Penetrations sealed with silicone instead of flashing. Aluminium rails bolted straight to steel. Modules sitting flat with under 100mm clearance, or clamps outside the specified zones.

What proper engineering does differently. Structural design starts with a wind load calculation to the applicable standard, plus a check that the roof can carry the added load. Fixings and clamp positions follow the module maker's loading instructions, penetrations are flashed, dissimilar metals isolated, and the array keeps a ventilation gap.

7. Batteries Installed in a Hot, Unventilated Room

What goes wrong. Batteries are the shortest-lived and most expensive part of a hybrid system, and temperature governs their lifespan. Lithium iron phosphate cells want roughly 15°C to 30°C, and sustained heat above that cuts calendar life sharply. Yet batteries routinely end up in a locked store room with no ventilation, beside the generator, or on an unshaded roof.

Mona runs a cold store outside Cairo in 6th of October City. Her installer put 24 lithium modules in a converted stock room with one small window and no extraction. She held an eight-year warranty and expected a decade of service.

At month fourteen the pack was down to 61% of rated capacity. Because ambient temperature had exceeded the manufacturer's stated operating range for most of its life, the claim was rejected outright. The batteries had cost more than the rest of the system combined.

Warning signs. No ventilation, extraction or cooling specified for the battery room. No temperature sensor logging the space. Batteries against a west-facing wall or under uninsulated metal roofing. No clearance around the enclosure.

What proper engineering does differently. Battery location is a design decision with a thermal calculation behind it: heat rejection from the pack, room volume, ventilation rate, and where the sun sits. Rooms get forced ventilation or cooling, temperature monitoring is wired in, and clearances follow the installation manual.

8. No Maintenance Plan

What goes wrong. Solar is often sold as maintenance-free, which is a marketing claim rather than an engineering one. Two things degrade output invisibly: soiling and loose connections.

Soiling is the bigger number here. In dusty, low-rainfall conditions, losses of 15% to 30% between cleanings are routine in the dry season. Nothing alarms; the system quietly makes less. Loose connections are the dangerous half: thermal cycling loosens terminations, a loose terminal creates resistance, resistance creates heat, and heat loosens it further.

Warning signs. No operations and maintenance proposal alongside the installation quote. No monitoring platform, or an app showing current power with no historical yield. No mention of cleaning frequency, thermal scans or annual torque checks.

What proper engineering does differently. The system ships with a maintenance schedule matched to your environment: cleaning intervals set by local soiling rate, annual thermographic inspection of DC terminations, and string-level current testing. Our guide on solar panel maintenance in dusty climates sets realistic intervals for the region.

9. The Installer Disappears

What goes wrong. Component warranties are worthless without somebody to administer them. A 25-year warranty is a claims process needing a registered installer to file paperwork and hold the commissioning records. This is the most common ending to a bad solar installation: a crew forms around one big job, wins on price, installs at volume, then dissolves when the callbacks start.

Osman ran a poultry operation outside Omdurman and lost his inverter to a storm in year two. The MPPT stage was dead and the manufacturer's warranty was valid, but the claim needed the commissioning report, the serial registration and an authorised installer to process it. His installer's phone was disconnected and the business had never been registered.

He paid full price to replace an inverter that was still under warranty, then paid again for a survey to learn why the first one died. Nobody had fitted surge protection.

Warning signs. No registered company details or office you can visit. No completed projects you are allowed to call. Under three years of trading. No liability insurance, and no named engineer on the design.

Lithium battery capacity after 14 months in an unventilated store room at a cold store in 6th of October City, Egypt; the eight-year warranty claim was rejected for operation outside the stated temperature range. Source: MIMAH field case records.

What proper engineering does differently. An established firm keeps commissioning documentation and serial registrations on file, and runs a service organisation independent of any one project. Ask how many systems the company installed five years ago it still maintains. The answer separates contractors from vendors.

10. Designing for Today's Load and Ignoring Growth

What goes wrong. A system sized exactly to today's consumption is obsolete the moment the business grows. You add a production line, a second cold room, more cooling. Because nobody planned for expansion, adding capacity means replacing equipment rather than extending it.

The constraint is usually the inverter and the DC architecture. A string inverter with both MPPT channels populated cannot take another twenty modules, and a bank built from a discontinued battery cannot be extended, because mixing old and new cells degrades the pack to its weakest module.

Warning signs. Zero headroom on inverter capacity or MPPT inputs. No conversation about your three to five year plan. Roof or ground space fully consumed with nothing reserved. Battery model with no stated expansion path.

What proper engineering does differently. Designs are modular from the start. Inverter capacity and MPPT channels carry documented headroom, conduits are sized for future circuits, switchgear has spare ways, and space is reserved for a second array block, with the expansion path written into the handover documents.

The Due Diligence Checklist for Choosing a Solar Installer

Take this into every vendor meeting. If a company cannot answer these comfortably, the price advantage is not real. Ask for:

The metered load study: a logged profile with daily kWh and peak starting current, not an estimate.

The voltage drop calculation, with cable cross-sections in mm² for DC and AC runs.

The wind load calculation for your site and roof height.

Which surge protection devices are fitted on the DC and AC sides, and what earth resistance was measured.

Serial numbers and factory certificates (IEC 61215 and 61730 for modules, IEC 62619 for batteries).

Written confirmation that the inverter and battery are a tested, communicating pair.

The battery room ventilation design, and what temperature it holds in August.

The maintenance schedule and the annual contract price.

Company registration, insurance, and the qualifications of the engineer signing the design.

Three reference sites installed three or more years ago that the company still maintains. Then call them.

The written remedy if the system underperforms against the yield estimate, and how it expands when your load grows by 40%.

Frequently Asked Questions

How long should a solar system actually last? Modules should deliver 80% or more of rated output at year 25. Inverters run 10 to 15 years, and lithium iron phosphate batteries give 8 to 12 years if kept within their temperature range. Anything failing well inside those windows points to design or installation, not technology.

Can a badly installed system be rescued, or do I start again? Most can be rescued, and the panels are usually the salvageable part. Typical fixes are re-cabling the DC side, adding surge protection and earthing, ventilating the battery room, and replacing an incompatible inverter. Normally far cheaper than starting over.

Why is one quote 40% cheaper than the others? Because something in the bill of materials is not what it appears to be: relabelled modules, B-grade cells, undersized cable, no surge protection. The saving is real on day one and gone the first time the system works under stress.

Do I really need a maintenance contract? In a dusty climate, yes. Soiling alone can cost 15% to 30% of production between cleanings, and annual thermal imaging catches the loose DC terminations that start fires.

Which of these solar installation mistakes causes the most early failures? Skipping the load audit, because it sits upstream of everything else. Get the load profile wrong and the inverter, battery bank, array size and cable sizing are all wrong by the same margin. No amount of careful installation can correct it later.

Why Solar Systems Fail Is a Design Question, Not a Hardware One

Look back at the ten causes and a pattern emerges. Not one is a technology problem. Undersized systems, counterfeit components, mismatched voltages, thin cable, missing surge protection, poor mounting, hot battery rooms, absent maintenance, vanished installers and ignored load growth are all decisions, made before the equipment arrived.

That is genuinely good news. Early failure is preventable, and you can assess a vendor by asking for the calculations instead of the price. An installer who has done the engineering will be glad to show you. One who has not will change the subject to discount.

Across 150+ projects in 12+ countries, the systems still performing are the ones where somebody measured the load, specified real components, calculated the conductors and protected the DC side first. That work adds a modest percentage to project cost. It is the difference between an asset and a write-off.

Have a system that is underperforming, or a quotation you want checked before you sign? Talk to our engineering team. We will review the design against the ten failure modes above and tell you plainly what we find, including when your existing system is worth saving.