Solar system maintenance: what is really necessary

How much maintenance does a solar system really need?

The honest answer lies between two extremes, both of which are in circulation. One extreme sells the system as «completely maintenance-free» – install, forget, collect. The other sells maintenance contracts with an annual roof appointment and cleaning subscription, as if an oil change were due. Neither does justice to the technology.

In fact, a photovoltaic system has no moving parts: nothing turns, nothing burns, nothing needs to be lubricated. Modules lie still on the roof and age slower than most warranties promise. Precisely this inconspicuousness is also the only real weak point of the system – a system that makes no sound can produce silently or stand still, and from the outside both look the same.

Therefore, a three-level model is sensible, which has proven itself in practice:

1. Observe – you do this. Fault messages active, from time to time a look into the monitoring app, once a year the comparison with previous year's values and a visual inspection from the ground. Effort: half an hour per year, without tools. 2. Inspect – the professional does this, on occasion. After hail or severe storm, in the event of unexplained reduced yield, before important warranty periods expire and within the scope of the periodic electrical inspection of the property. 3. Maintain – this is plannable. In a system life of 25 to 30 years, an inverter replacement is typically due, occasionally a small part. Those who factor this in from the start will not be surprised.

Anyone who keeps these three levels apart has the topic under control – and incidentally reliably recognises which service offers are worth their money and which primarily serve the provider.

Monitoring: the most important part of maintenance happens on the screen

It sounds unspectacular, but it is the core of this entire guide: most problems with solar systems are not discovered on the roof, but in the numbers. A failed string, an inverter in fault mode, a gradually growing shading – all this is visible in the yield curve long before you can see it from the street.

Three tools are sufficient for this:

The fault messages. Every modern inverter can report failures and errors via app or e-mail – provided the function is set up and the messages do not end up in spam. This is part of commissioning; check it at handover. A fault message that no one reads is not one.

The annual comparison. Note the annual production once a year – a simple table is enough, paper is sufficient. The yield fluctuates noticeably from year to year due to the weather; it is not the single weak year that is suspicious, but the trend over several years or a drop without a weather explanation. As a rough guide: deviations of more than around 10 per cent that cannot be explained by the weather are worth querying. The natural ageing of the modules does not explain such jumps – it is significantly below one per cent per year.

The comparison with the bill. Once a year, it is worth looking to see whether the feed-in on the grid operator's bill roughly matches the values in the app. Why the two figures never exactly match – the app measures at a different place than the utility's meter – is explained in the article Measurement and meters. Rough plausibility is sufficient; Rappen accuracy is not the goal.

How these steps together result in a fixed appointment, we have written down separately as an annual check with checklist – that is the implementation version of this chapter for ticking off.

This includes an often underestimated fourth component: the system dossier. Quote, acceptance protocol, warranty documents, your annual values and notes on special events, all in one folder. In warranty cases, complete documentation shortens any discussion, and at the latest when selling the house it is worth ready cash.

Which components age – and what breaks down when?

Not all parts of a system age at the same rate. A realistic picture looks like this:

Guideline values, As of: July 2026. The manufacturers' warranty conditions and the condition of the specific system are decisive.

Two points from this table deserve a second look.

The inverter is the wear part of the system – as the only permanently active component, it works every day and therefore lasts shorter than the modules. A replacement after ten to fifteen years is the normal case, not a planning error, and belongs in the profitability calculation right from the start. What it achieves and what types there are is explained in the inverter guide; the right time and the costs of replacement are dealt with in a separate article.

The modules are more durable than their reputation. The gradual loss of performance is real, but small – and it does not proceed linearly towards the bad: the degradation is greatest in the first year of operation and flattens out thereafter. More important than the ageing is the distinction between degradation and defect: a sudden drop in performance is never «normal ageing», but a case for the monitoring chapter above.

What does the operation of a solar system cost per year?

Little – but not nothing, and anyone who knows the items budgets more honestly. For a typical single-family home, the following apply:

Overview without amounts, As of: July 2026 – the amount of the items depends on the grid operator, insurance and system. The respective tariffs and policies are binding.

Noticeable is what is missing in this table: the regular service appointment. It simply does not exist as a mandatory item for a single-family home system. The operating costs of a solar system essentially consist of fixed fees and a provision – not of recurring work. Precisely for this reason, the operating costs section belongs in every serious cost calculation before the purchase, but not as a spectre.

A word on insurance, because the question comes up regularly: report the new system to your building insurance. How photovoltaics is covered – via building insurance, an extension or separately – differs depending on the canton and policy; the report itself costs nothing and prevents the unpleasant surprise in the event of a claim.

Cleaning, snow, martens, trees: what is there to the classics?

Four operating topics appear again and again in consultations – mostly with more concern than the matter deserves. The short version, with in-depth information in the respective article:

Clean? Usually not necessary. On pitched roofs, the rain does most of it by itself; yield losses due to normal dirt are small. Professional cleaning makes sense for stubborn deposits – agricultural dust, bird droppings, the typical dirt rim on flat-inclined modules – and is visible in the yield. A blanket cleaning subscription sells one thing above all: itself.

Snow? Leave it be. The winter months contribute little to the annual yield anyway, modules often clear themselves of snow due to their smooth surface, and clearance by laypeople causes more damage than it brings – not to mention the risk of falling. Details including the exception of roof avalanches: Snow on the solar system.

Martens? A wiring issue. Marten damage is rare, but real – and almost always a question of cable routing during installation. Anyone who notices tracks around the house should have the cables specifically checked instead of hoping for the best: Martens and the solar system.

Trees? The slow-motion problem. The shading, which was not yet an issue during planning, grows a few centimetres a year – too slowly for the eye, but fast enough for the yield curve. The annual comparison from the monitoring chapter is exactly the instrument for this; what to do then can be found under subsequent shading.

When does a professional belong on the roof – and is a maintenance contract needed?

First the rule that beats all others: The roof belongs to the professional, the ground to you. Cleaning modules, clearing snow, straightening cables, «briefly tightening» fastenings – the combination of falling risk, sensitive modules and warranty conditions makes any DIY work on the roof a bad deal. There is no exception that softens this rule.

There are clearly defined occasions for a professional check:

• after extreme weather – hailstorm, severe storm –, even if everything looks intact from below;
• in the event of unexplained reduced yield, when fault messages, string comparison and weather are ruled out as explanations;
• before warranty periods expire – a documented finding shortly before the deadline secures claims that expire afterwards;
• routinely as part of the periodic electrical inspection of the property to which the system belongs.

And the maintenance contract? For a single-family home it is usually unnecessary – the three levels from the beginning cover the need, and incident-related checks can be commissioned individually. The calculation looks different when more is attached: for self-consumption community (ZEV) and apartment building systems with billing obligations, for business systems where a failure costs money, or for difficult-to-access systems, a contract with defined response times can be worth its price. What is decisive is what is in it: guaranteed response time, assumption of monitoring and a clear scope of services – not an annual roof appointment as an end in itself.

Optimising in operation: how do you get more out of the running system?

Maintenance secures the yield – optimisation increases the value. Because very few systems are «finished» after acceptance: What a kilowatt hour of solar power is worth depends on whether you consume it yourself or feed it into the grid, and this lever remains movable over the entire life of the system.

First address: self-consumption. Self-consumed solar power replaces expensive grid supply, fed-in power only brings the feed-in tariff – the difference is the reason why behaviour and control are more worthwhile than any fine-tuning of the technology. The spectrum ranges from free (washing machine in the midday sun) to an investment (battery); the six levers including an honest assessment are in their own article. Whether in the end self-consumption or feeding in is economically ahead depends on your tariff area.

Second address: the tariff side. Feed-in tariffs change annually, guarantees of origin can sometimes be remunerated separately, and anyone who checks their billing for five minutes once a year recognises model changes and errors before they become expensive.

Third address: the expansion. Battery, wallbox, heat pump – what was called «later» during installation becomes concrete in operation. Whether the system is prepared for this (hybrid inverter, reserve space, empty conduits) determines the effort; the question therefore belongs in the consultation, but only pays off now.

In this order – first behaviour, then tariffs, then investments – optimisation costs the least and brings the most. The reverse order is the more expensive one.

Sources: EnergieSchweiz and Swissolar (operation and maintenance recommendations); empirical values from the installation and service practice of ecoEn GmbH, Zurich region.

Last updated: 9 July 2026 · Author: ecoEn editorial team