When a permanent lighting system stops working the way it should

One node dark, a section out, or the whole run off each point at a different fault. What actually fails on a permanent lighting system, and how it gets fixed.

The pattern of what is dark tells you most of what you need to know. One node out is almost always that node. A whole section out is almost always the feed or the first connection into that section, not fifty lights that failed on the same night. Everything dark at once is usually not the lights at all, it is power or the controller. Wrong colors on part of a run is a data problem rather than a diode problem. And a system that is lit but unreachable from your phone is a network fault with no lighting fault behind it. Those five patterns cover most of what goes wrong, and identifying which one you have is the difference between a short repair and a long afternoon of guessing.

Worth saying plainly before any of that. A permanent system is outdoor electronics fastened to a building, powered most nights, and exposed to everything a Nebraska year does. The honest question was never whether it will ever need attention. It is what needs attention, how often, and how fast it gets sorted. A well built system tends to go long stretches with nothing happening and then have a single event.

Read the shape of the failure first

Before you describe anything to anyone, work out which of these you are looking at. Do it after dark, with the system set to plain white at full brightness, because color and dimming hide things.

• One node dark, everything else fine. Usually that node, or the joint immediately at it. Nothing else in the system is implicated.
• The run is lit up to a point and dark from there on. The fault is at the boundary, not spread across the dark part. The first dark position is the address of the problem.
• Everything dark. Power or controller. Look at the ground before you look at the roof.
• A stretch showing the wrong color, or stuck on one color. Data reaching the nodes is corrupted or a node has failed in a way that still passes power.
• Flicker. Either a loose or corroded connection, or a supply being asked for more than it can deliver.
• Lights on and behaving, app cannot see the system. Networking. The lights are not broken.

Photograph the run at night with everything on white, and note where the light stops. That single photo narrows a diagnosis faster than any description over the phone.

A run is a chain, and that changes where you look

Individually addressable systems are why the section rule works. Each node carries a small driver chip with an address. Data enters the run at one end and is passed along, node to node, with each one taking its own instruction and forwarding the rest. Power is usually carried down the same run alongside it.

Two consequences follow, and they explain most of what homeowners find confusing.

The first is that a break anywhere on the data path takes out everything downstream of it while everything upstream stays perfectly lit. Twenty dark nodes do not mean twenty failures. It means one failure with nineteen innocent lights behind it. That is why the boundary between the lit part and the dark part is the useful information, and why a technician who knows the architecture walks to that boundary rather than working along the whole dark stretch.

The second is that a node can fail in ways that do not break the chain. It can go dark but keep passing data, which is the single-node case. It can sit stuck on a color while the rest of the run follows the app. Both are local faults, and neither says anything about the health of the system around them.

There is a related pattern worth knowing. If the far end of a long run has gradually gone dimmer and slightly off-color compared with the start, that is usually voltage arriving low at the end rather than lights that aged differently. Added resistance at a corroding joint does the same thing. Both are fixable and neither is solved by replacing modules.

Connections and water do the most real-world damage

Modules and channel carry ingress ratings, and those ratings describe the part as it left the factory. The places water actually gets in are the joints made on site: splices, the transition from the run into a soffit or wall, the entry into the controller enclosure, and anywhere a cable was cut and rejoined on a ladder.

Water rarely arrives as a flood. It arrives as humid air that condenses inside a connector body when the temperature swings, then sits on the contact. Corrosion adds resistance long before it creates an open circuit, which is why this failure usually announces itself as intermittent rather than dead. Fine on a dry October night, unreliable after a thaw, gone by March. Intermittent is the harder fault to chase, and it is the reason a system that "sometimes does it" should be looked at rather than waited out.

Two details separate installations that stay dry from ones that do not. Whether the cable is dressed so water runs away from a connector instead of down into it, and whether each seal was properly engaged rather than pushed together and left. Neither is visible from the ground, and neither has anything to do with which product was bought. This is workmanship. It is why the same hardware performs very differently on two houses on the same street, and it is the strongest reason to judge an installer on how they detail the work rather than on the catalogue they carry. The installation process is where a decade of reliability is decided.

What years outdoors do to the parts holding everything up

Over a long enough period the parts that need attention are more often the ones holding the lights than the lights themselves.

Ultraviolet exposure embrittles polymer. Clips, cable jackets and any plastic trim on a south or west elevation age faster than the same parts on a shaded north face, and a brittle clip is one that lets go in a wind event. Thermal cycling works on the mechanical side: aluminium channel and the board behind it expand and contract at different rates every day of the year, which slowly works fasteners and is why a long unbroken run needs somewhere to move. Freeze and thaw at the drip edge loads anything that protrudes, and it is where ice does its damage. How permanent lighting copes with Nebraska weather goes further into what to look for before buying.

For service purposes the useful summary is that fasteners, clips and cable are ordinary wear items on a ten-year horizon, and none of them are a sign the system was a mistake.

Everything dark usually means power, not lights

Work through it in order of likelihood, and the order runs from the ground up. Something upstream of the controller comes first: a tripped breaker, a tripped outdoor GFCI, or a receptacle on a wall switch that somebody turned off. Then the power supply. Then the controller.

The supply and the controller are the components most likely to need replacing first in the life of a system, because they are the parts doing the electrical work every night while the modules mostly sit there being switched. When either fails it takes the whole run with it, which reads to a homeowner as catastrophic and is in practice the repair involving the least work at height, because both parts live where a person can stand.

When the lights are fine and the app is the problem

This is the most common support call for any connected product, and it is usually not a fault at all.

Anything that changes the home network breaks the link: a new router, a new internet provider, a changed wi-fi password, a renamed network, a mesh system reconfigured. The controller is still holding credentials that no longer work. It keeps running its stored schedule perfectly well while the app reports it offline.

That combination is the diagnostic. Lights still coming on and going off at the right times, app cannot reach them, means networking and nothing else. If the schedule has also stopped behaving, something more than the network has changed.

Two other things get reported as faults and are not. A scene left active by someone else in the house explains a surprising share of wrong-color calls, as does a schedule that is doing exactly what it was set to do a year ago and has been forgotten since. Check both before calling anything broken. What the app is actually doing covers how schedules and scenes are stored and why that matters.

What you can safely check yourself, and where to stop

All of this is from the ground, and it is worth doing before you call anyone.

1. The breaker. Reset it once. If it trips again, stop there and get it looked at. A breaker that keeps tripping is reporting a real fault, not being awkward.
2. The receptacle feeding the power supply. If it is GFCI protected, press test and then reset. Outdoor GFCIs trip on moisture, and this is the most common reason a whole system goes dark after a storm. Check at the same time that it is not on a wall switch somebody flipped.
3. The app. Is the controller showing online. Is a schedule running. Set everything to plain white at full brightness by hand. If the system responds, power and the data chain are both alive and you are looking at something narrower.
4. The run itself, after dark, on white. Note where light stops and photograph it.

Where to stop is not a matter of confidence, and it has two parts.

Nothing involving the roofline. No ladder to reach a node, no pulling a module out of the channel, no lifting a clip to look underneath. The edge of a roof is where people get badly hurt, and it is the same edge the system is fastened to, so a slip damages the thing you were trying to fix.

Nothing involving an enclosure. Do not open the controller or the power supply housing, and do not unplug connectors to test them. A weatherproof enclosure seals against a gasket that was compressed once, at installation, in the right conditions. One opened and closed by hand rarely reseals properly, and that is how water gets into the one box on the house that has to stay dry. The same applies to connectors on the run: pulling one apart to look at it is often what starts the problem you were investigating.

How a service visit actually works

Diagnosis on a chain runs in one direction, from the supply outward, because everything downstream depends on everything upstream.

The sequence is: confirm the source is live and delivering what it should, confirm the controller is powered and putting out data, then walk the run to the boundary between working and not working and test there. Testing at the boundary rather than swapping parts along the length is the whole method. On a run that is dead from its first node, the question is whether power and data are arriving at that node at all, which points back at the feed. On a run that dies partway along, the answer is at one specific joint and the rest of the run is a bystander.

Component-level replacement is what makes any of this worth doing. A node or a short segment comes out, the joint is remade, and the channel stays on the house. That is the difference between a permanent system and a strand of lights, and it is worth confirming your system is built so a single position can be changed without disturbing the run.

One thing to expect at the end of a visit: the run should be driven through plain white at full output, each color on its own, and a low brightness level, before anyone calls it fixed. A fault that only appears on one color channel or only at the bottom of the dimming range is exactly the one that gets signed off and comes back three weeks later.

Have it serviced by whoever built it, where you can

A run's layout is not readable from outside. Where the feed enters, how many nodes are on each run, which stretch of roofline corresponds to which range of addresses, where a segment was joined behind a downspout, which corners were made with a factory part and which were made on site. None of that is visible from a driveway, and a company that did not install the system starts by rediscovering it, usually from a ladder, before it can fix anything.

There is a second reason to ask the question rather than assume. Find out whether having someone else work on the system affects your coverage, because that answer varies and it is easier to establish before you need it.

What to keep on file from the install

Ask for these at handover and put them somewhere you will still find them in five years. A service call on a well documented system is a different job from one on a system nobody wrote down.

• A drawing or marked-up photo of the run layout with feed points identified, and the node count for each run.
• The make and model of the controller and the power supply, and the supply's rating.
• Where those two are physically mounted, and which breaker and which receptacle feed them.
• The exact module type and color specification used. This matters more than people expect, because a replacement from a different production batch can read visibly different sitting next to the original run.
• The app account, registered in your name, and how to add or remove a user.
• The warranty terms in writing.

Ask what the coverage says before you need it

Get the terms on paper at the time of the project and read them once. The distinctions that separate one warranty from another in this category are specific, and they are worth asking about by name:

• Are the modules, the controller and the power supply covered, and for the same length of time? Electronics are frequently treated differently from the light-emitting parts.
• Are parts covered, labour, or both? On a roofline system, getting to the fault is most of the job, so this is the question that decides what a repair actually involves for you.
• Is workmanship covered separately from the product, and for how long? A leaking field joint is an installation issue rather than a manufacturing defect, and the two are often handled under different terms.
• Does coverage survive a sale of the house, and does anything have to be done to transfer it?
• What voids it? Work by other trades, third-party repair, and pressure washing near the run are the exclusions most commonly named.
• Who do you call, and what happens if the manufacturer discontinues the controller?

Every one of those has a real answer. Ask for it in writing rather than in a phone call.

The one maintenance event you can plan for

A roof replacement is the predictable interruption in the life of a permanent system. Lighting mounted on the drip edge or the fascia has to come off before the roofing crew starts and go back on after they finish, which means two visits scheduled around somebody else's work.

Decide early who owns that coordination and who does the removal and the refit, and ask the question when the roof is being quoted rather than on the morning the crew arrives. The alternative is a delay or a roofer improvising with your lighting, and neither ends well.

Gutter cleaning, ice removal and anyone hanging seasonal decorations are the other occasions when someone is working near the run. Tell them the channel is there. Most physical damage to permanent lighting comes from another trade who did not know it was live.

If yours is misbehaving right now

Work out which pattern you have, from the list at the top, and photograph the run after dark with everything set to plain white. Then check the breaker and the outdoor GFCI. Those two steps resolve a fair share of calls on their own and they make the rest of the diagnosis much faster.

TruLight Omaha installs permanent LED roofline lighting, accent lighting and the other outdoor lighting systems described on this site. If yours is not behaving, describe the pattern and say where the light stops when you get in touch.