LED Light Bar Overheating: Normal Thermal Derating or a Real Fault?

An LED light bar that dims or shuts itself off after 10 to 30 minutes of running is usually protecting its own LEDs, not failing. Automotive LED drivers are built to reduce current as temperature rises and to switch the output off when the temperature keeps climbing, so the behaviour that looks like a dying lamp is also the behaviour of a healthy lamp in an installation that cannot shed heat. An LED light bar overheating is therefore a symptom with two opposite meanings, and the difference is settled by measurement rather than by opinion.

Quick answer

Most owners who notice output falling after a few minutes are looking at thermal foldback, a deliberate control loop rather than a defect. The short answer splits into three cases.

If the beam fades gradually, returns to full brightness after the lamp cools, and the housing shows no smell, stain or deformation, the lamp is probably working as designed and the installation is the limit. If the output cuts out abruptly, comes back after a pause, then cuts out again, the driver is reaching a shutdown threshold and the thermal margin is small. If the output never returns, returns weak, or the housing is discoloured or deformed, the lamp has moved from protection into damage.

What you observe	Most likely state	First action
Fades over minutes, recovers when cool	Thermal foldback, installation-limited	Record time, ambient and surface temperature
Cuts out, restarts, repeats	Shutdown threshold reached	Stop long runs, inspect airflow and mounting
Dims and stays dim	Emitter or driver degradation	Compare against a documented baseline
Smell, smoke, blistering or deformation	Permanent damage risk	Stop energising and disconnect the circuit

Everything below is the evidence behind those three sentences: what to measure, where to put the sensor, which readings decide the case, and the point at which testing must stop.

What LED light bar overheating actually does to the output

A light bar does not simply get hot and fade. It gets hot, a control loop notices, and the loop reduces current in steps or in a ramp until the temperature stabilises, which is why the beam settles at a lower level instead of falling to nothing. That behaviour is a design feature with a name: thermal foldback, or thermal derating.

One documented implementation makes the sequence concrete. In the Texas Instruments application note for the LP8867-Q1 automotive LED driver, an external temperature-sensing resistor drives the foldback input: when the sensor reaches a set temperature T1, the driver begins to reduce LED current, and when the current has fallen to 17.5% of its nominal value the output turns off until the temperature returns to the operating range (Texas Instruments SNVAA22). The same driver family lists current derating, automatic LED current reduction with an external temperature sensor and thermal shutdown as product features, alongside LED open and short detection, in an AEC-Q100 Grade 1 part rated for a −40 °C to +125 °C ambient range (TI LP8867-Q1).

Two consequences follow, and both matter for diagnosis. First, the lamp has more than one temperature threshold, so “it dims” and “it switches off” are different events at different points on the same curve. Second, the recovery is part of the design: a lamp that comes back to full output after cooling has not healed itself by accident, it has completed a cycle the driver was built to run.

What the driver is protecting is described in the LED maker’s own application literature. Bridgelux states that as temperature rises, light output falls, forward voltage falls and the colour point shifts toward blue, and that these shifts are temporary and recoverable, while the absolute maximum ratings for case and junction temperature must not be exceeded because exceeding them can irreversibly damage the product and cause permanent performance shifts (Bridgelux AN10). Foldback exists so that the lamp lives inside the recoverable part of that description.

Why the junction decides the output and the housing only reports it

The temperature that sets light output sits inside the LED package, on the semiconductor junction, and it cannot be measured from outside the housing. What you can measure is the outside: the extrusion, the end caps, the bracket, the air around them. Those surfaces are the last links in a chain that starts at the junction, and the chain is why a housing that feels mildly warm and a junction that is far too hot can coexist.

The arithmetic that connects them is a single product: temperature rise equals dissipated power multiplied by thermal resistance, junction to ambient. Take a bar that dissipates 40 W through a path of 1.5 °C/W, an illustrative pair of numbers rather than a specification for any product: the junction sits 60 °C above ambient, so at 30 °C ambient it is near 90 °C while the outside of the extrusion, which is only one element in that chain, may read far lower. Double the thermal resistance by blocking airflow and the same 40 W produces twice the rise. This is the calculation behind the whole article, and the reason a single surface reading is weak evidence on its own.

The physics of that chain, including how a thermal resistance is calculated and how a housing and its fins can and cannot help, is covered in the site’s guide to LED light bar thermal management. This article deliberately does not repeat it: everything from here on is about reading a real lamp’s behaviour and deciding what it means.

The heat sources inside one light bar

Heat in a working bar arrives from four places, and they behave differently, which is why two lamps with the same LED count can fail to hold output in completely different ways.

Heat source	How it is generated	Where it appears first	Effect on output
LED junctions	Current through the emitters; only part of the input leaves as light	Emitter board and the metal directly behind it	Flux falls as junction temperature rises
Driver electronics	Switching and conduction losses in the converter	Driver board, often at one end of the housing	Foldback starts here if the sensor is on the board
Terminals and crimps	Resistance at every joint carrying the full current	Connector body and the first 100 mm of cable	Local hot spot, voltage drop, no foldback
Absorbed environment	Sun load, radiator and engine heat, hot air trapped in a cavity	Whole housing, evenly	Raises the starting point of every other source

The distinction that decides most diagnoses is between the first two rows and the last two. A junction or driver problem scales with time and with how hard the lamp is working, so it appears after minutes and recovers after cooling. A terminal problem scales with current and shows up as a hot connector, sometimes within seconds, and does not necessarily recover on its own. An environment problem scales with where the vehicle is parked and how fast it is moving, which is why some lamps only misbehave on a hot day in traffic and never on the highway.

HELLA’s technical literature puts the design requirement in one sentence: the LEDs’ power loss must be effectively dissipated from the LED chip and transferred to the surroundings, and a modern LED headlamp cannot do without sophisticated thermal management (HELLA Tech World). A light bar in a bumper cavity is the same problem with less room.

Five behaviours that look alike from the driver’s seat

Owners describe five different things with the same words, and the five point at different causes. Separating them is the first real diagnostic step, and it costs nothing but attention.

Behaviour	What is happening electrically	Most likely cause	Typical next check
Slow fade over 5 to 30 minutes	Current is being reduced in steps or a ramp	Foldback working; thermal margin small	Surface temperature versus time
Cuts out, restarts, repeats	Output is switched off at a threshold, then restored after cooling	Shutdown threshold reached; hysteresis cycling	Airflow, obstruction, mounting position
Flicker only when hot	Driver or connection becomes unstable at temperature	Solder joint, connector, or driver protection	Wiggle test cold versus hot; see the flicker guide
Colour shifts toward blue and stays	Junction temperature high but not yet at a limit	Sustained overheating, recoverable	Confirm against a cold baseline
One section dark, or lamp dead	A string or channel has failed	Emitter string, channel, or driver stage	Segment mapping, then the partial-failure guide

The fade and the cut-out are the two that get confused most often, because a fade that ends in a cut-out is one continuous process. Watching a stopwatch rather than the wall is what separates them: a fade has a duration you can time, a cut-out has a moment you can timestamp.

Normal protection or a real fault: eight criteria

The question “is it broken” has an answer that can be assembled from observations an owner can make without opening anything. Eight criteria, applied together, cover almost every case.

Criterion	Points to normal protection	Points to a fault
Recovery after cooling	Full brightness returns and holds	Output stays low, uneven, or absent
Time to onset	Repeatable within a few minutes across runs	Erratic, or progressively shorter each run
Ambient dependence	Earlier onset on hot days, in traffic, or when parked	Onset regardless of ambient or airflow
Symmetry between two lamps	Both lamps behave the same way	One lamp misbehaves and the other does not
Smell	None	Hot plastic, phenolic or burning smell
Housing and lens	Clean, no stain, no deformation	Discoloured, blistered, deformed, or crazed
Connector condition	Cool, clean, no discolouration	Warm or hot, brown or green, melted insulation
Output when cold	Matches a documented baseline	Below baseline from the first minute

The symmetry criterion is the most underused and the most powerful. Two identical bars on the same bumper share a supply, a ground and an ambient, so a difference between them is evidence about the lamps, while identical behaviour in both is evidence about the installation or the vehicle. The corollary is that a single lamp on a vehicle has no twin, which is why the cold baseline in the last row matters: without a reference, “dimmer than before” is a memory, not a measurement.

None of these criteria requires a tool, and none of them requires more than twenty minutes. They exist to stop the two expensive mistakes: replacing a healthy lamp because it behaved correctly, and continuing to run a damaged one because it still lights up.

The 20-minute test that answers most of the question

The most useful test is short, repeatable and does not require a workshop. It produces a time series, and a time series is what turns “it gets dim” into evidence.

Begin with the lamp cold, meaning at least an hour since it last ran and, ideally, in shade rather than direct sun. Record the ambient temperature with the same instrument you will use on the housing, so the two numbers come from one calibration. Switch the lamp on and note the time. Then record four things at 0, 5, 10, 15 and 20 minutes: the housing surface temperature at a marked point, the current drawn by the circuit if you can measure it safely, what the beam looks like against a wall at a fixed distance, and whether the sound of the lamp changes.

The marked point matters more than the instrument. Pick a spot on the main extrusion directly behind the emitters, mark it with a pencil, and use the same spot every time, because surface temperature varies by more than most people expect along a bar. At 20 minutes, switch off and let the lamp cool for 15 minutes with the ignition off, then repeat a single one-minute run and compare the brightness and current with the first minute of the original run.

Three outcomes are possible. A lamp that returns to its initial output and current has a reversible, temperature-driven behaviour, which is the signature of foldback. A lamp that returns weak, or uneven, or draws less current than it did cold, has changed permanently in some way and needs the segment and connector checks. A lamp that never dimmed at all in 20 minutes of stationary running but dims on the road is telling you that airflow, vibration or supply voltage changes with motion, which moves the investigation to mounting and wiring.

The record sheet worth keeping

A page of numbers per season is worth more than any single reading, because the useful comparison is a lamp against its own past. The columns below are the minimum that makes a later diagnosis possible.

Date	Ambient	Minutes to first fade	Surface at 20 min	Current cold / hot	Recovery after 15 min
Entry one
Entry two
Entry three

Two rules make the sheet usable. Always write the ambient temperature next to the surface temperature, because a 70 °C housing at 15 °C ambient and a 70 °C housing at 38 °C ambient are different stories. And always write what you did between rows, such as a wash, a bracket change or a new connector, because otherwise the series will contain a step change you cannot explain later.

The sheet also answers the warranty question that owners ask first. A manufacturer can act on “it dims after 12 minutes at 30 °C ambient and recovers fully” in a way that nobody can act on “it seems dim sometimes”.

Where to put the thermometer, and what it really reads

An infrared thermometer does not measure temperature. It measures radiated energy in a cone and converts it to a temperature using an assumed emissivity, which means the number on the display is only as good as two assumptions: the emissivity you set and the size of the spot you filled.

Emissivity is not a detail. Omega Engineering’s total emissivity table gives unoxidised aluminium values in the region of 0.02 to 0.06, heavily oxidised aluminium around 0.20 to 0.31, matte black paint around 0.96 and black oil paint around 0.92, and the table’s own preface warns that emissivity varies with temperature and surface finish, so the tabulated values are a guide and an absolute measurement requires determining the material’s actual emissivity (Omega Engineering). A polished or anodised extrusion is close to the low end of that range. Point an instrument set to 0.95 at it and the reading is not slightly wrong; it can be wrong by enough to hide the problem you are looking for.

There is a practical way around the problem that needs no emissivity table. Put a small piece of matte black tape on the marked spot, let it reach the surface temperature for a minute, and measure the tape with the instrument set to a high emissivity. You are now measuring a surface whose emissivity you know, and the tape is thin enough to follow the metal. The absolute number will still carry an error, but the comparison between two runs on the same tape is sound, and comparison is what this diagnosis needs.

The second assumption is spot size. Every infrared thermometer has a distance-to-spot ratio, and the spot grows with distance, so measuring a 30 mm wide extrusion from two metres averages the metal with whatever is behind it. Get close enough that the spot is smaller than the part you care about, and hold the instrument at the same distance and angle for every reading in a series. Finally, remember what the number is not: it is the outside of the housing, several thermal interfaces away from the junction, and it responds slowly compared with the junction it is standing in for.

Measurement point	What it tells you	What it cannot tell you
Extrusion behind the emitters	How well the LED board is coupled to the housing	Junction temperature
Fins and end caps	Whether the housing is shedding heat to the air	Whether the driver is in foldback
Connector body and first 100 mm of cable	Whether a joint is dissipating power	Anything about the LEDs
Driver end of the housing	Local driver heating, often the foldback trigger	Emitter temperature
Bracket and mounting surface	Whether heat is being conducted into the vehicle	Air temperature inside a cavity

Housing temperature to junction temperature: the arithmetic

Owners often ask what housing temperature is acceptable. The honest answer is that no universal number exists, because the relationship between the outside of a housing and the junction inside it is a property of that housing, and it is not published for most lamps. What can be done instead is to reason with the arithmetic and to compare a lamp with itself.

The relationship is a chain of thermal resistances: junction to solder point, solder point to board, board to housing, housing to air. Each interface adds. If the total is 1.5 °C/W and the lamp dissipates 40 W, the junction runs 60 °C above ambient. If a mounting change doubles the housing-to-air resistance, the junction runs 120 °C above ambient for the same power, and the surface may rise by much less than that, because the surface is upstream of the bottleneck rather than at it. This is why “the housing only feels warm” is not evidence that the junction is safe, and why a 20 °C rise on the outside can correspond to a much larger rise inside.

Two practical consequences follow. First, track the shape of the curve rather than a single value: a surface that stabilises in ten minutes is shedding heat as fast as it arrives, while a surface that is still climbing at twenty minutes has not reached equilibrium and will keep heating on a longer run. Second, use the second lamp as the reference whenever there is one, because it shares the ambient, the supply and the mounting, which removes every variable except the lamp itself.

Installation decides the heat balance before the lamp does

A light bar is usually installed in the worst place the vehicle offers: behind a bumper, under a grille, against a radiator, or on a roof where the sun load arrives before the airflow does. The same lamp that holds full output on a bench can fold back in fifteen minutes in a cavity, and the difference is not a fault.

Position	What the housing sees	What to check first
Behind a bumper or grille	Recirculating hot air, little cross-flow	Clearance behind the fins, open area above and below
In front of the radiator or intercooler	Radiator discharge air, engine heat at low speed	Whether the bar blocks the cooler it sits in front of
Roof or roof rack	Direct sun load when parked, good airflow when moving	Fade in traffic versus on the highway
Bumper, low and exposed	Mud, snow, salt film, water spray	Coating on the fins and lens, cleaning routine
Near the exhaust or turbo	Radiant heat from a surface far hotter than ambient	Clearance, heat shielding, cable routing

Two installation factors do most of the damage and cost nothing to check. The first is clearance: fins need somewhere for the air to go, and a bar mounted 10 mm from a solid panel has fins that cannot convect. The second is coverage: mud, snow and dried salt act as insulation, and a bar that ran cool all summer can fold back in the first wet snow of the year with nothing wrong except a dirty housing.

A third factor is heat soak, the period after the engine stops when the bay temperature rises because coolant and metal are no longer being cooled by airflow. A lamp parked in that condition and then switched on starts from a higher baseline than one switched on from cold, which is why some faults appear only on a second run.

Ambient temperature and the shape of a derating curve

Suppliers who publish thermal data usually publish a derating curve, and the curve is worth reading properly because it tells you what the lamp will do rather than what it can do on a bench. The shape is consistent even when the numbers are not: full output up to a knee, a falling section where current is reduced as temperature rises, a limit at which the output is switched off, and a recovery band before it comes back.

Four things belong on any curve you are asked to trust, and if any of them is missing you are looking at marketing rather than data.

What the curve must state	Why it decides the answer
The measured point, such as case, board or ambient	A curve against ambient and a curve against case are not comparable
Whether the fall is current reduction or shutdown	Foldback dims; shutdown cuts out; the symptoms differ
The recovery condition and hysteresis	Explains a lamp that restarts and cuts out in a cycle
The mounting and airflow used for the test	A bench figure does not transfer to a bumper cavity

The most common misreading is to treat the knee as a failure threshold. It is not: the knee is where the designer decided that reducing output was better than letting the junction keep climbing, and a lamp that reaches it is behaving exactly as specified. The second most common misreading is to compare two lamps by their knee temperatures when one curve is drawn against ambient and the other against case temperature, which is a comparison between two different measurements.

When the cause is supply, not heat

Before you accept that LED light bar overheating is the explanation, rule out the circuit, because a supply problem produces the same complaint with a different clock. A heat problem scales with time and ambient; a supply problem scales with current and load, and it can appear the moment the lamp is switched on.

The mechanism is worth following because it explains why a bad joint looks like a heat fault. A constant-current driver holds LED current steady, so if the voltage arriving at the driver falls, the driver draws more current from the vehicle to deliver the same power. That extra current flows through the same undersized cable and the same corroded crimp, which dissipates more power in the joint, which raises its resistance, which drops more voltage. The arithmetic is unforgiving: 0.4 V dropped across a joint carrying 10 A is 4 W of heat inside a connector body the size of a thumbnail.

Evidence	Points to supply	Points to heat
Timing	Immediate, or when another load switches on	After minutes, and later on hot days
Response to engine speed	Changes with rpm or charging state	Unchanged by rpm if airflow is unchanged
Connector after two minutes	Warm or hot while the housing is cool	Cool while the housing is hot
Voltage at the lamp under load	Below the battery voltage by more than expected	Normal at the terminals
Two lamps on one circuit	Both affected equally	One affected, or both only if both are obstructed

The measurement that settles it is a voltage reading taken at the lamp terminals while the lamp is running, not at the battery with the lamp off. A second reading from the lamp’s ground terminal to the battery negative gives the return-path loss, and the two together account for almost every supply-side case. If you find more than a fraction of a volt in either path, stop diagnosing heat and fix the circuit; the site’s guide to a bar that dims at idle covers the charging, wiring and load side in full, and the division of labour is deliberate: that article owns the supply, this one owns the temperature.

Colour shift, flicker and dead segments at temperature

Three symptoms that appear when a lamp is hot are often reported as overheating but mean different things, and treating them as one problem leads to the wrong repair.

A colour shift toward blue in a hot lamp is the documented behaviour of an LED junction at elevated temperature, and it is recoverable while the absolute maximum ratings are respected (Bridgelux AN10). The useful test is the cold comparison: if the beam returns to its original colour after fifteen minutes off, the shift is thermal and the installation is the thing to change. If the colour is different from cold even after a full cool-down, the emitters or the phosphor have changed permanently, and no amount of airflow will bring the original colour back.

Flicker that only appears when the lamp is hot is a different case again, and it is usually a connection rather than an emitter. A crimp, a spade terminal or a solder joint that has expanded with temperature can make and break contact at a rate you can see, and the same joint is often invisible when cold. The test is a gentle wiggle of the connector and the first 200 mm of cable, cold and then hot, watching the beam; if the beam stutters only in the hot state, the joint is the suspect. The site’s guide to light bar flickering covers the wider set of causes, including PWM and relay behaviour, and this article does not repeat them.

A single dark section is not a thermal symptom in the ordinary sense. It is a failed emitter or a failed string, and the correct next step is the segment mapping procedure in the guide to partial failure and dead segments. There is one thermal connection worth knowing: on designs where several strings run in parallel from one driver, losing a string can push more current through the survivors, which makes the remaining emitters run hotter. That is a design-dependent effect, not a universal rule, and it is a reason to stop using a bar with a dead section rather than a reason to keep running it.

The signs that end the test

Some observations convert a diagnosis into an evacuation. They are not subtle, and the correct response to any of them is to stop energising the lamp rather than to gather one more data point.

Sign	What it means	What to do
Hot plastic or phenolic smell	Insulation, potting or a connector is beyond its rating	Switch off, disconnect the battery negative, do not re-energise
Smoke or a hiss	Something is decomposing or venting	Switch off, disconnect, let it cool outside the vehicle
Blistered, crazed or discoloured lens	The optical assembly has been over temperature	Retire the lamp; the housing has been damaged
Deformed housing or softened bracket	Structural plastic has passed its softening point	Retire the lamp and inspect the mounting for heat sources
Brown, green or melted connector	The joint has been dissipating power	Cut back and replace the connector and the cable section
Cable insulation hardened or cracked near the lamp	Local overheating of the conductor	Replace that section of the harness

The rule behind the table is that all of these are irreversible changes. Foldback is reversible, colour shift is reversible while the ratings hold, and a hot connector is repairable, but a blistered lens or a melted connector body marks a lamp or a harness that has already been outside its design envelope. Continuing to test after that point adds risk without adding information.

Safe test boundaries

The tests in this guide are deliberately limited to what can be done with the lamp installed, the circuit intact and the vehicle stationary. The boundaries below are not bureaucracy; each one corresponds to a way people get hurt or turn a repairable lamp into a fire.

• Work with the vehicle parked, the ignition off for any wiring work, and the battery negative disconnected before you touch a terminal.
• Keep the circuit fused at the rating the harness was designed for, and never fit a larger fuse to stop it blowing.
• Remove rings, watches and bracelets before working near a live 12 V circuit, because a shorted ring or a dropped tool across a battery terminal can weld and burn.
• Do not test with the vehicle in motion or on a public road, and do not use the behaviour of other drivers as a test signal.
• Do not open a sealed light bar. The housing is bonded or potted, there are no user-serviceable parts inside, and opening it destroys the ingress protection that the rest of the design depends on.
• Route and re-route cable away from exhaust manifolds, turbochargers and catalytic converters, and never coil surplus cable in a hot zone.
• On a hybrid or battery-electric vehicle, treat the high-voltage system as untouchable. The orange cables and the pack are not part of this diagnosis, and only the 12 V auxiliary circuit is in scope.
• Stop and hand the job to a qualified auto electrician if you smell burning, find melted insulation, find damage to the vehicle’s own loom, need to work on a circuit shared with a factory module, or cannot measure safely.

Three of those boundaries deserve a sentence more. The sealed housing rule exists because the first thing an owner wants to do with a hot lamp is open it, and the second thing is to seal it again with a generic sealant; both make the lamp worse and neither is necessary for any diagnosis in this article. The fuse rule exists because a fuse that blows under load is telling you about the circuit, and fitting a bigger one converts an electrical fault into a fire. The 12 V rule exists because a vehicle battery can deliver hundreds of amps into a short, which is far more energy than the lamp ever sees.

Repair, derate or replace

Once the evidence is in, the decision is usually obvious, and it is worth writing down because the three options have very different costs.

Finding	Decision	Reasoning
Fades and recovers, no damage, one lamp of two	Change the installation	The lamp is behaving as designed; the mounting is the limit
Fades and recovers, no damage, both lamps	Change the installation or the duty cycle	A shared cause is environmental or electrical, not the lamp
Cycles off and on within minutes	Improve cooling before anything else	The margin is small; a bracket or clearance change may be enough
Dims and stays dim, uneven sections	Replace the lamp	Permanent change inside the housing
Dead section, rest of the bar working	Stop using it, then replace	Survivors may be carrying extra current
Hot connector, cool housing	Repair the circuit	The lamp is not the fault
Blistered lens, deformed housing, smell	Retire the lamp and inspect the vehicle	The lamp has been outside its envelope; the installation may still be wrong

The pattern in that table is the article in one line: fixing LED light bar overheating usually means changing where or how the lamp is mounted, and replacing the lamp is the answer far less often than owners expect. The exception is the row where the housing has already been damaged, because at that point the reversible part of the story is over.

Changes that lower the running temperature

The interventions that work are unglamorous, cheap and mostly about air. In rough order of effectiveness per unit of effort, they are clearance, airflow, shielding, cleanliness and duty cycle.

Clearance comes first because fins need somewhere for the air to go. Moving a bar 30 mm further from a solid panel, or rotating a bracket so the fins face the airflow instead of the bulkhead, changes the housing-to-air resistance without touching the lamp. Airflow comes second: a bar behind a grille insert often sits in a pocket where air is stagnant at low speed and turbulent at high speed, which is exactly the pattern that produces “it only dims in traffic”.

Shielding is the answer to radiant heat rather than air temperature. A metal heat shield between the lamp and an exhaust or turbo component blocks radiation without blocking convection, and it is the correct fix where clearance cannot be increased. Cleanliness is the answer to mud, snow and salt, all of which insulate the fins; a soft brush and water at every wash is enough, and a pressure washer aimed at the seals is not.

Duty cycle is the intervention nobody wants but everybody can use. A work light that folds back after 25 minutes can be run in 20-minute blocks with a short pause, which keeps the junction in the flat part of the curve and costs nothing but discipline. If the task genuinely needs continuous output, the honest answer is a lamp with more thermal headroom rather than a fan taped to a housing.

A seasonal routine that catches it early

Most thermal faults announce themselves long before they strand anybody, and a short routine twice a year catches them while they are still installation problems.

Check	How	What is acceptable
Fins and lens	Visual, after a wash	No mud, snow, salt film or insect packing
Clearance behind the bar	Hand and torch	Open path for air above and below the fins
Mounting hardware	Spanner and visual	No loosened bolts, no cracked bracket, no fretting marks
Connector body	Touch after a two-minute run	Warm at most, never hot, no discolouration
Cable near hot parts	Visual along the route	At least a hand’s width from exhaust and turbo surfaces
Baseline run	The 20-minute test, in the same season each year	Time to first fade not shorter than last year

The value of the routine is the baseline in the last row. A lamp that has always folded back at 18 minutes on a 30 °C day is a known quantity, and a lamp that starts folding back at 8 minutes is telling you something changed. Without the baseline, both lamps produce the same complaint and neither produces evidence.

Fleet and duty-cycle planning

Fleets hit this problem in a different way, because the question is not whether one lamp is faulty but how many hours a day a lamp can be asked to deliver full output.

Duty pattern	Thermal consequence	Planning response
Short bursts, long pauses	Junction returns to ambient between uses	The published output is realistic
Continuous 8-hour shift	Housing reaches equilibrium and stays there	Size for the equilibrium, not the peak
Alternating lamps	Each lamp sees roughly half the heat	Useful where two positions exist
Hot climate, stationary work	Lowest airflow and highest ambient together	Derate expectations and increase clearance
Cold climate, snow coverage	Fins insulated by snow, ambient low	Cleaning matters more than ambient

The planning rule that follows is simple and unglamorous: specify the lamp for the duty pattern, not for the peak number on the box. A bar that holds full output for twenty minutes and then settles to a lower level is a perfectly good lamp for a work truck that uses it in bursts, and a poor choice for a machine that runs it all shift. Writing the duty pattern down, and the observed time to first fade beside it, turns a recurring complaint into a purchasing criterion.

What this guide does not cover

This article covers the diagnosis of output that falls as a lamp heats. It deliberately does not cover five adjacent problems, and the divisions are worth stating so that the right article gets read.

Water ingress, fogging and droplets inside a lens are a seal problem, covered in the guide to condensation and water ingress. A lamp that never lights at all, with no dimming phase, belongs to the not working checklist. Blown fuses and tripping relays belong to the short-circuit diagnosis procedure, and nothing in this article should be used to justify a larger fuse. Radio interference is a separate electromagnetic problem. And the design theory behind the heat path, including how a housing and its fins are supposed to work, is in the thermal management guide.

What remains is the useful core: a lamp that dims, cuts out or shifts colour as it heats is usually doing what it was built to do, and the job is to find out whether the installation or the lamp is the limiting element. The parts chosen for a hard duty cycle, such as the Jungle Series 40 inch bar or a heavy-duty work light, simply move that limit further out; they do not remove it, which is why the measurement routine matters more than the specification sheet. The LED light bars category shows the range of housings available when a mounting position turns out to be the real constraint.

Frequently asked questions

How long should a light bar run before the housing gets hot?

In practice most housings are close to their final temperature within 10 to 20 minutes of continuous running, and the exact figure depends on the housing mass, the mounting and the airflow. A bar that is still climbing steeply at 20 minutes has not reached equilibrium and will continue to heat on a longer run, which is the observation worth recording rather than the absolute number.

Is it normal for a light bar to dim after 20 minutes?

Yes, provided it recovers. A gradual reduction in output after 20 minutes, followed by full brightness after the lamp cools, is the signature of thermal foldback working as designed. A reduction that does not recover, or that arrives with a smell or a discoloured lens, is not normal and should be treated as damage.

Can a light bar overheat while the vehicle is moving?

It can, and the usual reason is that the mounting position does not receive airflow even at speed. A bar set into a bumper pocket, or mounted hard against a panel with the fins facing the wrong way, can sit in stagnant air at 100 km/h. The diagnostic signal is a lamp that behaves better on the highway than in traffic: that pattern points at airflow, not at the lamp.

How hot is too hot to touch?

There is no published surface temperature that applies to every lamp, so the useful rules are relative rather than absolute. Treat any surface you cannot keep a hand on as a burn hazard first, then compare the reading with the second lamp and with the same lamp’s own baseline. A surface that stabilises is shedding heat; a surface that is still climbing is not.

Will a hotter light bar fail sooner?

Sustained operation at higher junction temperature is the mechanism that shortens LED life, which is why drivers fold back rather than let the junction keep climbing. The practical consequence is that a lamp which folds back every day is being worked harder than one that never reaches its knee, even though both are still functioning.

Does a bigger heat sink or an added fan fix it?

Sometimes, and less often than owners hope. Additional metal only helps if it is coupled to the original housing and exposed to air, and a fan only helps if it moves air across the fins rather than around the bracket. The interventions that reliably work are clearance, airflow direction, shielding from radiant heat and cleaning, in that order, before any hardware is added.

Should I keep running a lamp that folds back?

A lamp that folds back and recovers can be used, but it is telling you that the thermal margin is gone, and the honest response is to fix the installation or shorten the duty cycle rather than to accept the lower output indefinitely. If the foldback becomes a shutdown cycle, or if any of the damage signs appear, stop using it until the cause is corrected.

Is LED light bar overheating a warranty issue?

That depends entirely on the seller’s terms, and no general answer is safe. What you can do is make the claim diagnosable: record the ambient temperature, the time to first fade, whether full output returns after cooling, and photographs of the housing and connector. A claim supported by a time series is a different conversation from a claim supported by an impression.

Editorial note

Written by SAEC Lighting Engineering Team, Foshan.

Last updated: 2026-10-18

This guide describes diagnostic procedure and measurement logic for installed lamps. It does not replace the lamp manufacturer’s instructions or the vehicle service manual, and it deliberately stops short of any procedure that requires opening a sealed housing or working on a high-voltage system. Where a figure in this article is arithmetic or illustrative, the text says so: the derating curve, the heat-path resistances and the thermal-resistance examples are drawn to explain behaviour, not to specify any product.

Sources

• Texas Instruments, “Implementing Analog Thermal Foldback with the LP8867-Q1 and TMP61-Q1” (SNVAA22, June 2021) — https://www.ti.com/document-viewer/lit/html/SNVAA22/GUID-8F3431E7-A67B-4ABB-ACAF-CC9171A95CEF
• Texas Instruments, LP8867-Q1 automotive LED driver product page and feature list — https://www.ti.com/product/LP8867-Q1
• Bridgelux, “Application Note AN10: Effective Thermal Management of Bridgelux LED Arrays” — https://www.bridgelux.com/application-note-an10-effective-thermal-management-bridgelux-led-arrays
• HELLA Tech World, “LED technology: we answer the most important questions” — https://www.hella.com/techworld/uk/lounge/led-questions-answers/
• Omega Engineering, “Total emissivity table” (emissivity reference for infrared thermometry) — https://www.omegaengineering.cn/learning/tblofttlemiss.html