An LED headlight bulb retrofit is a three-part decision rather than a bulb purchase: the base has to match the lamp holder, the driver has to keep the vehicle’s bulb-out monitoring satisfied, and the optics have to keep the beam inside a cut-off the lamp was never designed around. Get all three right and you have a lighting upgrade. Miss any one of them and you have a dashboard warning, a flicker at speed, or glare you cannot aim away.
This guide works from published manufacturer data and from the two rulebooks that decide the question in the two largest English-language markets. The numbers below are the ones a buyer can check: lamp base designations, nominal wattage, maximum power input, nominal luminous flux with its tolerance, test voltage, and the circuit currents that follow from them at 13.2 V. Where a figure could not be verified from a primary document, this guide says so instead of filling the gap.
Five base codes cover most halogen headlamps sold in North America and Europe, and they are not interchangeable. H4 uses a P43t base and carries two filaments in one envelope, rated 60/55 W. H7 uses PX26d, H11 uses PGJ19-2, 9005 is the HB3 category on a P20d base, and 9006 is HB4 on a P22d base; all four of those are single-filament and rated between 51 W and 60 W in the OSRAM data used here. The practical consequence is electrical: a pair of 55 W halogen lamps draws about 8.3 A at 13.2 V, while a retrofit driver that consumes 25 W draws about 3.8 A for the pair. That current reduction is real, and it is the only thing the retrofit reliably changes. Beam pattern, glare, colour compliance and dashboard warnings are separate problems with separate answers.
A base code describes a mechanical interface, not a brightness class and not a vehicle. Each code maps to a lamp cap designation — P43t, PX26d, PGJ19-2, P20d or P22d — and that designation is the standard reference for the flange, the keying and the retention inside the holder. The manufacturer data used throughout this guide lists those pairings explicitly: H4 on P43t, H7 on PX26d, H11 on PGJ19-2, 9005 on P20d, and 9006 on P22d.
The two naming systems overlap and that is where most buying mistakes begin. 9005 and 9006 are order references; the regulated category names in the same datasheets are HB3 and HB4. OSRAM’s page for the 9005 lamp lists the global order reference as 9005 and the ECE category as HB3, and the 9006 page does the same with HB4. A listing that says “9005 HB3” is naming one lamp twice. A listing that says “9006 = 9005 compatible” is naming two different lamps once each.
The physical difference is easier to see than the naming difference. H4 lists a 17.0 mm diameter; the other four all list 12.0 mm. That single number explains why an H4 lamp will not enter a holder cut for H7 regardless of what the box claims. H7 and H11 are the pair most often confused: both are 12.0 mm, both are 55 W nominal, both are tested at 13.2 V — and the bases are PX26d and PGJ19-2, which is to say completely different retention geometry.
Length matters when clearance is tight. The same datasheets give 82.0 mm for H4, 59.0 mm for H7, 67.0 mm for H11, 76.2 mm for 9005 and 80.3 mm for 9006. A retrofit assembly adds a heat sink and often a fan behind that dimension, so the number that matters for fitment is not the bulb length but the depth available behind the holder.
The table below reproduces the manufacturer’s own fields. Nominal wattage is the number printed on the box; maximum power input is the higher figure the same datasheet publishes; nominal luminous flux is the rated output with the manufacturer’s stated tolerance. All five are 12 V nominal lamps tested at 13.2 V.
| Base | ECE category | Order reference | Nominal wattage | Maximum power input | Nominal flux | Flux tolerance | Length |
|---|---|---|---|---|---|---|---|
| H4 | H4 | 64193 | 60/55 W | 75/68 W | 1650/1000 lm | ±15 % | 82.0 mm |
| H7 | H7 | 64210 | 55 W | 58 W | 1500 lm | ±10 % | 59.0 mm |
| H11 | H11 | 64211 | 55 W | 62 W | 1350 lm | ±10 % | 67.0 mm |
| 9005 | HB3 | 9005 | 60 W | 73 W | 1700 lm | ±12 % | 76.2 mm |
| 9006 | HB4 | 9006 | 51 W | 62 W | 1000 lm | ±15 % | 80.3 mm |
Every field in that table comes from the manufacturer’s own product pages, which also publish the flux tolerance, the 13.2 V test voltage and the overall length: H4 64193, H7 64210, H11 64211, 9005 as HB3 and 9006 as HB4, with the base and wattage columns matching the ORIGINAL LINE family overview.
Two things in that table are worth more attention than the lumen column. First, the flux tolerances run from ±10 % to ±15 %, which means a 1350 lm lamp is a lamp the manufacturer will stand behind anywhere from roughly 1215 lm to 1553 lm. Second, the H4 figures are paired in the order high beam then low beam, so 60 W and 1650 lm belong to one filament and 55 W and 1000 lm to the other. Reading the H4 row as a single 60 W lamp producing 1650 lm in both switch positions is the most common misreading of that datasheet.
H4 is a two-filament lamp in a single glass envelope, and the second filament is not a spare. Hold one and the arrangement is visible: two filaments at different heights, one of them shielded. That is what lets a single reflector produce an asymmetric dipped beam with a cut-off as well as a separate main beam, and it is why the H4 datasheet carries two values in almost every row: 60/55 W, 75/68 W, 1650/1000 lm, and two B3 lifetimes, 250 h for the first filament and 400 h for the second.
The consequence for a retrofit is structural, not electrical. A single LED emitter in an H4 holder has to reproduce two different source positions, and most retrofit assemblies solve this by moving the emitter or by using two emitters with a shield. Whatever the mechanical solution, the lamp’s dipped beam was certified against a source at one specific position with one specific size, and the retrofit’s source is at a different position with a different size. That is a statement about geometry, not about quality: it is the reason a retrofit can look bright on a wall and still fail an aim check.
The contrast is useful when you are reading listings. An H7, H11, 9005 or 9006 lamp has one filament and one beam function per lamp, so a vehicle using them typically has separate lamps for each function, or a projector with a movable shield. An H4 vehicle has one lamp doing both jobs. If a listing offers an “H4 LED with high and low beam” it is describing two emitters in one body, and the dipped beam is the half that decides whether the result is usable.

You will find lists online that map models to base codes, and this guide deliberately does not add one. The reason is that the base code is only the first of several conditions, and the other conditions are not visible from a model name. Trim level, projector versus reflector optics, the presence of a bulb-out monitor, whether the lamp was certified as a replaceable-bulb headlamp at all, and mid-cycle changes within one model year all move the answer.
There is a verifiable method instead, and it takes about five minutes. Open the owner’s manual and find the lamp replacement table; it gives the base code and often the wattage. Then read the lamp itself. In the United States, FMVSS No. 108 requires the lens of a replaceable bulb headlamp to be permanently marked, in front of each light source, with the HB Type of that source (S6.5.3.4.1), so the lens tells you whether you are looking at an HB3 or an HB4 cavity. In the United Kingdom the equivalent evidence is the approval sticker described in the MOT manual, which must state the light source category, for example “H11 LEDr”.
| What you are checking | Where the answer is written | Why the model name is not enough |
|---|---|---|
| Base code | Owner’s manual lamp table, or the base of the lamp you remove | Trims and mid-year changes vary |
| Regulated category | HB Type marking on the lens in the US; ECE category on the lamp | A trade number does not identify the cavity |
| Whether the lamp accepts a retrofit | Approval sticker requirement in the UK; the vehicle maker’s own statement elsewhere | Certification is per lamp, not per model |
| Bulb-out monitoring | Owner’s manual, or a warning after fitting | Not visible from outside the car |
| Available depth behind the holder | Physical measurement with the cover removed | Body and lamp variants differ |
If any of those five is unknown, the retrofit is not yet a decision. That is the honest position, and it is more useful than a list that is right for one trim and wrong for the next.
A product listing is written by the seller. A lamp is marked by whoever certified it. The gap between those two documents is where most of the risk sits, so it is worth knowing exactly what each one is allowed to say.
In the United States, the letters DOT moulded into a headlamp lens are a manufacturer’s certification under 49 U.S.C. 30115, not a government approval of that individual part (S6.5.1). The same standard requires the lens to carry the manufacturer’s trademark, the voltage and the part or trade number (S6.5.3.1 and S6.5.3.2). In the European system the equivalent evidence is a type-approval mark with an approval number, which is why the UK MOT manual can ask a tester to read a sticker rather than dismantle the lamp.
The practical routine is to read four things off the part in your hand: the base designation, the regulated category (H4, H7, H11, HB3, HB4), the voltage, and the approval or certification marking. If the part in your hand has none of them and the listing has all of them, you are reading marketing copy rather than a specification. Our own guide to DOT and E-mark approval markings covers how to decode those symbols in more detail, including what neither mark proves.
Every lamp in the table above has a maximum power input higher than its nominal wattage. For H4 the figures are 75/68 W against 60/55 W, which is 25 % and 24 % more. For H7 it is 58 W against 55 W, only 5 % more. For H11 it is 62 W against 55 W, 13 % more. The two American categories sit at 73 W against 60 W and 62 W against 51 W, both about 22 % more.
That gap is not a manufacturer’s trick. It reflects the difference between a nameplate number and the power the lamp actually consumes at the test voltage, which the same datasheets give as 13.2 V rather than 12 V. A lamp designed for a nominal 12 V system is tested at a voltage a charging system actually produces, and the current follows from the power at that voltage.
| Base | Nominal | Maximum input | Increase over nominal | Current at maximum input, 13.2 V |
|---|---|---|---|---|
| H4 | 60/55 W | 75/68 W | +25 % / +24 % | 5.68 A / 5.15 A |
| H7 | 55 W | 58 W | +5 % | 4.39 A |
| H11 | 55 W | 62 W | +13 % | 4.70 A |
| 9005 | 60 W | 73 W | +22 % | 5.53 A |
| 9006 | 51 W | 62 W | +22 % | 4.70 A |
The buying lesson is that “55 W” is a label, not a measurement. When you compare a halogen lamp against a retrofit, compare maximum input power against maximum input power, and compute current at the same voltage for both. An LED headlight bulb retrofit that quotes only an “equivalent wattage” has told you nothing you can put into a fuse calculation.
Current is power divided by voltage, so every row below is arithmetic on the manufacturer’s own numbers at the 13.2 V test voltage. The last four rows are not product data; they are what a retrofit driver would draw if its input power were 15, 20, 25 or 30 W, which is the calculation a buyer has to do because published input power for retrofit assemblies is often missing.
| Circuit | Source type | Input power used | Current per lamp at 13.2 V | Current for a pair |
|---|---|---|---|---|
| H4, low beam | halogen | 55 W | 4.17 A | 8.33 A |
| H4, high beam | halogen | 60 W | 4.55 A | 9.09 A |
| H7 | halogen | 55 W | 4.17 A | 8.33 A |
| H11 | halogen | 55 W | 4.17 A | 8.33 A |
| 9005 / HB3 | halogen | 60 W | 4.55 A | 9.09 A |
| 9006 / HB4 | halogen | 51 W | 3.86 A | 7.73 A |
| Retrofit driver | LED, if 15 W | 15 W | 1.14 A | 2.27 A |
| Retrofit driver | LED, if 20 W | 20 W | 1.52 A | 3.03 A |
| Retrofit driver | LED, if 25 W | 25 W | 1.89 A | 3.79 A |
| Retrofit driver | LED, if 30 W | 30 W | 2.27 A | 4.55 A |
Read the last two rows against the first. A retrofit driver consuming 30 W per lamp draws half the current of a 60 W halogen, and a 15 W driver draws a quarter. That is the entire electrical case for the conversion, and it is a good case: less current means less voltage drop in the existing wiring, less heating in the connector, and longer relay contact life. It does not mean the lamp produces more useful light, and it does not make the installation compliant anywhere.
Lower current buys three things that a careful buyer can predict. The first is reduced voltage drop: with a fixed circuit resistance, drop is proportional to current, so a circuit that loses 0.84 V at 4.2 A loses about 0.4 V at 2 A. The second is reduced heating at every resistive joint, because power dissipated in a contact goes with the square of current. The third is reduced wear on switch and relay contacts, which arc less when they interrupt a smaller current.
It does not buy three other things, and these are the ones buyers assume. It does not buy beam pattern: the shape of the cut-off is set by the reflector or projector and the position and size of the source, neither of which the driver changes. It does not buy colour compliance: the MOT manual describes acceptable headlamp colour in words — white, predominantly white with a blue tinge, or yellow — so a Kelvin figure on a box cannot by itself demonstrate compliance. And it does not buy legality: as the next sections show, the rules attach to the light source and the lamp as a certified pair, not to the current drawn.
The awkward middle case is a retrofit that draws more than the halogen it replaces. That happens when a seller adds a load resistor to satisfy the bulb-out monitor. The resistor’s dissipation is real power turned into heat inside the lamp housing, and it is added to whatever the driver already produces. A resistor sized to replace a 55 W halogen load adds roughly that much heat in the same enclosed space the halogen used to heat — which is a thermal question before it is an electrical one.
Luminous flux is a real quantity and it is measured under defined conditions. A flux figure becomes comparable only when it comes with three things: the electrical condition it was measured at, the tolerance the manufacturer will stand behind, and a statement of what was measured — the source alone or the complete lamp.
The OSRAM datasheets above supply all three for halogen. Test voltage is 13.2 V, tolerance is stated per part (±10 % to ±15 %), and the figure is a nominal luminous flux for the light source. That is why the table in this article can quote a tolerance next to every lumen value. When a listing for an LED headlight bulb retrofit gives a single large number and none of the three, the number is not wrong so much as unanchored: there is nothing to compare it to, because the conditions that produced it are unknown.
There is a second trap that has nothing to do with honesty. Flux is total emitted light in all directions. What a headlamp does is collect a fraction of that light and aim it into a pattern with a cut-off. Two sources with identical flux can produce very different road illumination depending on how much of the flux the optics collect and where they send it. IIHS makes the same point from the other direction when it states that its headlight ratings are technology neutral and that using a particular technology does not guarantee a good rating.
If you want the vocabulary for the difference between the number on the box and the light on the road, our explainer on raw and effective lumens goes through it; the short version is that a flux number without a beam is a budget, not a result.
| Condition a flux figure needs | Why it changes the number | Present on the halogen datasheet | Commonly present on a retrofit listing |
|---|---|---|---|
| Measurement voltage or drive current | Flux rises with current and with junction temperature | Yes, 13.2 V test voltage | Usually absent |
| Stated tolerance | Turns a point value into a range | Yes, ±10 % to ±15 % | Usually absent |
| What was measured | Source flux is not lamp output | Yes, light source | Usually ambiguous |
| Beam or intensity data | Decides what reaches the road | Not on the source datasheet | Rarely useful when present |
A dipped beam is defined by where its upper edge falls, not by how much light it emits. The UK MOT manual expresses that as a tolerance band: for headlamps whose centres are 850 mm or less above the ground, the horizontal cut-off of the beam image must fall between the 0.5 % and 2.75 % lines; above 850 mm, between 1.25 % and 2.75 %. Those percentages are the whole regulatory idea in one line: light above the cut-off is glare, light below it is illumination.
The IIHS headlight research quantifies why that matters. About half of traffic deaths occur in the dark or at dawn or dusk, and vehicles rated good for visibility in its headlight test have 19 % fewer nighttime single-vehicle crashes and 23 % fewer nighttime pedestrian crashes than vehicles rated poor. Its test uses a pedestrian target at 140 ft and a deer target at 220 ft, which is a statement about distance, not about lumens.
The same source puts glare in proportion: it is reported as a contributing factor in only one or two out of every 1,000 nighttime crashes. That is a small share, and it is worth quoting because it cuts both ways. Glare is not the main safety problem — inadequate visibility is — but a small share of a very large number is still a real number of people, and the fix for glare is geometry and aim rather than a dimmer bulb.
CAN is a vehicle network standard, and its scope is narrower than the marketing use of the word suggests. The current edition, ISO 11898-1:2024, published on 24 May 2024, specifies the controller area network data link layer and the physical coding sub-layer. Its own abstract states that how the higher-layer functions are handled is not specified in that document.
That single sentence resolves a lot of confusion. There is no standard-defined category called a “CANbus bulb”, because the standard in question does not define lamp monitoring, let alone a lamp that satisfies it. “CANbus-ready”, “CANbus built-in” and “error-free” are seller shorthand for a lamp assembly that includes extra electrical load or a small circuit intended to keep the vehicle’s lamp monitoring quiet. What that circuit does, and whether it does it at 5 °C and at 60 °C, is a question for the seller.
The function it is trying to satisfy is described in HELLA’s technical documentation: many modern vehicles are equipped with bulb failure detection, and when a bulb is defective the on-board electronics detect the missing load and show the driver a warning. That is a load-based test, and it explains the whole problem in one line. A halogen lamp is a resistor that draws about 4 A; an LED driver is a circuit that draws about 2 A or less. The monitor sees a missing load, not a missing light.
Two different faults get called “CANbus problems”, and confusing them wastes money because the fixes are different. The first is a warning lamp: the light works, the dashboard complains. The second is visible flicker: the dashboard is quiet, the light strobes or pulses. HELLA’s documentation separates them in exactly that way.
For the first, the cause is the missing load, and the fix is an adapter that restores enough load for the monitor. For the second, the cause is the supply: some manufacturers control the power to the bulbs with a pulse width modulated signal, so the voltage is regulated by a higher-level control unit on demand rather than applied continuously. HELLA notes that this can make an LED retrofit flicker while driving, and that an additional adapter is needed for that case. HELLA’s technical documentation for its LED retrofit range separates them in exactly that way.
| Symptom | Mechanism | What actually changes it | What does not |
|---|---|---|---|
| Warning lamp on, light works | Monitor sees lower current than expected | Load adapter sized for the original lamp | Brighter LED, higher colour temperature |
| Visible flicker while driving | PWM-controlled supply interacting with the driver | Supply-appropriate adapter, or a different driver | A bigger heat sink |
| Both at once | Two independent causes | Both fixes, tested separately | Assuming one adapter covers both |
| Works cold, warns when hot | Load changing with temperature | A seller’s temperature specification | A generic “error-free” claim |

The diagnostic order matters more than the parts list. Fit the lamp, note whether the warning appears with the engine off and with it running, then note whether the light is steady at idle and at speed. Those two observations separate the load problem from the supply problem before you buy anything.
A claim of CANbus readiness can legitimately mean that the assembly includes a parallel load, a filter, or an active circuit that emulates a filament. Each has consequences the phrase hides.
A parallel resistor is the simplest and the most thermally awkward. If it restores the load of a 55 W lamp, it dissipates something close to 55 W minus whatever the driver already draws, and that energy leaves as heat inside the housing. A capacitor or RC filter addresses ripple and short pulses rather than steady-state load, so it can silence a flicker without satisfying a load test. An active circuit can do either, and its behaviour at temperature extremes is exactly what a datasheet should state and a slogan will not.
| Claim on the listing | Plausible meaning | Question that decides it |
|---|---|---|
| “Built-in CANbus” | Parallel load resistor | How many watts does it dissipate, and where does that heat go? |
| “Error-free” | Tested on some vehicles | Which vehicles, which model years, and at what ambient temperature? |
| “No flicker” | Filter or supply-compatible driver | Does it stay steady with the engine running and the alternator charging? |
| “Plug and play” | Same connector | Does the assembly fit behind the lamp cover, fan included? |
None of those questions is unreasonable to ask a seller, and a seller who cannot answer them is telling you something useful. Our troubleshooting guide on LED headlight flicker and CANbus codes covers the after-the-fact diagnosis if you have already fitted a set.
A lamp converts electrical power into light and heat, and the split is never total. Whatever is not emitted as light has to leave the assembly by conduction, convection or radiation, and in a headlamp housing the available paths are narrow. HELLA’s description of a retrofit assembly is a compact list of the answers the industry uses: a built-in fan for active ventilation, an anodised aluminium heat sink with a large radiating surface, an integrated overheating protection function, and a constant-current driver that dims the LED as temperature rises.
That last item is the one buyers underestimate, because it silently changes the product’s performance. A driver that reduces current at high temperature is protecting the LED, and the visible result is a lamp that is brighter in the first minute than in the twentieth. Any brightness comparison made on a cold lamp therefore flatters whichever design relies most on foldback.
| Heat path | How it works | Typical failure mode | What to check |
|---|---|---|---|
| Passive heat sink | Fins or a braid conduct heat to surrounding air | Inadequate in a closed housing with no airflow | Clearance behind the lamp, cover fitment |
| Fan-assisted | A small fan moves air across the heat sink | Fan stalls, clogs with dust, or is noisy at speed | Audible check after 10 minutes, dust ingress |
| Driver foldback | Current is reduced as temperature rises | Output falls where the buyer cannot see it | Compare output cold and after 20 minutes |
| Added load resistor | Converts current into heat deliberately | Local hot spot near wiring or plastics | Where the resistor sits, and what is next to it |
Two of those rows are worth stating plainly. A fan converts an electrical problem into a mechanical one: it moves heat well until it stops, and when it stops the assembly may be worse than a passive design that never depended on it. A load resistor turns a monitoring problem into a thermal one in the same sealed space. Both are legitimate engineering choices. Neither is visible in a headline lumen figure.
For the general physics of junction temperature and thermal resistance in vehicle lamps, our article on LED thermal management covers the chain from junction to air; the retrofit-specific addition is that the housing was designed to move a different amount of heat.
Foldback means the light you bought is the light you get for the first few minutes. Because the driver dims the LED as temperature rises, a fair comparison between two retrofit assemblies has to be made after both have reached thermal equilibrium, not on the bench.
There is a practical test that needs no instruments. Fit one lamp, switch it on, and photograph the beam on a wall from a fixed position at one minute and at twenty minutes with the same camera settings. Repeat for the other side. A design that depends heavily on foldback will show a visible drop; a design with margin will show little. Doing it on a cold evening rather than in a warm workshop makes the difference easier to see, because the starting point is lower.
The same logic explains a complaint that appears in service rather than in reviews: a retrofit that looked impressive in winter can look ordinary in a hot climate with the lamps on and the vehicle stationary. The driver has not changed; the thermal environment has. Anything a seller states about sustained output, rather than peak output, is therefore worth more than a larger peak number.

A headlamp is an optical system in which the light source is a component with a position and a size, and the reflector or projector is shaped around both. FMVSS No. 108 makes this concrete by defining a replaceable bulb headlamp as a bonded lens and reflector assembly plus one or two replaceable light sources, and by requiring the lens to be marked with the HB Type of the source it was built for. The certification belongs to the combination, not to the bulb.
Size matters as much as position. A halogen filament is a small, roughly cylindrical emitter; an LED retrofit is one or more flat emitting surfaces with a different shape and a different angular distribution. A reflector that was shaped to collect light from a cylinder does not collect the same fraction from a flat array, and the part of the beam that suffers first is usually the cut-off, because the cut-off is where the reflector’s precision is concentrated.
This is why HELLA’s installation guidance includes a set of twelve adapter rings: the retainer geometry differs between lamps, and the ring is what lets the assembly sit in the holder at all. Being seated is a necessary condition. It is not the same as being at the right optical position, and no adapter ring can change the emitting geometry of the source.
Aiming is the adjustment you are allowed to make, and it is powerful within its limits. The MOT tolerance bands quoted earlier show how narrow the acceptable window is: at 850 mm lamp height the cut-off has to sit between 0.5 % and 2.75 % of the test distance, which is a band rather than a target. Move a correct lamp inside that band and the result improves. Move an incorrect source inside that band and the beam may still have a shape the lamp was never certified for, with light scattered above the cut-off that no adjustment removes.
The scale of the problem is worth keeping in proportion. IIHS reports that excessive glare affected 21 % of the headlight systems it tested in the 2017 model year and 5 % in 2026, which shows that glare is an engineering outcome that manufacturers can fix, not an unavoidable property of bright light. It also reports that glare is a contributing factor in only one or two out of every 1,000 nighttime crashes — a reminder that “too bright” is a smaller problem than “not enough light where it is needed”.
For practical aiming procedure, including how to prepare a level test area and how to adjust one lamp at a time, our guide to auxiliary light aiming and glare applies the same method; the difference with a headlamp is that the tolerance is regulated rather than chosen.
In the United States the relevant instrument is FMVSS No. 108, and its scope is broader than many buyers assume. S1 states that the standard specifies requirements for original and replacement lamps, reflective devices and associated equipment. S3.3 applies it to lamps and associated equipment for replacement of like equipment on vehicles within scope. A replacement light source is therefore not outside the rule.
Two provisions do most of the work. S6.2.1 prohibits installing any additional lamp, reflective device or other motor vehicle equipment that impairs the effectiveness of lighting equipment required by the standard. And the definition of a replaceable light source ties conformity to 49 CFR Part 564: it is an assembly of a capsule, base and terminals designed to conform to the requirements of appendix A or appendix B of that part. Part 564 is the filing system through which replaceable light sources are documented, with appendices for replaceable light sources and for sealed beam units.
What that means for a buyer is narrower than “legal” or “illegal” and more useful. The question to ask is whether the light source in the box is a documented replaceable light source conforming to Part 564, and whether the installation impairs the lighting equipment the vehicle is required to have. A seller who cannot point to the documentation has not answered the first question. HELLA’s own US technical page is candid about the second: it states that the retrofit product described there is approved for off-road use only, and that use on public roads can lead to loss of the operating permit and insurance coverage.
Whether an added lamp may be used on a road is a separate question from whether the headlamp may be converted, and our guide to street legality for LED light bars covers the state-level layer that sits underneath the federal equipment rules.
The UK MOT inspection manual is the most explicit public document on this subject that this guide could verify, and it is worth reading closely because it shows both the rule and its exception. It states that existing halogen headlamp units on vehicles first used on or after 1 April 1986 must not be converted to high intensity discharge light sources. For LED it takes a different route: a halogen unit fitted with an LED replacement light source may be accepted where that source has been ECE approved for use in that lamp and is supported by a manufacturer-issued approval sticker.
The sticker requirements are specific, and they are the closest thing to a checklist a buyer can hold. It must be permanently affixed, readily visible to the tester, incorporate an anti-counterfeit feature such as a hologram, and display the LEDr category designation (for example “H11 LEDr”), the rated voltage, the ECE approval number and the manufacturer’s name. If it is missing, illegible, damaged, tampered with, or does not match the fitted source, the defect is “light source and lamp not compatible”, which is a major defect.
Two further details keep the rule from being over-read. The manual says testers are not expected to remove covers or light sources to verify approval marks, so the sticker is the evidence the inspection relies on. And it says the LEDr provision does not refer to complete replacement headlamp units, which are treated differently: a complete unit manufactured with LED sources must meet the other requirements for its type but is not failed for incompatibility.
| Jurisdiction | Instrument | What it requires of a retrofit source | What it does not establish |
|---|---|---|---|
| United States | FMVSS No. 108, S3.3 and S6.2.1 | Replacement equipment in scope; no installation that impairs required lighting | That a particular product conforms |
| United States | 49 CFR Part 564 | Replaceable light sources conform to appendix A or B | That a listing’s claim is documented |
| United Kingdom | MOT manual 4.1.4 | ECE-approved LEDr for that lamp, plus a compliant approval sticker | That the beam meets the aim band after fitting |
| Any market | ISO 11898-1:2024 | Nothing about lamps; it defines CAN’s data link layer | That “CANbus” describes a certified product |
The MOT is an inspection, not a type approval, and it is a UK instrument. Reading it as a general permission to convert would be a mistake in the other direction, and so would reading the US position as a blanket prohibition: what the US framework requires is documentation and non-impairment, and what the UK framework requires is an approved source plus evidence. Neither is satisfied by a lumen figure.
There are conditions under which an LED headlight bulb retrofit is the wrong decision, and they are common enough to state plainly.
There is also a softer condition that deserves its own line: if the existing halogen lamps are simply old, replacing them with new halogen lamps of the specified type restores the output the lamp was certified for, at a fraction of the cost and with no compliance question at all. A 1650 lm H4 that has lost output to age is not improved by an unverified retrofit; it is improved by a lamp that matches the specification on the lens.
The checklist below is ordered so that the cheapest disqualifiers come first. Each row has a pass criterion you can evaluate without instruments, and a fail action that is “stop” rather than “compensate”.
| Check | Pass criterion | If it fails |
|---|---|---|
| Base code | Matches the owner’s manual table and the base of the lamp removed | Stop; the lamp is not the one you think it is |
| Regulated category | HB Type on the lens, or the ECE category on the lamp, matches the product | Stop; a trade number is not a category |
| Approval evidence | ECE approval number and, where the UK rule applies, a compliant sticker | Stop; a claim is not evidence |
| Maximum input power | Stated by the seller, in watts, at a stated voltage | Ask; without it there is no fuse calculation |
| Bulb-out behaviour | Seller states what the assembly does and at what temperature | Ask; “error-free” alone is not an answer |
| Depth behind the holder | Measured with the cover off, compared with the assembly depth | Stop; a forced cover traps heat |
| Heat path | Fan or heat sink has clearance and the cover still fits | Stop; recirculated heat shortens life |
| Sustained output | Seller states output after warm-up, not only peak | Treat a peak-only claim as unquantified |
| Colour | Described in the words the applicable rule uses, not only in Kelvin | Re-read the rule before buying |
| Aim plan | You know where and how you will check the cut-off afterwards | Do not fit until you do |
Ten rows, and eight of them can be answered before money changes hands. The two that cannot — depth and aim — require the vehicle, which is why they belong at the end.
Sometimes the answer is that an LED headlight bulb retrofit should not happen at all, and the lighting problem is solved somewhere else. That is a legitimate outcome, and it is the one this guide recommends whenever the approval evidence or the beam data is missing.
Auxiliary lighting is the compliant outlet in most markets because it adds lamps rather than changing a certified one, and because auxiliary lamps are switched separately. The engineering questions then move to mounting, switching and aim, which are easier to answer honestly. If you are adding lamps to a vehicle rather than replacing a source, the relevant buying decisions are beam pattern, mounting position and circuit sizing rather than base codes.
For readers who reach that conclusion, the practical starting points in our own catalogue are the off-road lighting range and the light pod range, where the specification question is beam shape and current draw rather than lamp compatibility. A 40 inch bar and a compact spot pod solve different problems, and neither claims to replace a certified headlamp.
No. H7 uses a PX26d base and H11 uses PGJ19-2, and although both list a 12.0 mm diameter and 55 W nominal in the manufacturer data used here, the retention geometry is different. The base codes exist precisely to prevent that substitution.
No. OSRAM lists 9005 as ECE category HB3 on a P20d base, rated 60 W and 1700 lm, and 9006 as HB4 on a P22d base, rated 51 W and 1000 lm. They are different categories with different bases and different outputs.
It depends on whether your vehicle monitors lamp load, and on how it supplies the lamp. If the dashboard reports a bulb failure while the light works, the monitor is seeing too little load. If the light flickers while driving, the supply is likely pulse width modulated. Those are different faults with different adapters, so diagnose before buying.
Not necessarily, and the flux figure cannot tell you. Luminous flux is total emitted light; what reaches the road depends on how much of it the reflector or projector collects and where it sends it. IIHS states that its headlight ratings are technology neutral and that a given technology does not guarantee a good rating.
Because the driver is protecting the LED. HELLA’s technical description of a retrofit assembly includes an integrated constant-current driver that dims the LED as temperature rises. That is normal behaviour, and it is why sustained output is a more useful specification than peak output.
It depends on the market and on the documentation. In the United States, FMVSS No. 108 covers replacement equipment and prohibits installations that impair required lighting, and a replaceable light source is defined by conformity to 49 CFR Part 564. In the United Kingdom, the MOT accepts an LED replacement in a halogen unit only where the source is ECE approved for that lamp and carries a compliant approval sticker. Neither question is answered by a lumen figure.
Not automatically. A retrofit driver that draws less current produces less heat than the halogen it replaces, but a load resistor added to satisfy the bulb-out monitor turns current back into heat inside the same housing. The net thermal effect depends on the total of driver losses plus any added load.
Four things: the base code from the owner’s manual, the regulated category from the lamp or lens marking, the seller’s stated maximum input power at a stated voltage, and what the assembly does about bulb-out monitoring. If any of the four is missing, the purchase is a guess.
Written by the SAEC Lighting Engineering Team, Foshan. Last updated: 2026-10-17.
This article uses published manufacturer data for the five lamp bases rather than vendor claims, and quotes the United States and United Kingdom instruments directly where the wording matters. Three limits should be stated. First, this guide does not list vehicle applications: the manufacturer bulb-finder tools are interactive applications that could not be verified record by record, so the article gives a verification method instead of a model list. Second, the 15 W to 30 W retrofit figures are arithmetic examples, not product specifications, because published input power for retrofit assemblies is frequently absent. Third, the ECE regulation number behind the LEDr category is not quoted, because the MOT manual names the category without citing the regulation and the regulation text itself could not be retrieved for this article. Where a figure came from a regional variant of a manufacturer page, the source list names that page.