LED Bulbs, Really

An LED light bulb is two separate engineering problems stacked in one glass-and-plastic shell, and almost every consumer complaint about them — flicker, dimmer incompatibility, early failure, harsh color — traces back to only one of those two problems, and it’s not the diode. The semiconductor die that actually emits light is close to the most reliable component in the entire assembly, rated for tens of thousands of hours of operation with genuinely gradual, well-characterized decline. The part that breaks, flickers, buzzes, or dies outright is almost always the driver: the small circuit board tucked into the base that converts your wall’s 120-volt (or 230-volt) alternating current into the low-voltage, tightly regulated direct current an LED die actually needs to survive. Understanding an LED bulb means understanding those two systems separately, because they’re solved by completely different engineering disciplines and they fail for completely different reasons.

How a Blue Diode Becomes White Light

A light-emitting diode is a semiconductor junction that emits photons when electrons recombine with electron holes across it, and the wavelength of that light is set almost entirely by the material’s bandgap energy — a fixed property of the semiconductor, not something you can dial to an arbitrary color after the fact. The overwhelming majority of practical white LED bulbs don’t attempt to combine separate red, green, and blue emitters to make white directly; instead, they take a single, highly efficient blue LED die (typically an indium gallium nitride, InGaN, junction) and coat it directly with a layer of phosphor.

Phosphor is a photoluminescent material: it absorbs high-energy blue photons and re-emits them at a lower energy — a longer wavelength — a process called down-conversion, governed by the same physics (the Stokes shift) that makes highlighter ink glow under a blacklight. The standard “dichromatic” approach coats the blue die with a yellow-emitting phosphor, commonly a cerium-doped yttrium aluminum garnet (YAG:Ce). Some of the blue light escapes the phosphor layer untouched; the rest gets absorbed and re-emitted as yellow. The eye combines the leftover blue with the new yellow and perceives the mixture as white — which is why an LED bulb with its phosphor coating scratched or chipped often shows a distinct blue-white halo where the raw diode emission escapes without being converted.

                    ┌─────────────────────────┐
                    │   Phosphor layer (YAG:Ce) │
                    │   absorbs blue, emits     │
                    │   yellow (down-conversion) │
                    └───────────┬───────────────┘
                                │
        blue photons ──────────┤────► some pass through unconverted
     (InGaN LED die,           │
      ~450nm peak)             ▼
                     eye perceives blend of
                     leftover blue + phosphor
                     yellow as "white" light

The exact color temperature and color rendering quality of the resulting white depend on phosphor chemistry and thickness. Warmer, more “incandescent-like” bulbs (2700K–3000K) use a thicker phosphor layer and often blend in a secondary red-emitting phosphor to fill in the red end of the spectrum that a simple blue-plus-yellow mix leaves weak — which is directly why warm-white LEDs measurably lose luminous efficacy relative to cooler, bluer LEDs of the same die: thicker phosphor and more red-phosphor content means more photons get absorbed and lost as heat during conversion rather than escaping as usable light. Public research puts phosphor-converted warm-white output at roughly 80 to 85 percent of the raw photon output of the same die’s uncoated blue emission, and pushing color rendering index (CRI) up toward 95+ — closer to how sunlight renders color — typically costs another 15 to 20 percent of luminous efficacy on top of that, because a high-CRI spectrum requires broader, more even coverage across the visible range than a simple two-phosphor blend delivers cheaply. This is a real, physical trade-off baked into the chemistry, not a marketing artifact: a bulb boasting both extremely high CRI and extremely high lumens-per-watt on the same package is claiming two numbers that pull against each other, and one of them is being optimistically rounded.

The Driver: Where AC Becomes a Number the Diode Can Survive

An LED die is a current-driven device, not a voltage-driven one — it doesn’t behave like a resistor with a predictable voltage-to-current relationship; instead, once you cross its forward voltage threshold, current through it rises extremely steeply for small changes in applied voltage, and too much current burns the die out almost immediately through thermal runaway. That means an LED can never be wired directly to household AC. Something in between has to convert the wall’s 60Hz (or 50Hz) sine wave into a stable, tightly regulated DC current — and that something is the driver, a small switch-mode power supply crammed into the bulb’s base or heatsink assembly.

A typical driver chain looks like this: a bridge rectifier converts the incoming AC sine wave into pulsating DC, a filter capacitor (often, on cheaper bulbs, a small electrolytic capacitor) smooths that pulsing into something closer to steady DC, and a constant-current regulator — built around a dedicated driver IC — then meters out a fixed peak current to the LED array regardless of small fluctuations in the input voltage. On dimmable bulbs, an additional stage has to detect and interpret the incoming dimmer signal and translate it into a proportionally reduced current level, which is where most of the real-world dimmer-compatibility complaints originate.

1
2
3
4
5
6
7
8
9

# Simplified conceptual model of a constant-current LED driver stage
# (illustrative control logic, not a real driver IC's firmware)
def regulate_led_current(rectified_dc_voltage, target_current_ma=350):
    sense_resistor_ohms = 1.0
    measured_current = read_current_sense(sense_resistor_ohms)
    error = target_current_ma - measured_current
    duty_cycle = clamp(duty_cycle_from_pid(error), 0.0, 1.0)
    set_switching_mosfet_duty(duty_cycle)   # buck converter switch
    return duty_cycle

Why Dimmers Are the Compatibility Nightmare

Household dimmers built for incandescent bulbs almost universally use TRIAC-based phase-cut dimming: rather than lowering the voltage smoothly, the dimmer chops out a portion of each AC half-cycle, switching on only after some delay into the cycle (leading-edge cutting) and letting the rest of the wave through. An incandescent filament, being a simple resistive load with thermal mass, doesn’t care about this — it just runs a bit dimmer and cooler on the reduced average power, smoothly and predictably.

An LED driver is not a resistive load. It’s a switch-mode power supply with rectifiers, filter capacitors, and often power-factor-correction circuitry, and a chopped, discontinuous AC waveform confuses that circuitry in several specific ways. When the TRIAC first switches on partway through the cycle, the driver’s input filter capacitor draws a sharp inrush current spike to recharge; if that inrush current briefly dips below the TRIAC’s minimum “holding current,” the TRIAC can shut back off prematurely before the half-cycle finishes, then re-trigger, producing a visible flicker or an audible buzz at the dimmer or bulb. At low dimmer settings, where the chopped waveform delivers only brief slivers of each cycle, this misfiring becomes much more common, which is exactly why cheap dimmable LEDs are usually fine at 80% brightness and visibly flickery or unstable below 20%.

The fix that actually works is largely on the dimmer side: trailing-edge (reverse-phase, sometimes marketed as ELV-style) dimmers cut the end of each half-cycle instead of the beginning, which is inherently gentler on a capacitive, switch-mode load like an LED driver and produces dramatically less buzz and flicker than a cheap leading-edge TRIAC dimmer built decades ago for incandescent loads. On the bulb side, better drivers add dedicated EMI filtering and a minimum-load “bleeder” circuit specifically to keep the TRIAC’s holding current satisfied even at low dim levels — a feature cheap drivers routinely omit to save a few cents in bill-of-materials cost.

Why Cheap Bulbs Die Early — and It’s Rarely the Diode

The LED die itself doesn’t fail the way an incandescent filament does — it doesn’t snap. It degrades gradually through a well-characterized process called lumen depreciation: the light output slowly declines over tens of thousands of hours as the semiconductor and phosphor age. Manufacturers rate this with an L70 figure — the number of operating hours before output falls to 70% of its original brightness, the threshold at which a typical eye first perceives the dimming as significant. Quality LED bulbs commonly carry L70 ratings in the 25,000 to 50,000-hour range.

The catch is that this rating assumes the LED die is held within its designed thermal envelope, and junction temperature is brutally punishing to that assumption: LED lifespan roughly halves for every 10°C increase in junction temperature, so a die running at 85°C might see an effective L70 lifespan of around 25,000 hours where the same die at 65°C would see something closer to 50,000. Manufacturers who skimp on the heatsink — the metal or ceramic structure that pulls heat away from the die and driver and dumps it into the surrounding air — are trading a few cents of aluminum for a real, measurable cut in service life, and it’s an invisible trade-off at the point of sale because a bulb’s thermal design doesn’t show up on the packaging.

But in practice, the component that fails first in most early-death complaints isn’t the die at all — it’s the driver, and specifically the electrolytic capacitor in the driver’s rectifier/filter stage. Electrolytic capacitors fail primarily through electrolyte evaporation, a process that accelerates sharply with heat — and an LED bulb’s base, packed with a driver board right next to a heat-generating diode array in a sealed plastic or glass enclosure with limited airflow, is a genuinely hostile thermal environment for a component whose failure mode is heat-driven evaporation. Field data on LED fixture failures consistently shows driver degradation, not LED lumen depreciation, as the dominant real-world failure cause — a large share of reported outdoor and commercial luminaire failures trace back to driver electronics rather than the diode array. Cheap bulbs compound this by using smaller, lower-temperature-rated electrolytic capacitors to hit a price point, placing a component with a comparatively short thermal-stress lifespan directly next to the hottest part of the assembly.

Honest Trade-offs

• Driving LEDs harder for more initial lumens shortens their life. Manufacturers can push more current through a die to squeeze out higher initial lumens-per-dollar, but this raises junction temperature and directly cuts into L70 lifespan — a bulb that looks brighter on the shelf may be quietly trading away years of service life.
• High CRI and high efficacy are in real tension. A bulb advertising both a very high color rendering index and very high lumens-per-watt is describing two numbers that don’t naturally coexist in phosphor chemistry; one of them is likely measured under favorable conditions the other doesn’t share.
• Dimmer compatibility is a two-sided problem the consumer usually can’t diagnose. A flickering “dimmable” LED bulb might be a defective bulb, an incompatible legacy leading-edge dimmer, or — most often — both, since cheap dimmers and cheap bulbs are disproportionately likely to be paired together and neither manufacturer tested against the other’s product.
• The recyclability and repairability story is genuinely worse than the “20-year bulb” marketing implies. Because the actual failure point is usually a small driver board sealed inside a bulb never designed to be opened, a bulb that dies from one blown capacitor after two years is functionally unrepairable for a typical consumer and gets thrown away as e-waste, despite the diode array inside likely still having tens of thousands of hours of usable life left.

Verdict

The diode in an LED bulb is close to a solved engineering problem — a blue emitter and a phosphor coating doing well-understood, gradually degrading photoluminescence, rated for a service life most owners will genuinely never reach. Nearly everything that goes wrong with a real-world LED bulb — the flicker on a cheap dimmer, the early death after eighteen months, the slightly harsh color from a budget “daylight” bulb — is a driver-electronics problem or a phosphor-cost trade-off, not a failure of LED technology itself. Buying a good bulb means paying for a well-specified, adequately cooled driver with quality capacitors and, if it needs to dim, a driver actually designed to tolerate a chopped AC waveform — none of which is visible on the box, and all of which matters more than the lumens-per-watt number printed in large font on the front.

Sources

• CircuitDigest — Transformerless LED Driver Circuit for Reliable Low Cost LED Bulb Designs
• PMC — Active Thermal-Fuse for Stopping Blue Light Leakage of White LEDs Driven by Constant Current
• Tecolite — TRIAC Dimming for LED Bulbs: Compatibility & Risk Guide
• Monolithic Power Systems — AN080 TRIAC-Dimmable, Offline LED Controller
• LEDYi Lighting — Everything You Need to Know About TRIAC Dimming for LEDs
• AccessFixtures — Lumen Maintenance: The Inverse Relationship Between LED L70 Lifespan and Lumens Per Watt
• ScienceDirect — Modeling LED Driver Lifespan Through Capacitor Degradation Due to Thermal Cycling
• PMC — From Physics to Fixtures to Food: Current and Potential LED Efficacy
• Nature Scientific Reports — Design of Highly Efficient Phosphor-Converted White LEDs With CRI ≥ 95