Why do LEDs generate heat? (2025 Update)

Simply because of the voltage drop across the junction. The blue LED that allows white light is the worst dropping over 3 volts. Watts (heat) is equal to Volts x Amps. The actual voltage drop will depend on how the internal LEDs are wired but it is generally over 4. Think of your cheap LED flashlights, they connect the LED directly to

It doesn't have to be heat though. Some or all of the energy could be changed into light.

That would break the LEDs. There has to be a current limiter, even if it's just a resistor.

Not a cheap flashlight, the battery is the current limiter.

Those LEDS are wired directly to the card that connects directly to 3 AAA cells.

The "light" still ends up being dispersed as heat eventually but it does get radiated away from the LED. It is also a pretty small part of the total energy consumed.

Those must be tough LEDs. Most LEDs are designed to take say 20-30mA. A battery is capable of way way more than that. Even the cheapest "flashlight" (I think you mean torch, they don't flash....) I've seen has a resistor.

Must be a european thing. I have never seen one with a ballast resistor and I have at least a half dozen different types from little key chain lights with 4 button cells in them to an assortment of single LED and multi LED 3 AAA cell torches like the one in the picture. Next time a get a few minutes I will see what the operating current of these various lights is.

I can't believe that can work. An LED requires a constant current. A battery will not do that, it will overload the LED very quickly, unless you happen to have a battery very very close to the required LED voltage. Try connecting an LED to a bench supply set to a constant voltage. It may run fine at 3.2V. Now give it 3.3V and it'll draw double the current and die within seconds.

Millions are doing it every day so I think your predictions of imminent death are inaccurate. I showed you a picture of the circuit card of one of the most popular designs (9 LEDS) but I can show you some different designs if you like or you can take one of yours apart and show me a picture of the resistor.

and it'll draw double the current and die within seconds.

I took my USA sold by Walmart made in China 9 LED 3 AAA cell light apart. On the circular plate where the LEDs are soldered in there is a SMD resistor labled 5R1. I believe that would be a 50 ohm resistor. I also doubt that the battery powered lights are operated without some device to control or limit the current to the LEDs. There are others that use an electronic control that vry the brightness and will pulse off and on.

The couple of 60 watt equal 120 volt bulbs I have opened up have an electronic control of some kind. They do not use a ballast resistor. The bulbs are plastic instead of glass. Most of the heat is from the LEDs. I don't recall the number,but there are about a dozen small leds that glow very bright.

I'm not going to dismantle my expensive CREE cycle lamp, but my father's cheap torch (from Tesco - a UK supermarket) which broke and I repaired, had a resistor. It's a basic fact that LEDs require a constant current, and that if you apply a very slightly larger voltage, you can end up doubling the current, as an LED tends to stay at a specific voltage, no matter how much current goes through it - the voltage doesn't need to be doubled to double the current like a resistor. Just try it with a bench supply. Connect it to an LED and monitor the voltage and current. You'll get something like:

It is an interesting question, to which I don't know the answer.

Run some current through a copper wire and it will get hot. Why does that happen? I don't know. The individual atoms are moving faster, and that is caused by the loose electrons moving slowly through the wire, but the exact mechanism is lost in past memories of physics classes.

We hardly ever get that deep into the details. For all practical purposes, knowing iR^2 is enough.

How does a mirror catch light and throw it back at you? Same idea.

The first electrical formula I learned, P=IE.

The voltage drop across a LED (while it is lit) will not exceed the LED forward voltage. If it did, the LED would destroy itself trying to conduct infinite current.

The one I examined had a couple of thin lithium cells (about 3v each) between the LED leads. A bit of springy material made a "switch".

These cheap lights use a small battery with sufficient internal resistance to drop the voltage. If you measure the battery voltage while the LED it lit, it will equal the LED forward voltage. A bigger battery would require a resistor.

Using a resistor to operate a single LED on 120v (dropping around 117v) would waste a lot of power. You could improve that by putting several LEDs in series (as is done in Holiday lights, most I've seen have 25-35 in series). Also, a current regulator may be used.

There will always be some heat. No energy conversion is 100% efficient, and the waste will be heat.

But not always a real resistor, as in my previous post.

BTW, I once had a LED break when I accidentally connected it directly to a power supply. A POP and half the LED (plastic package) was gone.

With a typical LED voltage drop of 2-3V, so the power is 40-90mW. I've heard of 1W LEDs.

The resistorless light I saw used little coin cells, not AA or AAA cells.

Surely P=IV? E is energy, as in E=mc^2.

I've never known a torch use anything less than an AA battery, which can produce 20 amps easily with very little voltage drop. This would kill an LED rapidly.

Cheap ones use a capacitor to drop most of the voltage (as does the thing I rigged up to power LEDs off the mains to indicate which zone my central heating is powering). Expensive ones use a SMPSU.

Why can there not be 100% efficiency? Even if not, we should be able to improve on the 40% we have now.

Yes I have also blown them up. I got something more dramatic when I connected mains across a power transistor intended to switch some lighting on a disco light. I was repairing a dud solder joint with the soldering iron while the circuit was live, not realising that the soldering iron was earthed. There was a very very loud pop and half the transistor was found at the other side of the room. Another reason earths are bad! I've de-earthed my microwave so when my pet parrot stands on it and chews an electrical cord, it can't get a shock through its body. Although now when the microwave is running and I touch it and the sink, I feel a light tingle, so there is some current which must have previously gone to earth. That reminds me - where I worked 20 years ago, a microwave oven tripped the earth leakage circuit breaker for half the building, that was most amusing. My friend had his house breaker keep tripping with one aswell. I guess microwaves leak current, and/or ELCBs are far too bloody sensitive.

It's too variable for this to be practical. You need a battery that can't supply more current than is safe for the LED.

Not double, the limit of the supply as it will try to pull that 3.3V down to it's threshold voltage. The time I tried this, it died so fast I never saw any light.

Yeah I had some of those in a CREE (s**te) bulb. 4 of them trying to make 50W equivalent in a spotlight. It got too hot to touch and lasted 6 months. The lights I use now have about 40 LEDs consuming 9W, so a quarter of a watt each. They're just under 4 volts (70V across all 40, in pairs), so 64mA.

The threshold voltage isn't that constant. You can get it a fair bit up, probably as much as half a volt, before the thing overloads. And if you turn it up slowly, you'll see the LED get brighter and brighter, then it'll start to get dim and just stop working. An explosion only happens if you crank it up fast.