In searching for the correct replacement LED for a project, I've noticed that, more often than not, high power LEDs seem to be primarily categorized by their wattage. Why is this the standard?

It would seem that voltage and amperage would be more appropriate for classification and searching, since multiple volt * amps combinations can result in the same wattage, but obviously LEDs of the same wattage are not necessarily interchangeable between circuits.

I find myself endlessly diving into specs after finding an appropriate wattage LED, just to determine that the voltage & amperage are not compatible for my circuit. But then again I'm new to this. So just curious.

  • 5
    \$\begingroup\$ Post some datasheets of these LED's with crazy voltage you speak of. LED's are devices that operate between 2-4 volts, virtually no exceptions. If you are finding things wildly different, they are ARRAYS of LED's, not a single LED die. \$\endgroup\$
    – Kyle B
    Jun 10, 2021 at 6:38
  • 1
    \$\begingroup\$ @KyleB Especially for residential lighting, packages with more than one diode junction in series are extremely common, so you can get white "LEDs" with 1x3v, 3x3v and 6x3v, etc. Reason for this is that in a cost optimized bulb with no transformer, the ~10-15 packages have to sum close to rectified lines voltage, so 9v in 120v countries, double that elsewhere. \$\endgroup\$ Jun 10, 2021 at 14:04
  • \$\begingroup\$ @user1850479 Haha - Yes, I know that. I was trying to get the OP to put up one of his datasheets so we could help explain to him how to read it. ;) \$\endgroup\$
    – Kyle B
    Jun 10, 2021 at 15:16
  • \$\begingroup\$ Thanks. We’re trying to replace the LED array in an exterior fixture. There are 10 fixtures, 5 of which have burned out, and it’s likely more to go. Wanted to replace the LEDs (Build a replacement platter or use a different LED type altogether) in some fashion rather than buy replacement fixtures, which would be more expensive. Metering a working LED platter reveals there are 2 parallel paths in the circuit, each containing 3 series LEDs, each with a Vdrop of ~6V. Power supply is 14-20V 350mA. Searching for similar 6V LEDs led me down this path. I’ll try to re-find some of the odd LEDs & post. \$\endgroup\$
    – b. insler
    Jun 10, 2021 at 19:12
  • \$\begingroup\$ Those 6 volt LEDs are very likely two LEDs in series. A 6 volt LED would be emitting light well into the far ultraviolet, which this is definitely not. \$\endgroup\$
    – Hearth
    Jun 11, 2021 at 1:04

5 Answers 5


I somewhat disagree with the other answers. The power class of the diode is actually important, since low power diodes can be used with cheap FR4 PCB, higher power can be used on metal core PCBs, and very high power COBs typically are directly bounded to metal heatsinks. Since you usually decide (based on budget and manufacturing capabilities) what type of PCB and thermal solution will be used before selecting specific parts, the power classes are logical.

That said, unless you're looking at Amazon/eBay (where the specs are made up anyway), most vendors list by voltage, nominal current and power. Digikey for example will let you pick all the 9v white LEDs with a certain nominal current.


Impossible to answer conclusively, but a good reason is that a lighting designer will select an "LED" (or array/cluster of LED's) that provides the illumination required, THEN design a circuit to drive it. You're kinda doing it backwards.

  • \$\begingroup\$ Interesting. So in the case of repairing a lighting fixture rather than designing one (aka factory LEDs burned out but we want to keep the fixture), you essentially just have to dig deeper into all the specs to find a working match, or design a new circuit within the fixture for your replacement? \$\endgroup\$
    – b. insler
    Jun 10, 2021 at 8:16
  • 4
    \$\begingroup\$ Sometimes there are LED substitutions available for existing fixtures. Usually the path of least resistance would be to continue to use the original bulbs (fluorescent or whatever). In a factory, probably they'd replace the whole fixture outright - Time is money. They can spend hours & hours reverse engineering something, or just have a new fixture purchased and installed in about an hour. Its likely gonna be WAY cheaper to the company to replace the fixture. \$\endgroup\$
    – Kyle B
    Jun 10, 2021 at 15:18
  • \$\begingroup\$ Sometimes it's optimal to buy an identical fixture and replace the failed parts in situ, as that can be easier than removing +replacing the fixture. Having replaced the failed driver/power supply twice in one fixture, getting compatible parts is doable but a pain - and it's usually the drive circuit that fails rather than the actual LEDs \$\endgroup\$
    – Chris H
    Jun 11, 2021 at 14:58

Yes, they are bulbs after all: you buy bulbs by wattage.

The light intensity is more or less proportional to power, so that for some products you see the real wattage and the higher one "as if" (compared to e.g. incandescent bulbs). An old incandescent bulb had about 50W and 500-600 lumens (measurement of luminous flux). Similarly this is obtained with a 5W LED more or less.

Then, as @kyle-b says, you will design (or purchase) the driving circuit to regulate them. It is a good approach to feed some in series, because you come up with more manageable voltages, and you feed them with the same current.

Originally LEDs had different threshold voltages depending on color; but high-power LEDs for lighting try to use a efficient light emissions and then phosphors/filter to adjust the emission to the desired type of color and color temperature. For lighting you may desire a white (bluish) LED, so the voltage should be around 3.5-3.6 V. The yellow ones (e.g. for large displays on a highway or parking lot) are still 2.3, maybe 2.5V.

Add-on: regarding phosphors, you may see other question Why are most COB LEDs physically yellow?


LEDs are rated by wattage because that's a clear statement of how much power they can dissipate without damage. In general, the larger the LED physical size, the more power it can handle. Larger size means the LED can shed more heat.

LEDs are also rated in:

  • light output vs. forward current
  • spectral output
  • continuous current
  • transient current
  • forward voltage (Vf) vs. current

For example, this 75mW, 30 mA rated LED that can withstand 180mA at 0.1ms pulse and 1/10 duty cycle.

Forward voltage (Vf) depends on the LED material and its emitted color. Shorter wavelengths will have higher Vf than longer wavelengths.

Forward voltage (Vf) shifts upward somewhat with increased current. This results in a higher Vf drop, increasing the power dissipation in the LED even more.

The design flow is choose the LED (size, color, light output), then design a driver that sets the right current for it, and add a cooling system if needed. Forward voltage becomes a constraint if Vf is higher than can be reliably supplied (for example, a 3V LED with a 3.3V supply is not a good idea - not enough margin.)


more often than not, high power LEDs seem to be primarily categorized by their wattage. Why is this the standard?

  1. Institutionalised stupidity.

    People keep doing stupid things because their predecessors did stupid things and they expect other people to be confused if they stop doing stupid things.

    It dates from the days when light was produced as an inefficient tiny side effect of heating tungsten filaments almost to their melting point. At that time the amount of power consumed or heat produced was a rough analogue of the amount of light produced because there were few alternative technologies in general contemporary use. The general public became familiar with the amount of light expected from a 15, 40, 60 or 100 watt incandescent tungsten-filament light-bulb. Maybe the majority of people encountered and understood watts more than candela or lumens.

  2. To confuse their customers.

    See "music power". Manufacturers of inefficient products like to hide that inefficiency. Thus they promote watts as if it were a virtue and hide luminous flux (lumens), lumens per watt, illuminance and other more useful measures.

Don't get me wrong, power consumption is an important piece of information - but it shouldn't be the primary characteristic of a light source.


It has been suggested that, in the context of wiring up lights in a house, watts are the most important characteristic. This is not so.

Many years ago, when I moved into my new house, the electrician had provided light bulbs in every socket. Since these were mostly 60 watt incandescents and not not 15 watt fridge bulbs I can tell that power consumption was not the primary selection criteria.

Like almost all UK houses, mine has a 6A lighting circuit for all lights on the ground floor. That is standard. The electrician who wired up that circuit followed the general standard and gave no thought at all to the wattage of bulbs that might be used. Nor should they.

That circuit has 10 light sockets, the kitchen has additional under-cabinet lighting but I'm pretty sure those come off a separate circuit. But lets include them anyway and pretend I have 15 sockets where the householder can insert a light-bulb.

If I go to a supermarket today, the most powerful light bulb I can buy is 12 watts. I would need to fit 70 of those to my 15 light-sockets to approach the rating of the circuit. Clearly the power rating of typical domestic light-bulbs is irrelevant.

Electricians are in fact much more concerned with amps than watts. Indeed bulbs now often have both marked on them. UK klighting circuits are 230 volts nominal (EU standard) though usually 240 volts in practice (old UK standard). I looked at one light bulb which was marked 12 watts, 85 mA. If you do the sums you'll see the electrician can't use the watts to check load on circuit accurately. Clearly there are complicating factors in the driver - power factors, inrush current for a capacitor or some other reason. The electrician should disregard the power rating in watts and use the current rating in amps.

As I said, when I moved into my new home, there were light fittings and light-bulbs already in place. There were no lamp-shades - these were regarded as a purely decorative item that the new owner should select themselves. But it is the lampshade's ability to handle heat that is the only time you need to refer to power consumption of the bulb. Lamp shades are marked with a power rating (e.g. "40W max").

However currently it is not possible to buy, from my nearest supermarket, a light-bulb that will approach even 40 watts. So this is, for me at least, also a non-issue nowadays.

As a final piece of evidence of this stupidity I offer the following

enter image description here enter image description here

This is one of the least egregious examples, many such items omitted the actual power rating or had it in tiny print on the other side.

I hope I do not have to explain the gross stupidity of what is shown above.

In summary

  1. When selecting a device whose sole purpose is to produce light, its most important selection criteria is, in fact, the quality and quantity of light it produces.

  2. In a domestic context, meeting a current budget is a secondary consideration and in most cases totally inconsequential. Most lighting circuits are grossly over-specified for todays high-efficiency lighting.

  3. When choosing a light source for a fitting where heat dissipation is an issue, power ratings are important. This is almost never an issue in a household context with LEDs in fittings designed for incandescents. But in any case it is third on my list, not first.

  • 3
    \$\begingroup\$ I would strongly argue that providing the primary metric of value to someone actually responsible for wiring up all the lights in a house is not by any means stupid, especially considering that there are many safety considerations that hinge directly on that value. And, of course, most people don’t care how much light a light source provides provided it’s ‘enough’, and even if they did they would need to care about the wattage for the aforementioned safety reasons. \$\endgroup\$ Jun 10, 2021 at 17:58
  • \$\begingroup\$ But you can calculate wattage (and cost) from voltage and amperage, but not the reverse. So function is being obscured by the need to present cost instead. Seems... well, exactly the same way everything else in the world works actually... \$\endgroup\$
    – b. insler
    Jun 10, 2021 at 21:15
  • \$\begingroup\$ @mishan do you mean that voltage is a given from the utility? My question was more within the scope of circuits driving LEDs that are being transformed down through a power supply to lower DC voltages. \$\endgroup\$
    – b. insler
    Jun 10, 2021 at 21:25
  • \$\begingroup\$ @Austin, thanks for the comment. I disagree. At least in the context of me buying light bulbs for my home. YMMV. Answer updated to clarify. \$\endgroup\$ Jun 11, 2021 at 9:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.