LED effiency Watts=> light and Watts => heat

[The Dane]

So when i apply 3,75Volts and 1000mA to a Cree XR-E it euqates 3,75Watts of consumption from the source, BUT:

What amount of the 3,75Watts ends as light that exits the front of the light and how much ends as heat?

I tried to search these forums, and because I'm not the sharpest knife in the drawer, i ended up without an usefull answer :shrug:

 

[jtr1962]

A rough but useful guide for cool-white phosphor white LEDs is ~330 lumens per watt of emitted output. High-CRI LEDs will be somewhat less than this, perhaps in the high 200s. Anyway, if for example your LED consumes 3.75 watts at one amp, and emits 250 lumens, then light output will be 250/330, or 0.76 watts. Since heat = power input - light emitted, then this means the LED will give out 3.75 - 0.76, or 2.99 watts of heat.

Note that up until when the Cree XR-E came out a few years ago, most people didn't even bother taking the light output into consideration. Most LEDs were less 10% efficient so it didn't affect calculations by much. Now the best LEDs are 30% efficient or more. It's starting to get to the point where it makes sense to take this into account. It's also interesting to note what happens as efficiency rises further. Suppose we manage to make LEDs which are 80% efficient. That means 80% light, 20% heat. Now one might say it doesn't pay to try to get to 90% as that would only mean 12.5% more light for a given power input. However, note that this increase cuts your heat in half compared to 80%. And going from 90% to 95% cuts it in half again! Practically speaking, this means smaller heat sinks, or more light output for a given size heat sink.

 

[LukeA]

A rough but useful guide for cool-white phosphor white LEDs is ~330 lumens per watt of emitted output. High-CRI LEDs will be somewhat less than this, perhaps in the high 200s.

Source?

 

[jtr1962]

Source?

I read it in a technical paper from Nichia I think. Also, I remember seeing this info in some LED technology roadmap papers.

 

[LukeA]

I read it in a technical paper from Nichia I think. Also, I remember seeing this info in some LED technology roadmap papers.

You forgot to put the link in your post.

 

[jtr1962]

You forgot to put the link in your post.

Didn't forget-it was a while ago and I just don't remember exactly where I read it.

 

[LukeA]

Didn't forget-it was a while ago and I just don't remember exactly where I read it.

Then your prior assertions are false.

 

[carbine15]

Then your prior assertions are false.

This is not a conclusion you can logically draw from his statement. If you know otherwise, then inform us.

 

[jtr1962]

Then your prior assertions are false.

Why? Because I don't time to hunt through the Internet for hours for my source(s), assuming they even still exist? I read lots of things online. Some articles I save on my hard drive, others I don't. I generally don't bother saving links because more often than not they disappear within months. This 330 lm/W figure I alluded to has been common knowledge around here for quite some time. Do a search of CPF and maybe you'll find someone who already linked to a source if that's really so important to you. Or better yet stick a white LED in an integrating sphere where you'll get a nice output of the spectrum, and can use that to calculate the lumens per watt of emitted light. I bet it comes pretty close to the 330 lm/W figure I posted. In fact, someone around here a while back integrated the area under the spectrum of a Luxeon and ended up a figure in the neighborhood of 330 or 340 lm/W.

On another note, one of my pet peeves are people constantly hounding others for "sources". Sometimes an online source doesn't plain exist. Other times maybe the person is the source as they originated the knowledge in the first place. Besides all that, when I'm online I'm usually more concerned with actually learning things rather than building a directory of where this knowledge came from. I'm not a librarian.

 

[IMSabbel]

Then your prior assertions are false.

Well, you are not helping either.
He is right, btw. But i wont give YOU the link, pal.

 

[monkeyboy]

What amount of the 3,75Watts ends as light that exits the front of the light and how much ends as heat?

I tried to search these forums, and because I'm not the sharpest knife in the drawer, i ended up without an usefull answer :shrug:

The Dane, this is a common source of confusion and certainly not a stupid question. The answer is that it doesn't make sense to quote a percentage efficiency for a light source.

You can quote a percentage efficiency for radiated power to power consumption but this has no relation to what the human eye sees and also includes radiated heat; infrared radiation (as opposed to heat conducted from the LED).

The lumens per watt rating is the best way of quoting efficiency as the rating is weighted with the sensitivity of the human eye to a particular wavelength or spectrum of wavelengths. This means that the maximum possible lm/w is dependent upon the spectrum of the light as jtr1962 pointed out.

Some examples:

A pure IR source that was 100% efficient in terms of radiated power would have zero (0) lm/w and similarly for a pure UV source.

The human eye has a peak sensitivity at 555nm with a maximum possible 683 lm/w.

Wikipedia quotes percentage efficiency based on the max 683 lm/w value which is pointless and confusing IMO.

 

[SemiMan]

Nice thread...... first question was good and valid and was followed by a fairly accurate reply. Not sure what happened after that, but it was not very productive........ oops, I just did it....

 

[likeguymontag]

Nice thread...... first question was good and valid and was followed by a fairly accurate reply. Not sure what happened after that, but it was not very productive........ oops, I just did it....

Indeed. I found the question and the answer interesting and useful. Then this happened; dunno where the aggression came from. :thumbsdow

You forgot to put the link in your post.

Then your prior assertions are false.

 

[LukeA]

Indeed. I found the question and the answer interesting and useful. Then this happened; dunno where the aggression came from. :thumbsdow

If you can't separate intellectual rigor from aggression, then I know what list you belong on. :thumbsup:

 

[Marduke]

Source?

You forgot to put the link in your post.

Didn't forget-it was a while ago and I just don't remember exactly where I read it.

Then your prior assertions are false.

Just because a fact is not referenced explicitly does not make it false.

If you can't separate intellectual rigor from aggression, then I know what list you belong on. :thumbsup:

There is a fine line between intellectual rigor and being a *******. I'll let the reader decide which side the above posts are.

BTW, typically the burden of proof falls on the accuser, so why don't YOU post a link citing your source for doubt. :thumbsup:

PS, I have seen a paper as described above, so I do know the statement is TRUE. For those who are not simply lazy, 5 minutes of Googling will give you at least half a dozen such papers.

 

[likeguymontag]

There is a fine line between intellectual rigor and being a *******. I'll let the reader decide which side the above posts are.

I respectfully disagree. I believe that the two are totally independent of each other. That won't stop me from quoting from The Big Lebowski,

Walter Sobchak: Am I wrong?
The Dude: No you're not wrong.
Walter Sobchak: Am I wrong?
The Dude: You're not wrong Walter. You're just an asshole.
Walter Sobchak: Okay then.

BTW, typically the burden of proof falls on the accuser, so why don't YOU post a link citing your source for doubt.

In science, at least when we're talking about new information, the burden of proof falls on the person presenting the new information. Clearly that standard can't apply when the information is (or could be) common knowledge.

Aah, well. I probably shouldn't have revisited this thread anyway.

 

[Marduke]

I respectfully disagree. I believe that the two are totally independent of each other.

True, I guess you can also be both.

In science, at least when we're talking about new information, the burden of proof falls on the person presenting the new information. Clearly that standard can't apply when the information is (or could be) common knowledge.

In the LED world, this is very OLD information, and is in no way "disputed". Common knowledge indeed. Saying it's false is basically calling basic physics "false", such as if someone writes a basic equation (think V=IR) without also providing the proof for that equation.

 

[likeguymontag]

Saying it's false is basically calling basic physics "false", such as if someone writes a basic equation (think V=IR) without also providing the proof for that equation.

I feel a Wikipedia "citation needed" joke coming on.

 

[BillyNoMates]

OK. it appears to me to be useful to put a stake in the ground. By typing 'led lumens per watt' into a well known search engine, I found this press release close to the top of the list...

http://www.cree.com/press/press_detail.asp?i=1227101620851

This says 161lm/W it is dated as Nov 2008, so it is reasonably recent stuff.

I also found a reference in the luxeon Rebel data sheet stating 275mW of radiometric power for the royal blue led at 350mA drive. Now assuming that 350mA and a 'typical' voltage of 3.15V (value from data sheet http://www.philipslumileds.com/pdfs/DS56.pdf), the electrical input power is 1.1W. Efficiency of this LED is close to 25% (better than I thought it would be) therefore 75% (825mW) goes as heat.

#RANT
Now, that took about 5 mins and I hope I didn't annoy anyone in the process..... ..... I only hope it is helpful, or even better stimulates some interesting discussion
#END RANT

 

[The Dane]

Thanks alot guys.
Many fine comments.