Watt to Lumen conversion

stuartgmilton

Newly Enlightened
Joined
Nov 6, 2008
Messages
15
Hey guys,

I found a conversion that said 1 lumen = 0.0003Watts. Surely this is wrong?

You can buy a 3Watt flashlight, which has an output of 65L, so is there anyway to compare the 2 values?

Stuart
 

HKJ

Flashaholic
Joined
Mar 26, 2008
Messages
9,715
Location
Copenhagen, Denmark
Hey guys,

I found a conversion that said 1 lumen = 0.0003Watts. Surely this is wrong?

You can buy a 3Watt flashlight, which has an output of 65L, so is there anyway to compare the 2 values?

Stuart

That may be correct at some specific color, but:
Lumen is defined according to the eyes sensitivity, i.e. the watt<->lumen conversion is depend on the color of the light. One extreme is UV/IR light, your can have lots of watts in it, but the lumens are 0, because we can not see it.

And with flashlights (and anything else), your can not expect a 100% conversion, a lot of energy is wasted in heat.
Some typical values are:
incan: 5-30 lumen/watt
led: 60-100 lumen/watt

incan will have improved efficiency when really hot, i.e. very white light.
led will best efficiency at low output.
 

Marduke

Flashaholic
Joined
Jun 19, 2007
Messages
10,110
Location
Huntsville, AL
No, there is no conversion between lumens and watts. The "wattage" of a flashlight gives you absolutely no idea how bright it actually is.

When a manufacturer lists watts, you have no idea if they are stating the LED draw, the battery consumption, the max rating for the LED, or simply pulling a number out of thing air.


Furthermore, it depends on the LED efficiency. A current generation can produce over 100 lumens on 1 watt of power. Dial the time machine back a few years, and a previous generation might only do 40 lumens for 1 watt of power.
 
Last edited:

HKJ

Flashaholic
Joined
Mar 26, 2008
Messages
9,715
Location
Copenhagen, Denmark
Lumens is what matters. Watts is marketing.

Watt may be marketing in flashlight specifications, but there are conversions between watt and lumens.

One conversion is lumen output from bulb with a specific watt input. That depends on lots of factors.

Another one is how many watt a specific light "contains", that is a exact measure, but depends on the color of the light.
See here for a formula http://en.wikipedia.org/wiki/Luminosity_function
 

TranceAddict

Newly Enlightened
Joined
May 28, 2008
Messages
189
Location
Bristol, United Kingdom
you're talking about 1 apple equal to how many oranges.

watt is a derived unit consist of work done(energy) over time.

while lumen itself is the basic fundamental unit, not derived from any unit but exist as a standalone unit in nature. you can't break it down further or you will be another Einstein.
 

HKJ

Flashaholic
Joined
Mar 26, 2008
Messages
9,715
Location
Copenhagen, Denmark
you're talking about 1 apple equal to how many oranges.

watt is a derived unit consist of work done(energy) over time.

while lumen itself is the basic fundamental unit, not derived from any unit but exist as a standalone unit in nature. you can't break it down further or you will be another Einstein.

Not exactly correct.
You have "radiant flux" that is measured in watt, this is light and other radiation without any compensation for the eyes sensitivity.
If your know the spectrum of your "radiant flux", your can convert it to "Luminous flux", that is adjusted to the eyes sensitivity.
The absolute best conversion is with 555nm light, there your have 683 lumen/watt and the worst is anything outside the eyes range, this will have 0 lumen/watt. Anything in between is possible, depending on the spectral contents of the radiation.
 

TranceAddict

Newly Enlightened
Joined
May 28, 2008
Messages
189
Location
Bristol, United Kingdom
we are focusing on the visible electromagnetic spectrum, but no datasheet provides good info regarding within which spectrum it emits radiation and its strenght, while you're trying to include full spectrum of EM, consisting RF and even gamma ray. too complicated to derive power from flux within the range.
 

HKJ

Flashaholic
Joined
Mar 26, 2008
Messages
9,715
Location
Copenhagen, Denmark
we are focusing on the visible electromagnetic spectrum, but no datasheet provides good info regarding within which spectrum it emits radiation and its strenght, while you're trying to include full spectrum of EM, consisting RF and even gamma ray. too complicated to derive power from flux within the range.

It does not really matter if your include full spectrum or not, anything outside the visible spectrum is bvery easy to convert to lumen (it is 0 lumen).

Getting the spectrum of a led is not that difficult, in fact your can find the spectrum of many lamps on the ledmuseum, see here: http://ledmuseum.candlepower.us/seventee/itpop.htm

Anybody wishing to do the math, can get a rater good estimate of the optical power in watt, from the lumen value.
 

Marduke

Flashaholic
Joined
Jun 19, 2007
Messages
10,110
Location
Huntsville, AL
HKJ,

I think the OP is talking about the "on the package" wattage rating vs how bright a light is, not really going into the physics behind the energy involved.

ie. Application, not theory
 

TranceAddict

Newly Enlightened
Joined
May 28, 2008
Messages
189
Location
Bristol, United Kingdom
It does not really matter if your include full spectrum or not, anything outside the visible spectrum is bvery easy to convert to lumen (it is 0 lumen).

Getting the spectrum of a led is not that difficult, in fact your can find the spectrum of many lamps on the ledmuseum, see here: http://ledmuseum.candlepower.us/seventee/itpop.htm

Anybody wishing to do the math, can get a rater good estimate of the optical power in watt, from the lumen value.

just find the function of the line and do integration with given limit, or simply use approximation method like power series, not really hard in mathlab, but the problem is, different wavelenght(frequency) has different energy, we can't say given same intensity, energy of 10mhz is same as 10ghz.

now you have one more changing variable with respect to the frequency domain, how you gonna solve it in integration ?

more over you have to take the non-linear efficiency curve as well, as energy is wasted as heat form. another changing variable. unless manufacturer provide info of Vf and If on different brightness setting, if the flashlight uses MCU for infinite brightness level and PWM is deployed .....fainted....:sick2:
 
Last edited:

HKJ

Flashaholic
Joined
Mar 26, 2008
Messages
9,715
Location
Copenhagen, Denmark
just find the function of the line and do integration with given limit, or simply use approximation method like power series, not really hard in mathlab, but the problem is, different wavelenght(frequency) has different energy, we can't say given same intensity, energy of 10mhz is same as 10ghz.

I do not now waht your mean here, it is the formula that i linked to in my first entry?

now you have one more changing variable with respect to the frequency domain, how you gonna solve it in integration ?


I am not going to solve it, but splitting the light into discrete frequencies and use some weights for each frequency it not that hard.

more over you have to take the non-linear efficiency curve as well, as energy is wasted as heat form. another changing variable. unless manufacturer provide info of Vf and If on different brightness setting, if the flashlight uses MCU for infinite brightness level and PWM is deployed .....fainted....:sick2:

Now your are getting to calculating the efficiency of the led. Typical values of If and Vf and other led parameters can very often be found in the data sheet (Cree has curves for many different values).


But how much work your want do do, depends on what your want to know:

Input watt for output lumen: Just use the lumen/watt specification.
watt in light: your need spectrum and lumen.
 

MrGman

Flashlight Enthusiast
Joined
Feb 6, 2007
Messages
1,777
you cannot determine the lumens output of any LED source based on watts of power consumed as has been said before.

If you know exactly how many lumens you actually have, I believe there is a formula that tells you how many watts of power that is equivalent to but not how much is being drawn from the source to make it which is quite another thing.

You can't say because I have 3 watts of energy at the source, I therefore have X amount of lumens going out. There is absolutely no formula and no one has ever shown a way to do this.

Example: a 100 watt light bulb may give off 1100 lumens. That 1100 lumens of actual light output is not 100 watts of power, it may actually be the equivalent of 10 to 15, maybe not. A lot is lost as heat in the conversion. there is no direct mathematical formula to say 100 watts in is X lumens out. Lumens simply have to be measured by a system specifically designed and calibrated to do it. Not open to debate.
 

HKJ

Flashaholic
Joined
Mar 26, 2008
Messages
9,715
Location
Copenhagen, Denmark
you cannot determine the lumens output of any LED source based on watts of power consumed as has been said before.

Why not? Most led has a lumen/watt specification, for a Cree Q5 is is minimum 107 lumen/watt, then your have to do a derating, depending on temperature and current (All these numbers are in the datasheet).

But if your want total efficiency, from battery to torch lumens, your have a problem, because your are missing some specifications (DC converter, optical system).


You can't say because I have use 3 watts of energy at from the source, I therefore have X amount of lumens going out. There is absolutely no formula and no one has ever shown a way to do this.
(Note: the red words are the original text, I assume that he means the green words)

This again depends on what specification your have

Example: a 100 watt light bulb may give off 1100 lumens. That 1100 lumens of actual light output is not 100 watts of power, it may actually be the equivalent of 10 to 15, maybe not. A lot is lost as heat in the conversion. there is no direct mathematical formula to say 100 watts in is X lumens out. Lumens simply have to be measured by a system specifically designed and calibrated to do it. Not open to debate.

It is is a standard 120 volt 100 watt light bulb it will give about 1700 lumen, and it will be about 2.5 watt radiation energy.
Many bulb types has specifications, that allows estimation of lumens and radiation energy.
 

Marduke

Flashaholic
Joined
Jun 19, 2007
Messages
10,110
Location
Huntsville, AL
HKJ, you're missing the point. If a flashlight says it's "X watts", there is no way for you to convert that to lumens.

You don't know where they are measuring (if they are measuring at all), you don't know the specs of the LED and driver specifically, and you don't know the thermodynamic properties of the specific light in question. Presumably you don't have the flashlight in hand, so you cannot even take these measurements if you wanted to. You would also have to fully understand the optic system used to derate those effects also.

And assuming you COULD do all of the above, you would still end up with an approximation of an estimation, which may or may not be even close to the real value, as can be measured in an integrating sphere.
 

HKJ

Flashaholic
Joined
Mar 26, 2008
Messages
9,715
Location
Copenhagen, Denmark
HKJ, you're missing the point. If a flashlight says it's "X watts", there is no way for you to convert that to lumens.

You don't know where they are measuring (if they are measuring at all), you don't know the specs of the LED and driver specifically, and you don't know the thermodynamic properties of the specific light in question. Presumably you don't have the flashlight in hand, so you cannot even take these measurements if you wanted to. You would also have to fully understand the optic system used to derate those effects also.

And assuming you COULD do all of the above, you would still end up with an approximation of an estimation, which may or may not be even close to the real value, as can be measured in an integrating sphere.

That is correct and I have not said otherwise.

And your last point about estimations are very valid, an estimation might not be close to the real value, but can show bogus specifications.
 
Top