Torch lumen vs emitter

i still confuse the differences about torch lumen versus emitter. can somebody here explain to me about this??? thank you

When light hits something, some of it is absorbed, some of it bounces off in a direction you want, and some of it bounces off in a direction you don't want. Emitter lumens refers to the potential lumen output of the actual emitter, be it LED, hot filaments, various gases, etc.. Torch lumens refers to what actually comes out of the whole system.

So, torch lumens usually has to deal with:
1. Losses at the reflector.
2. Losses through the lens.
3. Losses at the bezel.
4. Actual output of the emitter, in the case of LEDs.

Woah, what's number 4? LEDs are usually specified at room temperature, or some other way that means the rated spec will not be reality. But that spec is used anyway.

If you catch all the light coming out of the bulb/LED that is emitter lumens.

If you subtract out :

- light which goes out the hole in the back of the reflector
- light which is used to heat up the reflector (ie it doesn't bounce off)
- light which is used to heat up the front lens (ie it doesn't pass through)
- light which bounces back to the emitter and heats it up more
- a couple of things I probably forgot

... you get torch lumens. You wouldn't expect them to be so far apart but apparently the "torch = 2/3 of emitter" rule of thumb isn't far off. Obviously the ratio for a given light varies depending on the details of the reflector, lens, bulb/LED etc...

Determining emitter lumens is easy with LED lights -- measure the current going into the emitter, look at the LED spec sheet, find the spec'ed output, done. Determining torch lumens is hard - you need specialized measuring equipment and even then there is debate about the accuracy of one integrating sphere vs another, particularly with very tight "throwy" beams.

To further complicate things, some companies (most notably Surefire) try to spec "real world torch lumens", so they use partially discharged cells like a real user would have rather than spec'ing only what you get in the first few seconds of use with brand new cells. I think the "2/3" rule of thumb may actually be for "emitter lumens to Surefire lumens" but not sure.

bridgman said:

To further complicate things, some companies (most notably Surefire) try to spec "real world torch lumens", so they use partially discharged cells like a real user would have rather than spec'ing only what you get in the first few seconds of use with brand new cells. I think the "2/3" rule of thumb may actually be for "emitter lumens to Surefire lumens" but not sure.

Click to expand...

Do they?
For some light the specifications say "Max output" and runtime is measured until the light is down to a few lumen output.

But at least they do deliver, at least, the specified lumen out the front, with fresh batteries.


To see some real torch lumen measurements see here:
http://candlepowerforums.com/vb/showthread.php?t=211402

For the Fenix light it is easy to compare to the data sheet specification and see how much loss the optic system has.

L2D turbo: specification 180, measured 145, i.e. 80% percent of the bulb light comes out the front of the flashlight.

I'm not 100% sure about Surefire using partially depleted cells, but (a) I hear it a lot here from knowledgeable people, and (b) that lines up with the "stupid Surefire always under-specs their lumen output" comments.

When I see measurements of Surefire lights with fresh cells the measurement is frequently well above the rated spec, but not sure if that is just because the Surefire specs are generally conservative to cover manufacturing variations or because they measure with partially depleted cells to give the most realistic specs.

bridgman said:

I'm not 100% sure about Surefire using partially depleted cells, but (a) I hear it a lot here from knowledgeable people, and (b) that lines up with the "stupid Surefire always under-specs their lumen output" comments.

When I see measurements of Surefire lights with fresh cells the measurement is frequently well above the rated spec, but not sure if that is just because the Surefire specs are generally conservative to cover manufacturing variations or because they measure with partially depleted cells to give the most realistic specs.

Click to expand...

Either way they look good,and their a$$es are covered:twothumbs What's not to like?

Take care

From HKJ's link:
Mag-LED 4D 50
Mag-LED 4D 80 no reflector
So 50/80 or only 62% of the light got out.

I assume MrGman just took the head off so the losses are from both the reflector and the PLASTIC lens.
The reflector of my Mags seem quite shiny to me so I think a lot of the difference between the Fenix (80%) and the Mag (62%) is due to the difference between AR coated glass vs plastic.

LEDninja said:

From HKJ's link:
Mag-LED 4D 50
Mag-LED 4D 80 no reflector
So 50/80 or only 62% of the light got out.

I assume MrGman just took the head off so the losses are from both the reflector and the PLASTIC lens.
The reflector of my Mags seem quite shiny to me so I think a lot of the difference between the Fenix (80%) and the Mag (62%) is due to the difference between AR coated glass vs plastic.

Click to expand...

Wow, if that is true then that means that a stock maglite with incan is probably less than 50% efficient.

Incans will almost certainly lose more than LEDs because a lot of light disappears into the "hole" in the back of the reflector, and there willbe more reflection loss as a greatre fraction of the lumens from an incan strike the reflector compared to an LED.

AR coating itself isn't going to make much difference in lumens, at least not compared to plain glass, which will reflect about 4% from each interface (2 interfaces total) and absorb almost nothing. The plastic might be MUCH worse in efficiency than even plain glass however, especially if it is scratched and dirty. However, in a reflectored light, a large fraction of the light reflected from the glass will strike the reflector once or twice and eventually make it back out. So 8% reflection doesn't necessarily mean 8% lumen loss.

However, lumens isn't the whole story. In most cases, 60 lumens collimated into a beam will be more useful than 100 lumens coming from a bare emitter with no reflector or window. There's also the issue of how much of that light is actually landing somewhere useful.

This is why I am a fan of TIR optics. These are able to collimate ALL the light (not just the fraction that strikes the reflector surface -- which is quite small in some LED lights), and it can do so potentially far more efficiently than an aluminum reflector and glass window combo.

A premium TIR reflector with the right optical coatings could actually be nearly 100% efficient.

What I'd be interested in seeing is a TIR lens on a flashlight that is connected directly to the LED with index matching gel -- this should eliminate the insertion losses, the main cause of loss in cheap plastic TIR optics in flashlights.

This optic could be improved further by AR coating the front surface (ie the surface where light exits from the optic). Note: I do NOT mean AR coating the entire optic, as obviously that would defeat the purpose of total internal reflection...

Theoretically, such an optic could well exceed 90% efficiency.

There has been studies on stock Mags with incan bulbs. They came out with 65% transmission (that is where my rule of thumb number came from).

I have mixed feelings on optics.
Fraen he 1st optic I came across is very nice - wide hotspot and a dimmer but wider spill than most reflectors.
But the early Cree optics collimated the beam too much - a single hotspot no spill.
I live downtown, do not go out at night much. So most of my usage is indoors where a tiny hotspot no spill is not useful. I use a lot of Glad Press n Seal to diffuse my lights.
I hold my 3C-P7 like a caveman holds his fire on a stick, pointing straight up. That way I do not get blinded by the reflection of the hotspot from my off white walls, and with the SSC-P7 almost as bright as a 15W CFL, I am not lacking in light.

Ok, lets really confuse things. Off course incan formulas do not work with LED designs. In general I see it is covered here, beam pattern, flood v. spot influences usefulness and the appearance of "brightness".
So does color. LEDs seem to have more color variance than incans which are generlly more in the warmer shades. But the human eye is more sensative to certain shades and that influences perception too and takes away from comparison of pure numbers.