The LightBox Experiments... Comments please!

All -

With all the problems of using LUX as a measure of light output - namely that it is drastically affected by lens and reflector shape - I set out to make a "poor man's integration sphere". Now admittedly it isn't that scientific, but it IS very reproducable. Plus the readings I am getting from it only vary + - 5% depending upon the focus of the light. So it is giving a good RELATIVE reading of the output of the lights.

Here's the box:



I had to find a box that was very white on the inside so that the readings were only of reflected light. How about a milk carton? I cut a hole in the end for the meter, a hole in the top to shine the light through, and covered it all with aluminum foil to prevent external light from getting in and from too much light from leaking out.

Readings are very consistent. I can remove and replace the meter, hold the light by hand, and as long as it is pointing straight down with the bezel front even with the opening, I'm getting nearly the same numbers every time. The meter sensor has a groove that mates with the hole I cut, so it goes in the same spot every time.

I added a 3"x1" strip of paper to the edge of the hole near the meter on the inside to prevent any direct beam artifacts from striking the sensor.

Now these numbers allow for REALATIVE comparisons between light. They are Lux received at the sensor reflected from the internal surface.

Here are a few of the readings so far.

Surefire G2 (P60) ........................................ 5200
Surefire M2 (P61) ........................................ 7000
Pelican M6 .................................................. 5700
E2 (MN03) ................................................. 4000
Surefire KL1 Green (1 LS) ............................. 2080
Streamlight Tactical TL-3 LED (1 LS 5 watt) ... 4100
Streamlight Tactical TL-2 LED (1 LS) ............. 2160
Surefire KL3 White (1 LS) ............................. 2450
Surefire L4 Digital Lumamax (1 LS 5 watt) ...... 5800
Surefire L1 Digital Lumamax (1 LS) High ......... 2200
Streamlight TaskLight Luxeon (1 LS) ............. 1550
Surefire KL1 White (1 LS) ........................... 2150
InReTech White (Alkaline AA) (1 LS) ............ 1150
CMG Reactor 3 (1 LS) ................................ 1032
CMG Reactor (Alkaline AA) (1 LS) ................. 465
Surefire L1 Digital Lumamax (1 LS) Low ........... 70
Surefire A2 Incand + LEDs ........................... 5750
Surefire A2 (3) ........................................... 520
Dorcy 2-C (8) .............................................1400

It's not an integration sphere, but the light hitting the sensor is reflected, so it should be giving a general idea of the overall light output as compared to Lux readings which only give the reading of the part of the beam hitting the sensor.

I'm getting some pretty wild readings - the A2 3 LEDs read higher than the CMG 2AA reactor. The Inretech 2AA LS reads higher than the CMG 3AA reactor! The Green KL1 consistently reads LOWER than the white KL1 (!). The E2 and TL-3 LED are putting out about the same amount of light.

Comments? Shall I continue?

Quickbeam, I'd say it beats the heck out of simple Lux measurements which are almost meaningless.

The color anomolies might be from the imperfect white insides of the box absorbing some wavelengths more than others.

Perhaps lining the box from the INSIDE with the foil might reduce these effects ? It would increase "light" readings somewhat, but all wavelengths might reflect more evenly.

The readings are all relative anyway, so the differences "might" be more accurate.

Which brings up another point: Is the Lux meter evenly sensitive to all wavelengths ? Probably not, don't know how to factor that in.

Well, Lux readings are useful in comparing throw of the lgihts, just not overall output.

I'm afraid that lining the whole box with aluminum foil may result in direct reflections striking the sensor. The object here was to try to have muted, reflected light just like in an integration sphere.

IIRC, Lux meters are designed to have sensitivity similar to the human visual range.

More comments, Please!

Foil lining is not the answer. For many years, Omega used color enlarger heads that used white styrofoam to integrate the light. It has very little effect on the spectrum, and its high reflectivity combined with non metallic finish yielded a soft featureless light.

I think this is an awesome idea Will. I would love to see a more comprehensive list developed over time. Way to go! /ubbthreads/images/graemlins/smile.gif

David<><

Hmmm, printer paper (non-glossy) or some other "more" white inside surface ?

Anyway, I like the idea.

"I think this is an awesome idea Will."

Ummm... It's "Doug" /ubbthreads/images/graemlins/smile.gif

Keep em' coming!

Sorry about that Doug /ubbthreads/images/graemlins/blush.gif
Brain was in left gear or something. LOL

Right brain checking to see if left brain is okay.../ubbthreads/images/graemlins/poke2.gif

David <><

Quickbeam,

Great idea and a good list of data.

I had thought of doing something like this, but like a lot of other projects, haven't started it yet. My plan was to make a cardboard box with enough cardboard baffling to prevent directly illuminating the sensor. The idea was to put the flashlight hole and the sensor hole on the same side of the box. I even bought a can of white spray paint for the indside.

However, the best setup would be made of a stiffer material -- wood or metal, because measurments would be more repeatable. Since I haven't put it together yet, I presently use the method of bouncing the light off the wall or ceiling. But that doesn't have the repeatability that a box has.

As you probably already know, the more floodlike the beam, the more repeatable the measurement and the less sensitive it is to flashlight orientation. For flashlights with narrow beamwidth (long throw), I suspect orientation is the biggest problem. That is, the measurement varies wildly as the light is moved. I would guess that putting additional baffles in the box would reduce this effect. Ideally, it would be nice to be able to move the flashlight to any angle and still get the same measurement.

In theory, you want to add together all the pieces of total output in every direction from the light. Different patterns of baffles will give emphasis to different angles. Reflecting directly off a wall, as the simplest of examples, emphasizes the beam center -- the better the throw, the higher the reading. Devising an optimum baffle configuration is a real mental challenge

Well, you wanted more thoughts. You got some rambling. Hope you can use some of it.

Paul

" Ideally, it would be nice to be able to move the flashlight to any angle and still get the same measurement."

Good point. Orientation changes do affect the readings in my set-up, but as long as the light is straight up and down I get the same readings, so I'm happy enough on this point.

BTW, the 8C UK produces a reading of 12,000

Since this seems to be an OK idea so far, I'll do a few more reading and post a chart.

But please, if anyone sees a real issue with these comparisons or some numbers seem wayyyyy off, let me know before I get ALL of the lights done! /ubbthreads/images/graemlins/smile.gif

I've been thinking about this idea all morning, and I think it is a SPLENDID and well executed idea.

I think it will give us something to chew on that is better than what we have at present.

I do think a more pure white inside inside the box, however "dull", would reduce the different wavelength problems.

Here is the chart so far - FYI, it will be removed when a complete chart is placed on my website.

(Removed - chart is up on site now! http://flashlightreviews.home.att.net/reviews/lightbox_output.htm)

I'm stopping here for now for more comments.

I'm not so sure the inside of the milk box is that bad for wavelength differences. If in reference to the KL1s, perhaps I just have a really good white and it just outperforms the green...

Here's the list alphabetically so you can find it in the chart easier (it's a gif to save CPF bandwidth /ubbthreads/images/graemlins/thumbsup.gif )

(removed - see link above)

Edit - added Mag Minimag and 3-D lights with emphasis since these are very common and well known

hmmmn ive been looking for something like that makes it easy to see at a glance about flashlights capibities

I see a lot of real world agreement with the lights I have, and have looked at closely.

You've got a good thing here.

Quickbeam, do you have any 3 cell surefire lamps?

Thanks, Quickbeam! This is fascinating, educational and useful. /ubbthreads/images/graemlins/thumbsup.gif

Now I'm really looking forward to my SL TL-3 Xenon arriving!

BC

Quickbeam,

This is a terrific, valuable project, and the chart is especially helpful. It helps to validate my own tests because my results generally coorelate with yours even though I use two different techniques. One is the simple ceiling reflected light test performed under controlled conditions.

The second is my own "integrating arc". It consist of matt-white cardboard 28"x22" bent into a curve and taped to a table with the walls of the curve perpendicular to the table surface. The flashlights are exactly positioned in an improvised jig so that the lens would touch an imaginary line drawn between the two vertical edges of the curved cardboard. The meter head is fixed in place slightly to the rear of the flashlights and elevated a few inches. This arrangement ignores differences in beam type or focus as shown by virtually identical readings between the Stinger and the Stinger HP, two lights with the same output but totally different beams.

I have found that not only do the results from these two testing methods correlate closely with each other, but also with your figures which leads me to conclude that when determining relative flashlight outputs, most carefully controlled and consistantly repeatable methods of measuring total indirect light will correlate closely with each other. That is a conclusion based on comparative results from three entirely different methodologies.

Brightnorm

hmm it seems from everything i've heard about the peli m6 that reading's kinda low.. i always thought it was noticeably brighter than the scorpion and the a2.. the nite fighter has a nice rating too.. but i dunno a thing about the light.

then again, i have no technical merit when it comes to flashlights.

Quickbeam,

That's quite a collection you have!

I think what you've really come up with is a set-up which could be reproduceable by others if a standard mat white photopaper could be specified. And, not counting the lightmeter, at very low cost. A "community light" could be mailed around for relative calibrations.

Larry

How about a 3 gallon white plastic pail from Home depot? No corners/angles and possibly uniformity? Just a thought. Cool experiment you have going.

[ QUOTE ]
BC0311 said:
Thanks, Quickbeam! This is fascinating, educational and useful. /ubbthreads/images/graemlins/thumbsup.gif

Now I'm really looking forward to my SL TL-3 Xenon arriving!

BC

[/ QUOTE ]

I think you in for a pleasant treat when your TL-3 arrives... Got mine the other day and boy does it throw out a very bright white beam that is focusable and it can be focused to a smooth beam for tactical purpose. It does have a textured reflector which makes the smooth beam possible, hat's off to StreamLight!