Reflector Efficiency discussion/observation, LEDs, Incans, defocussing effects etc...

I know that various aspects of this have been discussed in the past in various threads. I'd like to start fresh. I'd like to present my thoughts on the subject, a few example diagrams, and a few hypothesis and offered methods of proving or disproving my hypothesizing. I would appreciate any constructive input to this thread. If anyone has a light-box of sorts that could runs some tests, proving or disproving me, that'd be awesome. Hence forth, anything said here should be read as my hypothesis and not fact.

---------------------------------------

The generally accepted rule of thumb for converting to "torch" lumens in a flashlight is to simply multiply the bulb or emitter lumens by 0.65. This approach has been proven to be a very reasonable figure when discussing flashlights as a whole. There are obviously many variables involved, but the 0.65 rule seems to hold true quite frequently. However;

I have personally found myself questioning the validity of the 0.65 rule when applied to particular flashlight designs. There are 2 cases where I hypothesize that the 0.65 rule may be off by enough to warrant discussion:
1. In flashlights utilizing LEDs as their source of light.
2. In incandescent flashlights who's bulb envelope size to reflector size ratio is severely disproportionate to typical flashlight designs.

It should be noted, that I entirely realize, that this doesn't really matter much in the end. As it takes LARGE differences in output to change what useful applications a light has. Finding out that a light is 10-20% brighter or dimmer on paper based on a conclusion here will not have any realistic meaning to how useful the flashlight is and should not change anyones mind about what to buy. So with that in mind, lets keep any discussion on the subject friendly and positive because, it doesn't matter that much anyways.

------------------------------------------


First, a very simple test that almost any flashaholic can do themselves in about 30 seconds.

1. Go grab a maglite with the adjustable focus, a 2-3 cell stock one will work perfect.
2. place the flashlight on a flat surface pointed straight up onto the ceiling. in the middle of a room or hallway or bathroom, ideally.
3. Adjust the light to a tight focused beam. Observe illumination level of the room as a whole.
4. Defocus the beam, such that the bulb is moved forward in the reflector, (may need to double check which way the bulb is going by turning off the light and looking at the bulbs motion). Observe illumination level of the room as a whole.
5. repeat the focusing and defocusing and try to observe changes in the total illumination of the room you are in.

With my eyes, I can see a difference, when the beam is defocused with the bulb forward in the reflector, the room appears slightly more illuminated. This is substancial becuse in order to even be able to see a difference, the difference has to be noteworthy. Eyes are not very good light meters.

Here's some diagrams exploring what I believe is happening:




The filament of a bulb emits light in all directions. Different arrows in the diagrams represent different things happening.

The dark blue arrows, and all the area in-between them below the reflector opening, is a region of heavy light loss. As the position of the filament is moved up or down in the reflector, the angle of those dark blue arrows changes, and so does the amount of light headed into that "region."

This is not a region of 100% loss, because, as an observant user may notice, there are often materials in this region that will reflect some light, just not as well as a reflector. For example, the potting material in a bulb might be white in color, this would reflect some light, the PR base on many bulbs is a metallic material the lip around the top of the pot will reflect some light. People who use PR-bi-pin adapters may notice that the top plate of the adapter is a metallic material that should reflect some light. I do not know what the transmission efficiency of light emitted into the hole of the reflector could be estimated at. I could only guess perhaps that's it's between 20% and 50% depending on the particular flashlight in question.

The area between the long yellow arrows, is light that is not reflected in any way, it's only loss is whatever the lens absorbs, in the case of modern lenses, this can be anywhere between 1% and 10%, usually closer to the 1% if it's clean

The area between the blue arrows and yellow arrows is light that is "captured" and reflected forward as part of the "beam."

You may notice, that as you defocus the bulb forward into the reflector, not only does the angle between the blue arrows shrink (reducing losses), the angle between the yellow arrows grows, this means more light is being transmitted out the front, without taking any reflector losses.

The only reason I have included this excursive in my "hypothesis" to to attempt to prove that the area between the 2 blue arrows is in fact a region of significant loss. Many people seem to be in denial that this region exists, I believe it does and I believe it is a factor that incandescent lights have to deal with.

-------------------------------------------

However: many things can be done to help reduce the effects of the evil blue arrow region!

As you may be able to deduce, a smaller hole in the bottom of the reflector (smallest possible for the bulb size), will increase the amount of light that hits the reflector, and reduce losses.

Installing a large bulb in a smaller reflector, will often cause more waste emissions, while installing a small bulb in a huge reflector, has the opposite effect. Here's some examples:




Ideally speaking, if a bi-pin bulb, with a reduced size "base" (many of them have a smaller rectangle shaped base that support the pins), were installed through a hole matching the size of that base through the top of the reflector, a great deal of losses could be eliminated. Good flashlight designs will have the smallest possible opening in the reflector for the bulb.


---------------------------------------------

Now I'd like to take a look at LEDs:



LEDs emit light on a "half sphere" and their lumen rating includes what is emitted on that half sphere. Imagine a plane intersecting the LED at it's base, nearly nearly 100% of the emissions occur above this plane. The grey double arrow line in the diagram above represents this cutoff point. There is no light going below this line, therefor, nothing is being lost "out the back" into an area of poor reflectivity.

Again, the long yellow arrows represent the cutoff for light that is being reflected, in-between those arrows, is light that is just coming out on a free ride, taking only a loss through the lens. Light emitted between the gray arrows and yellow arrows is light that must be reflected, and takes a loss due to reflector surfaces not being "perfect."

LEDs are different than filaments in many ways, not only do they emit all of their light on one side of a plane, they also do not emit light evenly in all directions above that plane. You can pull up a spec sheet on various types of LEDs available and they will almost always have a "radiation pattern" chart. Showing the intensity of light as measured from different angles in front of the LED. In many designs, the most intense light is found in the general range of about a ~90 degree angle (45 degrees either side of dead center), but dead center is not usually the strongest point. There are peaks in output at either side if dead center. So depending on the reflector design, deepness, shallowness, etc, or the type of emitter chosen (there are often different radiation patterns available), the amount of light that is reflected, compared with the amount of light that comes out as "spill" (free ride through lens, not reflected into the beam) can vary widely. Different designs will take a different amount of loss on the reflector depending on how much of the light is captured by the reflector compared with how much just comes out the front. Trading spill light for more throw can be accomplished by producing a deeper reflector, here's and example:



In the same way as we did with the incandescent examples we can compare the angles to get an idea of how much loss is taken in different parts of the emitted light. Comparing this example, to the LED in the "standard" reflector above, we see that the angle between the 2 long yellow arrows has decreased, while the angle between those yellow arrows and the grey ones has increased; This means that less light is coming out the front only taking lens losses, and more is taking reflector losses. But this type of design, would put more of the available light into the center beam, which will illuminate objects farther away and reduce the wide angle of the spill found in the "standard" reflector. The same design changes can be made for incandescent flashlights to achieve different beam intensity/spill tradeoffs.


--------------------------------------------

Discussion ideas/questions:

1. Do you believe that the region between the 2 blue arrows exists? explain.
2. What do you consider to be the generally accepted reflector surface efficiency? (not the reflector as a whole, just the surface of the reflector) I think between 70-85%
3. Assuming you are not in denial about the region between the 2 blue arrows: What do you feel is an acceptable general reflector efficiency to be assigned to the region between the blue arrows in most flashlight designs? Any ideas on how to maybe estimate this? I think between 20-50%
4. What transmission efficiency can be agreed upon to use as an "average" or "typical" in a broad discussion of flashlights? I like 95%, since 99% only applies to flashlights in a clean room with no dust or dirt on the lens, lol.

With some general consensus on these numbers, I could run some numbers and come to some sort of conclusion, or at least present some more ideas for fun.

5. Would "Sort of Conclusion" make a good name for a band? explain.

-------------------------------------------

Thank you for looking,
Eric

A most excellent post. I had a bunch of thoughts and question of a quick read, but don't have time to look at it and respond properly now. I did want to ask if you have links to those old threads posted by some of "The Godfathers of CPF" where this 65% assumption was introduced?

Again, many thanks in advance for what obviously has taken you a long time to post. :thumbsup:

Yet another awesome thread destined to be a sticky.
:thumbsup:

Here is the post about the 65% being confirmed by a 1274 in integrating sphere.

When the small-hole reflectors first arrived, the calculation was for ~7% more theoretical output (may be seen as coming from smaller "blue arrow angle area" loss in your example).

Looking at your diagrams for the incans, what I beleive would be a better approach than the typical "mag" reflector shape, would be one that is deeper and more recessed into the light (for D-cell mags there is plenty of room to recess a reflector) with the focal point further away from the "hole" at the base, so as to minimize the amount of light lost there. The deeper shape would also refer to a greater spot-to-spill ratio, which in many cases is more important than overall lumen output.

On the opposite extreme from my "idea" reflector, the light loss is very great for lights using the Osram lamps, such as the 64623, where the hole is larger, and the filament sits relatively low in the reflector.

The area between the long yellow arrows, is light that is not reflected in any way, it's only loss is whatever the lens absorbs, in the case of modern lenses, this can be anywhere between 1% and 10%, usually closer to the 1% if it's clean

Click to expand...

When you are talking about 10% window loss, do you mean from reflections at the interfaces? In a non-AR-coated lens, that will actually complicate your drawings a bit. The amount of light refected from each air-glass interface will be about 4% for light striking normaly, more otherwise. So 10% would be a good estimate for the amount of light reflected by the window. Much of that light will in fact re-strike the reflector, then re-exit. Raytracing that will be complicated. Some of the light reflected off the window might bounce off the reflector again once, twice, or even disappear down the "hole" at the base. Either way though, I'd bet that the majority eventally makes it back out as spill.

So, adding an AR coating will primarly improve the "spot to spill" ratio, not so much the overall output, though even that will depend on innumerable factors.

The other case is that for a good TIR optic, which implemented AR coatings at certain interfaces should theoretically be able to achieve nearly 100% light transmitted out the front. (Note: before anyone replies, I'm aware you can't AR coat the entire thing, as it would cease to be a reflector... I just mean coating the interface between the LED and optic, and the front window surface)

Hi 2xTrinity,

As for window losses, I've heard of certain completely clean and new glass windows providing as much as 99% transmission. Like you said, some coatings may cause some reflection back into the system, which like you said, will have another chance to try to re-emerge with further reflector losses (and over and over again). But provided the lens is perfectly clean and free of defects or imperfections, 98-99% is still *possible.*

As I understand it (could be wrong on this) What ends up happening, is lenses get dirty, scratched, etc etc. The light is partially absorbed and converted to heat in such an example. Some lens materials will just have less transparency and absorb more light.

As I understand it, the reason a "black" object is black, is that it is absorbing light rather than reflecting it, a black shirt on a sunny day is a perfect example, heat conversion. So any lack of transparency is probably a heat conversion going on at the window itself.

Eric

I was just realizing something that has baffled many on the forums for years..

I think my post above explains why Surefire rates the P91 at 200 lumens, and the MN16 at 225, even though it is widely accepted that these 2 bulbs are practically identical based on all measurements of electrical behavior recorded here at the forums. The difference probably comes from the relationship between the filament position within the reflector.

Eric

that "loss area" with bulbs certainly exist. imho it steals a noticeable part of the light and out of that "stolen" light just 10 % (at best) might come back as undirected reflections from base material.
(Thats why reversely mounted led never got a foot on the ground - the losses are too high (+ the cooling is more difficult, too much space needed, ...)

as to general losses:
imho its totally the same with bulbs or led!
any media transition (air/glass, glass/air) is said to steal a good 5-7 % (less with coated materials)
bulb has: inside bulb air/bulb (1), bulb/inside the light air (2), inside the light air/lens (3), lens/outside air (4) ... makes 4 transitions + reflection at reflector
led has: inside dome/dome (1), dome/inside the light air (2), inside the light air/lens (3), lens/outside (4) .. again 4 transitions + reflection at reflector/optic
... so both have the same losses ... with about 20 % just from the transparent parts (or less, depending on quality) + the losses of focusing device.

as to device:
very good reflector should be around 85-90 %, as we read in here.
imho optic is the same or worse, I dont belive that TIR 99 %. If these 99 % were real, then the led base "behind" the optic should not be this illuminated, as is it is - with an optic.

so whatever used, the difference between emitter lumens and otf lumens is that 65 % You mention in Your excellent 1st post.

For LEDs, I say 75% is accurate.

Take, for example the Pelican 7060:

McGizmo tested one in an integrating sphere and got 174 lumens. Pelican says the light draws 4.4W. At 3.7V, that's about 1200mA. At that current, a Cree P4 (the light was made a year ago, so it's not a high bin) puts out about 225 lumens. Multiply that by .75, and you get 169 lumens out the front. That's 4% error, scientifically acceptable almost everywhere.

169 lumens is about 72.7% transmission, so 75% is accurate enough for our estimations.

And that's with a really deep reflector (compared to its diameter).

nice thread, but definately beyond my scope and my language

yellow said:

any media transition (air/glass, glass/air) is said to steal a good 5-7 % (less with coated materials)
bulb has: inside bulb air/bulb (1), bulb/inside the light air (2)

Click to expand...

These reflections shouldn't lead to hardly any lumen loss, since the vast majority of the time the bounces off of one of these interfaces, it will make it back through and still strike the reflector, just not quite as focused -- the actual envelope will look like the light source, rather than the filament. You can even see the effects of this in the beam of some incandescent flashlights, where the coil of the filament is focused as a hotspot, and there is a corona around caused by light scattered off of the bulb envelope.

inside the light air/lens (3), lens/outside air (4)

Click to expand...

These reflections more than half will make it back out as a fuzzy corona, or as spill. Most of it will bounce of the reflector once, less of it will bounce off the reflector twice, some of it will disppaer into the "hole of doom". Though, admittedly, what does make it out of here will be randomly distributed "spill" moreso than the bulb-air interfaces (I have compared before and after of lights with and without AR coatings and can see differences in spill -- which has a more abruptly defined edge with the AR coating.)

... makes 4 transitions + reflection at reflector
led has: inside dome/dome (1)

Click to expand...

There is no air inside the dome, it is connected to the die with optical gel to eliminate this interface

dome/inside the light air (2)

Click to expand...

This is actually what I suspect leads to a lot of beam artifacts. Lumens do make it out though. You can tell as in the case of the original XR-E, when there was phosphor all around, you could see brown fringes around the edge from the reflected light striking the phosphor AGAIN before being re-emitted. Now only the die itself is coated, so color is more consistent. This is the interface responsible for muhc of the "cree ring" artifacts which many peopel would rather do away with.

very good reflector should be around 85-90 %, as we read in here.
imho optic is the same or worse

Click to expand...

A perfect aluminized coating will be highly efficient like that, tI hav always read that reflectors inside our common lights may be much worse efficiency than that, like 75-80%, with reflector efficiency representing the vast majority of losses, not interface reflections.

I dont belive that TIR 99 %. If these 99 % were real, then the led base "behind" the optic should not be this illuminated, as is it is - with an optic.

Click to expand...

TIR could be nearly 99% transmittance if kept perfectly clean, and if they applied an AR coating to the interface between the LED and the inside edge of the optic, and the front window. Or they migh make one with the gap beteween the LED dome and the optic flooded with optical gel, along with optically coated front window -- though this would exhibit more of the traditional reflector with spot and spill pattern, not all throw like current TIRs.

What happens is light is reflected off of that first interface, then when it goes back though the optic again, it's not at the approporiate angle of incident to undergo internal refelctor. Same with light reflected off the front window -- unlike an aluminum reflector system the light that reflects off the front window won't necessarily be reflected back again, depending on incident angle.

so whatever used, the difference between emitter lumens and otf lumens is that 65 % You mention in Your excellent 1st post.[/quote]

I have been finding the old threads to see what points and work has been done, and examine how determinations were made previously. **

This is the main thread I was looking for. As I suspected, the basis for people making up the "torch lumen" term, and assigning it with an approximate 65% number is riddled with assumptions, limited measurements, and questionable methodologies.

• 2/6/06 - bLu vs. tLu: IS confirms 65% conversion factor (js - History of torch lumens, 65% figure, limited testing, and many assumptions. - 125 posts)

A few posts in earlier threads grew from ungerminated data seeds into a consumately intertwined mesh of giant beanstalks. The topic appeared rooted deep the in the earth of "near-Biblical" certainty. I even checked to see if the postulations of Torch Lumens were also being considered for the "Fourth Law of Thermodynamics." Alas, it was not. OK, I'm pushing the envelope on my style of humor.

Realistically, IMHO, there were some early CPF heavyweight endorsements, and a steady tincture of repetition that seemed to create a standardized description and a 65% multiplier for Torch Lumens. The number and concept was further incorporated into AWR's widely disseminated Hotrater Spreadsheet and multiplied times the manufacturer's stated default bulb lumen ratings. I began to question most of the theory and data models when I saw other erroneous data & assumptions presented in the Hotrater.

Whether or not this (well intentioned) approximate 65% number & standardized "torch lumen" concept turns out to be consistent or accurate is unknown to me. There is not yet sufficent data or scientific method being used to say. Notice: I am not saying it is right or wrong...but I am saying that the information should be questioned at a fundamental level. I salute what mdocod is trying to accomplish.

The story of "Torch Lumens" and a 65% multiplier starts out with a series of posts by PaulW back in 2003. It uses some theoretical methods & data obtained with a consumer-level Meterman LM631 Light Meter. I do not believe this used a reliable/scientific method, nor were his Lux readings of various light setups legitimate using his creative corner bounce test setup. The remainder of these links are what I found as chronological references that have other useful tidbits or reinforcements/justifications for subject theories.

• 4/9/03 - What kind of light meter should I get? (PaulW - Getting a Meterman LM631 & talk about making a light box - 32 posts)

• 7/15/03 - Performance of Bright Lamp Configurations (PaulW - Initial bounce Lux test results for several bulbs - 19 posts)

• 1/25/05 - Mag85, how many REAL lumens (juancho -21 posts)

• 9/4/05 - Lumens Estimates for Aurora2 & Mag85 (PaulW - Bounce/corner testing & first mention of 65% - 23 posts)

• 10/6/05 - another M6-R with MN61 (kalengkong - More references and other rating formulas discussed - 142 posts)

• 11/5/05 - L2==>M4==> ? M6 Help a new member (LuxLuthor - My embarassingly auspicious arrival at CPF, and where I now remember Luna being the source of my journey into incan hotwires...more relevant information starts about 30-40 posts into thread. before it crashes and burns into antiquity.- 78 posts)

• 8/10/06 - A testament to CPF's measurement resources...kinda (Delvance - 1 post)

• 8/18/06 - How much light actually ends up where we want it? (Reflector idea) (Flummo - 31 posts)

• 3/29/07 - Incan vs. LED ? (7histology - 122 posts)

(** Personal Note: Given the comments by some in above threads, this reference post is intended to be a constructive resource. So if my poor attempts at humor send you careening off the side of a building, please consider the source, this disclaimer, and assume the best of intentions.)

Just a thought on that 65% figure. Some of our members with IS's have measured the Malkoff M60 at 225-235 lumens, a drop in with a Cree Q5 at about 1 amp to the LED, using an optic. A 225 lumen figure would equate to 346 lumens, before losses. I don't think the Q5 Cree is putting out 346 lumens at the LED, and the improved lumen performance of the M60 is due to a very efficient optic, and losses are much lower than 35%. Just one example, I know.

Bill

Oops nvm

LuxLuthor!!!

Nice work gathering that up. I'd say it appears you have about as much work in that summery of the history as I have in the OP!

OT: Your humorous approach is greatly appreciated, keeps the spirit of positive discussion loose and fun and alive. I'd like to keep this thread fun and not have anyone throwing any pesky facts into the mess. I like the concept of just accepting that the facts don't exist yet, we can only speculate. With that thought in mind, I'll move on to some speculation:

-----------------------------------------------------------------

I remember back when I was in science classes and math classes in high school (6-10 years ago I think, lol), the teachers rambling about constants, variables, etc. If I took anything from it, it was that you had to establish some constants to work from in order to determine the variables, or maybe it was the other way around, or who really cares what I learned in high school anyways...

ON TO THE SPECULATION!
(Oh, FYI: I've decided to call that area between the dark blue arrows the "black Hole" region)

I wanted to define a few speculative constants. (to be semi-established from answers to the discussion questions from above!)

For the following equations I'm going to try the following constants:

1. bulb/emitter lumens: 100 (duh!)
2. lens loss: 5%(x0.95)
3. "black hole" loss: 80% (x0.2)
4. reflector loss: 20%(x0.8)

The variables are what percentage of light is falling into each category based on the position of the bulb/emitter and the shape of the reflector.

I'm going to do the first speculative equations based on the small-bulb/large-reflector and large-bulb/small-reflector diagrams from the original post, assigning approximated values of each piece of the pie and see what I get.

------------------------------

Large bulb/Small reflector:

1. spill region: 90 degrees or 25%.
2. black hole region: 120 degrees or 33.333...%
3. reflected region: 150 degrees or 41.666...%

Now I'll apply the losses that apply to each region. In the order the losses are taken.

1. spill: 25 x 0.95= 23.75 lumen
2. black hole: 33.333 x 0.2= 6.666, 6.666 x 0.95 = 6.333 lumens
3. reflected: 41.666 x 0.8 = 33.333, 33.333 x 0.95 = 31.666 lumens

61.75 lumens make it out the front, or you could say that the multiplier is 0.6175, or you could say that this system as a whole is 61.75% efficient at converting bulb lumens to torch lumens.

-------------------------------

Small bulb/Large reflector:

1. spill region: 90 degrees or 25%
2. black hole region: 60 degrees or 16.666%
3. reflected region: 210 degrees or 58.333%

As I did before....

1. spill: 25 x 0.95= 23.75 lumen
2. black hole: (16.666 x 0.2) x 0.95 = 3.1666
3. reflected: (58.333 x 0.8) x 0.95 = 44.333

71.25 lumens make it out the front on this one.


--------------------------------

Same as above but with an LED instead of a bulb:

1. spill: 90 degrees or 25%
2. black hole.. err . nope
3. reflected: 270 degrees or 75%

1. spill: 23.75 lumen
3. reflected: (75 x 0.8) x 0.95 = 57

80.75 lumens out the front here, or 0.8075 conversion factor... hmmmm....

--------------------------------

Going from a large black hole to no black hole has increase net output by about 30% here. That's noteworthy. We've kept all the major constants here the same between the designs, just changed the arrangement of things.

I'll play around with some more ideas in the future, maybe play with some different constants from the get go and see how this factors in. But this is really an example I wanted to get across.

Eric

What LL & mdocod said.

One caution I would like to point out - for some LED flashlights I've seen there IS a black hole, as the bottom of the reflector cup is somewhat above the plane of the LED. In a well designed flashlight, this can of course be totally eliminated, but many cheap or homemade torches will have this defect.

And for reflector-less aspheric setups, by the above maths the "torch lumen factor" (TLF) for any light source will be abysmally low, despite the insense hotspot.

The Laws:
0: The banker is fair.
1: You can't win.
2: You can't break even.
3: You can't even stay out of the game.
4: Your torch is not as bright as you think it is.

LEds don't have half sphere beam shape. Most light goes out in forward direction less goes to sides. So less light hits the reflector and more light goes directly to the spill. This is imho the reason why LED lights with reflectors have more than 80% OTF lumens. http://www.mtb-news.de/forum/showpost.php?p=4320770&postcount=127
(McR 17 with Cree - 89%, IMS 20mm with SSC - 85%, Fenix L2D CE - 109,7 OTF lumens -> 81,3%).

Hello lctorana,

that is an excellent point. I too have observed a few LED flashlights that fail to take advantage of "gathering" up all available lumens from the side of the emitter... but as jirik_cz said....

Hello jirik_cv,

that's exactly correct, in my simple formula about I did not take this into account but did make mention of this behavior in the original post. It would be an interesting undertaking to assign a number of regions in the output a average illumination intensity for the region, and then re-calculate, as you have said, and this is especially true of more standard "wide" reflector LED setups, the spill light is brighter than comparable output incans, but the beam is less intense usually. I have observed this in use and agree 100%

Thank you both, excellent points

Eric

I have been thinking about basic concepts before moving into the complex mechanical behavior & measurement of various light sources in reflectors or behind lenses. Some of my questions began long ago when I got one of these $10 Amazon.com spinning Crooke's radiometers (a must have for any true flashaholic). :thumbsup:

Not knowing how they worked, I found it interesting that NONE of my LED's would spin the vanes, but all of my incans & HID's would. This led me to finding out difference between measurement of radiation (radiometry) & light (photometry).

Once I understood that this Crooke's Radiometer was based on heat transfer, I began to wonder about how Light Meters and their porous white sensors were mearuring different light sources. When I bought the Meterman LM631 Light Meter (because most other CPF members bought it), I noticed it measured light in either Footcandles or Lux...using the "CIE Photopic Spectral Response." I thought that sounded impressive, but had very little idea what it all really meant.

I assumed the LM631 would be equally proficient measuring LED, Incan, Fluorescent, or HID lighting, but then noticed more expensive meters have a variety of light source menu profiles &/or sensors--so now I'm not sure. Until about a year ago, like many CPF members, I just went along pretending to vaguely understand various terms like Lux, Lumens, Candlepower....and since I knew what a candle flame looked like, and that 12 inches were in a foot, how complex could a "Footcandle" be?

It seems that the more I read....the more questions I have, and the more previously presented information I trusted, no longer appears to be following fundamental concepts. I started doing the destructive testing when learning that a lumen predicting formula from the Welch Allyn website had been taken and used as valid for any brand of bulb or degree of overdrive. Suffice it to say that it is not an accurate general predictive formula.

If lumen predictions were being hyped, in addition to the 100 quadrillion spotlight candlepower claims, I thought it useful to go back to the building blocks of light description & measurement. For example, I'm not sure if most people understand that a lumen is based upon a specification of a single, green wavelength (555 nanometers) of light. Conversion between other units of light measurement and color wavelenghts have limits and strict definitions. There are many further refinements related to human color perception & scales. Reading and understanding all of this information is interesting but time consuming.

LL,
I really like that point you are bringing up about using the re-rating formulas. When I put together my incan/li-ion guide, I had to use something to make approximations, but was very clear that they are ONLY approximations. When I started trying to use the formulas on various bulbs I realized that something was out of whack and actually adjusted them slightly and re-did the entire chart based on an adjusted constant. The ^3.5 for bulb re-rating seems way off in almost every case, in some cases, 3.3 is more accurate, and in others, 3.0 or less. I've studied your destructive testing quite heavily and have worked the equations backwards and found that every bulb has a very different approximate set of re-rating exponents to be used in the formulas.

There is evidence all around that the formulas are not universal. A perfect example is driving a an MN61 on 3 li-ions. Seems like very often they survive, but the generally accepted bulb life formulas (and knowing how to read the results) says that it's impossible for the bulb to survive. This means that either the exponents are wrong, or the original bulb specifications were wrong, my feeling is that they were all wrong.