Asperities on the real wall, the LED not being a point, and the surely imperfect alignment will conspire to show the projection regardless, I think. It’s finicky though.
1 Thank
It looks like @Enderman figured it out some 8 years ahead of us and put the equations in Desmos for us to use (it can even be modified there).
6 Thanks
Yeah that makes sense! So basically the beam shape is the constant that links lumens to candela for any specific light setup.
Different power levels same multiplier, that’s why the percentages track together.
Thanks for clarifying, this helps a lot with understanding throw measurements.
Although slightly oversimplified, this is a nice way to think about beam shape. Technically there exist some complications: two different lights can have the same total output, same maximum intensity, but different beam profiles.
In general, any single-valued measurement like cd/lm can only describe aggregate information like maxima and averages, but not the entirety of the beam shape.
Fortunately, these complications are almost non-existent if you compare lights with the same type of optical setup, like reflector versus reflector, or TIR versus similar TIR.
Which brings me to an interesting problem: what would be the smallest set of values that could adequately if approximately describe a flashlight (with the emphasis on the smallest)?
I figured that if I know
- min lm
- max lm
- √cd/lm
- lm×hr
I can tell quite a bit about the flashlight usefulness based on these four. One can even argue that min lm is one too many, but then you may buy Sofirn HS43 for reading.
Sure, there is a lot of info missing, but then the list becomes long if included (and one thing leads to another: if you include CCT then CRI, R9, duv and others come begging etc.). What’s your minimal set?
2 Thanks
Very interesting problem, I’ll need to think more on this one! This problem has the same flavor as “describe a probability distribution on [400,700] using just 2 values (CCT, duv)”, but is more complex. I expect any sensible candidate solution to be similarly nontrivial.
For simplicity, assume optimal engineering, in the sense that there is no way to make one attribute better without making another worse. Also assume that the emitter is a white LED, which has well-known electrical and spectral behavior. Maybe one can do a ranked list like a power series expansion, starting with the most important terms. Mine might go something like the following, with max/min taken over all modes:
- Max electrical power (W)
- Max irradiance at 1m (W/m^2)
- Total energy (J)
- Min electrical power (W)
- Max sustainable electrical power (W)
…
List very much subject to change. I decided to go with W rather than lm, and avoided colorimetric measurements, to avoid penalizing preferences of CRI vs efficacy. Electrical power is used to enable high-precision runtime calculations, which I find more important than high-precision output calculations.
One might argue that output vs throw is another an instance of preference, but I argue that output is more important than throw in the following sense: a very high output guarantees a decently high lower bound for throw, but very far throw does not guarantee any nonzero lower bound for output.
2 Thanks
Not the same, and also I know these flashlights well enough already, but an attempt to practically but numerically compare a few lights (range, cd/lm, and Max lm can be reduced to any two, but it may be easier to read the way it is?; battery type is nice to know, but Im×h metric covers that and the LED/driver performance):
Also, using klm and km will make it more compact?
2 Thanks
More values definitely makes the table more readable! And easier for pattern-spotting. For example, one might notice that the (max lm)/(lm*h) ratios all stay in the same order of magnitude, and one might interpret it as some sort of “hotrod coefficient”.
klm probably does compactify things, very nice to be living in an age when klm is a reasonable unit to use! Perhaps the folks in the future who use Mlm would disagree…
Power demand is an interesting twist. Though given the the flashlight purpose lm may be more natural, despite drawbacks.
Again, I’m worried about penalizing the preference for color quality. Given the ratio (current at a given mode : current at which max output is achieved), the efficiency of a white LED should be pretty estimable, and typical spectral efficacy of various low/high CCT/CRI sources can be estimated too, to give a ballpark figure for the luminous flux.
Restricting to white LEDs also rules out some crazy exceptions like SST20-DR, which has pretty low efficacy but apparently up to 70% efficiency, way higher than the typical white LED.
With you and @QReciprocity42 knee deep into this discussion about beam shape, I’d like to ask you about adding a complimentary discussion that’s long overdue, that being what I call “functional throw” versus the now ridiculous marketing specification for throw (max candela).
Back in 2009, the ANSI/NEMA-FL1 (and later by ANSI/PLATO-FL1 in 2019) formally defined and adopted several specifications, including one for beam distance so that flashlight manufacturers (and others) could provide consumers with a way to make apples-to-apples comparisons of their products for how far they projected their beam down range. The only distinction of this standard has been how far, not for what purpose.
The problem that I’ve had with the candela standard is that portable lighting and LED technology quickly outgrew the distance that the human eye could see of value (a target), without the aid of binoculars. I mean what’s the point of a flashlight that can light up something we can’t distinguish. “Hey man, is that a deer or a bush” …“seriously dude, I can’t make it without my scope”. LOL!
It’s great that we can easily compare flashlight theoreticals by reading manufacturer’s specs on their box or website, but how helpful is the FL1’s “throw specs” anymore when the distances involved require sighting optics to see, especially when the FL1 Standard is only on a tiny spot of an equivalent of moonlight?
In my personal opinion, the specification standards from 2009 have never really been helpful because the “BEAM SHAPE” & DISTANCE specs don’t account for actual “functional usage”. Thrower flashlight specs need to help us choose how well it will illuminate an entire subject down range (a person or animal), but not with a speck of moonlight intensity.
So I’ve been gradually drafting a letter to the ANSI/Plato organization about creating and adopting a more user-friendly sub-category that would give consumers an additional and better way to compare and choose a long-throw flashlight for real world purposes, what I’m calling FUNCTIONAL BEAM DISTANCE (FBD). Where I need help is with the physics and mathematics that go into defining a functional throw standard. It’s long overdue.
With your assistance, I’ll begin a thread focused on developing and defining this new meaningful specification standard for “functional throw”. Hopefully ANSI/Plato will consider formally adopting this additional specification standard.
So what do ya’ll think about this? Please discuss.
2 Thanks
1,000,000cd could be useless if the beam is broad enough to reflect/refract off everything else.
There needs to be some defined limiting factor, such as beam angle.
Because the LED isn’t a point source, you’d need to account for the widest possible angle in conjunction with any optics. A deep/narrow reflector doesn’t necessarily mean a narrow projected beam, it just disperses the beam differently. It just so happens that because the exit aperture is narrower compared to a wider reflector, the directly projected beam “shrinks”. The reflected portions still come out at weird angles.
If you shine a bare LED down a narrow tube vs. A wider reflector, the reflector is still likely to throw more light forward past a certain distance. I’m curious that if two reflectors that are scaled proportionately with the exact same deflection angles would have similar beam characteristics.
Say there is a target object 100 m away. We illuminate it to some level (say 1 lux, or any other value). Then say that this same target moved to 200 m. We also illuminate it to 1 lux (to do it we need to quadruple the output or halve the beam divergence angle). This part is easy to calculate and argue.
The target having the same illumination will also appear to have the same luminance, mostly regardless of distance (it should look equally bright or dim). But the target farther away will have smaller angular size (we’ll see it as twice as small) and therefore will be arguably more difficult to detect.
Militaries and astronomers tried to quantify for centuries how the target size and its contrast from background affect detectibility - the oldest I know of is Ricco’s ‘law’ and there are also US AF studies during and after WW2 along the same lines, but that involves quantifying what humans react to - which, despite numerous studies, it’s debatable (or more like ‘it depends’). This may not be ideal for the flashlights specs.
So far I found the metric √cd/lm (or half ANSI distance divided by √lm) quite ‘illuminating’. It’s independent of the output and describes how efficiently the flashlight utilizes the lumens to form a beam on an intuitive scale. A mule is some 0.6-1, a flood maybe 1-2, moderate throwers like T6 perhaps around 6, a proper thrower like 3x21D some 12, and LEPs with little spill may be 30+.
I like it, but I recognize that it may not be ultimately what you’re looking for - knowing √cd/lm suggests the flashlight character but needs √lm to evaluate how illuminating it may be at a distance (target angular size still ignored).
If it was easy it would have been done already :-)
Or more specifically, the non-focus…this includes the foreground objects that may be more intensely illuminated, including dust, webs, and shrubs…
Maybe the solution is so easy that we’re overthinking it?
Start the thread, but give it a funnier acronym…Functional Aggregate Range Test
**it would help if the test consisted of angles of the beam, the hotspot, the spill, and the corona. Separate designations for smooth flood (diffused beams) verses uniform flood (mules) would help also.
For standard beams with a hotspot, corona, and spill, in addition to the angles, perhaps a % allocation of lumens and candela rating per section of the beam.
I.e.:
SUPERTHROWER 9000:
Profile: standard
Lumens: 5000
Hotspot, 20degrees, 500kcd, 85% of lumens
Corona, 70degrees, 3kcd, 9% of lumens
Spill, 140 degrees, 100cd, 6% of lumens
Superflooder 9000:
Profile: diffused beam
Lumens: 5000
Center, 10kcd
Lumen Gradient from center, -x% cd/degree
Terminal edge: 174degrees, 50cd
Superflooder 9000b:
Profile: mule
Lumens: 5000
Center, 3kcd
Lumen gradient: see LED specs minus transmission losses of lens
Terminal edge: see LED specs
This is an attempt to use collimators like TIR as a baseline, and reflectors etc should follow nicely.
This is especially pronounced in headlamps - in a fog or a drizzle I’m lucky to see my feet. A bit off topic, but the only method to reduce backscattering that I know is to keep the light away from your line of sight and hit those particles farther from the eyes and at a greater angle so that they backscatter less. Definitely take it off your head.
2 Thanks
Hi - thanks for the nice discussion. You bring up many important points. If I may add my humble opinions:
-
Agreed 100% - it’s a measured number for comparison, no more no less: “throw distance is specifically measured at the point where the beam’s intensity drops to 0.25 lux at the target distance”. It doesn’t define “this distance is what a human being could see object x, with size y.”
-
As you already pointed out, what is the definition of the “target” and “could see”? An elephant or a fly? And what does “could see” mean? A silhouette of an elephant, or the white of its eyes?
“Could see” at an optometrist means we could read letters of the Snellen chart. This standard, defined subjects and more importantly, SIZE of subject, is not part of ANSI throw. I believe this is why you/we have difficulty with discussion such as “throw number is x, but actually you could only see 1/3 or 1/2 of x, etc.” That is flashlight hobbyists’ interpretation, not ANSI.
2 Thanks
Definitely!
Possibly!
1 Thank
When I’m ready to formally open a thread, these ideas can certainly be included, but a separate thread would be better since the primary discussion needs to be focused on the “possible” development and creation of a new ANSI/Plato standard for functional throw.
Your right that reducing fog or dust back scatter requires keeping your light source far enough away from from your eyes, such as simply carrying your light at waist-level.
It’s also helpful to use a narrow beam source.