A funny thing happened on the way to the forum........

Off-topic Chatter
1

Hi All,

I have been thinking about flashlight/torch (I'm going to use 'torch' from now on - it's shorter) outputs, as you do. I have noticed quite a difference in stated lumen output across many torches of differing size, capacity, design, etc. Then I get reading how some people question the manufacturers stated output. Delving further, I read how some people setup their test equipment and rigs etc, and then rattle off the measurements they made and how these measurements vary (considerably I may add) form the manufacturer's stated figures.

Surprise, suprise, their measurements are significantly less than the manufacturer's stated claims. Then a clamour arises on forums like this about how manufacturers make unrealistic or inflated claims about their torch's output. How laboratory testing varies from real world conditions, blah, blah, blah, etc, in an effort to discredit the manufacturer's stated figures. This made me doubt the claimed outputs of my rather modest collection torches.

Well now, I thought time to do some testing of my own to see if the manufacturer's claimed output of one of my sample torches was true. So I looked in to how to test light source outputs. Well, what do you know! It turns out that to do accurate testing (and the key word here is accurate), you need some real expensive gear and a reasonable amount of know how. Hmmmm! This doesn't look like it is going to be a feasible excercise. So then I revisit how many of the regular Joe's out in the real world actually did their testing. I saw all kinds of jury rigged set ups involving pipes, opaque and translucent sheeting, black masking/packing tape, - a real Macgyver operation. After all this, I read in many places just how inaccurate these testing setups and practices are in obtaining accurate output figures.

If this is the case, don't these people realise how inaccurate their reading are? Don't they realise that they are on quicksand when trying to base an argument on how a manufacturer's claimed figures are inflated/unrealistic when their own figures are much more inaccurate than the manufacturer's figures? The manufacturer has to test their products to the ANSI/NEMAFL1 standard to be of credibility. If people doubt the manufacturer's claimed outputs, then they should test the same torch using the same or higher standard (if there is one), which means using the proper expensive test gear in an exacting environment. Who do I believe? Regular Joe's who have every right to be suspicious of any manufacturer in whose interest it is to inflate claim outputs or do I believe the manufacturers who have to test to an international standard and if actually tested to that standard, the claimed figures achieved? Of course, I'd like to check figures for myself.

So, how do I accurately (to within +/- 0.5% tolerance) test a torch to see if the output actually is what the manufacturer claims it to be? Any suggestions/ideas?

2

this is how I do it;
to test a light’s brightness
from a distance , point the light at wife’s face
/
the louder shell yells at you = the brighter the beam :person_facepalming:

3

What decibel meter would you recommend…and what light should be used to calibrate it?

4

Without spending thousands of dollars on an integrating sphere, you can't achieve that kind of accuracy.

You can, however, get a flashlight from some of our BLF members who own one of the expensive integrating spheres. That light (which has been tested on the integrating sphere) can then be used a reference light - which can then be used to make comparisons to other lights - or to check the accuracy of one of the home-made measurement contraptions, like the Texas Ace Calibrated Lumens Tube, which is generally accepted as accurate to within 5%. But .5% is a whole nether level of accuracy.

5

I don’t get caught up in all the numbers. All I really have is one light that has a known output, as tested by someone else’s very expensive equipment. A dark room and a light meter allows me to see how much, percentage wise, one light is as compared to the known light. I’m sure that still isn’t even perfect. It doesn’t really have to be.

I am fairly certain that tolerances in mass produced lights from China will cause varying degrees of brightness from one light to another of the same model.

If I have a light that claims 1000 lumens but only really puts out 900. By eye, I won’t be able to tell the difference. I’ll just be the weird guy in the dark closet trying to prove it with a light meter. :slight_smile:

6

Just go by reviews. A few lumens here or there isn’t going to matter.

7

Yes, people building their own integrating spheres tend to be aware of the limitations.
Yes, the acccuracy is often good enough to discredit makers who fake their numbers. It depends on how well is the sphere built.

Myself when I see some performance measurements I treat them with a big grain of salt when they are done:

  • by most members (there are literally several members who I remember as experienced owners of good spheres - I trust their measurements to be within 10% and one that has a pro sphere with outdated calibration - I trust his measurements to be within 5% though I expect them to be much better than that)
  • by manufacturers
  • by non-members

Yes, manufacturers frequently have equipment that’s better than what we build on a budget. At the same time they have an incentive to cheat - and even some high profile manufacturers have been reliably caught doing that.

BTW, yes, FL1 is the measurement standard. It is very outdated and unsuitable for measuring many modern lights.

8

So you are saying I’m on quicksand if I claim these lights have unrealistic figures without having measured with expensive professional gear?
https://www.amazon.com/Skywolfeye-Tactical-Flashlight-Waterproof-Flashlights/dp/B07D389848/
https://www.amazon.com/Flashlight-Zoomable-Case，5-Mode-Waterproof-Performance/dp/B01MRIGLH9
https://www.amazon.com/Flashlight-Libison-Handheld-Resistant-Rechargeable/dp/B07HYPHZK7

9

Welcome to BLF, whiteheat1963! :slight_smile:

0.5% tolerance in light measurements is surrealistic, even 1000ths of dollars worth of equipment will not get you that, and it requires extremely rigid procedures. Light measurements are notoriously difficult to do well.

A 10% tolerance is realistic I think if you do everything right. It requires not so much expensive equipment (although expensive does help) , it has a lot to do with clear thinking, clever procedures and strict monitoring if those procedures keep working well. Not just test your light sources but also keep testing your equipment and procedures.

Do not get fooled by McGivered test equipment, if thought out well and used in a smart way, they give very usable results. I suspect that a fair amount of flashlight manufacturers have very expensive light measurement equipment but do not properly use it, their results may be less accurate than what a clever BLF-member can achieve with styrofoam balls, cellotape and proper procedures.

10

What makes the OP think that the Flashlights themselves all same model coming off same assembly line on the same day are within 0.5% tolerance?
They’re not due to the fact that the LED itself does not have that tight of spec.

From Zebralight to Convoy this is just not going to happen.

11

Which is why crying over a 10% output difference, or even 20% is stupid

Not only do mcgyver setups have variance with actual lab setups and conditions …

Even IF the difference is real, you really dont notice it in real world usage

The difference between 1000 vs 1100 or even 1200 lumens wont be noticeable if you close yr eyes for a few seconds :wink:

12

Exactly. Isn’t there a phrase that has floated around, called “tint lottery?” The idea being that the same emitter and bin designation can have variation. I can’t locate the topic, but I have a distinct memory of a Zebralight discussion about emitters regarding an SC700W. There were some complaints about emitter tint. Someone posted about having two of these, and side-by-side the beam tint from the same emitter type in each light was slightly different.

13

You don’t. Nobody does. The LEDs are only specified to a 7% tolerance. Basically, all spec numbers in the entire flashlight industry can be assumed to have at least a 7% margin of error. Add in the variation from other components, and it’s probably more realistic to assume at least 10%. And that’s before taking into account that there are multiple different definitions of what a lumen is.

People can get pretty accurate measurements, even with cheap gear, if it’s designed well and operated well. But none of the numbers are very precise. If a light is specified to produce 1234 lumens, that means it’s probably somewhere between ~1110 lm and ~1360 lm… and the actual value is expected to be different even for two lights of the same model.

Getting higher precision in this field generally requires many data points to be able to determine where the signal is among all the noise of random variation.

14

OK people, thank you all for your sage words of wisdom. I feel suitably schooled. Forgive my ignorance. However, I have a couple points to raise.

First, I didn't realise that there is no absolutely precisely measureable torch output, either between identical examples of the same torch or in any observers/testers ability to measure any torch output that accurately. A reasonable tolerance posited was 10% I think. Well, on a 1000 lumen torch, that's upto 100 lumens either way, which as pointed out will hardly be noticeable. However, on a 100,000 lumen torch, that's up to 10,000 lumens either way. I don't know if that's noticeable by eye but I'd be concerned that my stated 100K lumen torch is actually only delivering 90K worth of lumens. Of course, I'd be smiling more if that same said torch was actually outputting 110K worth of lumens. But this brings me to my second point....

Most of you have asserted that either you or others can and have made relatively accurate torch output measurements, perhaps more accurate than the manufacturers themselves. Well, how do you actually know that to be true? If there is no absolute method of referencing and the testing procedures and testing equipment can vary so much as to substantively affect the results, how do you really know that your own results are that accurate or more accurate than anybody else's results? My point is that without a scientific procedure carried out according to a proven recognised standard, that is repeatable by others and with results obtained being either identical or almost identical between different testers, any claims to having obtained accurate results seems rather erroneous, no matter who the tester was, what they used to test and the method they used to test.

Therefore, it matters little if people claim their testing has produced 'more accurate' results because none of that is verifiable in absolute terms. However, when a manufacturer states their figures which have been obtained under proper lab conditions, assuming they claim the ANSI/..... standard, a casual observer has to believe those figures over those claimed by others who's methods and equipment cannot be verified as meeting or exceeding the conditions stipulated by the international standard. So, why wouldn't the casual observer believe the manufacturer's figures over and above those made and claim by others without knowing whether the recognised standard's conditions had been observed? This assumes that the manufacturer: a) knows what they're doing and b) are truthful about the results they obtained, either of which I'll grant cannot necessarily be taken for granted.

15

It would really help the group to envision whiteheat1963’s comments as coming from Sheldon Cooper of Big Bang Theory, makes it kinda fun that way at least.

Now then;
They would and should believe the manufacturer, if they followed the protocols for ANSI however flawed that system may be, it’s all we have.
Your points are squat.

Nowhere do I read that anyone said their system was
“perhaps more accurate than the manufacturers themselves” as you have said.

You are Trolling and you have been found wanting.
Troll on somewhere else.
Try again Skippy.

16

New members often seem to be the ones that make the most noise… then while the majority of them later settle down, a few decide to try knocking lamps off tables & see what happens. :person_facepalming:

How can we know any specification is real if we can’t test the manufacturer’s claims with similar equipment? Hmmm…
And now I’m worried if my car really has 205hp, or that it might just be 199hp. Won’t be able to sleep now…

17

Whiteheat1963, are you referring to posts in which the poster has directly challanged a manufacturer’s output claim because of measurements the poster made on homegrown apparatus, or have you seen posts about the inaccuracies of home-testing as well as posts refuting manufacturers claims, and linked the two?

If the former it would be useful to see the posts which directly challenge manufacturer’s claims as a result of home testing to be able to comment on them.

If the latter, well, if a manufacturer claims that a torch with optical and electrical losses has the same output as the bare LED under lab conditions then you don’t need to measure anything to know that the claim is bogus.

And, welcome to the forum :slight_smile:

18

Yes, it was the latter. I don't exactly or rather, that precisely know what a manufacturer is claiming, so I don't know what is real, unreal, bogus, really close with minor inaccuracy or just imprecise for whatever reason, and thanks for the welcome to you all who have offered a welcome.

19

Hahaha! Love it. No, not trolling, never have, never will. However, I get the trolling label. Even so, my question/statement was legitimate and stated in a perfectly civil and balanced way. Well, I guess the Asperger's syndrome in me comes out as Sheldon Cooper - who ironically I think would never troll. But thanks for the welcome nonetheless.

20

Hi mate. Thanks for the welcome and all this extra info that I ws unaware of. Appreciate it.