I did some testing over at CPF to compare how the individual components affect output . Maybe some of you over here at BLF might find the info useful.
Nice quantifiable evidence you have there! Thanks for publishing this, because even though I don’t use maglites, I like that you tested reflectors and lenses which is a lot more useful to me.
Although you called the reflectors ‘KD reflector’, what is KD an abbreviation of?
Oops I knew I forgot something, KD stands for Kaidomain.
Edit: fixed over at CPF.
Ah, I thought it meant Kaidomain - just wasn’t sure.
EDIT: Also, I think I have a theory behind why the OP reflector had a higher reading than the SMO. The SMO was more focused ray of light, and less dispersed, meaning less light would have hit the light sensor. The SMO had a wider scattering of light, meaning more of the light would hit the light sensor. This could probably be solved with an integrating sphere.
But other than that, the lens numbers are good (and expected).
Thanks. That should ease everyone's mind about needing a super lens. Only 3% difference between the lot of them. Not enough to ever worry about and certainly not enough to ever begin to see.
Thank you for sharing your test data. Very good information. Well Done lilkevin715.
I have measured the difference between a plain lens and a proper multicoated lens at around 7-8% on several different lights/lenses… the same gain that you get with a LED flux bin increment.
BTW, a proper multicoated lens is an optical grade lens. Typical double coated lenses sold for flashlights may only be around 5% better than a plain lens.
Very interesting results, thanks.
I wonder if you’d have seen the same percentage difference on High. I would assume so (and understand why you didn’t use it). Still curious though.
Assuming I could keep the LED at a stable and relatively constant temperature on high (yea right… fat chance 
) then I would say with 99.9% certainty the percentage loss would be the same. I’ll leave the 0.1% chance out there for someone to test this and prove me wrong. :bigsmile:
Nice data! Thanks.
Need? 
Absolutely?!
I can see it! The colored reflection off AR, that is… 
Love it. It adds bling!
Thanks for doing this testing, and feel free to post your results here as well. I have often wondered what the actual losses were from reflector and glass. I have used the ‘rule of thumb’ 25. Your data looks like the rule of thumb should be closer to 17.
EagleTac publishes their D25 Lumen both emitter and ANSI OTF: they show between 27-34%. I wonder why they are so much lower? Is the smaller reflector a factor?
Thanks for bringing some hard data to the discussion.
For me personally I wouldn’t use a “rule of thumb” for OTF percentage for LED flashlights. As you can see from my data various combinations of components can yield varying results. A percentage range would probably be more appropriate.
I’m guessing the low percentage is probably due to heat related output sag. I’m assuming the reflector and lens they are using are of decent quality. On a related note I find it hard to believe any manufacturers specs anymore (even ANSI numbers). When I was calibrating my lightbox I ran into a discrepancy with one of my lights… it was a Eagletac as well (T20C2 MKII “S2”). You can read about it in my Lightbox thread (post #7). After that experience I always look at manufacturers specs with a grain of salt.
Why, oh why, are you mucking around with a lightbox? A 16” sphere can be had for around $40 shipped… http://www.smoothfoam.com/product/10073.html
Also, there is not much need to worry about external light getting into the sphere. My sphere sits on my kitchen floor. It is not covered or lined. The kitchen is lit by 10,000+ lumens of LED lighting. Around 0.5 lumens of stray light makes its way into the sphere. Not likely to be a problem except for very low level lights.
This is my data acquisition system: Any interest in a LED/Battery analyzer device?
The control software can compensate for background levels by subtracting the minimum light readings (i.e. the background light) from the data. For the real finicky tests, I turn off the lights. I can also mount the sphere light sensor onto the outside wall of the sphere and use the sphere wall as an attenuator. This increases the maximum measurement capacity of the sphere from 5000 lumens to around 600,000 lumens.
If you send me $40 texaspyro then I would consider using it.
Keeping the Scientific Method in mind… if at all possible it is always best to eliminate any variables from affecting your test results. While the syrofoam sphere is convenient it does have some drawbacks as you have mentioned. If a variable (even small) such as external light can enter the test area, then test results can be compromised. Sure you can try and compensate for external variables but it adds another element of inaccuracy. Also on a related note if light can enter the test area externally, then light can also be lost from the inside going out. This is why I chose to use aluminum foil on the inside of my lightbox as it acts as a control buffer.
With my lightbox I can consistently measure “moonlight” lumen levels well below 0.5 lumen irreguardless of the ambient external lighting conditions. Theorhetically my lightbox can measure as low as 0.008 LBL, of course I haven’t found a LED flashlight that actually goes that low. The lowest so far that I have measured (and verified by others) is a Thrunite Neutron 2A that does 0.1 lumen.
Next up is some Solarforce L2 testing. The test results for this upcoming experiment should be more accurate as the testing range (lux reading) will be lower compared to what was used for the Maglite testing. With my light meter the lowest resolution from 200-2000 lux is in 1 digit increments. With lux below 200 the resolution can go down to 0.1 lux increments. I’ll post a new thread for the Solarforce results.
If you are that concerned with accuracy, then you should definitely NOT us a rectangular box. It will introduce a lot of errors in the readings. A sphere is the only way to go.
You could coat the sphere with foil, etc (on the outside) to help block external light, but for most lights it is not needed. Same for blocking the access port. I can use the plug that I cut out of my port (4”) as a filler, but again testing showed that it provides little real benefit. I have used it (along with turning off the kitchen lighting) for small/low lumen lights. The software “auto-zero” capabilities works just as well for subtracting out the ambient lighting.
When I was using the Rohm light sensor, I could get around 0.001 lumen resolution. I have switched to the Taos sensor which gives around 0.01 lumen resolution.