Hello BLFers, what brands do you consider to have good output rating, the actual output closely matching with the advertised number?
I have been trying to calculate a multiplier for my DIY sphere and I have troubles getting a consistent multiplier.
Even for the different modes of the same light (turbo and moonlight already excluded), I got wildly different multipliers. I have the same problem for both copies of my Nitecore TINI2. The medium mode has a much higher multiplier than low and high modes.
The next light I tried is a Lumintop EDC01, I got better numbers, 0.264 on high and 0.260 on medium. However it only has 3 modes and low is not useful for calibration.
I will mostly be measuring warm white/ high CRI lights. So I started looking for a “calibration” light, with reasonable output (maybe more than 300 lumens) , has multiple modes, preferably, has a warm white high CRI emitter. And most importantly, accurate lumen specs.
I am considering Skilhunt E2A (SST-20 high CRI) and Lumintop Tool AA 2.0 (219C). They also don’t break the bank.
What brands are known to have reliable output specs?
Every light will be slightly different. You need one or more reference lights, measured with calibrated equipment. @maukka sold them in the past, but it seems nobody continued his work.
No manufacturer provided specification will be accurate enough. The brightness depends on battery voltage, driver components tolerances, LED binning (even the same flux hin will have differences) etc.
Thanks for the reply, I understand that the manufacturer specs are not 100% accurate, and LEDs from the same flux bin have quite a bit of difference.
I am not looking for absolute accuracy, I just want to have something to base my measurements on (to have a multiplier to turn the measured raw lux number into a ballpark lumen value). It would serve as a useful comparison between my own measurements.
A light from a reputable manufacturer would probably be better than me building a light and guesstimating the output from the LED datasheet, no?
Hmm… Not really. I don’t know any manufacturer that has so tight tolerances and will give an exact specification of the brightness. @maukka used inexpensive Convoy S2+ flashlights, measured them in his sphere and sold them together with a certificate including the brightness, color temperature, CRI etc. You’ll need something similar, someone who measures a light for you.
Might be a daft question as I’ve not built a lumen tube myself, but is the lux meter you’re using behaving in a linear way?
Also, what’s wrong with averaging a bunch and discard any outliers. If you record the raw data and use a multiplier separately to calculate output, you can always change the multiplier in the future.
There’s a lot of variables that affect output, but I’ve tested over 100 lights from various manufacturers on calibrated test equipment. Here are the consistently most accurate brands:
These are within 10% (my results) of the advertised output in Lumens and throw (candela). Olight is #1 in the list, followed by Nitecore. Note that these are on the higher end of the price scale so you’re paying for consistency and quality. The two go together. For the buget brands, Lumintop is okay and so is Thrunite.
That’s a good point. I was thinking that about 2 things, either the lux meter ia not linear or the modes of my light is not at the specified level.
The original meter I used is a Tasi 8132. I compared it to my Opple LM3 and it reads like half the value . I opened up the Tasi and sure enough, the optical filter is super cloudy, cleaning doesn’t help either. So I went ahead and got a HS1010A, now the reading of the Opple matches pretty well with the HS1010A.
I have considered filtering out the outliers, but since only 3 modes are used, taking out the data of 1 mode would remove 1/3 of the total data I have. And since both samples exhibit the same behavior, I am thinking that it is the problem of this flashlight model.
Good idea on the raw data part. I started to record the raw lux reading instead of the multiplied lumen number. This way I can recalculate in case multiplier is changed.
Edit: even though the Tasi meter doesn’t give correct lux value, the linearity seems alright. All 3 meters show that the medium mode has a much higher multiplier than the other modes.
Thanks for the input. May I ask what type of light you usually test, keychain, single cell, soda can?
The Nitecore model that I used (TINI2) is relatively low price (compared to the rest of their lineup) and it has a built-in battery. Maybe that’s affecting the accuracy.
I have searched for test results of the light, haven’t found much data. 1lumen tested the light, low and high are pretty on point, but medium is 75 lumens (Nitecore specs: 65 lumens). However, if I use 75 instead of 65 to calculate the multipler, I would get a even higher multiplier, which is more off.