Maukka, Could you explain why my SC600w HI Mk3 measure 1680 lm@30 seconds in TA’s tube (Zebralight’s spec: H1-1126 lumens), but I got Exactly same M1-61 lm @30 seconds in TA’s tube as Zebralight’s claim??
Texas_Ace BLF Calibrated Lumen tube / Sphere No math skills needed - Several spheres still available
The M1 on mine is actually 46 lumens.
Thanks! now It makes sense!
You can try putting a neutral density filter inside the sphere over the sensor. I then took a few readings (under 1000 lumen) both with and without the filter to get a correction factor. Then I could measure my brighter lights. At least the ones that would fit in the hole.
The JoshK sphere is also set up to put a high powered light on the outside of the ball to shine in. There is a shield inside for this purpose. It also uses a different method of calculating the output which I’m not familiar with. I need to re-read the thread and find those instructions.
I’m getting high readings too.
Might have a solution to the problem.
I tried it without using any of the rings. I just hold the flashlight centered right above the lip of the tube.
I didn’t try all my lights but the reading seems to be a little more accurate.
Zebralight light - spec 1163 - reading 1803 - w/o ring 1233
astrolux c8- spec 1300 - reading 1650 - w/o ring 1121
mt09-35 - TA tested 7000 - reading 8090 - w/o ring 6980
Surefire Fury - spec 500 - eading 883 - w/o ring 615
after more testing. I notice that small lights under 1” dia cannot be tested without using a ring.
with smaller flashlights, you’ll need to use a ring that will give you about 1” of opening on all sides to get a decent reading.
this is only mine finding. might not work for all.
I tried it without rings too, that’s what gave me the idea of coating the underside of the ring black.
I don’t think it’s going to be a practical solution because it won’t be consistent between floody lights and throwy lights.
I think we have to have that reflective tape to help trap all of the light.
Sorry I was not around last night, still catching up.
I have suspected the readings are a bit on the high side before, which is why I always round down but never even considered them being this far off. I figured maybe 10%.
I went back and looked at my notes on the calibration process for my sphere from 2 years ago and did realize a few important details I completely forgot about.
1: I was having a horrible time getting readings to agree with basically anything, the more lights and LED’s I tried, the worse it seemed to get.
2: I ended up getting so frustrated I decided I trusted the Cree spec sheets more then light ratings and ditched basically all of the light ratings except for basic comparisons.
3: I started testing all the LED’s I had on hand, which at the time were basically all from china sources like Fasttech.
At this point I think many of you can see where this is going. I was still fairly new to the raw LED market back then and put far too much trust in the ratings from fasttech and others. One thing that stood out to me is I see some notes about XP-L HI V6 LED’s (I was/still am a big XP-L HI fan).
I now know these do not exist but these played heavily into my testing at the time as they were the LED I was buying and using the most.
So I was using the cree data sheet for a V6 emitter when they were at best a V2. This also explains why other LED’s read much too low.
According to my notes I ended up averaging the numbers and settling on that as my final calibration. This put the few stock lights I had around the expected output as compared to other BLF numbers. For example my A6 at the time got just over 1600 lumens on an HE2 battery, although the notes do not say if the springs were bypassed.
Basically I think I royally screwed up the calibration on my sphere but due to lack of comparable numbers could never figure this out.
It reads lower then many I know for sure so I never expected it to be so high.
This would be great if we could get a direct comparison. I am really bothered that the readings are not even close to being correct. Feels like I mislead people when that was the last thing I wanted. I simply didn’t have anyway of getting better numbers.
I know one went to Sweden and 1 to the UK but that is all to Europe.
Maybe I could send you a few lights to get real lumen readings, then you can send them back and I can compare? That way I would have standard lights to calibrate the ones I still have and re-calibrate any that people want recalibrate.
I have wanted to do this for some time, even looked really hard for a real sphere to get a “calibration light” when I was building mine but could not find anything.
TA, in order to make this project work as expected, is it safe to assume that:
(1) after measuring several lights from respected manufacturers, each of us could come up with our own correction factor, and end up with relatively accurate readings?
(2) that correction factor should be about the same for everyone?
Based on four of the lights I've measured so far (from respected manufactures), my correction factor would be .72
Using a correction factor only takes a few seconds to compute, really no big deal.
Yes, that would work but a “standard” correction factor for everyone would keep all the numbers comparable. For that we need as large a data pool as possible and then to average the results.
If I could get some lights measured by a real ANSI sphere for calibration proposes that would be great, then for the first time I would have a standard I trust to calibrate my sphere and figure out a correction factor for everyone.
As I am sure many have noticed, lights vary a lot from the ratings. They may be ok in some modes but it is quite rare for them to be correct in all modes with all batteries.
I have many more lights I want to measure, to fine tune my correction factor.
For those of you who have been into flashlights longer than me, which manufacturers could I trust the most for accurate lumens outputs?
Someone mentioned Fenix. The three I have are PD35, FD40 and FD20.
Or any specific models that have been tested by a reliable source?
I may be on to something. I remember reading somewhere about someone using tissue paper in there lumen tube. Maybe it was TA?
I grabbed an extra waxed paper from donut section of the grocery store this morning. One piece over the sensor dropped it a little, two pieces were closer and three pieces seem just right.
Now the tube is reading pretty much the same as the JoshK sphere or just a tiny bit higher. Assuming that’s what we’re aiming for. Also, the amount of time the light is on before taking the reading does make a large difference and might explain why it’s not exact.
Here’s how I measured my BLF GT. I used a C8 reflector. I then stuck it in the appropriate sized ring. LOL. 
GT in TA tube:
Stock tube - 3300 lm - obviously too high
3 layers - 2200 lm - pretty realistic
I know it’s not the same without the big reflector and lens, but that doesn’t fit on the 3” tube. No biggie. It’s the only light I have that doesn’t fit.
The reading is about 1/3 higher than what is expected. We’ve seen this value before. The 4” tubes were measuring 1/3 lower, which is related to the fact that a 4” tube
has 1/3 more surface than a 3” tube.
Is there perhaps some kind of mix up?
The final calibration on all the tubes should be the same +/-~5%.
Although I do agree that it is looking like the calibration that I am using is indeed about ~30% too high.
I just need to get some hard data / more data and we can come up with a hard value for a correction factor.
After look at my notes from 2 years ago and applying what I have learned in that time, I am not as surprised that the readings are high, although I am VERY surprised they are so high.
Luckily it is easily correctable once we figure out a good correction factor.
Yes i will send some lights to Finland to get them measured.
… just curious, perhaps others in the know could chime in on this: If these spheres are reading high, does that mean the HaikeLite group buy lights rated around 20,000 lumens after TA’s mods, are are all overrated? This stuff gets confusing 
thank you
It would seem that way, although when compared to other lights in the same class they compared favorably, so I am not sure what to make of it.
We will need to wait for more data to make nay educated guesses.
agree, it’s just a number, and I couldn’t be happier with my MT03 and TA driver. The driver added some ~ 4000 lumens I think, a wonderful UI, and it’s bright as the sun.
It may just be a number but I still feel horrible that I might of unknowingly mis-represented these lights.
Once some more people get the spheres we can see how things compare with a wider selection of lights.
I am also going to see if I can get some lights measured in an ANSI sphere.
I just need to figure out some very consistent lights for this. Most of the lights I have here are not regulated anymore or not very consistent.