Good question - it was only meant as a temporary ‘collaboration’ solution. Individual ones will hopefully be replaced by one community calibration. I will go remix the choices around the latest values. Give me a few minutes.
Here’s what we have now. I should have done it right away.
Zebralights are not the only light that doesn’t stabilize for me. Nearly all of my lights have some degree of continuous fall off with my current device. I’ve tested probably 20-30 lights and high/turbo usually start falling from “on.” There’s maybe a light or two that don’t, but I can’t think of them off hand.
Interestingly enough, there are some lights, a rare few, that the lumens actually climb for the first 30 to 60 seconds before stabilizing. The light I built yesterday and today is one of those. Even taking amperage readings, I had to wait a few seconds for the number to quit climbing instead of falling. Sheer cooling mass I think, as most start heating up and dropping virtually immediately. I’ve taken lumens readings from probably a couple of hundred lights. The one thing I’ve always seen is that it’s difficult to duplicate the first reading from a build. Just never get the same readings again, I don’t know if that’s the emitter wearing in or what. Probably so, as I tend to push em for all their worth. 
I want to say the X5 and X6 are those rare lights where they seem pretty stable during the first 30 seconds. I did readings on those recently, and I don’t remember there being a falloff.
Received my sphere, and all is well. Josh you did an excellent job on this project! Well planned/executed. Efficient use of materials. Very nice. You most certainly didn’t make any money on this, so thanks for the big contribution to this forum. I appreciate it very much.
I started looking around my house for a light to test, and had a hard time finding any lights that weren’t modded. I did however find my old Zebralight SC600 MK2. This was the first version of the MK2 that used the XML, instead of the L2, that used the XML2. I believe the max output on my light was factory rated 900 lumens. My sphere, using the Fenix calibration settings, showed max output at 687 lumens. I used an 18650B4.13v. Since then I noticed an update on the sphere calibration. When I enter the highest possible number (.0064) my calculated lumens jumps to 887.
Next I tested the only other stock light I could find, a Zebralight SC52 with the XML. Used an AW14500@4.13v. I believe this light is rated at 500 lumens with the 14500 cell. If I enter the .0064 multiplier into the sphere calibration box I end up showing an output of 537 lumens.
Once again, excellent piece. Not sure how I should proceed with sphere cal. inputs?
I think the general consensus is in the .0059 range at the moment. We need some more people to toggle between them and give their opinion. I know the spheres agree with each other. Now we just need to find the best multiplier.
Hopefully I can contribute next week. I have 2 fenix lights and a Zebralight that I can use.
Thanks mhanlen.
And as a reminder to all, be careful of ratings that are not ANSI, ratings that are not Out-The-Front, weak batteries, or a leaky sphere hole. Then post the multiplier you trust the most in this thread.
Ok, got mine today and started messing around with it. It’s easy to use and put together, so thanks Josh. Now onto the readings. My light meter calibration is 550. I tried a Nitecore light and the readings all seemed fairly low on that, by quite a bit. Like I’m getting 820 on High with a fully charged battery. My old device read 1084 lumens. I started out using .557. I suspect that all of the modes on the MH20GT light are not properly rated… so maybe 557 isn’t the proper multiplier. I feel like that’s low. Plus I don’t trust Nitecores well rounded numbers compared to Fenix or Zebralight.
I have a Fenix LD15… it’s an 8 lumens on low and 117. So I was getting 6 lumens on low with the new device (1050 lux) and 115 lumens on high (17760 lux). I had to use the highest .0064 multiplier and I’m still short.
Alright, I increased the range like crazy.
Thanks! I’ve got some more testing to do, to find that sweet spot. I will say this, even ANSI rated lights can’t be right 100% of the time on all modes. I’m coming across some discrepancies within lights on some modes. Right now I’m concentrating on a zebralight and a fenix on some of the lower modes and trying to find some consistency.
Yea I don’t know what to believe at this point. ha. Graphing all the results in one Excel graph shows how crazy it all is.
Does anyone know an ANSI approved lab I can pay to rate one of my lights?
I think the key to wrangling any of this stuff in is getting lots and lots and lots of data. I mean Cree’s efficiency bins span what, a 15% range? So any two factory lights, all other things being the same, could vary by as much as 15% and still be completely within the specs with nothing funny happening.
I agree. Now if there were just a few ANSI lights that everyone owned. I am currently making cutout templates that are foil backed for head sizes that are different than the 3 included. My wife has been asking me questions all day about what it is I’m doing. My Zebralight and Fenix unfortunately do not fit those 3 sizes.
You think that’s hard to explain, try explaining 11 of them in 154 different pieces to your wife! That takes all the square feet in my house. 1 sq ft for each sphere half, and 1 sq ft for the blick box and 1 more for the new uline box. That’s like 144 sq feet!
I received my sphere in tip top shape. And I agree with everyone else that got there’s….this really is a well made and thought out system. I like the easy to change templates for the different size lights, the bracket that holds the light meter, and the bracket that even holds the sphere is even cool. Nice, cheap and efficient way to house the sphere, the cardboard box allows me to finally do some light testings in a bright sunny room rather than go in pitch dark closet or bathroom anymore haha.
Everyone that seen the sphere are so curious and says the same thing, “What is that? It looks expensive. Are you doing some kind of experiment?
Well anyways, here are my readings to the only stock lights that still have.
I have meter cal. 555, and chose 0.0060 as the sphere cal. because that gave me the closet reading to the zebralight with only a 1 lumen difference.
zebralight sc62 - manufacturer reading: 1000 lumens blf sphere: 999 lumens
thrunite tn11s - manufacturer reading 830 lumens blf sphere: 852 lumens
olight m20 - manufacturer reading 340 lumens blf sphere: 258 lumens
olight s30r jav -manufacturer reading 900 lumens blf sphere: 969 lumens
sunwayman f30r -manufacturer reading 880 lumens blf sphere: 1069 lumens
sunwayman c23c -manufacturer reading 1000 lumens blf sphere: 1146 lumens
Here’s some info on the math the app does:
Meter cal - Corrects for your meter errors. Don’t fiddle with it.
Sphere cal - The slope of the calibration line. (aka the multiplier.) Choose one, and always use it.
hole reference & wall reference - There is a wall multiplier, but this ratio corrects it for each light’s geometry. It ensures a hole reading and a wall reading match. You are creating a matched set of reference numbers.
I found one more stock light that I need to test.
Meteor M43 with dedomed xpg2’s
As funny as this sounds, I honestly still don’t really understand how to do the wall readings with these larger and higher lumen lights.
I know u showed us in the video, but that’s with a smaller light and it may just be the same thing, but I’m not getting it.
Can you make another video with a larger light that does over 1000 lumens?
Or can you or someone explain in brief steps the order I needa do please?
Thanks!
You still need to choose a meter-cal and sphere-cal. Then just set the light to a medium or low and do the same measurement through the hole and through the wall. Those two lux numbers are the reference numbers the app needs to prepare. After all that is ready, you can do brighter wall-only measurements for that light.