Indeed, the only reason I trust Maukka’s numbers more then most is that he has an expensive standard lamp to compare against. This is the single most important factor to a quality sphere.
I am very curious to see what the numbers work out to. Some early data says it should agree with what most others are seeing in the high .6x range.
This is probably what they’re actual numbers are showing. It doesn’t mean it’s super exact. So you have to take it with a grain of salt. If it says 842 lumens then you know it’s probably 840 ish. Maybe even think of it as the mid 800 range.
Don’t beat yourself up about this. It’s not that big a deal.
Yeah, I know that and it doesn’t bother me directly so much as what the novice members among us think when they see those numbers. It propagates a false impression of exactness that we simply don’t have the capability of achieving.
For me the biggest attraction to your sphere project is the fact that a group of people here will have a standard to compare with each other - not some exact number.
Getting a lumen number is complicated… everyone
get different reading and different output…
I just realized that you don’t need to test the “Maukka measured” lights across your several tubes to find an average. Since those are all based off your original tube (plus or minus 5%) it means the original tube is the average. Duh, why didn’t I realize earlier. :person_facepalming: lol
No two identical flashlights ever give the same exact reading for a single integrating sphere as well. Cree bins their emitters plus or minus 7%, plus you have other smaller factors. :partying_face:
Maukka, since your test lamp sits in the middle of your sphere, how does its output compare to having it shining in through a port with a white background around it similar to how you would measure a flashlight shining into a port?
I guess it’s not very relevant for our purposes, but I was just curious.
Also, what did your standard light cost?
Does it also need a costly power supply to operate it?
The standard calibration light was under $100. The power supply doesn’t need to be super accurate as long as you have an accurate multimeter to verify the current stated on the calibration certificate when calibrating.
I used the calibration lamp inside the sphere in a E27 holder. After that I put the Emisar D4 in there at 100% 7135 for flashlight calibration (subtracting the effect of the D4 itself using an auxiliary light). After that I moved the D4 outside the sphere and measured the lux through a diffuser on the input port. I’ve noticed that a diffuser is the best single thing to improve the integration.
Now I had a calibration light for measuring other lights. I also measured the Haikelite SC02 as a second calibration light because its readings are very stable. Every time I measure lumens, I first check that the two calibration lights produce a calibration factor within 0.5% of each other. So I know if either of them has drifted for some reason.
For measuring lights I use an auxiliary light again to subtract their effect on the readings. With the diffuser in place, the small (<30mm) lights don’t really affect the readings, but for bigger lights the effect is clear (bigger readings when the reflector bounces light back to the sphere).
Great info and also glad to hear that I am on the right track with the diffusers, I noticed the exact same results when I added them to my sphere.
I finally was able to get caught up with all the flashlight mods, I have some other things to get caught up on over the next few days but I hope to start playing with ways to make a correction factor for the spheres after that. Hopefully the lights from Maukka will be here by then.
Has anyone tried sixty645’s method for calibration?
I have not tried that method exactly but I did try calculating the lumens early on in my sphere development and quickly dismissed it as the spheres we use are not nearly precise enough to mathematically calculate the lumens.
With a very high end “perfect” sphere, it should be possible though.
Do you have any results from the testing?
He did send me some preliminary results and they were right in line with everyone elses numbers. Around .67 - .68. This is looking to be the target range.
I am going to go hunting today for some options to reduce the readings by that much. They have window tint film that blocks 30% of the light that should get the numbers really close, I am going to see if any local shops have some scraps I can play with.
Ok everyone. Good news, the lights from maukka showed up! Huge thanks to maukka for these!
So after testing them along with some other lights the numbers are lining up very well.
Like most I found .68 is the magic number.
So now to correct the spheres. I spent most of the morning looking at options and trying to get this to work in the simplest method possible. Including going to window tinting shops and getting some samples.
After a lot of testing I think I found a method that should work, I still need to test it on several more but the random sampling of 5 spheres I have been using are working well with it.
Even after sitting for awhile and being moved around several times and having these correction disks added they are all reading within my 5% goal (although 1 of them was right on the edge). I think I am going to raise the official specs to +/- 7% tolerance simply because I am sure there is an outlier that falls in that range and that is the Cree tolerance range anyways so why pretend that these can be more accurate then cree.
So the setup I am working with now are 2 of the diffusing sheets I used inside the sphere.
Depending which way they are orientated they can have a correction factor of .70 or around.66. So either 2% high or 2% low.
I think I am going to recommend going with the .70 correction factor and then simply rounding down your numbers since I recommend rounding them off anyways.
Plus I find that my sphere tends to drift a bit low every few weeks just due to dust and such. So that should balance out the readings to make them basically spot on since these will get dirty no matter how hard you try to not let them.
I am still going to try some other things but so far this is the closest I can get without having to stack 4-6 layers in just the right order and orientation of various materials. That is simply too much room for error for my tastes. Even how clean/dirty the disks are plays a noticeable role in the results.
The opaque diffusing sheet gets it really close with a single sheet but it is still around .65~~.66 one direction or .70~~.71 the other. It is simpler but slightly worse off.
Plus by sending everyone 2 of these sheets the spheres could be dialed in a bit better on a per sphere basis if a good “standard” light is ever made for individual calibration.
So opinions, how does everyone feel about a .70 correction factor?
Would you prefer .66?
Should I keep searching and find a .68 correction factor even if it is much more complicated leaving room for error?
If everyone agrees on a .70 correction factor I can get another diffusing sheet and get it laser cut this week and have it shipping out by next week.
Personally in my own sphere I will be using .70, it is good enough for me. I generally find that the simplest option is the best option. Plus since I round down anyways the numbers should be good.
Thanks for everyone having the patience to wait for a fix.
Close enough and I like simple
Good news TA! :+1: 
I’m okay with .70 as we can round down a bit if necessary.
Do these sheets your talking about fit in front of the light meter? If so, it seems very important that it be thin. As the meter gets pushed out it effects the readings.
I just thought of something. If you use that sheet that gives a .70 correction factor we can add a tiny spacer to the meter to push it out a tiny bit and get .68. At least I can. Even a thin 1mm thick clear lense might give that bit of extra drop we need.
Yes, they sit in front of the meter and stop against the small lip inside the fitting.
They are thin enough to not effect the meter placement, the meter does not reach all the way to the back of the fitting anyways.
Although I did just realize that the 4” spheres could have an issue getting these disks to stay put as there is no lip on those fittings to hold them in. I will have to think on that.
Could possibly hot glue the disk to the meter itself but that is not as simple.
I am going to go grab a 4” and do some testing.
