I use a 0.34 factor with mine - same exact setup as you have Dale, meter and everything... I know I don't take temp into account whatsoever which is a problem. At this time of year, my office may be 63F when I start, then my room heater warms it up to 70F or so, so there's a lot of variation. Sometimes my readings of the same exact light/cell varies quite a bit from one eve to the next morn.
I basically cal'ed mine against the same lights as rdrfronty and manxbuggy1 used/tested with their's, so my cal banks on the much better array of lights tested with their PVC lightbox. We kind of came up with the 0.34 factor after some collaboration.
I spent this morning dialing in a conversion factor. The first problem I had was that most of my lights are modified...
I managed to find several stock lights, & corresponding numbers from experienced members/reviewers here on BLF. It's not a large data set by any means, but it seems to scale well from small to large lights.
Below are the lights I used, & lumens from my pipe-meter after conversion factor, followed by other member's data;
This has been the main reason why I decided against a permanent mount for the sensor. I find the readings can be off by >10% due to temp fluctuations (not insignificant for higher powered lights). Will be making some additional tweaks to this and also some aesthetic mods for a "storage" mode for when I'm not using it for measurements.
EDIT: @Tom E - Apparently you can't be quoted properly because of the space in your username...
EDIT2: Weird, had to go to HTML to fix it...
Not sure if you saw djozz's thread on his rather high end light meter - real interesting stuff. Seems like any meter under $2-3K will not measure tints accurately - think warms measure lower thena they really are, cool whites measure a tad higher, etc. Here: https://budgetlightforum.com/t/-/28689. I was gonna buy a higher end meter ($150 or so Extech) but didn't after reading this - no point accept for features. The accuracy seems like it will be no better than my $35 meter.
Guess I should thank you TurboBB for the PVC elbow light box idea, copied it a year so ago after I saw it here… here’s my rig:
I’m using 1.5” elbows as my largest light is only 40mm. As luck would have it, there is no conversion factor on my meter’s x10 setting (lumens = lux) and I only need to mentally move a decimal point for sub-/low- lumens (meters down to 0.01 lms). For some reason, I feel my rig starts understating increasing more as lumens increase - notably over 300 lms - but I’m not really into max output or bright lights anyways. I’m a sub-/low- lumen enthusiast and primarily use my light box to test low lumen efficiency in side-by-side output/runtime tests which I compare results on a lumen-hours basis (lumens x hours). I like to know which manufacturers lie, and which can be trusted (and marketing integrity, warranty and CS seem pretty correlated to me).
For calibration, I use Quark XMLs - my QPA2-X specs seem the most accurate across most modes. Also use a QP2L-X, which is a little “off” from spec, but I match the various outputs that CPF reviewer “ti-force” gets. Ti-force is the only reviewer I’ve seen that claims ANSI accuracy using laboratory tested calibration lights, and based on my own collection of lights, I agree with his scale. Selfbuilt is far too lenient (IMHO) but I do try to reconcile to SB’s RELATIVE results. Here are the accuracy statements of the two reviewers:
NOTE: I only use lower modes to calibrate since I find max far too dependent on sample variation and especially battery condition (age, state of charge, time into run, etc).
Also wanted to post some of my earlier light meter contraptions, in case people are interested in lower cost alternatives with existing tools…. all are significantly more objective and accurate than using the naked eye (IMHO).
I started with a DSLR and variation on the ceiling bounce. By greatly closing the distance between the bounce and ambient sampling surfaces you have more light to work with improving accuracy and low lumen metering.
Then onto the PVC elbow lightbox
Then onto a free photography app in my smartphone (“Pocket Light Meter” for iPhone)
FWIW, the photog method is calibrating to, and solving for, shutter speed - your lumen output is then the reciprocal shutter speed. The 1/3rd photog stops can seem like wide increments, but in practice you find it’s closer to 1/6th since you can tell where the midpoint is between two stops (two settings satisfy the same exposure). Alternatively, I have a formula to convert EV (exposure value) to lumens. The light meter app reports EV down to 1/100ths, and the conversion formula works quite well vs my lux meter reading, except at higher lumens - it doesn’t seem to understate like I feel my lux meter does (although overstatement may be a possibility). In any case, it’s at least going to be accurate on a relative basis.
I notice pipe dream users are using lux to lumens conversion factors of around 0.34 to 0.38.In my sphere , I’m using a factor of around 0.034(calibration is an ongoing thing).That is to say , about 25 lux per lumen , and this seems to correspond to other sphere users.My 480 ansi lumen SWM V20C records 12710 lux.You guys must only be capturing about 2.5 lux per lumen produced , that’s 10 times less!Am I missing something here?
Billy - We use glass on the ends, and the PVC is painted black on the outside, and we block the light around the head of the flashlight, as shown in post #33. Is yours the same setup?
I built one of these because this thread inspired me. I’ll post pictures once I finish making the base and mounting it. It’s looking like my conversion is going to end up being .8-.9 of the lux measurement. One difference is that I sanded the inside of all the pipe fittings to reduce reflections.
Complete unit, with some handles so it’s easy to move around.
The sensor in place in the end cap. I modified it on my lathe after I solvent welded the large reducer to the small one so that the sensor is unobstructed from the sides.
A nice friction fit holds it nice and firm.
A better view of the cut down end for the sensor.
Sanded a flat spot on the glass with a silicon carbide belt so I can easily pop it out with a flat head screwdriver. It’s hidden under the trim ring, so not visible when installed. Using 1/4” plate glass for now. I’ve got a disc of low iron glass on order with a local glass shop. This one was $4 and I had it in a day.
Thanks for the inspiration guys, this is going to be handy to measure the freakishly bright lights I just finished up.
Oops, yep! Spheres are a different animal, but with all the hundreds of lights we got published #'s on, rdrfronty/manxbuggy1 and TurboBB, the PVC pipe fixtures seem pretty consistent.
I don’t suppose that data has been collected into a single spreadsheet or thread has it? It would be massively helpful if there was a single archive we could all use and update. Maybe with a notation of who’s pipe measured the light. Just thinking it could be a pretty useful resource.
Got my glass from a local glass place. I’m using 1/4”(about 6mm) plate glass, 4.5” (114mm) in diameter. Cost me $4 + tax, and I had it in my hand 16 hours after I asked them to cut me the piece. I’ve got a piece the same size on order out of low iron glass when they get some in. Will be more expensive, but I’m not expecting it to be more than $6-8. Low iron glass is supposed to be more transparent (less green hue)
Some numbers from some lights I’ve measured. No solid (in my mind) calibration numbers yet, as I don’t have a good enough cross section of lights to feel comfortable with anything I’d call accurate, but .61 seems to be in the ballpark. (that’s the factor used to get the estimated lumens from the max lux in pipe values)
I’d love to get some feedback or suggestions on where it seems accurate, or way off base.
For reference, the conversion factor was 0.61, at max measured lux. I re-did the equations to reflect output at 30 seconds and they seem reasonable, but perhaps the factor needs to go up just slightly so the stock lights are closer to the average that other members have measured.
Edit: I take it back. It’s possible the number needs to come down a bit. Perhaps about 0.55 or so. That would put the dedomed ThruNite tn12 2014 at around 1000 lumens, which seems more reasonable.
