Currently we have the ~270Lm Maukka reference.
I have wondered how that translates to measuring a 10K, 20K (or more) Lumen light.
Just how linear are our Lumen Tubes/Spheres/Ice-chests?
Several designs have problems with the spot vs flood beam conundrum.
There are places one can send a light for evaluation to be tested in a mega$$ setup.
And no, I don’t know what this actually costs….
But what is needed is a light that produces a constant output at various higher Lumen outputs.
I envision something that uses perhaps a CPU cooling block and a constant current power supply that can be set to various outputs.
Once one is tested, it could then be used to “calibrate” other clones of itself.
The “at home” test setup need not even be very close to the fancy rig. All it needs to do is be repeatable as far as comparing one light build to the next.
Perhaps a pass-it-foward kind of deal could be worked out so various members could test their setups with a known high Lumen light.
A really serious integration sphere setup is precise and expensive. Most stuff we do as hobbyists is approximate substitutes for that. For a light as bright as 10k or 20k lumens, I think it’s enough to just set it up in a room with a white ceiling, do the test at night with the curtains closed, use a camera exposure meter (incident light meter = has a diffusion bubble over the photocell) in a few places around the room, and average the readings together weighted by distance from the light source. Then use the formula about lumens = lux per square meter to get the lumen measurement.
^ We have Zak Wilson’s ceiling bounce app for that. The conversion from lux to lumens is what Jeff is heading at.
Edit; Many of us have some light meters also. But the throwers and/or the high lumen lights don’t correlate with the lower valued Maukka lights. So I believe Jeff is proposing an exchangeable ‘pass around’ pre-calibrated light for our boxes/tubes/whatever.
Oh I see. A bunch of XHP mounted on a big aluminum plate, powered by a computer supply? Some lights were made like that back in the day. If getting regulated power I wonder if the leds would change characteristics as the phosphors age, or whatever. My feeling is that if the readings are repeatable to within 10% or so, we are doing fine. Maybe even 20%.
I built my own light box. Extrapolating a low lumen light calibration factor to a higher lumen light isn’t linear. Need a better reference and it’s the user that draws the line between the low / medium / high – subjective but many of us are somewhat perfectionists, so we’ll dabble here and there and find an acceptable compromise.
A high lumen light source could be a multi-array on a CPU fan-cooled heatsink, could be one of the more popular lights already being manufactured.
Any of the Imalent XHP-powered lights RC90, MS18, MS12, etc. There aren’t too many lights that make 10K Lumens that can hold that output reliably enough to use for serious benchmarking or evaluations. I am sure you could build such a testing device with a cooled heatsink (water cooled ideally) and a constant-current power supply and get high output reliably.
- 10k light with XHP50 or 70s requires some 15+ amps PS. Most bench Power supplies are no more than 20 Amps. Check, that works.
- Water-cooled (preferably Prestone 50/50 mixture) can be jerry-rigged from some of the gaming PCs. Pump and fan-cooled exchanger. Check.
Shipping cost of such a set-up – most probably $30+ (to and from). Reasonable within CONUS; not so for Canada.
- Would I be inclined to make such? Above my paygrade, resources and time.
- Imalent Flashlight: I have my doubts about these over-rated power lights. I don’t think they sustain high lumen output. And they are advertised in the 20k (RC90), 100k (MS18), 50k (MS12) lumens. Also, quite an expensive initial investment.
- Maybe more of a Q8 w/ 4 x XHP50s – above 10k and Narsil/Anduril could be set an upper limit for sustained (limited) output. This is more affordable and I do have the host/MCPCB. 3 Volt version leds are available now (Sofirn has in stock).
Shipping cost would be less than a multi-emitter / liquid-cooled set-up. Also less chance of transport damage. Self-contained, minus the cells (which I presume members have high current variants). Also, a deposit against lost/damaged/outright stolen would be more reasonable than a $600 light or $200 set-up.
- There are the SunLike COBs that can be purchased, complete with fan-cooled driver w/ or w/out reflector. Works off either 12 or 24 V. Would have to dig that link from my bookmarks.
This is my problem also. Most of my lights are different size throwers with various led and 5000 to 6500 range. A big problem is this variation in lights that react differently in a lumen tube. I have struggled with this inconsistency unsuccessfully. Maybe now we can get something working in our favor.
That luxmeters are not linear is an assumption, but no one on BLF has actually checked that, I believe. I did quite a few things with luxmeters and my impression is that they, even the cheap ones, are pretty linear over the complete measuring range. But I never did a proper test. What I can say is that most luxmeters use a common silicon photo detector (because they are cheap and work well) and that all the data that I can find for those detectors show high lineairity over a wide dynamic range.
Me, I use low lumen calibration lights (one is 200 lumen, one is 550 lumen) and assume that the calibration is just as valid for high power flashlights.
Still, I plan to do such a linearity test at some point for a range of luxmeters, and will post that on BLF.
Oh, the lux meters may be linear, but the lightbox/tubes aren’t picking it up quite well. I get consistent lower values out of my box with the higher powered throwers than the mid-level “blasters”. I’m using your data graphs as my reference as you are thorough and the graphs are quite well done, with comparisons to similar LEDs.
But I see no linearity problems with reflecting devices either, what reflections happen to 10 photons, happen to 1,000,000 photons in exactly the same way, it is not that photons will be “in each others way”, or that reflecting surfaces can become “saturated” and stop reflecting at higher light intensities.
My lightbox gives 5200 - 5500 lm with Q8 XPL-2 (5000ºK) - depending on cell. So the meter / box is within the forecasted range. But I shine a hot spot thrower, the lumens are ‘lost’.
Direct lighting on a picture versus inclined or diffused lighting makes quite a difference to my eye.
Not doubting the physics here, but what’s not working with my box, and seems others also.
Beats me too what is going on. I can not think of a reason for non-linearities with light boxes/pipes. They can be bad integrators and thus favor throwy or floody beams, which is a different parameter, but for the same type of beam I can not figure why they would not be lineair.
The light need not be steady state for long. As long as the thing can maintain a constant output for say 5 seconds(?) at most (maybe less) - that’s all that is needed to make a reading on a lumen test setup. Then adjustments or calibration tables can be modified as needed.
It just has to do it every darn time it’s used.
I suspect there is already something out there in the commercial/Photo lighting world that can do this. Unfortunately, I suspect, the cost might be prohibitive.
Or not the least bit portable. Hence my thoughts about a DYI reference.
FPJ that worklight and similar ones are not easily mounted to the majority of out Lumen Tubes/spheres/boxes. But perhaps some sort of mod could be done.
I have been researching a bit, but have yet to find the magic bullet. The Pro Photo world seems to be the best bet at this time for an off the shelf solution.
Photogs are particular about the quality and quantity of their light.
I’ve done the incident light reading thing with a high dollar Sekonic meter. I’d still like to try one of the commercial testing places, but have yet ot do any research along that line.
An yes, my thought is to have some sort of reference that could be used to calibrate additional lights against. Then have a light of known value(s) to be shared among BLF-ers. Or perhaps have a place members could send a light to be tested against the high lumen reference to be used to calibrate their own setup.
Jeff introduced a concept (or should I say an experiment) using a high lumen reference light.
Theoretically, a regular light should give the proper lumen equivalent, but CNCman’s and his own analysis, do not correlate. Jeff is working with a tube integrator, CNC’s with boxes and tubes. My box works well in as much there isn’t a distinct hotspot (fried egg).
I am happy this is getting some traction. I tried HERE to get this started earlier.
If there is anything I can do to contribute, put me on the list.
Below are my non-experienced and unsuccessful efforts.
.
6” × 4” Lumen Tube with diffusers at every joint.
.
.
A 6” donated Lumen Tube and Dr Meter from Rick, a friend of DB Custom
.
.
8 inch tube, the middle diffuser has 3 coats of matte white paint and the inside diameter.
.
.
.
.
.
Nice setups of much effort integrators. The first one, the 6 to 4” with multiple diffusers, what would be some numbers – offhand. Don’t go digging, but rather did the higher lumen lights not come close to the expected results?
Still waiting for the lumen count on your modded Q8…
