Thanks on all of your emitter tests on this forum!
I just cant believe we aren’t see any progress on this field… I am the guy who likes Single cell hosts so seeing 20A buck driver for certain emitters is just insane…
I don’t like anything that burns skin out of my hands lol
That’s why that Osrams are fascinating thrower vise although not lumen vise of course… Doing that at max 4-5A(ideally on FET driver from Led4Power and other guys (similar to direct driven yes but still a driver since they have mcu, lvp and rvp)…
Maybe there is some new LED from China on Aliexpress or somewhere with a good potential? Like those Yinding circular leds? While I had tried them and some of them were good if you pick good out of a lot… Cause it seems that they have greater performance inconsistency or performance variations than premium LEDS.
Anyway hope we’ll find something up? Maybe there are white laser on 3535 footprints as an example and we still don’t know? I doubt but just as mind food…
Enderman LOL bro! I think I changed your for Koef3 that’s why “supp on TLF” But you also did a lot of test for us
I recall reading somewhere that the 2.32mm diameter was supposed to be a typographical error, and that the “official” figure should be around 4.2mm^2 as the name suggests. But as koef3 pointed out, the phosphor pour extends past the die, and I can see a bit of unevenness in phosphor distribution, so it’s very hard to get a good estimate for how big the effective LES is.
I do think that this method is much more translatable to actual usage. With a lens, I think one potential difficulty is that some modern LEDs have very rough phosphor (or uneven distribution), so the projected image might have significant fluctuations in intensity at a local scale.
A reflector is probably better because it excels at blending away such irregularities, though it seems very hard to find reflectors that have enough precision to focus an LED on the order of 1-2mm^2.
To summarize: it is basically impossible to get exact (± 5 % or so) luminance values.
That is why I am seriously considering no longer “measuring” luminance in future tests. The reflector measurement method does not work with some LEDs (too large a footprint, focus cannot be adjusted properly, etc.), and without a reflector, light emitted/reflected to the side is just as much of a problem because it distorts the luminance values. I cannot and will not spend hours finding the right focus for each LED (however that is supposed to be done in a semi-automated way) in order to obtain reasonably useful values, this is simply not possible in my daily schedules.
Luminance measurements are simply too prone to (measurement) errors, no matter how thoroughly you work. Not to mention the time involved, as I said.
Another problem: LED design is changing over the course of their production span. The chip itself is changed or the package is adjusted, maybe the phosphor mixture is also changed (for example because of another supplier etc). Even though the PCN repeatedly states that “optical properties remain unchanged,” they can change to such an extent that previous measurements become useless. Another problem I had not too long ago: reflectors are not always the same. I had used a Convoy C8-SMO reflector, which at some point became too dirty and scratched. However, the replacement reflector was different, and it was much more difficult to create a clean beam (I would have had to rebuild my entire setup), which would have made it even more time-consuming… It’s a shame that I can only be of limited help in this regard, but that’s just the way it is.
Thank you for sharing your thoughts on this–it makes complete sense.
Even +/-5% accuracy is extremely impressive and more than we realistically need, given that the threshold for difference detection (without side-by-side comparison) is much greater. I probably couldn’t tell a 20% difference in intensity for the same light; with different lights and different beam profiles, perceived intensity can depend strongly on factors other than true intensity.
I think my best guess(timate) is at around 10-20 percent accuracy for my measurements. Since I ditched the reflector measuring method some months ago (for the reasons I explained earlier) this is the best I would guarantee for.
Luminance is unfortunately so difficult and time consuming to measure. Maybe I really should only do basic measurement with brightness in 1 m distance and calculating luminance from LES size, but only giving numbers rounded to 10 like 220 or 120 cd/mm2. Even with plausibility checks and attempts to reduce the error rate, luminance measurement is an absolute nightmare.
To make somewhat accurate luminance measurements, I would have to build a complex rig from scratch, probably out of metal machined parts or aluminum profiles (and with very tight tolerances), and it would likely require quite sophisticated electronics to achieve a reasonably automatic focus adjustment just to save time. This is not feasible in terms of time or my skills, and there is also a space issue in my apartment, as such a rig would probably have to be installed more or less permanently somewhere—my current setup is not like that.
My E10 with green osram looks almost like a LEP, No led with tir or reflector has as much throw or beam as tight. it is so tight, and has so little spill, it is very impractical, unless you mount it on a rifle scope and use to illuminate a target , and only a target, hundreds of yards away,
This is still very good. Plus, as long as you use the same method from a certain point onwards, often times consistency is sufficient for relative comparisons, without the need for absolute accuracy.
I agree with all of the above. More fundamentally, we never really defined what luminance even is!
For example, it’s unclear whether we care about peak luminance across the LED or the average luminance. If we care about average luminance, do we take the LES to be just the phosphor? If there is sideways light leakage like Cree’s domeless HI series, does the non-phosphor area that lights up also count toward LES? All of these questions have no nice, clear, simple answers.
For all practical purposes, just a reasonable estimate of the LES area suffices; luminance can be calculated from there. Though as the previous paragraph mentions, even estimating LES is an ill-defined and difficult task…
I’ve been using low tech methods for relative light “measurements”, and would like to remind you of multi-pass diffusers and pinholes…with a decent dark room, you might be able to catch cross sections or manageable sample sizes of luminous flux or irradiance…extrapolation is your friend if you can compare enough samples for a ballpark…
**(you can manually trace the rays/differential brightness of a hotspot simply by putting a pinhole in a dense translucent piece of material, then recording the pinhole location, checking the Cd, then recording where the pinhole is on the translucent hotspot.
I was thinking of more like an off-axis parabolic or long FL aspheric lens
Definitely not a forward facing reflector
Focal distance will be the same for any LED
If you then take the lux measurement at a far distance (like 10-20m minimum) the spot size will have grown to where you can easily measure the full intensity of the LED and ensure the full luxmeter sensor is covered
This method makes it so that you can calculate the intensity based on the front area of the optic instead of the area of the LED, and the front area of an optic you can know very precisely
In addition to that, your lux readings will be scaled up significantly because now that area is hundreds of thousands of mm2 instead of only a few mm^2, effectively increasing the signal to noise ratio
And this makes the area of the LED die not factor into the calculation
The down side is you have to factor in the efficiency loss of the optic, but if you have a good quality one the manufacturer should have that value
My initial concern is that the local intensity fluctuations on the LED (e.g., due to rough phosphor) might make the readings highly dependent on small differences in measuring location, but the long focal length and small image size should solve this problem as the sensor now covers a substantial patch of the emitter’s image, so local fluctuations are smoothed out by averaging, per the law of large numbers.
This is nice, and essentially equivalent to measuring average intensity over the portion of the LED’s image covered by the sensor. Such a measurement indeed does not depend on other parts of the LES, such as overall size and locations of irregularity (e.g., phosphor over-pour, sideways leakage).
This method also allows the back-calculation of “effective LES”, effective in the sense that it allows the naive equation “output equals intensity times LES” to hold, similarly to how nominal voltage of a battery allows “energy equals capacity times voltage” to hold.
Interesting, can you name some examples that exhibit the performance you are referring also please explain Convoy reflector precision, which flashlights with this issue.
With a bad batch of L21B reflectors, the floodier SFT40 out-throws the SFT25R despite significantly lower surface intensity.
It’s hard to say which lights you get might be impacted because there seems to be quite a bit of inconsistency in manufacturing precision between batches, there’s no telling which batch you might get. I once bought a pair of C8 reflectors sealed in the same package, and one was 0.5mm shorter than the other, enough to cause a rattle. I’ve also had bad luck with all S2+ smooth reflectors, which can’t correctly focus anything smaller than a SFT40.
last time I checked there were LED’d that produced an 8 degree light without a reflector. For comparison a helicoptor spotlight is focused to 4 degrees.
that’s why “supp on TLF” 
But you also did a lot of test for us 
