180 is based on a reliable, measured value at 3A (145-151cd/mm^2). So I adjusted it for a higher current (same as with lumens, see djozz tests). But of course, it is not perfectly precise because with FET drivers the current changes all the time. So here it is just a ballpark figure.
The 250cd/mm^2 of the Osram is also a reliable measurement, it might actually be a little higher bcause not alle the optical losses of the test setp where accounted for. So this value just shows what is definitely possible. Not all Black Flats will be this good. Some might be worse, some might be better.
Yes, here. sma measured an XP-G3 S5 @5.05A with and without dome (he shaved the LED with a shaving blade and some silicon lubricant).
With dome: 88cd/mm^2
shaved: 186cd/mm^2
So itâs actually much better than people think. Especially if one takes into account that it can be driven at much higher currents.
I did a similar measurement of the bare die and got similar results. The problem is when the dedomed XPG3 is actually put in a flashlight (reflector or aspheric) the resulting lux is much lower than the dedomed XPG2. See some measurements on this thread: Cree XP-G3, testing a S5 3A emitter - #71 by FmC
But the fact is that no one has modded a light with the dedomed XPG3 that throws nearly as well as the dedomed XPG2, even with the increased current (lower Vf) of the XPG3.
The problem here is that we donât know how those people built their lights (maybe the focus was off?). We donât know how they de-domed their LEDs (maybe they didnât remove enough silicone or maybe they damaged the phosphor). We also donât know how consistent the luminance is after de-doming (probably not a problem though).
Maybe sma made a mistake in his measurement.
So in the end we donât really have enough real evidence either way. I can say that sma has a lot of experience with luminance measurements and that in the past his measured values pretty much always correlated with measured throw of different lights.
I think the problem is that a few people tried this LED and then wrote negative things and everyone else gave up.
Also a lot of people have started building unregulated lights with FET drivers. This poses the risk of damaging an LED when turning on a light because the startup current can be extremely high when using these new low-Vf emitters.
Controlled tests must in my eyes be done with constant current.
The_Driver, those are plausible explanations. I would love for the XPG3 to meet or beat the XPG2, but Iâm doubtful. There are several examples of experienced modders getting similar underwhelming performance results with the XPG3.
I did extensive testing on the XP-G3 when it first came out, with MEM assisting. I found it to not meet the capabilities of the XP-G2 in a reflectored light although MEM found promise with aspheric builds.
I do not like the beam profile from the G3, domed or sliced or de-domed, so I donât use it.
See what I mean though? The blueish purple aura and odd spill color? I donât like that, so I donât use it. Not fond of the XHP-50.2 either, for similar reasons. Will eventually pull the three from my scratch built light and go with something else.
Back about 3 years ago, when the HD2010 was giving me these problems, I did this to solve clearance issues. Made it easy to dial up the focus without worrying about touching electrical contactsâŚ
Yes, I agree with that.
Above I was just talking about center beam lux/candela, not all the other problems.
There might be a solution for this though. Somebody here on BLF did it with an XP-L2. After de-doming it he cut away all the excess phosphor. This way he got rid of the yellow-green corona.
The blue spot can probably be fixed by starting with a lower color temperature LED before de-doming.
Itâs not so bad, the first âtrickâ is to get the rectangular shape of the thermal pad cut into a copper box, then match it width and height with some sheet for the electrical contacts. I cut slightly shorter strips from a coke can as aluminum doesnât solder. Placed like 3 strips between the sheet and pillar, held them with a small clamping tweezer (might have been a hemostat) and placed the assembly on the Noctigon with solder paste present, solder paste on top of the copper as well. Then I put the emitter on top, re-flowed all of it simultaneously.
I did it to the XP-G3 in a Courui D01 when they first came out, it helped, but it did not improve throw not one bit. This had .005â silicone left over from shaving. I made up ground washers with different thicknessâs .005ââ increments, to shave as close as I could.
This is what happened when I went to the next .005â smaller washer! There was still a bit of soft mushy phosphorus left, kind of thin though! :person_facepalming:
So close and yet so far! What was amazing back then, when they first came out, was these G3âs were pulling 8+ amps, it got very exciting, but yet I was only able to achieve 270-280kcd IIRC in the Courui.
Iâve bought two lights from Amazon and I build both lights with BLF Fet driver. One light is with domed xp-g2 s3 1a and the second one is with dedomed xp-g2 s2 1d. The dedomed led is reflowed on Maxtoch 26mm with custom build centering ring and throws as well like my modded jax x1 with same Setup. For this small money, I think you canât get better light.
How precise is your test setup? How does the lens look? Does it have the colored reflection at high angles? Are you accounting for all the variables (full batteries, exact same measurement distance, room temperature etc.)?