Sliced XHP50.2

I’ve been slicing and dicing a number of these new cree LEDs and so far am reasonably liking the results, unfortunately not the same throw as certain other older LEDs but for the lumens pretty respectable. For a comparison, all in a Jaxman Z1, XP-L2 ~190kcd, XP-G3 ~200kcd, XHP-50.2 ~190kcd. Although compared to the 400+kcd the SoH does in the same host they are not that impressive.

Could be the manufacturing process, then.

Old way, apply phosphor, then mount the completed chip on the substrate.

New way, mount the chip inverted, spraypaint the phosphor after.

Just guessing…

That’s the way it looks to me. Phosphor after assembly would probably result in a significant reduction in rejects.

Nice work. Please let us know your results with the xpg-3’s and if you can scrape around the die on those as well.

XP-G3 close shave to about .005’’ silicone remaining.

Yes you can scrape the edges…

And both XP-G3 L2 50.2 and soon to be 70.2 are crap for throw IMHO! And they can’t be de-domed with out pulling the phosphorus off, I have tried!!

But you can de-dome the old XHP70-50 and 35’s!

I can confirm that all of the new Cree emitters have a clear some, and the yellow appearance is just because the phosphor covers the whole chip. I have XP-G3, XP-L2, and XHP50.2, with XHP70.2 on the way. I thought they were distributed phosphor as well until I had them in hand. Looking at any of them from the side the dome is crystal clear.

I like the new emitters, better efficiency. Never cared much about throw though, and I like warm tints so the color change in the beam is less pronounced.

Hey guys so here’s some proper, measured and error corrected tests of the following.
Convoy C8, smooth reflector, sliced and scraped xhp50.2 80+CRI 3000k, direct drive, VTC5’s @ 3.98v(I stupidly forgot to charge them)
Convoy M2, orange peel reflector, sliced XM-L2 ?? 4C, direct drive, fresh VTC5
A bone stock TN4A XPL Hi NW. Fresh eneloops.

Some testing info:
Tests were done at 3 meters.
My lux meter states that under 100k lux it’s accurate +/-4%.
I assumed that it over read by 4% so corrected as follows.
(Reading / 104) X 100

The results;
C8: 82890cd
M1: 17982cd
TN4A: 54684cd

I feel confident in this correction as the TN4A spec states 52340cd and I measured and calculated 54684cd
Without correction the TN4A result was 56880cd.

My XPG3’s have been shipped. When they arrive I will add their results into here aswell and make sure I get some good comparison shots.

Anyway I hope this comparison proves useful to someone.

Nice result with the C8. Thanks for the measurement.

Just some personal notes I made testing these last year. My measurements are simple relative measurements.

Surface brightness comparison
XPG-2 S3 8.6
XPG-3 S5 7.1

Corrected surface brightness (like for like binning). Shows effect of differing phosphor application.
XPG-2 9.8
XPG-3 7.1 = 38% decrease in surface brightness with dome on.

All dedomed
XPG-3 modded with the one chip subdivided into three distinct sections.
Stock shaved dome 3.3 - 57% dec over XPG2
Shaved plus thin silicone coating 4 - 48% dec over XPG2
Transplanted XML phosphor 4.7 –39% dec over XPG2
XPG-2 S4 7.7

Same here, big intensity decrease by shaving my XP-G3 dome.

Im confused? Dedoming these LED’s reduces surface intensity? How’s that work then?

Total intensity decreases. but since the apparent size (magnifying-glass removed) is smaller, intensity per unit area is higher.

Eg, if the chip looks only 50% of the dome-on size, but loses 30% of the total lumens, you’re getting 0.7 of the light in only 0.5 the size, or 1.4× the surface intensity (40% higher). You’d have to lose fully half the total lumens for it to be a wash.

Just pulling numbers out of my… hat… so keep that in mind.

No.
The “intensity” already IS per unit area. That’s literally what intensity is.
The total “flux” (aka lumens) is what probably decreases.
I don’t know because I don’t have an integrating sphere.

.
The 50% less intensity is compared to a XP-G2 which is almost identical in size.
I measured both intensities through a lens so any differences in size are irrelevant because a luxmeter only measures a small point.
The XP-G3 sliced dome had far less lux than an XP-G2 dedomed, or an XP-L HI.
All were tested at 6A.
.
I don’t know why it’s so bad, all I know is that me and everyone else who tested dome shaving got horrible lux compared to the G2, and dedoming these chips doesn’t turn out well so I didn’t even try that.

Aha, was misreading it, then.

Though in general, dedoming an LED does reduce its overall lumenage somewhat, but the smaller area (and thus hotspot) does concentrate what’s left into that smaller area for a net increase in intensity (and throw). That’s why dedoming is all the rage over here in blf. :smiley:

Yes that’s how it usually works, except dedoming the new cree LEDs isn’t really possible (I’ve seen one person do it successfully) and all the people that have tried dome shaving have also gotten a lux decrease.
Dedoming doesn’t really affect the total lumen output, but dome shaving does a little.
The smaller area and similar lux is what increases the intensity, even with dome shaving, but for some reason it doesn’t for the new cree LEDs, something to do with the new phosphor.

That’s why I wondered if flip-chip LEDs were running distributed phosphor, at first. Seemed odd that they’re hard to dedome, will likely get ruined if the attempt’s made, and other Bad Things™ happen to them.

Sticking to G2s for my dedoming needs. :smiley:

I’ve seen the odd report of the phosphor coming off when dedomed. I suspect that’s the heat and knife method but it was never explained. I have a bottle of MEK sat here for when mine turn up so I will have a look and see how it goes with that stuff.

At 6amps the xpl-hi and xp-g2 have a higher volts forward and thus are using more power. What would be the surface Brightness when compared to equal total power?
Xp-G3 s5: 21.9watts
Xpl-hi u5: 23.4watts
Xp-g2 s4: 27watts

I don’t know, I don’t have any constant current drivers higher than 6 amps for a single led.
But the lux decrease is definitely more than the 20% power difference.

Anyone have any before-shave/after-shave beamshots for the XHP50.2?

thank you.