LED test / review - New Cree XLamp XM-L2 U3 color kit 51 (≈ 6200 K, design change after PCN issued)

I can think of an alternative explanation.

How do you get your lumens numbers? Have you previously described your test setup somewhere?

Just write what you think: my measurements are wrong.

When you want to make better tests, go for it. Maybe I described my testing method before, but I can say that I am using a light box and a 2π configuration for the port hole. This box is calibrated to six different Cree emitters from different batches/datecodes and different types (which are also domed/domeless), which are only tested on Noctigon 20 mm DTP boards. For cooling I use an active one, to hold the Tsp at around 20-25 deg C. In general, I test only on copper DTP boards with temperature monitoring (Type K temperature sensor).

Thanks for describing your testing methodology, and thanks again for the test. It will put a smile on my face to see the post PCN 5154 XM-L2 U3 putting out 3800 lumens on a cold start.

Is there anywhere to get these leds in 3-4000k?

Edit : just ordered some from Kaidomain.

1 Thank

Yes, it is quite impressive.

However, yesterday I discovered a potential flaw with this LED (or this LED platform, I dunno), although I have to prove this later on - I tested the XP-P, some kind of CSLNM1 competitor. I don't know if this emitter is based on a same platform like this new XM-L2, but the XP-P is very sensitive to temperature and also driving current which results in very high color deviations and a relatively significant light flux loss with rising current. More I will describe in the final test of the XP-P, but this seems not very promising if Cree is establishing a new LED chip platform. For the moment I do not yet really understand the background that could explain such a performance.

Interestingly the post PCN 5155 XP-L HI datasheet lists a 0.5 °C/W thermal resistance. XM-L2 says 0.75 °C/W.

With this kind of conductivity, low temp solder paste reflows should be viable, maybe even when running the LED well beyond max spec. From what I have heard, some output capability is lost due to high reflow temperatures, so there could be a benefit to low temp reflows. The thermal conductivity of Indium may be beneficial to output as well.

I suspect that the size of the thermal pad and especially of the LED chip have a huge impact on how well the heat will dissipate from the LED chip and the package. I think there is a practical limit to how much heat can be dissipated over a given area in one second, and therefore very small LED chips like in the XP-P are rather at a disadvantage there than the big ones, especially if the thermal pad is smaller due to the LED package. Maybe this is one of the reasons why smaller LEDs are in general not as powerful as bigger ones, despite the thereotical low thermal resistance.

Hello, could you test the new xpl hi or hd? It seems that I changed a lot to the old version

The test of the new design XP-L HD/HI is planned, but I have to order them from Mouser, and moreover I have to hope that the LEDs Mouser is selling are the new design ones (the order code is the same with the predecessor as with the XM-L2 tested here).

Kaidomain sell the new versión of the xpl hi and hd in highest bins

1 Thank

I prefer to buy my test emitters from global distributors like Mouser or Digikey, to minimize the risk of getting counterfeit models or LEDs of unknown supply chains.

On KD's page there is also no image of the reel or other evidence, so I can't say for sure if the emitters are really of the right binning or even real XP-L HD, especially if they're sold as 'highest bin'. To date at least, I don't heard that KD sold counterfeit emitters, although the prices shown are already very cheap compared to wholesalers.

Brilliant test koef3!

The real question is who will push this one more than 10A to get decent performance? So far it seems that Cree is like VW in car industry. Lot of problems, older technique, and living on the old glory.

Fantastic info, thanks!

I had the same thought. (And I also can't imagine how more than 10 A current will be practical in a portable light.)

I recently tested the XP-P, for which I will publish the review soon. The performance of this LED is also not outstanding, the Black Flat or NM1 still have a higher luminance and slightly better performance. A big disadvantage of the XP-P is the strong temperature dependence of the luminous flux and the strong color distortion at overcurrent. So Cree is already using a new technology/platform for the LED chips, but it doesn't really seem mature to me yet.

Yes koef3… Funny and sad at the same time. Because it is Cree after all. Same like VW. New MQB platform. VW Passat B8. I am very disappointed in this car. Old VW Passats(up to 2004.) were much better cars. Same like Old XP-G2 S4 2B Dedomed is still better than any new Cree emitter.

Did you hear about new Yinding 5050 LED

I would like to know who is manufacturer of that “Yinding” LED because it seems like it owns Osram type of technology so I think that “Yinding” is just OEM for certain Ali seller. Anyway… That is one serious performer. Especially Glass version of this LED. With 5A current draw on the Glass one I got slightly higher emitter light intensity(about 15%) than with the best pick of old g2s42b dd, and flat glass dome acts as a precoolimator(gives free magnification without lux drop) so it is good for zoom flashlights.

1 Thank

A quick note. I bought the U3 5A3 and U2 7A3 xml2 (new) from Kaidomain. They are the new design as stated. There is a definite increase in throw and the hot spot is tighter. I have not done any real testing but overall the results are positive. The tints are good, perhaps a touch of green but not bad. The 5A3 is quite clean and the 7A3 has a hint of green in lower modes.

I’m happy as they are a replacement for some ugly cool white sst40. They do not compare to the nicer Nichia options but are a suitable replacement for sst40 which tends to be ugly for me. The 5000K I had was nauseating.

1 Thank

I heard of this LED but I can't find a trustworthy source - no one can guarantee that the LED ordered at Aliexpress is the real one I really get... I wouldn't be surprised if Chinese LEDs are counterfeited just like other manufacturers are.

Maybe I order some and hope for the best. :D

With dome, the new design is not too bad in terms of luminance and color uniformity, where the older ones are still better. The luminance is a little bit higher due to the smaller LES, and the (in most cases) higher efficiacy which results in higher light flux at same current.

But dedoming is not worth it, unlike in the past where this brought a significant gain in luminance.

So I just ordered six pieces of the "Yinding 5050" LED, estimated arrival around mid of January. I have no idea if I can even get these LEDs or what type I end up with. It's like gambling now. Like I wrote - I should hope for the best. :D

Would you try also dicing the emitter (cutting away material surrounding the die) after dedoming? Currently a lot of the output is scattered by the surrounding transparent material, so I would think that cutting it away would bring that scattered output back to the die and increase luminance. Notice that in all of the top performers in throw (Osrams, SFT40, etc.) the material surrounding the die is completely opaque.

Dicing is a standard operation for 2nd gen XHP emitters, partly for this reason. (Also to address the tint shift toward green as a result of slicing alone, without dicing.)

There are 2 main types. “Glass Domed” and “Egg Yolk” type.
Egg Yolk has 2 types: cold white and warm white, while glass domed version should be neutral white. I ordered 1x Glass nw and 1x Egg cw tint and I got better result with the glass one at lower current draw(it pulled only cca 5.5A) than with the Egg version(it pulled 7.5A). Tested in real condition with 1x18650 Samsung INR30Q and fet driver everything bypassed.