Three Oslons tested: 1) latest gen. Oslon Black Flat 2) SSL80 4000K 92CRI latest gen. 3) SSL80 4500K 96CRI 1 gen. before latest


First the Osram Oslon Black Flat. I bought it from Mouser, the partnr. is LUW HWQP-8M7N. It only comes in cool white. Despite the many emitter tests that I do, I'm often guessing what generation a led is and how long it has been on the market, manufacturer websites are usually unclear about this, and sometimes someone more knowledgable than me has to correct me. In this case I think that this the second generation of Oslon Black Flat, and it came on the market not long ago in 2015.

This led is interesting because it has a small 1x1mm die (like the XP-E2), it does not have a dome, and the thermal resistance looked very promising: 4.3 K/W, which is very good for such a small die and way better than the XP-E2 Torch which is the led with the highest surface brightness that I had tested sofar (9K/W for the XP-E2 Torch) (edit:I was wrong here, after a suggestion by chouster, a quick calculation learned that an S4 XP-G2 reaches a higher surface brightness when at max).

Some initial observations. The Oslon footprint is a bit smaller than the XP-leds, but can be reflowed on XP-type boards, but not too much solder paste must be used because the led does not seat itself and excess solder can not be pushed out from underneath the led.

(Oslon left, XP-led right)

The die size is equal to the Cree XP-E2:

The light comes very cleanly from just the die, around the die it is totally dark (die at 0.5mA):

A very annoying feature of Oslon Black emitters is that the led-minus is electrically connected to the central thermal pad. This requires, when the led is mounted on a DTP-board, that the ledboard is electrically insulated from the flashlight body.

The test was done like all my more recent emitter tests. I described it in detail in my XP-L test. In summary: 1) just one led was tested, reflowed on a DTP copper board (XP-Noctigon 16mm) 2) I used my larger version II integrating sphere with much higher quality luxmeter, but that should not matter for results, 3) the output numbers and voltages were measured with the led close to 'steady state' for each current, so warmed up and settled, you should be able to get these numbers in a well heatsinked flashlight. Mind that these are output numbers of the bare led, in a flashlight there will be losses from light obstructions, lens and optic, 4) output is in 'djozz-lumen' defined as 1/550 of the output of my Sunwayman D40A on high setting, which I hope is close to the real lumen, but at least is consistent over all my emitter tests done in integrating spheres.

Here's the results, presented in a graph. I plotted the led together with my test results of a dedomed XP-E2 Torch U5 bin, that test was done earlier:

Now that looks well! The output of the Oslon Black Flat up to 2A is almost the same as the 'Torch', and above 2A it leaves the Torch behind, maxing out at 4.5A, compared to 2.8A for the Torch, the brightness at max is 50% more than the Torch. It looks like there is a new surface brightness leader among my tested leds (edit: also higher than a S4 XP-G2)! Because the maximum output is obtained at over 4V, the led can be run direct drive on a single Li-ion cell.

Much encouraged by these results, I did a Brinyte B158 mod with this (very) led and a direct drive AK-47A+FET driver (ledboard glued to the pill with an as thin as I could make it layer of Arctic Alumina Adhesive to get the required electrical insulation between led-minus and pill) in the hope that it would crush the throw of the dedomed XP-G2 S4 2B led, but unfortunately it did not, I got 'only' 285kcd from the Oslon BLack Flat, with 322kcd for the S4 2B. I'm not sure why it does not get that high, the amps were good (3.65A), perhaps these small dies are not imaged well enough by larger aspheric lenses to get the spot brightness of a good XP-G2.

Time will tell what this led will do in practice for thrower flashlights.


Next in line: An Osram Oslon SSL-80 4000K 92CRI typ. of the latest available generation, partnr. GW CS8PM1.CM-KSKU-A535-1 . I bought it from RS-online. The rs-online webpage does not list the CRI, but the datasheet does.

SSL-80 leds are special in their emission angle: 80 degrees to half intensity compared to 120 degrees for almost all other leds. This is optically achieved by using a very high and sharply curved dome and a small die, I still have to check the die size by shaving a dome off, but I expect a 1x1mm die, like the Black Flat Oslon and XP-E2. The small emission angle gives much lower light loss if used in an aspheric flashlight, and a very nice defined hotspot in reflector lights. The small die however implies that it is not a high current high output emitter, the OslonSquare/XP-G2/Nichia119/219 have a more than twice as large die, let alone the XM-L2/XP-L die size.

The led is interesting compared to one generation earlier because according to the datasheet, the thermal resistance is lower, 5.3 K/W compared to 7 K/W for the earlier generation. Not a great difference but it should be noticable I guess. Further, the Vf is 0.2V lower, considerable. The package of the latest generation looks very much like the generation before but can be visibly distinguished by the way the phosfor is deposited, it looks like it is 'poored' after the die is assembled on the base, covering part of the base too, while the older SSL80 (the last led in the first picture) has the phosfor neatly confined to the die area.

At some point I guess also the 96CRI version will be available, in the meantime this 92CRI one is pretty nice too.


Last led is an Osram Oslon SSL-80 4500K 96CRI typ. of one generation earlier, LCW CR7P.CC-KSKU-5J7K-1 , also bought from RS-online.

It is 96CRI 4500K, so almost directly comparable to an earlier version of 96CRI 4000K that I tested before a few years ago. Unfortunately the test method was different at the time, no integrating sphere measurements but ceiling bounce measurements with a reflector on top of the led. So the output numbers are not directly comparable. It is the led on the right in the picture at the top of this post.

I have put the test results of these two emiiters in the same graph, together with two recent but earlier tested high CRI leds: a gen.2 Oslon Square 4000K 92CRI typ. (there's a gen.3 already, want to test that too but I have to order through work and can do only so many orders for private use), and the Nichia 219B V1 4000K 92CRI.

*although the older gen. 96CRI SSL-80 test from two years ago can not be compared because of the different test method, with some 'educated mind extrapolation' it is clear that the output went up, and the voltage went down since.

*As you can see the two SSL-80 leds have not the high current and output capability of the Oslon Square or even the Nichia 219B V1, also the Vf is higher, although the newest generation comes close to those up to 2A, both in output and voltage.

*I would use these leds at 2A, at 80-90% of the maximum output, the output at that current is in the same ballpark as the bigger die leds and you have the benefit of the small beam angle in particular builds.

*the newest genaration has indeed a 0.2V lower voltage, great for use with smaller batteries and longer regulation. My first build with it will be a small 14500 zoomie :-)

Conclusion. These two high CRI SSL-80 leds perform clearly better than the previous generation high CRI SSL-80 that I tested two years ago. I think these leds are great for many high CRI flashlights, I really like the 80deg beam angle for several types of flashlights. A very important test has not been done yet: the judgement of tint and colour reproduction. I'm sure both will be fine, but because I can not measure tint and CRI quantitatively, it will be a subjective test when I have build the leds into a flashlight. I will update in this thread with that impression.

Thanks for reading!

Great review and tests djozz. I was wondering what this “black” Oslon was all about. I’ll be following all your Oslon emitter testing as closely as I can. I have some various oslons on the way from mouser. Fairly random, I’m not too good with spec sheets.

PS. When reflowing the oslons onto an XP-G noctigon, does the emitter snap into place, or do you have to hold it in place with tweezers?

Thanks djozz. Great information.

I wonder how hard it would be for us to make a DD that switched positive insteed of negative. Would not only help with this flat emitter, but with all emitters. Would not need to run a negative lead to the driver. Bases reflowed to drivers will provide an excellent low resistance electrical connection.

While I was at mouser, I ordered some single emitter boards foe oslons. They’re aluminum with no DTP. No need to isolate the boards negative thermal path. I’d be interested to know how far we can push one of these emitters without a DTP. At what point would it be a problem?

My guess is that you start noticing a difference above about 1.5A, and up to 3A you are still fine (for this Black Flat led that may be the maximum on a non-DTP board) but with a little less output.

As a fan of HCRI LEDs, I really appreciate these tests!
I would love to hear your thoughts on how these new Oslon 96CRI compare tint wise to the Nichias.
IIRC, in one of your previous tests you mentioned that the Oslons were less rosy than the the Nichia.

The narrow emission angle of these newer Oslons remind me of the older Cree XRE so I’m guessing they will be relatively throwy.
These might work well with some narrow optics for good throw with minimal side spill.
I’m of the opinion that for distance ‘throw’ lighting less side spill is better.
Side spill can cause close up glare which hinders visually what we can make out from the throw beam and whatever it’s lighting up.
Lights with minimal side spill there is less distraction for our eyes and it makes for better focus on what we’re lighting up at distance.
Thanks djozz!

You're welcome!

Just finished the post of the OP with final thoughts sofar. A flashlight build, and tint judgements, will have to wait untill I have time again.

I had forgotten how high that new Square can go. If only we could get that ~850 lumens at 3amps instead of 5.5amps…

As always your efforts are appreciated Djozz

Thanks for another great Djozz-test!

Are you sure the XP-E2 Torch has a higher surface brightness than the best performing XP-G2?

When I saw the graph of the Oslon Black Flat I thought it has to be the new alphadog… Strange that is was clearly outperformed by the XP-G2 S4 in your test.

That three-solder-pad-design without one beeing neutral is somewhat weird. I think they lowered thermal resistance that way…

Amazing work as usual djozz. Thanks.

I have actually not compared surface brightness for the latest bin xp-g2 and 'Torch'. Here we go:

-maximum output XP-G2 S4 2B: 1420lm

-maximum output XP-E2 'Torch': 530lm

-die size XP-G2: 1.48 mmsq.

-die size XP-E2: 1 mmsq

So (for the single two emitters that were tested) the surface brightness of the XP-G2 is actually 1.22 times the surface brightness of the 'Torch', if both are driven at max. In practice on a single li-ion the XP-G2 S4 2B does not get past 4.6A while the Torch can be driven at max, so the fair difference is 1.12 times, still the XP-G2 is brighter.

I did not expect that, thanks chouster for making me calculate that!

The Oslon black never had a electrically neutral thermal pad, this is not a recent change.

Could you post a mouser link to the other two oslons that you plan to test? Just trying to keep up my homework. Someone said you were going to be giving us a test.


The XP-E2 “torch” is still a great choice, especially for smaller single cell lights when heat management and runtime matters.

Yep, just took a look at datasheets of Oslon black versions 1.1 and 1.5. Two cathode pads on both. However they must have had a reason to do it that way…

Great work, thanks.

Are there Oslon Black Flat emitters with larger die surface?

This small is pretty close to good old type of S4 2B…

Probably it would be the same if mcpcb could be hard pressed to thermal paste or soldered directly to pill, but as Djozz described; it has small issue (led-minus is electrically connected to the central thermal pad).

Hey djozz, someone directed me again to your amazing tests when I asked about the black flat LED :stuck_out_tongue:

So I’m just wondering, you said the S4 XP-G2 reached 1420lm max, and with a side length of 1.48mm and area of 2.2mm^2 it is only 648lm/mm^2 correct?

How is that higher intensity than the nearly 800lm/mm^2 of the oslon flat black? In fact isn’t it maybe over 800lm/mm^2 due to the cut out corner?

And the 7N bin is supposed to get 400lm at 1A based on the datasheet, why were you only getting 300?


Wouldn’t he like to know as well!
I can tell you that Djozz is a bit frustrated about that but no worries: he has ordered some Oslon black flats for some mods…… :sunglasses: so I’m sure the will do some further testing.

A member of the German forum has modded his Eagletac M25C2 Turbo to over 1Mcd.
Of course the spot is tiny and the modded light uses two cells in series.
Measurements were taken with a calibrated Gossen Lux meter.


I’m planning to modd my Fenix TK61 with this led, and perhaps one of my Thrunite Tn31’s and the Olight M2X UT Javelot.
Focussing will be a bith thought, with a led this small :wink:


Ok, well I hope he does a bit more testing to figure out exactly how intense the flat black is, because I would really like to have an alternative to the XP-G2.
This is the graph I’m looking at, and I’m not sure if the S4 2B in this graph was dedomed or not…

If it was not, then the lumens coming from my XP-G2 are probably significantly less than 1400 since dedoming loses some lumens.

Since the flat black doesn’t need to be dedomed, it would still have 800lm/mm^2 which would be a lot more than the 500-600lm/mm^2 of the XP-G2 (because ~1000-1200lm / 1.45^2)

djozz does not measure luminance of LEDs, he measures lumens. Two totally different things. For throw you need to know the luminance [cd/mm^2]. Lumens per square mm of DIE-area should be a good estimate though, but only if the actually value (also the size) measured is actually precise and representative of all the LEDs of this kind. For Osram LEDs this isn’t the case, because you can’t buy the Black Flat 2 in a single BIN. Osram only sells you a range of Bins. In the German TLF forum multiple people have found out that there is a rather large variance between these LEDs. If you want a “good” Black Flat that will actually produce the super high luminance values, you either need to be lucky or buy multiple LEDs and test them individually under the same conditions. Testing means that you test with a lux-meter if it gets brighter until 4A or even 4.5A. If it doesn’t, you got a “dud”.

Here is a thread in the German TLF forum, where somebody tried to precisely measure the luminance of a Black Flat. His LED got brighter until 4.5A where he stopped measuring. He didn’t actually calculate the luminance from his measured values, but sma and I did that for him. His LED managed 250cd/mm^2. That is around 10-25% better compared to any XP-G2 that has ever been measured in this way.
Just like djozz he only tested one sample though…

cd/mm^2 should increase linearly with lumens, no?
so by just doing lumens / area I should be able to compare between LEDs

Also, yeah I see it actually is 8M7N so the LED could be anywhere from 8M (280lm) to 7N (400lm)
The LED itself is really cheap, only $1.5
Problem is that I don’t have stars for it, nor do I know how / have the equipment to reflow one of these onto a star every time to figure out which one works best.

Even just assuming that the 1420lm is with a dedomed S4 XP G2, the 800lm on 1mm area die should still be better than the 1420lm on 2.1mm area of the XP-G2, right?

EDIT- ok so I read djozz’s post wrong, the one that is NOT higher than the XP-G2 is the XP-E2, the black flat IS higher than the XP-G2 :stuck_out_tongue:
When talking about XP-E2: “(edit:I was wrong here, after a suggestion by chouster, a quick calculation learned that an S4 XP-G2 reaches a higher surface brightness when at max)”
When talking about the flat black: “(edit: also higher than a S4 XP-G2)”

So that clears things up…
Now the question is, can I pay someone to reflow a bunch of flat blacks onto copper stars? plz :slight_smile: