some measurements and observations on the Osram Oslon SSL 80deg. 4000K 96 CRI typ. led (nov26th: added a sk68 mod)

I had the opportunity to buy some Osram Oslon leds (at RS-Online, through work, I managed to incude it in an order for some other parts, at work, I did pay for them myself by the way). I am curious about any led that is not a Cree because in flashlight country there is a narrow focus on Cree (apart from that one high CRI Nichia led), and perhaps not always rightly so. One of the leds was a Oslon SSL 80deg. 4000K 96 CRItyp. ( partnr. LCW CQ7P.CC-KRKT-5L7N ). It was the closest I could get to the well-known Nichia 219 4500K 92CRI. A newer led with a bit more output that I'd rather had was the Oslon Square 4500K 96CRI typ. (LCW CQAR.CC-MPMR-5J7K-1) but I could not get that one from RS-Online unfortunately. (I did however get a Oslon Square 3500K 80CRI in this order, I'll report about that one some later time).

But now on to the Oslon SSL 80deg. 4000K 96 CRI typ. Although this one is not the closest match to the Nichia219 that Osram makes, it has some interesting features; a led die that appears to be bit smaller than the xpg or Nichia219, it has a narrow emission angle (80 degrees), a bit narrower even than the good old XR-E (XR-E: 90deg. XP-G2: 115deg. Nichia219: 110 deg. Oslon Square: 150deg.), and a CRI even better than the 219 (I am sure you will see the difference ;-) ).

First I did a reflow onto a 16mm xp-g Sinkpad. The solder pads of the two do not really match, but are close enough for a succesfull reflow (see for a picture of that in my thread on the Oslon Black IR 850nm). The mismatch was such though that the led was not really sucked tight onto the board, as happens usually in reflows, so thermal transfer may be less than optimal in this test (but I will ignore that ). This is the test set-up:

The led-board is mounted on aluminium, the led is directly connected to a power supply, the current is measured by the power supply, the voltage is measurend over the led wires by a DMM, the lux was measured ceiling bounce, with a OP-reflector placed over the led (no lens). the lux-numbers were converted into lumens using comparison flashlights with known outputs. For really precise measurements this set-up leaves a lot to be desired, but this is what I have and I am sure the general picture of the findings is valid. In the graph I compare the results with similar measurements on the Nichia 219 on copper I did a few months ago. Only real difference is that during the Nichia measurements the reflector had a AR-coated lens on top, so for a precise comparison you should add 1 or 2% to the Nichia values (that are already higher).

The first two things that strike from this graph is that at any current the output of the Oslon is a bit less than the 219 (at 1.5A it is 21% less), and that the maximum output is reached at a lower current than with the Nichia. The output of this high CRI Nichia is considered already utter meh by the cool white fans, well, this one is worse, at 1.5A just about half the output of a cool white XP-G2. On a brighter note, the Vf is considerably less than the Vf of the Nichia. Considering that the high Vf is one of the major limitations of using the Nichia219 in single cell flashlights, this lower Vf of the Oslon (it is close to the Vf of the XP-G2) is a real improvement, at 1.5A, using a lineair driver with a li-ion battery it stays into regulation during most of the drainage of the battery.

As a conclusion from these measurements I think this led will not shock the flashlight community, but because of the narrow emission angle it will be a nice led in a small high CRI aspheric light, e.g. a sk68 type of light.

Because I had it still lying about not doing much, I swapped the xpg of a Ultrafire M5 with this led. It will be the last time that I sanded a 16mm Sinkpad down to 10mm (and made it thinner too, to fit it into the M5), I have some of the new 10mm Sinkpads in the mail from VoB

Here is the result, compared to a Nichia 219 mod (bottom=Maratac AAA copper with Nichia219):

No beamshots because I can't get the tint right on picture, but the Oslon looks quite a bit warmer than the Nichia (and not as pink), the difference seems to be even a bit larger than I expected from going from 4500K to 4000K. The beam pattern of this mod is really nice by the way, emitters with narrow beam angles give very distinct hotspots (even with the OP-reflector) with a good bright spill. The light coming from the flashlight is beautiful, the tint is nice and it seems like the 96CRI really shows.

I put a small projection lens in front of the flashlight and made two pictures of the projected led, at two focus points, these pictures give an idea of how the hotspot would look if this led is put into an aspheric light:

It is the most messy phosfor-layer I have seen in a led thusfar, there is even a bit of blue light leaking from the edge of the die! But look, the spot is more or less round, I like that in an aspheric! I have one of these leds leftover, I will try it in a small aspheric and report about the results.

Thanks for reading, I am afraid I will be the only one ever to use this led in a flashlight, but who knows this post will inspire someone :-)

Thank you for posting this! ^^ Do you think it's more thermally efficient than Nichia 219? 96 CRI, I think I want to try these instead of Nichia for my halogen replacements. LCW CQ7P.CC datasheet is interesting, it has gold-plating inside. These can't be cheap!

How did you open the Maratac head?

Interesting post .

Thanks for this .

Very interesting LED. 4000K and 96 CRI. Makes me drool all over the keyboard. ;) Too bad about the output though. Cant have it all..

Great post and testing!

Thanks djozz. Amazing work as well as interesting.

interesting post, thanks for testing and sharing!

great review and good to know they will reflow on xp boards alright. i do wonder if there is a significant loss in thermal footprint.

it looks like this same emitter is available bare and on Al star at led-tech.de

thanks for the appreciation :-) !

@Chloe: This led has a thermal resistance of 9.6 K/W, the Nichia 7 K/W, so the Nichia is supposed to shed heat a bit better, the graph seems to support that. I paid 2.32 euro for this led, but shipping for the order (7 leds) was 6.75 euro, so that is an extra euro/led.

@texaspyro: the maratac AAA copper had no glue, it was a straight swap, just reflowed the Nichia on the existing board. The mod was done a year ago, I would have used a copper board now. I have also modded the copper Maratac AA, that one was nasty: lots of silicone gooey (turned green from the copper, apparently not acid-free) inside the head, had to smoke the pill out with a blow torch and brute force, but the driver survived.

@islisis: hey thanks, that is a nice source for this led, rs-online only sells to companies (in my case: schools). So with that more easy source this led may after all find its way to more than one flashlight :-)

(I found that Osram has an improved version of this led, with a white backing instead of transparant, which is a bit more efficient. but that is not the reason for this post.)

I had one of these leds leftover and one cheap sk68 clone, so I made a sk68 clone with this led, should be nice with the 80 degrees angle. Very simple led-board swap, with the stock driver, the led on a copper Sinkpad, bit of Arctic silver around the edge to improve the heatsinking:

The results did not disappoint.

The current draw did disappoint a bit: 1.2A compared to 1.7A for a stock sk68-clone with XR-E led. I should have changed the driver after all, the led needs around 1.8A for best performance. I will put in a different driver eventually :-) .

The output zoomed out is 156 lumen and that is very nice: if you compare that to the test nubers above you see that 1.2A with a reflector and no lens gives 185 lumen, so even though this is an aspheric, it gets close to the output of a reflector light. This is one advantage of a narrow beam :-) .

Zoomed in there is still 113 lumen leftover, so that is 72% of the output of zoomed out, and that is really good (another advantage of the narrow beam); the stock XR-E version has 62%, and my dedomed XP-E2-mod does only 55%.

Considering only 1.2A current, the throw is ok: zoomed in 9.8 klux at 1 meter (zoomed out it is 0.83klux at 1 meter), compared to 14klux for my XR-E version. If the current would have been 1.7A as well, throw at 1 meter would go up to 12klux and that is close to the XR-E (but in 4000K and 96CRI ! :-) )

White wall beamshots, all taken at 4 meters with white balance on daylight and the flashlights zoomed in, the stock sk68-clone with XR-E is there for comparison, first overexposed to show the rings of the XR-E:

now exposed a bit less to have a look at the led-dies, this particular oslon led also has some blue, at the same spot as the other one, the die-size is somewhere in between the XP-G and the XP-E/XR-E:

and finally for a nice surprise, if you put a thicker o-ring where the lens part screws into the head (so that the lens is a tiny bit further from the led), the focus is a bit above the led-die in the cilindrical part of the led-dome, resulting in one of the holy grails of aspheric flashlights: a round hotspot, fairly even illuminated :-) . (Throw is not affected by this focus shift)

Concluding, this led does very nicely what I expected from it in an aspheric: because of its narrow beam less losses than other leds, with a relatively nice throw for the humble output of the led, with as a bonus a round and even hotspot. Osram makes the SSL80 in a newer version as well and also in much better efficiency than this one if you do not care about the high CRI. I am sure those will kill the old XR-E in both output and beam!

Also worth trying in an aspheric is the Osram Oslon Black (also available in all sorts of tints and CRI), with the same beam angle as the XR-E (90 degrees) but with more 'modern' output. I am planning to buy some of those when my funds allow it again :-( and perform some tests.

Nice! :)

Thanks for the reply RaceR :-)

perhaps a glass lens to reduce the chromatic abberation would do more justice to the hi-cri >_<

Not sure about that, but correct me if I am wrong :-) :

Chromatic abberation, at least what I can think of, causes hardly a problem in 'led die projection' (which is what aspheric flashlights do) because the object to be projected is a more or less evenly emitting led surface, so color shifts mostly cancel each other out, except at the edge. At least in this flashlight I can not see much gradually shifting of tint going from the center to the edge of the beam, zoomed in or zoomed out, it looks worse on the picture now that have another look at them. Actually the zoomed out beam at a short distance produces a more beautiful illumination than the led in a reflector because of the even illumination.

The glass aspherical lenses made fore flashlights that I have seen thusfar have worse optical properties than PMMA ones, to reduce chromatic abberation you need achromats (at least a doublet made of two different types of glass), this may be way too fancy for a flashlight, where you mostly do not not want a sharp image of the led die anyway.

very interesting post, thanks for sharing your playing experimentation. :bigsmile:

that’s interesting info, thanks djozz

a while back i bought a glass aspheric to replace a pmma one which i wasn’t happy with… yes it is the edge aberrations which are most noticeable to me, and only when zoomed in

the focal quality of the lens wasn’t good, so i gave up after that :confused: it did however lesson the aberration a little.

i guess glass optics are much more difficult to manufacture, but in principle i still am interested in sourcing some since they should exhibit less dispersion

Just random thoughts regarding output/efficiency based on your posts Djozz.

Your graphs are great by the way :-)

And they show that (If I´m reading this correctly)

Vf A Lumen

Nichia 3,6 1 205

4,1 2 340

Osram 3,2 1 160

3,5 2 250

And if you about watts and L/W you get

Vf A Lumen W L/W

Nichia 3,6 1 205 3,6 56,9

4,1 2 340 8,1 41,9

Osram 3,2 1 160 3,2 50,0

3,5 2 250 7,0 35,7

Which somehow makes the difference seem so much less to me.

Thank you for the insights to you testing. Interesting emitter for sure.

Do you know of any good PWM-based drivers for this LED?

The Vf is not much different from a XP-G(2), so this led is very suitable for using a 7135 based lineair driver in a 1xli-ion set-up. I'd say 1.5-2A is a good current to get the most out of the led, so perhaps a Nanjg 101-ak (1400mA) will do the job wel, or the 6x7135 driver from Fasttech.

Thanks for your interpretations, and for comparing the efficiency of these leds, it is only fair of course to take into account the voltage difference as well. To complete the comparison, you should also compensate for that during the Nichia 219 measurements the reflector had a AR-coated lens on top, and the SSL80 measurements were done without lens (why do americans call a flat piece of glass a lens by the way???), that should add 1 or 2% to the Nichia values.