OSRAM KW CSLNM1.TG (1mm2), KW CSLPM1.TG (2mm2) & CULNM1.TG 1mm2 version of Oslon boost HX
Some of us prefer to call them White Flat 1mm or White Flat 2mm.
This is edited thread so first available version that we managed to test is White flat 1mm (CSLNM1.TG) so first part of edited post starts with it:
Led4Power brought us this interesting emitters and I had opportunity to test this led in direct comparison with good old XP-G2 S4 2B dedomed.
Used same host(uniquefire 1503, 50mm aspherics with selected lenses), same driver(Djozz FET driver), same 18650 cell. Testing equipment uni-t clamp meter and uni-t lux meter.
Test subject Uniquefire 1503 fitted with XP G2 S4 2B has 255kcd performance. 950FC or 10225Lux at 5 meters.
Mine test started with low current 4.35V cell LG18650BD1 and I roughly got this results:
4.2V 5.35 A 1180FC, 4.1V 5 A 1150FC, 4.0V 4.8A 1150 FC, 3.9V 4.6A 1150 FC, 3.8V 4.5A 1150FC, 3.7V 4.1A 1100FC, 3.6V 3.8A 1050FC, 3.5V 3.3A 950 FC with new OSRAM KW CSLNM1.TG. It equals performance of the best G2 at very low current draw which means it will kill in lux performance any XP(E2 or G2) based emitter!
Now torture test. Torture for this emitter will be any cell in Samsung 30Q rang or better if used in FET DD setup which of course will pull a lot of current. So in mine test emitter survived over 7A of current draw for more than 3 minutes(for couple of times). After cool-down it did not loose initial properties(performance remained the same) which implies that emitter is very robust. This were mine results. I also found that mine test subject had best performance at around 5.8-6A (1200FC)
So in mine test OSRAM KW CSLNM1.TG over-performs old G2 significantly in everything except die size This small emitter is like Viagra. It does not falls in performance that easy
My advice for such single cell FET DD setup is that you use low current cell. Sanyo GA and mentioned LGBD1 will work OK but there are plenty of others that could work without harming emitter.
Crucial things could be reflowing: Use as thinner layer of solder paste as possible (needle syringe application) since on this emitter it is very hard to push excess solder off.
Or you could just buy plug and play version on L4P DTP boards. So you can pick 20mm or 16mm Led4Power uses lead free solder in re-flow process which according to him should have better properties than solder paste.Led4Power tested it:
Test results:
I calculated/estimated lumens from my lux numbers and datasheet bin specification.
Yellow marked lumens calculation are based on old black flat measurements at 1A as reference (made by koef3 I think), non-marked lumens are based on ~middle of bin lumen number at 1A from datasheet (330lm, actual middle is 335lm).
Djozz tested it:
My son and and a friend who was coming to play at our place had way too much energy so I kicked them onto the streets to play in the playground a few streets away. So that freed some time for the led test.
![]()
![]()
Note that I used my standard test method with all the added disclaimers about how it is done. Especially note that I keep using the djozz-lumen although I know that it is between 7% and 10% too high, I do this to be consistent, so results can be compared to all my earlier led tests.
I used a DTP 16mm led4power board for reflow because those pads matched the smaller Oslon pads of the led just a bit better than a Noctigon. And indeed the reflow went easy.
As I noticed before my maximum is at a lower current than other testers, the cooling seems less well done in my set-up. Part of it may be that I usually apply a generous amount of solder during the reflow, a bit more than the minimum needed (with Oslon-sized leds on 3535 boards, excess solder can not be squeezed out by tapping on the led). This results in a somewhat (not much though) thicker solder layer that may add to a slightly higher thermal insulation of the led.
I stopped the test at 7A. The led emitted angry blue light then, but I kept it 2 minutes at 7A to see if it could handle it, and it did. Afterwards I turned the current down to 4A and kept it there for half an hour, the output was only 1% less than the initially measured output at 4A, so the led seemed to have handled the 7A maltreatment with no problem.
![]()
Led4Power told that this variations can happen because of excess solder during reflow process:
@djozz,
thermal path definitely looks not optimal, I got peak at 5.75Amps. Excess solder could cause that when power density is high like with this LED. Also, I’m using lead-free solder which has a little bit better thermal conductivity.
We already have builds with that emitter:
Also posting this here, so it may reach more people that are interested.I just put the new OBF successor in my BLF GT Mini. The max I measured was 282k cd. But with a protected battery at 4.1V the current is already at 6A at the tailcap. I’ll see if I modify the firmware to reduce the max output to 80% PWM. But the LED scaled all the way up to 100%, so I’m wondering what the actual limit is.
Even at 3.65V battery resting voltage, I get 260k cd. So it was probably overdriven a lot.
Beam looks like this. Ideally the 2 rays limiting the hotspot should be parallel right?
![]()
Yesterday the new Osram CSLNM1.TG leds arrived from led4power. Putted in one to my modded Convoy S8 and one into my GT mini. Some pics and beamshots:Convoy S8, S2+ SMO reflector, AR lens, 3,15A. 9×7135. It makes 573 lumens and 30800 cd.
![]()
It is slightly smaller than XP leds
![]()
GT mini with Emisar D1S reflector. You can see the beam thinner as going away from light so it need to raise the reflector a bit. Making 219000 cd with unfocused reflector and GA cell to not overdrive the led.
![]()
![]()
![]()
![]()
BLF D80 with KW CSLNM1-TG led. (Should this led get a dedicated thread btw?)Some more checks on theFirst, I was not happy with the quite random centering that twisting a 3535 centerpiece provides, so I modded the centerpiece: with a scalpel I cut new corners 45 degrees from the original corners, fixing the 3030-sized led much better. So the flashlight has two centerpieces, one locked around the other, both sanded very thin at the underside for exact focus, now also exact centered.
![]()
With the led perfectly centered in 3 dimensions the beam was superb.
![]()
Now I measured throw again, first on the almost full 30Q (4.08V): 140 kcd (as I measured before), then let the light run for half an hour (it got hot, I estimate 70 degrees! A biscotti driver has no thermal stepdown, and a 3.9 amp flashlight produces some 15W of heat), the battery was now 3.59 V, so wel over half empty. Measured throw again and still got 125 kcd.
I did the same test with the same battery on a C8 with dedomed XP-G2 S4 2B. Started at 4.08 V with 141 kcd, and with the cell at 3.59 V I measured 90 kcd.
So the low voltage combined with the much needed current regulation of the biscotti driver makes the output and throw of a flashlight with the new Osram led more constant than a light with a direct driven S4 2B.
Plus the current is a bit lower for the same throw.
Now we will talk about White Flat 2mm(CSLPM1.TG)
I also had opportunity to play with this larger 2mm white flat maybe even sooner than Led4Power. So this is mine pre- test with classic light meter and clamp meter (so test is just for reference before someone with better equipment test it). Full Samsung INR 30 Q. Single cell 18650 light with FET Djozz driver.
4.2V, 13.4A, 930 FC
4.0V, 12.5A, 950 FC
3.9V, 11.4A,1050 FC
3.8V, 10A, 1250 FC
3.7V, 9.20A, 1200 FC
3.6V, 8.20A, 1150 FC
3.4V, 7A , 1100 FC
Somewhere at 8-9A will be just enough…
The most beautiful emitter I ever had in my aspheric light. When searching at night in zoom in mode with aspheric light it looks like wide screen TV 16:9 format
But my conclusion this ain’t emitter for single cell fet drivers but you can put it in if you insist and wanna play. Regulation is must if you don’t wanna live on the edge
Also very robust emitter. It can withstand some serious current draw. Add as thinner layer of solder paste as possible.
DJOZZ TESTED IT:
Thanks for offering doing the test but I wanted to find out myself too now.Reflowed a new led on a 16mm L4P DTP board, using enough but sparse solder paste, clamped it well and flat onto the test mount with bit of AS5 in between, and this is what I got:
![]()
So in my emitter test set-up I find the led peaking lower at 7.5A, and not at 10A as l.i. found. It may be my reflow, it may be the amount or type of my solder paste (same chinese 63/37 solder paste that many use), I don’t know.Regardless, even in my test this led puts more light out than the XP-G2 S4 2B (the led from the test is not even dedomed, substract 10% output for after dedome), with a maximum output of over 1500 djozzlumen (probably about 1400 real lumen), which because of the low voltage can actually be reached in a single li-ion flashlight (unlike the XP-G2 S4 2B, that you will not get over 4.5-5A).
I do not know why I did not get better performance in my mini-GT, even though I thought that I had done the mod well, but as said before: so many variables…
![]()
![]()
![]()
(I always treat my emitters very well)
Some of us got larger numbers than Djozz but there is so much variables that could influence on max performance of mentioned emitter but one is sure:
You can’t go wrong if you follow Djozz tests and make everything according to that test.
Because emitters will be driven just fine and they will give better lumen and lux performance than the best dedomed XP- G2 S4 2B or any XP-E2 emitter.
Other additional important stuff and test will be added here. So lets start with this:
LEDs, the dedomed XPG2 (new style), close sliced SST20, and white flat 2mm^2. The white flat is from Mouser and the item has a range of possible bins so I don’t know the exact bin. I measured the output by measuring the lux directly above the LED (80cm away) and assuming the light emitting surface is a Lambertian surface, which is a good approximation. See here some more information about this method. I use a Tondaj LX-1010B lux meter. The lumen numbers are approximately close to others’ measurements, but assume the usual 10-20% possible error.I did some measurements on 3 of the best 2mm^2
![]()
I then measured the beam intensity of each LED in an EE X6 head powered by a constant current power supply. dedomed XPG2: 95.2kcd at 4.5A, sliced SST20: 92kcd at 4A, white flat: 111kcd at 6A. I calculated the luminance by dividing by the reflector area of 620mm^2 and assuming 90% reflector and glass efficiency. Then I scaled the output graphs according to this information and the luminance vs current is displayed below.
![]()
The white flat and dedomed XPG2 curves have close to the same relationship in the output and luminance graphs, which is expected since their dies have very similar areas. But the sliced SST20 luminance curve is high compared to its position in the output graph. I think this is explained by a non-uniform luminance across the SST20 die. The SST20 has a close slice, but obviously some silicone remains on the die, as pictured here. Near the edges of the die light can reflect off the silicone/air interface and exit the die to the side. This light doesn’t contribute to the light coming directly from the die; it bounces to the side and a higher fraction of it gets absorbed or emits to the side where it isn’t useful. You can see by looking at the lit die that the edges are less bright than the center, pictured below.
![]()
The other LED dies are much more uniform in brightness over their surfaces. I think this is the reason for the SST20 having a lower output than the other LEDs. But the measured luminance is high, even higher than the other LEDs for most of the curve. This is because measuring the luminance by measuring the beam center intensity from a reflector only measures the LED die-center luminance.Conclusion:
These 3 LEDs perform relatively close to one another at a given current, but in real world use there are big differences. The dedomed XPG2 still performs well, but the high forward voltage means the current can be low, especially in a mulit-emitter light. The SST20 improves upon the XPG2 in this aspect with its lower voltage, but dedoming difficulties hold it back. The domeless white flat with its very low voltage and higher max output is the clear winner in some applications (multi-emitter), but driving a single emitter optimally might be more difficult.
ENDERMAN teseted it:
RLT collar.Did another test, this time with a small
![]()
The LED reached its peak sooner, at about 5.3A instead of 5.6, likely due to the additional heating of the die.
About a 2x improvement from the small collar.Both graphs compared:
I should also mention I do all my tests continuously, without letting the LED cool down at any time, I just keep increasing the current until there is no improvement.
None of that “X performance at turn-on” stuff.For the record, I just did a test with the large RLT collar and got 708 (small collar max was 689) so 667cd/mm^2.
There is almost no improvement from the large collar over the small, it’s still ~2x intensity increase.
Newest CULNM+1.TG *1mm2 version of Oslon boost HX L4P review below:
KW CULNM1.TG is 4.0×4.0mm version of KW CSLNM1.TG (or 1mm2 version of Oslon boost HX), with lower thermal resistance, I tested two 5Q bin samples under same conditions (fan cooled heatsink), results:
So we got the first white LED that broke 1000lm/mm2 barrier (by a decent margin in case of higher performing sample)
What can be seen from numbers is two samples are within 5Q bin rating (710 – 800lm@3A), but one is at bottom and other is closer to upper level. 5Q bin is comparable to 6N bin of CSLNM1.TG (315-355lm at 1A).
Compared to CSLNM1.TG (white flat) peak current is just a litlle bit higher 6.0-6.5A vs ~5.75A. I was surprised by this, I expected more like 7A, but it seems droop effect at these current densities becomes so dominant that output stops increasing even if you cool LED die well.
But CULNM1.TG’s lower thermal resistance does help and max. output is higher at high currents.
Also the other good thing about lower thermal resistance is lower junction temperature, at 6Amps die temperature is about 75C higher than ambient, CSLNM1.TG temperature at 6Amps would be ~96C higher than ambient. So CULNM1.TG should be more reliable /longer living at high currents.
Disadvantage is considerably higher price compared to CSLNM1.TG and the fact that it has been already “scheduled for obsolescence “ by OSRAM due to lower than expected demand.
Osram KW CSLPM1.TG (2mm2) is available now:
https://led4power.com/product-category/leds1/osram/
Kills XP-G2 in throw, while having 1500lm+ output, which is great for factory undomed 2mm2 LED with pure white tint. Vf is ridiculously low, 3V@3A,3.13V@5A,3.3V@8,5A.