LED test / review Osram SYNIOS P2720 DMLQ31.SG white - Small chip, but big in luminance - at low current!

Based on looking into the front of the lens from a distance, nearly the full 62mm diameter is active. This area estimation seems to give a consistent measure of the dedomed xpg2 luminance at 510Kcd at 4ish amps.

:+1:

I assume that “luminous intensity” is the cd measured directly above the LED. That does not necessarily indicate a higher die luminance. The fact that the 219c and 219d have different proportionalities between the intensity and lumen output means they have different emission patterns, which is supported by their different quoted “directivity” angles in the links above. So the 219d simply has a different shaped dome which directs more light forward.

Also your estimated luminance of 154cd/mm^2 for the 219c d320 seems high. My measurement in the UF1504 puts the effective luminance at 100cd/mm^2 at 5.5A. Assuming 90% lens transmission that is 111cd/mm^2 die luminance. And according to Texas ace’s test of this LED the max output is only 5% greater than the output at 5.5A.

Good point regarding the emission angle.

Agro calculated the 154cd/mm^2, not me.

The front site shows that the LED typically reaches D320 bin, but they list D340 in the datasheet.

After posting I got a hunch the result was too high. Probably I made a mistake somewhere.
219D won’t be so good then….

Important notice:

I updated the luminance values for all LEDs shown in the test chart in first post, including SYNIOS DMLQ31.SG!

To ensure more realistic values for 'real-life' conditions (flashlight use / secondary optics) I established a new testing method to determine the luminance with a SMO reflector. All upcoming and already existing tests will use this method!

Updating of my other tests will be done in the next hours / days.

The SYNIOS are still very nice LEDs with high luminance (at max current 195.0 cd/mm²), but at much lower current as other known LEDs (Black Flat, XP-G2 S4 dd).

I’m doing some calculations…
All flat LEDs, emitting light only on the top surface should have the same cd/lm.
I compared several of koef3 tests and found:

Dedomed Luminus SST-40 does 0.324 cd / lm
Cree XHP 35 HI E2 does 0.318 cd / lm
Blackie does 0.294 cd / lm
Cree XHP 35 HI C4 does 0.293 cd / lm

About 10% variance, not bad.

But now let’s do the same calculations for Synios P2720 DMLQ31.SG. Result? 0.213 cd / lm.
Can anyone explain that?

ADDED:
If I get wiki right, the correct result is 1/π cd/lm or 0.318 cd/lm. So all LEDs other than Synios are within 8% of it.

ADDED:
So if this LED delivers as many lumens as it was measured to - it should do about 292 cd/lm².
Your test without reflector has shown 295 cd/mm²……what’s up?

I think you might be overestimating the accuracy of optical measurements. koef3 goes to a lot of effort to increase it, but there are practical limits.

That’s 33% too low. When all 4 others are within 8% of the expected result.
No, I don’t think this is a measurement error. Or even several, adding up.
I think the emitter doesn’t do what we expect or there’s some error in the method.

There was some talk awhile back (and I can’t remember if it was confirmed or just speculation) that the XP-G3 had a 3D (not flat) phosphor coating, and that maybe that’s why it doesn’t get good throw performance, even though it has better cd/mm² than older generation emitters.

Anyway, I don’t know if any of that was ever completely figured out. But, if the Synios has a non-flat phosphor coating, that could explain the low cd/lm compared to “standard”.

The SYNIOS has some white areas around the die.

There's light shining through or in this area and this has a big impact on the luminance, what the real-world measurements (with reflector) shows. See also my pictures from the test. I don't know that the SYNIOS has a three-dimensional die like the new generation Cree LED platform.

The synios die is completely flat, but it is indented so that the white area around it with a “bowl” shape acts as a mini reflector.
So the light is coming from not only the die but also the area around it.

This effect means that direct intensity measurements are higher than they should be and should have no impact on measurements with a reflector. They should have marginally negative effect on measurements in an integrating sphere.

What I describe means that the output measurements with the integrating sphere are too high compared to intensity measurements with a reflector.

This is correct.



Integrating spheres measure flux output.
This small indent in the surface of the LED does not affect the flux output, it remains the same.
Just because the light is directed forward a bit more does not change the measurement done by an integrating sphere, as the sphere is designed to perfectly diffuse the output and ignore almost all directionality.

That’s pretty much what I wrote. The small indent absorbs some light, so it reduces flux. But that’s purely marginal.

If the emitter is 0.503 mm² flat Lambertian surface which generates 461.2 lm, it should do so with about 292 cd/mm². Real 292 cd/mm², not abnormal readout caused by incompatibility of measurement technique with construction of this particular LED.
The measured intensity is 195 cd/mm², way lower than prediction.

So either:

  • it is not Lambertian (from the datasheet - it is actually not Lambertian, but extremely close)
  • it is not flat (seems it is)
  • it is not 0.503 mm² (I guess it is)
  • it does not make 461.2 lm
  • it really does about 292 cd/lm² and for some reason the test with reflector can’t show it. As well as results from the few people who tried this LED in flashlights.
  • there is something else?

Raw data

Osram SYNIOS DMLQ31.SG white

Amps lm Vf
0,05 22 2,76
0,10 40 2,81
0,15 59 2,85
0,20 75 2,89
0,25 93 2,92
0,30 108 2,94
0,35 122 2,97
0,40 136 2,99
0,45 151 3,01
0,50 164 3,03
0,55 179 3,05
0,60 191 3,07
0,65 203 3,09
0,70 216 3,11
0,75 228 3,13
0,80 240 3,14
0,85 251 3,16
0,90 263 3,17
0,95 273 3,19
1,00 283 3,21
1,05 295 3,22
1,10 304 3,24
1,15 315 3,25
1,20 325 3,27
1,25 333 3,28
1,30 342 3,29
1,35 352 3,31
1,40 360 3,32
1,45 368 3,33
1,50 375 3,35
1,55 383 3,36
1,60 390 3,37
1,65 398 3,39
1,70 404 3,40
1,75 411 3,41
1,80 417 3,42
1,85 422 3,44
1,90 428 3,45
1,95 433 3,46
2,00 439 3,47
2,05 442 3,48
2,10 447 3,50
2,15 450 3,51
2,20 454 3,52
2,25 456 3,53
2,30 459 3,55
2,35 460 3,56
2,40 461 3,57
2,45 462 3,58
2,50 461 3,60

If you create diagrams / comparisons with this data, I would ask you to publish these charts in this topic and also to specify the source of the data used.

I see there’s a new Osram SYNIOS P2720 DMLQ33.SG on the Osram website.
Weirdly, it seems exactly the same DMLQ31.SG. I compared datasheets page by page and failed to spot a single difference.

EDIT: fixed link

Higher maximum forward current, lower thermal resistance, binned at 700mA, lower forward voltage (new 800mA at 3V, vs 400mA at 3V old), lower tint shift (at constant temp. and changing current), bigger die

I posted a wrong link above, sorry. You’re quoting DMLS and not DMLQ. Which is both 31 and 33 now as well. I haven’t compared those.
The correct link:

33 is a new generation, not a big deal, just slightly better.

I noticed that the datasheet shows a cutout through the LED. It shows that the die with phosphor is about 0.06 mm thick mounted on a pole.
This allows me to calculate the side area which may be a part of the LES.
That extra area is about 0.18mm².
If sides had the same luminance as the front this would account for ~27% luminosity gap against theoretical calculations. In the actual measurements the gap turned out to be ~33%.
Now…I would be surprised if the sides were the same luminance as the front. But that would explain the discrepancy.