Nichia 519A

[QReciprocity42]

Djozz, before tossing the damaged emitter, could you slice the dome? I wonder if the new glass+silicone design makes it easier. I would also expect the amount of decrease in output, cct, duv to be different from emitters with homogenous dome composition.

[THE_dAY]

@djozz

]
I remember Bob’s test but can not find it now. What were his numbers? (I tested in the hotspot of a smooth reflector, not on the bare led, did Bob do that too?).

Btw, the Duv is dependent on current, at lower than 2A the CRI goes up a bit and the duv also, getting even slightly positive, at higher than 2A the duv is lower and CRI goes down a bit.

Here is a 519A test by him from reddit:
https://www.reddit.com/r/flashlight/comments/rmzqe9/ned_armytek_wizard_c1_pro_with_nichia_519a/

Note that his sample is 3 step sm453 and Simon has 5step sm455 according to thefreeman mentioned earlier in this thread.

[djozz]

@THE_dAY

@djozz

]
I remember Bob’s test but can not find it now. What were his numbers? (I tested in the hotspot of a smooth reflector, not on the bare led, did Bob do that too?).

Btw, the Duv is dependent on current, at lower than 2A the CRI goes up a bit and the duv also, getting even slightly positive, at higher than 2A the duv is lower and CRI goes down a bit.

Here is a 519A test by him from reddit:
https://www.reddit.com/r/flashlight/comments/rmzqe9/ned_armytek_wizard_c1_pro_with_nichia_519a/

Note that his sample is 3 step sm453 and Simon has 5step sm455 according to thefreeman mentioned earlier in this thread.

Thanks! Yes, so all sorts of reasons why his tint differs from mine, batch variations, tighter colour bin, perhaps a different current, and a different optic (a reflector hotspot that does not use the light from the cone straight above the led tends to increase the duv more than a TIR optic (that mixes light from all directions).

[thefreeman]

@Agro

Do you happen to have the raw data behind this chart at hand or did you dispose of that already? I’d like to compare efficacy and luminance as well…

3535 90CRI comparison

[Agro]

Thank you. I’m not into pixel counting now and I see that I don’t actually have apparent die size of 519 but at least I can compare efficacy.
And LH351D wins big time, needs ~10W to reach 1000 lm while 519 needs ~12.5W. At other currents the difference is similar, though slightly lower at the extremes.

That’s sad, I hoped 519 would be at least close. But it is what it is.

[thefreeman]

Yes there is a non negligible difference:

But the LH351D is annoying with its greenish tint, I only use it sliced, then the efficiency is slightly lower.

[thefreeman]

@Agro

Thank you. I’m not into pixel counting now and I see that I don’t actually have apparent die size of 519 but at least I can compare efficacy.

I got 6.37mm2 using webplotdigitizer

[djozz]

@QReciprocity42

Djozz, before tossing the damaged emitter, could you slice the dome? I wonder if the new glass+silicone design makes it easier. I would also expect the amount of decrease in output, cct, duv to be different from emitters with homogenous dome composition.

I wil do some surgery on the damaged emitter, it may not be before this weekend though.

[SKV89]

@thefreeman

Yes there is a non negligible difference:

But the LH351D is annoying with its greenish tint, I only use it sliced, then the efficiency is slightly lower.

Excellent table. Thanks!

[Haukkeli]

These seem to run hotter than LH351D’s. I have two Convoy S2+ triples that are built with same parts. Only difference is that other uses Nichias, other Samsungs. Nichia one gets hot really fast with full power.

[JaredM]

@Haukkeli

These seem to run hotter than LH351D’s. I have two Convoy S2+ triples that are built with same parts. Only difference is that other uses Nichias, other Samsungs. Nichia one gets hot really fast with full power.

What driver?

[Haukkeli]

H17Fx.

[thefreeman]

@Haukkeli

H17Fx.

That’s direct drive right ? The 519A has a lower Vf so the current will be higher, generating more heat.

Play with this and you’ll see : Graphing Calculator - GeoGebra
Info : LED direct drive current calculator

[Haukkeli]

@thefreeman

@Haukkeli

H17Fx.

That’s direct drive right ? The 519A has a lower Vf so the current will be higher, generating more heat.

Play with this and you’ll see : https://www.geogebra.org/graphing/raa8dkdm
Info : LED direct drive current calculator

I just measured tailcap amps with clamp meter. Nichias drew little over 16A, Samsungs 8,5A. I eyeballed forward voltages from your diagram and did some math: Nichias used about 53W and Samsungs 28W. No wonder they built heat faster.

[thefreeman]

@Haukkeli

I just measured tailcap amps with clamp meter. Nichias drew little over 16A, Samsungs 8,5A. I eyeballed forward voltages from your diagram and did some math: Nichias used about 53W and Samsungs 28W. No wonder they built heat faster.

That much lower with the LH351Ds ? Are they 70CRI (higher Vf) ?

[Haukkeli]

They are 5700K’s from Convoy store. Both have 20awg led wires and brass pillar on driver instead of spring. Driver is soldered to pill in both lamps and I used same battery in both measurements. From your diagram it seems that light output is actually quite close even though Samsungs drew that much less power. Edit: Nichias are also from Convoy Store

[Haukkeli]

Now I redid Samsung measurements. After a little wigling and cleaning I got 9,5A current draw.

[Hikelite]

Does anyone know the exact light emitting surface dimensions?

[djozz]

16 amps on a triple, that is serious current, and serious output, I guess over 3000 lumen of prettiness

[malkoff_junkie]

what will be the CCT range of the 519A?
I wish a 6500k was available in hcri
thanks.