Future development of the maximum luminance of LEDs

After quick look it looked like a winner.
But looking closer: it’s rated for mere 1.5A which seemed suspicious (White Flat is 3A).
Output scales nicely with current but Vf grows fast. So efficiency drops quickly.

I don’t expect it to challenge Osrams but I’d like to see it tested as well.

According to manufacturer, Getian GT-FC500X3-3.0 does 100-120 cd/mm². :slight_smile:

Can someone share some ways to measure or calculate [cd/mm2]?

I know how to measure cd and know my area size.
I have lumen tube, lux meter, laser power meter, and other equipment.

I never got around to actually calculating this value but now it is quite important to know because I am building LEP

Regards Xandre

Several ways.

  1. There are calculators from Enderman that let you estimate throw. You can use them in reverse, adjust luminosity so it matches your actual measurements:
    Reflector Type 1 (metric)
    Lens Type 1 (metric)
  2. A more direct way is to ditch optics and measure cd over bare die. This however requires to know your light emitting surface area. Which is frequently hard to establish with many LEDs emitting not only from the top but also from the sides and can lead to large errors.
  3. Similarly, you can use lumens instead of cd over the die but you still need to know the size of LES.

Thanks for the information.

Sorry, I am still a little confused on how to put everything together.
Like how you can get throw from know Lux and Distance alone, is there a way to know die illuminance in a similar way?

@Xandre
I have a lumen tube that can get OTF lumens. However, I do not have an integrating sphere so I am unable to get exact lumen measurements from the bare die I am trying to measure.
Do I enter 180deg as my apex angle?

@Agro
If I were to measure a bare die with a Lux meter, what would the general equation be?
Would I need to measure at different angles to know my emission pattern in an equation?

For example:
Lux meter @ 1meter = 150lux
die area = 0.04mm2

150 lux @ 1m = 150cd
150cd / 0.04mm^2 = 3750cd/mm^2 (very unrealistic, only short arcs can get this intense, LEDs are 10x less)
Anyway, you multiply that by the front area of your reflector/lens (again in mm^2) to get cd again.
(For example a 25mm diameter lens has area 491mm^2, which would be 1.8 million cd with that impossible tiny LED)

Let’s ignore the optic and just focus on the 150cd from earlier.
You use cd to get the throw using the inverse square law of light.
(150cd / 0.25) ^0.5 = 24.5m
This is the distance where your luxmeter would read 0.25 lux (ANSI standard for throw)

You can also check in reverse:
24.5^2 = 600, which means the lux at 1m is 600x higher than that at 24.5m
And 600x 0.25 = 150 lux @ 1m

Wwow.
I just realized after reading your explaination. All I literally had to do was divide by my surface area! cd / area! Lol

Thank you so much for showing the relations between lux and cd. Everything finally makes sense.

Yup!
No prob :slight_smile:

Note that accurately measuring the luminous intensity (cd) of a bare LED is actually quite difficult in practise. You need to make sure that there is no light going to the sides that is bouncing off of objects/walls and then hitting the lux meter sensor. This would artificially increase the measured values. A matte black pipe and some pieces of black foam could help.

Back calculating from the luminous intensity of a LED plus reflector is easier to get right.

It’s not that difficult, you use some light absorbing felt paper to form a cone around the LED with just a small hole in the top that lets light out, then place the luxmeter 1m away.
By changing the hole size you can make the spot of light coming out whatever size you want, so nothing else in the room is getting any reflected light, just the luxmeter.

Calculating from a reflector is inaccurate unless you know the exact fronal area.

Well in each case you have to know the area, and I would say it’s easier to measure the area of the reflector then the light emitting surface. The real problem with measuring the luminance of the LED directly comes when the light emitting surface is not completely uniform, like when you have light emitting from the areas to the side of the die. Then I would say it’s practically impossible to back out the luminance of the die itself.

Based on what I am reading so far, it seems measuring with a reflector may have an issue. What if the emitter has a very wide emission pattern that benefits the use of a reflector? Compared to an emitter that has a narrow emission pattern and most of the light does not effectively use the surface area of a parabolic reflector?

Do most LEDs have similar emission patterns?
Do you think LEP phosphors also have similar emission patterns?

“emission pattern” What is the correct/shortest term to call: “the distribution of light based on angle of incidence”?

_

I will try to measure with the bare die method because I really need to know the emission pattern of LEP phosphors. If I design a collar that is not 60 degree aperture, but possibly 70 or 80 degree, it would be beneficial to know how much light I can capture.

You don’t need to measure any area when the specifications of an LED already tell you the light emitting area to a few hundredths of a mm.
For flat LEDs of course.
But nobody uses anything except flat LEDs or dedomed LEDs when they care about throw this much.

Flat LEDs like the black flat, white flat, CFT90, SBT90, etc. all have pure Lambertian emission patterns.
LEDs with silicone domes basically act like a magnifying glass on the LED so their apparent area is larger, causing the decrease in intensity that inclines people to slice them.

Nichia now lists a new headlamp LED, NV3W470A.
It’s a high output, high intensity quad die.
4.9 mm² and about 4800 lm ain’t bad though the LES is much farther from round than most. And it needs a 7A 12V driver.

That is quite impressive lumen density. But 1.4x3.5mm LES is not ideal for reflectors. It will have the hotspot size of a typical 1.4x1.4mm LED but with a large corona.

From page 4 of the datasheet it looks like it is 3 dies in series.

Yeah, I think it would be interesting in aspherics. And bike lights. But in reflectors not really.
After a closer look at the Vf curve (both low current and change with temperature) I agree - it’s pretty surely 3 dies.

I wonder if some XHP70 driver would work well with it.

Just found a 1.3 mm² Nichia NCSW321F.
The regular power rating doesn’t look impressive, merely 1.5A. But thermal resistance of 3.8K/W suggests it may overdrive well…

To me, it seems like it might top out around 4-4.5A at around 2.7x binning output. So 1050-1100 lm maybe? About 25% more LESA than NM1 Osrams, so maybe 20% short on illuminance.

The NCSW170F-SA however is another story. 0.95x0.95mm >> 0.9mm²

Typical flux at 1000mA is 380 vs the 1.3mm² 410. Relative flux curve looks the same. This could really compete, but I hesitate going below 1mm² except in small diameter optics where the beam angle can stay above ~3°.

EDIT: NCSW131F-SA shares the same specs as the 170, but uses a 3 pad footprint instead of two.

Also, just caught on that the 321F is meant to be epoxied to its heatsink and electrical connections from the top. Very interesting. Makes a lot of sense for automotive.

Hoping for a NCSW270F-SA to come along

There’s a new Osram LED family, Ostar Projection Power.
There are single emitters as well as 2~~, 4~~ and 6- die arrays.
The dies are 2 mm². Single die variants have lower thermal resistance than Boost HX.

So far there are only:

  • amber
  • green
  • blue
  • deep blue

colours available.
So no apples to apples comparison, Boost is only white and these are only colour. They offer higher currents (12A pulsed for blue compared to 8A of Boost HX) but at least a part of the increase is the fact that it comes without phosphor.

So….white may come eventually and beat the current luminance records. Or it may not.
I fantasize about applying some high CRI phosphor over them to make a gorgeous thrower…but that’s far out of my limits. :wink: