4x E21A 9080 CRI output test by Texas_Ace - Amazing tint and good output

Here is a test I meant to do a few weeks ago but just never had the time.

Clemence sent me a few E21’s to test and I have to say that the tint is really nice indeed, I have not put it in a light yet but test bench says it is one of the best tints I have ever seen if not the best.

The output on a single die is not that great as expected but still impressive for what it is. The output with 4 dies on the other hand is very impressive considering the 90CRI and that it is a 3V LED setup.

This would be perfect for smaller light such as an EDC where raw output is not as important as color.

Thanks TA excellent work.

Wow thanks for your effort. Looking forward to seeing this emitter in flashlights.

Thanks TA, looks superb.
Would it have a donut hole in a light with condensing lens placed much closer than focal length? I mean Klarus Mi1C, UTorch S1 Mini and the like…
And when comes to output - there are even more efficient variants with lower CRI, though I can’t complain about this one either. :slight_smile:

With a beaded TIR optic the beam is very nice and smooth but then that is how those optics are supposed to be. I am using them in most of my builds lately.

In a reflector light, yeah, there will be a donuts hole for sure.

For the condensing lens, I assume it would have a dark spot in the middle but I have not tried it myself.

Assuming CRI70 works as expected, this LED can deliver 2400 lumens at 100 lm/W. Not close to XHP50.2, but nevertheless very good.

I did some quick ray tracing of how it would look in a lens way off centre. If I imagine another LED on the opposite side of the lens, looks OK to me…

Great test as always Tex! :+1:
With condensing lens as Agro simulated the beam is perfectly uniform, depending of the focal lenght and position. I tested it (put the lens directly on top as if it’s the dome).

So now we know the R9080 peaks at lower ~2A. My tests (and also Jensen567) with R70 peaked at 3,1A

- Clemence

……BTW, look at that extremely low voltage at high current. In my Armytek Tiara mods, this equates to longer & flatter runtime.
In 12V config it’s more than 0,5V difference at 2A compared to Cree XHP50. With 3V parallel the voltage would be even lower than most 3V LEDs.

- Clemence

Yes, this is the lowest Vf LED I have tested so far IIRC.

Thanks for doing this test. Very impressed with output and low Vf compared to 219C.
Please keep us posted on how you end up using these LED’s in your future mods.


I really can’t decide where I want to use them, I want a beaded TIR optic to smooth the beam, which limits the options somewhat.

Low Vf in not a good sign when Vf almost stops rising with current, that is sign of overheating (Vf drops by ~2mV/C). Over 5~6A Vf rises too slow compared to lower currents.

Overheated or not, any LED in parallel will usually have lower voltage than individual. The internal resistance divided. I was talking about the already low Vf became so much lower when configured in 4p.

The low Vf is consistent from low current to max compared to a single 219C (it should be around 12A with the D320 R70). I know what you meant when the die overheats the voltage will drops. Judging by the voltage vs. output curve it’s not “yet” overheated for BLF standard. It is indeed much hotter than the R70.

Below is 6V configured, same board but with 4x D320 R70 E21A. The max current for 6V is approximately half the 3V configured. Texas_Ace stopped the test I believe because of his PSU could not feed it no more. In below 6V setup, it was peaked at 6A or 12A if we were to test it in 3V config.
FYI the phosphor temp will always hotter than the die (Only for E21A case). The difference can be as high as 30C in peak current.

- Clemence

My max current recommendation for this 3V quadtrix is 9,6A. As I said somewhere in the E21A thread. But you know, people like Texas_Ace and DBcustom will always enjoying the LED’s near death bright light experience. :cry:

- Clemence

Do you have measuremens for CRI90, how much hotter they run compared to CRI70?

How are you measuring actual die temperature?

- Normal heavily sinked and cooled the output peaked at 3,1A with very marginal output rise from 2,9A to 3,1A (it’s not really worth it)

- From 3,1A and above the output dropped sharply.

- At 3,4A all the silicone in the phosphor layer started to burn

  • Phosphor layer scraped, watercooled, the blue light kept increasing until the whole die completely burnt at ~7A (this is single E21A)
    This evidence is a sure thing that the die would not overheats at the output peak current, only the phosphor layer. The phosphors itself can withstands ~200C but the silicone binder doesn’t.

We didn’t post the R9080 measurements due to inconsistency in measurement. But it was indeed hotter than the R70. I think Jensen567 still keeps the data.
E21A has special construction unlike ceramic based CSP LED (XD16 for example). Large portion of the phosphor layer are is isolated from the base due to the use of reflective barrier. It’s the white silicone-composite-something stuff surrounding the die.

The above measurement taken by Nichia was using normal MCPCB with very high thermal resistance unlike what we use in our boards (I believe we use the same technology :wink: ). With VR16SP4 board we can push the E21A far beyond it’s rated wattage. This also means the temperature difference also lower. Me and Jensen567 did many tests and found that the limiting factor is the phosphor layer not the die. We also found the phosphor layer max temp at max output always somewhere near the 180C surface temp.

- Clemence

Are there any drivers that could take a quad E21A to 10 amps? Over on reddit it was pointed out that with the Vf so low it seems FET would be a bad idea.

I’m imagining a high amp quad E21A build. Like a high cri BLF A6 competitor. Or maybe even a quad quad E21A build (four clusters of four) with a regulated driver that could push 15-20 amps. Maybe in a body similar to the S41.

I know about a-lot-of-stacked-7135 or LD-A4

So the E21A single-die here is really 4500K (Clemence doesn’t offere those in the store…)? Is it 70CRI or R9080 like the others?

No, it’s a mix between 4000K & 5000K. The real pure sm453 will be there together with the 219Bs shipment

- Clemence