Lumileds Luxeon V, test of a 4000K 70CRI emitter

I got them reflowed on 4040 boards too.
Very well done reflow indeed.

That’s the weird thing.
When you test them with very low current, it looks fine.

No mechanical damage on mine either.
The picture does show a small slightly blueish area, but that’s because of a sloppy, hastily put together aspheric test set up.

No, it’s just a sloppy set up, the lens was at a slight angle. Prism like tint shift effect.

Seem like the failures are most likely a manufacturing defect (as opposed to a design defect) as mine have been tortured with excessive handling and heat from reflows and high current. Took a pic in Turbo mode under welding googles:

EDIT: I just reread the above posts at the problematic emitters. I missed that one was working fine and then went bad. Seems unlikely to be a manufacturing defect when that is considered. Djozz's hypothesis makes a lot of sense. Would be informative to see how evenly the failed emitters lighted up with the domes completely removed.

If I remember right maximum current decreased also when the LED got damaged. Don’t have a flawless LED currently for comparison.

^

When you look at that emitter from the sides that the thermal pads extend out to, do you see indication that there is adequate solder between the emitter's thermal pad and the base?

EDIT: That may not work. The thermal pads don't extend all the way to the edge. Probably the best that could be done is to unreflow the emitter and see if it appears there is an adequate amount of solder on the thermal pad.

Checked 2 boards, there was no excessive solder on the visible ends of the center pad. So I unsoldered the LEDs and there is solder on the center pads of LED and board.

^

Thanks for checking. We expected that based on who reflowed the emitter, but it would have explained the drop in current and damage to your emitter if you had found otherwise.

If you don't think you will ever use that emitter in its current state, you could try a hot dedome. I would do it by reflowing it back onto a base. Before it cools (but after solder locks up), lift off the dome with a pick. The dome might break up. If it does, just try to lift of as many pieces as you can while trying to not remove phosphor. If you do remove phosphor, that is ok, but you will be left with a UV or Royal Blue emitter. After the emitter cools and you hook it up, put on some UV/Blue eye protection (if you lost phosphor) and power the emitter up. If the light is now evenly spread across the die, djozz's theory is correct (i.e. There was some separation of the dome from the phosphor) and you now have a dedomed Luxeon V.

Because?
What if the chip wasn’t properly soldered to its 4040 base?

You can hypothesise all you wish, but the LEDs were not mechanically abused, only saw 3 Amperes max.
The blueishness is only there in the picture because i had the little aspheric a bit askew.
You can see on the opposite side the yellow sticks out.
(I thought i explained that already…)

@ImA4Wheelr:
I might try that when I find the time but that wouldn’t explain the lower maximum current after the damage, I guess. I don’t know how this LED is build but I rather suspect some broken electrical connection in the LED. My damaged LEDs look exactly like the one shown by Jerommel, and with a separated dome I would expect some variation.

‘Expert’ is very relative here, everyone here is a good meaning amateur, just as you are. No one is claiming to know what really happened, and every one hopes that we will find out together how it can go wrong because this is one of the more interesting leds at the moment.

Edit: you re-phrased your post :wink:

I'm detecting a bit of frustration/anger Jerommel. If I sounded like I was trying to say you're to blame for the damage to your emitters, I apologize. I'm not. I just really, really like this emitter and want to know what is causing these failures. I wasn't counting the reflow as part of the manufacturing process of the emitter. I completely agree that if the base's thermal pad is not adequately connected to the LED's thermal pad, bad things will result.

If I come across as trying to sound like I know what I'm doing, I again apologize. I think I tend to write in a way that makes me come across as sounding authoritative. I'll try to watch out for that more.

I already deleted “experts”, it sounded too offended.

Both LEDs look like new, mechanical damage like you proposed would be visible to the naked eye.

Yes.
My first reaction when i noticed a problem was: Jerommel, what did you do wrong this time…
Then i doubted the reflow on the 4040 PCB, so i desoldered it, only to find out it was reflowed perfectly by Neven.
I had a somewhat similar thing with the XP-G3 in the Manker E02, only that one was half lit on low current and fully lit on high current (by the way, i replaced that one with a 219C).

I’m over it already, don’t worry.
Yeah, i was a little annoyed…
I would never claim something to be a faulty product if i wasn’t certain i didn’t damage it myself in any way.

I can understand you don’t want this product to have problems.
Neither do i.

Your suspicion sounds quite plausible, but lack of variation does not seem to rule out separation. If there is some defect in the dome adhesion process, it could result in similar looking failures.

EDIT: Fixed typo. Didn't include the word "not".

Cool. I don't think djozz was trying to say you caused the damage. I thought he was just asking a reasonable question to gain data. I imagine dome separation could be caused by a defect in the manufacturing process. That could explain how I could beat the hell out of mine with no failures, while other emitters failed for no apparent cause.

Ok, I was impatient (as mostly) and “cold” dedomed the LED including phosphor using my fingernails and tweezers.
Was hard to take close pictures with my low end smartphone but finally I put it on top of the lens of my 3rd hand which also (hopefully) helped to keep my eyes undamaged since I don’t have UV protection glasses. This is the (still ugly) result:

The orientation of the 2 pictures might be different since I moved the MCPCB after taking the first picture.
It’s only the inner square of the LED surface which is lit and I can only see a small difference between the halfes. But I powered the LED with a couple of milli amps only since otherwise my smartphone camera had been blasted. Might be there is a bigger difference at higher currents.

Edit: in the 1st picture the LED is not powered and the inner square doesn’t look patterned as seen in the picture, it’s even.

Actually the inner square is not completely even. Take a look at my first picture, you see a vertical line, a tiny bit left from the middle of the inner square. You don’t see it clearly with naked eyes, but with my tiny microscope there is very narrow line visible. It doesn’t appear to be a crack, more like the remaining kink in a paper sheet which has been folded and unfolded again, somewhat elevated. Don’t know at all the meaning of this, perhaps a conductive trace underneath the kink has been blown.

So it looks like the fault is in the die? The Luxeon V may be a double die led? Or so? And you can (partly) blow up one and the other keeps going.

Maybe…
But the strange thing is that when you do a continuity test (polarity check / “beep” it) with a DMM the whole thing does light up like any other LED.
Only when you push more current through it (like 150mA or more), the problem becomes evident.

Now we're starting to cook with gas. Thanks to Flashy Mike's and Jerommel's last contributions we know a lot more. It looks kind of like a layer is not properly lined up. Here is a diagram of some other type of chip scale package LED. I like it because it gives a sense of how the electrical pads connect to the layers that create light and, therefore, how come these types of emitters are so much harder to fry (due to connection with no bond wires).

What doesn't jive with that possibility is that Jerommel had an LED that initially didn't manifest the failure.

EDIT: Maybe the phosphor doesn't hold up well to the layer it is resting on that should normally be covered by another material. So over time the defect reveals itself.

EDIT2: I unpackaged my remaining 6 emitters and just momentarily tested them at 200mA. None appeared to have the defect. All 10 of my emitters are the 4K tint from Mouser. I want to again thank Flashy Mike for taking one for the team to help gain insight into this issue.