Here is a quick picture of my C8F with Luxeon V’s . You can see that one is only lighted half. I did not know that one-quarter and three-quarter were also possible. You can not see individual dies though, I do not know how the led is built up.
That’s fair enough. It’s just that as i was unknowingly tearing the dome off an XHP50.2 it looked like one quadrant was failing, then that three were failing with a similarity to your picture.
Physically it’s clearly one die (see page 5) but it seems internally / electrically it’s 4. This is from the STEP file provided by Lumileds, you can’t see it in this projection but these pads are on the top plane of the substrate.
I hadn't read through this whole thread, but I just skimmed most of it. I'll definitely limit my remaining Luxeon Vs to below 6A (assumed in the case of the D4) to prevent killing them. None have seen one-on-one time with a FET yet, so they should all be fine.
So out of 4 lights, 2 look like the lower one and 2 look like the first one. But one of the LEDs that look good is reading much lower amps then it used to.
Luxeon V has extremely low Vf so I will presume that direct drive is the culprit, and that such degradation might not occur even when overdriven as long as the LED does not see unregulated high voltages.
But from my experience, the failing Luxeon V’s is not a one-to-one result from overcurrent. The failing Luxeon V in my C8S was already the third one failing, after three reflows (on the challenging triple 3535 board ) and again a failing led I called it a day. And from the output (a bit over 4000 lumen when all 3 leds still worked) it is drawing under 15A (I could measure actual current again) so not really extreme for the V. On the other hand I have one Luxeon V in a VG10 with BLF-A6 driver and all bypasses doing over 2000 lumen, and still going strong!
I have wondered if, for super low Vf LEDS, high cell voltages may induce damage even if the continuously delivered current is not excessive. Also, LEDs are sensitive to reflow temps and duration, maybe some more than others, and possibly even moisture content as per the datasheets.
My little OTR M3, though modded to FET+1+bank, is running on a 16340 so can't achieve high amps, still damaged the LED. There may be something else goin on.
At least I do pay a bit of attention to my reflows, not in a frying pan on the stove but on a “regulated” heatblock that does not exceed 240 degreesC and with the board sitting on a separate small piece of aluminium placed together on the heat block so that the heating up and cooling down is not as instant as with the bare board on the heat block. Of course this still does not follow the reflow-instructions.