Luxeon V die voltage

[koala]

So Luxeon V is dated? now I got one to play with. It is my first Luxeon V so wired it to my rig and did some testing. It is a WYOS and cruises at 6.01V@700mA which is pretty low forward voltage. This also means each of the 4 die is 3V@350mA. It's hard to find a Luxeon I that low vf , I haven't seen one before except for the Rebels.

So I did some low current/voltage testing...

0.13mA @ 4.43V

0.30mA @ 4.73V

0.64mA @ 4.84V


I did manual adjustments to the camera so it looks like what I saw. I don't know what are the long term effects when it is driven at full power. If someone has posted this before, I am sorry I gave up on the search function...

have a nice day
Vince.

 

[yclo]

Interesting, it seems the top two dies has a lower vf than the lower dies?

 

[evan9162]

I have a WWOS Lux V that also does this. At extremely low current (less than 1mA) one or more (can't remember at the moment) of the dies don't emit light at all. At 30mA, all 4 are more or less evenly lit.

 

[koala]

Hey guys, that's right I also suspect that Lumileds does not closely match the vf of the LEDs. I am not sure what is the effects at high current but I know that the top row will be slightly overdriven...

 

[evan9162]

That's not quite correct.

Look at your photos. See the bond wires? Current flows straight from one bond wire to the other. The Lux V is a series/parallel setup. In your first photo, the "top" dies are in parallel with each other, and the "bottom" dies are also in parallel with each other. The top set is in series with the bottom set. Since the top dies are equally lit, they are sharing approximately the same amount of current. And since the bottom dies are in series, they also have the same current flowing throught them.

You shouldn't have to worry about current balancing, due to which dies are lit vs which are dim.

[edit]
I've modified your first picture to show the die layout:

 

[BayMoe]

The first diagram showing the two diodes dimly lit appears exactly like what I see when the emitter on my U2 burned itself out. What's strange is that the new emitter that replaced it (WX1R) died the same fashion.

 

[evan9162]

While individual dies might go dark in a damaged Lux V, these are not comparable situations.

The pictures above are taken with fractions of a mA flowing through the device. If you ramp up the current even into the 10s of mA, then all 4 dies appear evenly lit.

Now, if you have dies that stay dark while running 10s or 100s of mA through the device, then the device would appear to be damaged. I've seen this mostly caused by excessive current.

 

[koala]

Sorry, my assumption was wrong. This is getting strange...

 

[evan9162]

I'll see if I can get some shots of my WW0T exhibiting the same behavior

 

[evan9162]

Here is my WW0T at 0.6mA:

At about 2mA, the bottom dies start to light up.

At about 3mA, strange things happen. The bottom dies flicker a bit brighter, while the top dies flicker a bit dimmer, all the while, the current varies from about 2.4mA to 3mA.

I think what is happening is that the current flow through the bottom dies starts to intermittently "jump" across the junction, failing to create blue light when it does. At the same time, when the current jump happens, the Vf of the lower dies drops, raising the Vf of the top dies, allowing more current to flow. Then (possibly due to miniscule amount of heat, or the increased current), the current flows across the bottom dies properly, creating blue light (making them brighter) and increasing the Vf (reducing the current). This must be happening at some kind of critical threshold where this behavior happens quite easily.

Finally, at about 6mA, all 4 dies are about evenly bright, and there's no more flickering.

 

[koala]

I saw the flicker too I thought it was my connection or power supply!!!
Hey if the voltage is right, we could make a low power strobe out of the Lux V!

 

[evan9162]

I've thought about this a bit, and I think I might have an idea of what's causing this behavior.

This seems to be limited (or at least, happen the most frequently) with Luxeon Vs. I think their serial/parallel arrangement may allow this phenomenon to happen.

I think what is happening is that a high resistance "short" is allowed to form between the anode of the downstream diodes (which is also the cathode of the upstream diodes) to the heat sink slug. Then, the "-" terminal ESD diode allows this current to flow to the negative terminal by forward-biasing the ESD diode.

Here's a diagram:

At low current, the combination of the forward voltage of the "-" terminal ESD diode, plus the voltage across R(leak) when very low current flows is actually lower than the Vf of the downstream LED diodes, causing the current to flow through that path, rather than through the LED part. But, when enough current flows, Vz2+V(leak) equal the Vf of the second set of diodes, so current starts to flow through them. R(leak) might be variable/intermittent, so higher currents (1-10mA and higher), R(leak) could increase substantially; and at some critical point become intermittent, causing the flickering we observe.

I tested my WW0T Lux V and measured the voltage between the slug and "-" terminal (measuring Vz2). At ~1mA, Vz2 was about 0.6V. However, once I hit about 2.5mA, Vz2 started to drop, and continued to drop until I hit about 20mA, at which point all 4 dies were evenly lit. So the current through the "-" ESD diode was dropping, probably because R(leak) was increasing, no longer bypassing significant current from the downstream LED diodes.

This may be a byproduct of the way the luxeon V is constructed. On the silicon submount of all other luxeons, there are 2 contact pads. These contact solder balls on the bottom of the led die. Each contact pad has a zener diode to the slug (the ESD diodes), and a bond wire to the power leads.

However, on the luxeon V, there are 3 contact pads on the silicon submount. The "outer" ones serve the same purpose as the 2 pads on regular luxeons, and have the same connections. However, the central pad serves as a bridge to join the two series connected dies. That pad is supposed to be completely floating and isolated from everything, but I suspect that high resistance connections to the slug can form, causing the behavior we see above.

This is probably another reason why the Luxeon V will probably eventually be dumped, as it seems there are quite a few challenges to the packaging that you don't see in a single die LED.

 

[SemiMan]

For any diode, there are two variables that are going to determine the current\voltage relationship. When Lumileds makes a lux-5, or when you buy binned LEDS, they are binned at a very specific current. I would expect that at the rated current for a Lux-5, the forward voltages for the two diode strings are quite close. However, as there are multiple variables to arrive at the voltage/current relationship, there is no guarentee the forward voltage and current will match at other currents, though it is likely to be closer at higher currents where one variable dominates.

Semiman

 

[SemiMan]

Before the purests respond, of course there are more than two variables, but you get the point....

Semiman

 

[koala]

evan, that took me quite a while to understand..

 

[VidPro]

i have got another style of series/parellel emitter led packages, and it does the same thing, one entire set of series emitter units lights up first, before the next set of gates begin to output visable light.

(i have no idea why, i am just saying it doesnt seem to be exclusive to that led)

I love them pictures :twothumbs it certannly summed up 1000words.

if its a slight resistance differential in the curcuit used to assemble the package, i really doubt it effects it at the operating power, after all this effect is seen at ~15ma, so increase the voltage and the resistance drop with another .5v would probably be 20ma at most, so 20 of 750ma aint nothing. OR, if its 15ma "different" what is that to 750ma total?