My Seoul P4 has turned yellow!!!

Hi,

I have a 4 LED custom made caving light running two sets of two seoul p4s in parralel driven by b-flex.

I just noticed last night that one of my seouls has turned quite strongly yellow.

Any idea why this happened (i think it was white) and is it likely to deteriorate further?

The heatsinking should be massivly overspeced assuming it has not come loose, so that shouldn't be a problem. as it is being running in series/parralel with other LEDs the electrical connections should be good.

Can anyone suggest why it broke, and what I should look for if I open up the unit (water sealed it recently).

Sods law as well, I originally bought 5 led's in case one failed. as the torch had been running fine for a few weeks I tested the spare LED to destruction.

Any advice, or this I should check?

I would like ask another question while I am here.
My friend says that solid state lighting devices should not be underdriven to prevent damage.
Does anyone know what classifies as underdriven of CR-e or Seoul p4?

Could this be why my LED failed?

I should also add that the faulty LED is generating more heat than the others.

It's quite the opposite, given everything else the same underdrive will increase LED life.

1. As for your failure, how are you isolating the SSC-P4?
2. Are the Vf match? PERFECTLY?

Those should answer your FA.

Hi Peter
This wasn't the light that got filled with water at one stage was it?

Cheers
Dom

There were similar problems with the Amlite T5 Seoul turning browinsh-yellow. Supposedly there was a bad batch of SSC P4's in which you may have one of them.

Your friend needs to know what he's talking about before giving his 2 cents.

The only explaination for your Seoul is it is defective.

Let's clarify something. Is the emitter physically yellowed (looking at it when it's off), or has the beam turned yellow?

If it's the emitter turning yellow, I've seen this happen on one of my first Seoul mods, on a AAA flashlight hardly ever run. It didn't affect the quality of the beam or the output. And it was age, not usage that seemed to be the most significant contributor. This emitter came from the Sandwich Shoppe's first batch.

PeterScowcroft said:

I should also add that the faulty LED is generating more heat than the others.

Click to expand...

Is that the LED itself getting hotter, or the heatsink it its vicinity?
Are you using emitters or stars, and were they the ones you got from Dotlight?

When the unit is open, it may be worth measuring the volt drops across the various LEDs, since that might give some kind of hint.

Have there been any past incidents of accidental over-running the LEDs, even briefly?

EngrPaul said:

Let's clarify something. Is the emitter physically yellowed (looking at it when it's off), or has the beam turned yellow?

If it's the emitter turning yellow, I've seen this happen on one of my first Seoul mods, on a AAA flashlight hardly ever run. It didn't affect the quality of the beam or the output. And it was age, not usage that seemed to be the most significant contributor. This emitter came from the Sandwich Shoppe's first batch.

Click to expand...

Yup, one of the early SSCP4s I got from the Sandwich Shoppe physically turned brownish yellow. It works fine though. I wondered at first when I noticed it, but since the output still looked the same I figure it probably isn't anything to worry about.

Hi,

Thanks for answering me.

It was got from dotlight.de, the Vf are close but not perfect, the dual series set up and the fact they are all driven quite low should make up for that (max 500mAatm)

It is the colour of the light that is yellow.

The main heatsink itself is hotter on the side with the yellowing beam chip than the normal (all are electrically isolated).

Any oother ideas?

Peter

Forgot to add tehy are on stars

The only incident I've read about is Newbie's SSC P4 testing where the die kinda peeled a little and the beam became bluish...

I'm getting a little worried about the quality of Seoul's now, especially when putting multiple SSC's in a expensive mod. Hope its not a widespread issue!

I am thinking that my original suspitions were correct and the XR-E is a better designed chip.

I am worried at how hot the heatsink is getting, normally it is cool to the touch but something is horribly wrong making it painfully hot!!!

I should point out that it's just dam typical is fails after being watersealed!

I would like to know why a led would not be damaged by being underdriven, do you actually have a strong scientific basis to your knowledge or is it just a basic estimation due to reduced themal constraints?

I should add that my friend is a PhD in electronics who has worked in custom lighting design for a number of years, so assumably quite clue up on the particular subject.

For the Luxeon chips I believe being underdriven was a problem, I am curious to know if this problem was maintained with the XR-E/P4

Feel free to be quite technical in your answer as I do have a pretty good understanding of molecular theory.

PeterScowcroft said:

I would like to know why a led would not be damaged by being underdriven, do you actually have a strong scientific basis to your knowledge or is it just a basic estimation due to reduced themal constraints?

I should add that my friend is a PhD in electronics who has worked in custom lighting design for a number of years, so assumably quite clue up on the particular subject.

For the Luxeon chips I believe being underdriven was a problem, I am curious to know if this problem was maintained with the XR-E/P4.

Click to expand...

I haven't heard of LEDs being damaged by being underdriven, and it's not obvious what mechanism for damage there would be.

Luxeon datasheets recommend PWM for low power operation, and seem to cut off their output vs. current graphs at 100mA (200mA for LuxV), but they don't say LEDs will be damaged by low currents, and there are no minimum currents given in the ratings. I understood that at low constant currents there were likely to be (temporary) colour shifts, which using PWM avoids.
Cree and Seoul datasheets show output/current graphs with the currents going down to zero or near-zero.

I had a seoul in my hds turn brownish-yellow. I only noticed it after I accidentally got some arctic silver on the dome. After wiping it off with some scotch tape, I noticed that half of the dome had turned color. I turned it on to see how it affected the beam, and realized that I really like the yellow color, so I spread more arctic silver over the whole dome turned it on high (thinking maybe something in the arctic silver burned off, causing the color change), and wiped the arctic silver off. The results aren't pretty, one side of the beam is more yellow than the other, but I love it! It makes the beam seem more like an incan, similar to a UV1J luxeon.

edit** took a beamshot... it's not really useful since it's not before/after, but i compared my svoh turned yellow brown hds to the mclux pd-s, (swoh?)




I know most people on this board will probably prefer the tint of the swoh, but I really love that urine-yellow tint!

my friend is a PhD in electronics who has worked in custom lighting design for a number of years, so assumably quite clue up on the particular subject.

Click to expand...

no clue at all if he did not point the following out:
he is correct if it is neon tubes or Arc lamps (where it is not good if they see too much or less power than needed),
but LED and hotwires greatly expand their lifetime with lower power,
even we hobbyists know this
(f.e. going from DC to AC means at least a 10 time increase in lifetime for bulbs, a bit less voltage means an greatly expanded median here also (but I dont have numbers available by now))

PS: there were some SSCs on star that were awfully bad mounted, touching the metal only on one side and the other with an air gap. Maybe thats why Yours quit --> overheating (but then the sink should not get hotter) :thinking:
(thats why I now only place my emitters directly on the sink)
IMHO that emitter gets/got too much current. Have You double checked for shortings?

You could also check if the current somehow changed, just from batts to driver should give a 1st good idea.

PPS: what Do You mean with "isolated"? Are the (isolated) stars mounted with an additional isolation? Maybe this one is too much and the heat just could not be moved away --> overheated

It will be interesting to see if the emitter on that star was properly reflow soldered to the star base, or not.

And was this the light that was filled full of water and bashed on the concrete -or was that another light.

PeterScowcroft said:

The brighter picture is at 500mA, When my new board comes from fab then it should be able to peak at 2-3A for short periods, enough to flood a large cave chamber!

Click to expand...

PeterScowcroft said:

max current output is 1A and the seoul is set up 2x2 so 500mA each on max.
My friend has just designed the worlds best driver board so that will soon increase to a meagre 4A each (short burst due to heat)

Click to expand...

it this the light with the color problem?
--> is the boost still actual?
If "yes" are You still wondering? This much on overdrive might damage the emitter, no matter what heatsinking is done

Yesterday I lighted up my quadCree bike light the 1st time. Emitters mounted to a round base 5cm in diameter, 1 cm thick. Just for the test mounted to a Pentium II heatsink. At some 900 mA the whole assembly got really hot within some mins. The sink in the pic is hardly to see, but imho way to small/thin (if this is just a standard aluminium housing.
You can never overengineer a heatsink for high current led application!

PS: sure it might work for the application and sure You do more and heavier caving than I do, but that looks so AWFUL! Its a good very 1st mod but ...
And the sink is not enough, except if the light is submerged.
Why dont You use a Maglight head as a host? Its not larger than this here. If cut from the back also not longer, so not more in way.
Perfectly looking with the three/quad Cree/SSC setup and still within financial limits argueable. Can even be made fully submersible with some silicone.