Cree XTE Natrual white - Damaged phospher / accidental dedome now emits Royal Blue?

I'm building a fish tank LED lighting array from 10x 5watt Creee XTE LED's Natural white bin purchased from AliExpress for about $1each.
Power supply is 65watt laptop PSU connected to a Voltage and current adj. buck driver with PWM input (for planning micro-controller light level control).
I have two arrays of 5x LED's each mounted with thermal paste and screws to old Pentium III heat sinks (currently with no fans, I was hoping to have passive cooling)

They seem to perform exactly to the CREE spec sheet, so they are likely not counterfeit?

Anyway, I was putting the rig through a bench test too see how warm it ran at close to ~1.5amp (Max rating)
The heat sinks get hot to the touch, but not hot enough to be uncomfortable, which unscientifically is in the 70-80degC range I wanted to check the temp of the LEDs and stuck my finger on one (as I had done before) and the Silicone dome / phosphor disintegrated ! :sigh:
I will be ordering some spare LEDs, but I actually needed more blue in the array as it was a bit too green for my taste.

1) I was concerned that it is now emitting UV light, as it has that hue to it, but after some reading it seems that it is emitting Royal blue?

2) Is their something I can apply to the die to protect it?

3) are 90deg lenses worth the cost in this application? (was planning acrylic cover)

I might write up something on the PWM Drivers once I have it working with my micro-controller (Ti stellaris launchpad) as there is not much information on them around.

if the temp is only uncomfortable, it's closer to 40-50C. i can't keep my finger on 70C.

those LEDs are fried IMO

Royal blue, within a few nanometers. Intense royal blue looks very unnatural, but the coral reef guys like it.

You should probably leave the die alone. The optical silicones that Cree uses are about 98% transparent and very expensive. Anything less clear (90%, etc) will heat up and go opaque. Anything with "VOCs" (Volatile Organic Compounds) will heat up and degas (emit fumes) that go opaque, too.

The 90 degree lenses shape the beam. How far down are the plants/tank you're lighting? That's more important, since the acrylic will cover things anyway.


I take exception to your 'touch the LED dome' measurement method. At this output level, about 1 watt of photons are going to hit your skin in an area about the size of this 'o'. That will heat your finger a lot more than conduction from the warm LED. You could point a non-contact infrared thermometer at a lit LED and get hilariously wrong results (140C in my case). That's because the photons warm up the temperature sensor in a non-contact thermometer. Same for your fingers.

+1, and +1.

My 10W XM-L light will notably warm your hand from several inches away. If you want to do the finger test, touch the star or the heatsink near the LED. And "not hot enough to be uncomfortable" is definitely below 60C. 60C is about my threshold of 'can touch it, can't hold it long".

Keep in mind that the die will be warmer than the star, and depending on your thermal bond, may be considerably warmer than the heatsink.

Also, the fact that it seems to act like a CREE doesn't mean it is genuine. Fakes will generally have lower reliability, shorter lumen maintenance, and probably lower quality lens material. None of those would necessarily show up in a small sample size and short run time, unless you put your finger on the lens and destroy it :devil:. But even then, you'd need to know what it takes to destroy a genuine lens, for comparison.

Thanks for the replies, sounds like I am best to leave it alone and see how it goes, at least until I source a replacement.

I take exception to your 'touch the LED dome' measurement method.

Click to expand...

Fair call, To clarify I was trying more to judge if there was a temperature difference between the stars and the heat sink, rather than gauging die temp. You can't really look at these LEDs to see what you are doing though. Totally forgot to consider radiant heat from the LEDs though, that is a really good point and prob why I was getting vastly different impressions between LEDs on and off.

I really need to find a temp sensor or thermometer, I have some, in a box ... somewhere ... the joys of moving house.

The 90 degree lenses shape the beam. How far down are the plants/tank you're lighting?

Click to expand...

Bottom of the substrate would be about 42-45cm below the lights with the current setup, though I might adjust the height to get optimal coverage. At this stage I'm trying to mount the Light fitting as close to the top of the tank as possible to minimize light spill. I think I found some Lenses at a reasonable price (~$0.60 each), so I may just order a few and see what they are like, unless someone can advise me otherwise? (before the only ones I could find where twice the price of the LEDs :thinking

Also, good to hear I'm likely overestimating the heat sink temps, as the temp differential in my workshop is much greater (~8degC vs ~18degC) than the living room where the rig will live.
Fans and or PWM dimming to reduce temps still on the card I suspect.