# XLAMP MK-R by Cree power source

I was wanting to use the MK-R for a project of mine but after consulting the datasheet I burned up my first 10\$ LED and wasn't sure why.

I had an LDO providing 11.5V anode to cathode via gator clips. I ran the anode line through my multi-meter and verified it was pushing 450mA before the LED smoked and burned out (stayed lit for about 15 sec).I was trying to disconnect the clips real quick as I saw the smoke it was indeed burning extremely hot. I'm wondering what I am missing here from the data sheet as these LED's should take 1250mA with a Vf of up to 14V and dissipate 15W of power. Am I not understanding something in datasheet correct?

• Did you have any heatsink attached? Jun 17, 2013 at 20:59

As I see, there are two problems:

1. Your LED is getting 11.5V @ 0.45A. This means it's consuming ~5W on energy. Part of this energy is converted to light, part of it to heat. What part of the consumed energy goes to light depends on the LED, it's usually 15% ... 25%. The rest of the consumed energy is converted to thermal energy. This thermal energy has to go somewhere. If the LED is mounted on a PCB, the PCB draws the heat away. If it's not the LED melts itself into selfdestruction.

High power LEDs have special pads on the case itself designed to transfer heat away. The datasheet of you LED states: thermal resistance: 1.7°C/W. This is the thermal resistance between the PN junction inside the LED and the thermal pad.This means that per every Watt of energy going through the pad , the temperature difference between the junction and the pad is going to be 1.7°C But the thermal resistance between a pad not connected to anything and the air surrounding it is very high, so the LED and pad quickly heat up, not being able to transfer the heat away. For this pad to be able to draw heat away it needs to be thermally connected to something that has a low thermal resistance between it and the surrounding air. The simplest way to thermally connect a pad is to solder it (since metal conducts heat very well) and the "something" that draws the heat away is usually a PCB. FR4 is not a good heat conductor, so usually the thermal pad is soldered to a polygon, and if that's not enough special aluminium-core PCBs are used.

Here are some articles that go into more detail on this subject: a b c. This wikipedia page also has some good links.

2. A LED needs to be powered by a current source, not a voltage source. Each LED is unique, its precise forward voltage changes from device to device, varies with temperature (it usually decreases as the temperature rices, this is why connecting diodes in parallel to increase forward current is useless), and for one LED it may be 11.5V, for another one (of the same type) it may be 11.3V. Again, a LED should be driven with a current source, NOT a voltage source.

• Thank, I know I should have used a current controlled power supply but I was just trying to get an idea how bright this thing was and figured with the large Vf range I could test safely in the middle... I didn't understand the thermal resistance part of the datasheet but thanks for the info that really will help me when identifying future components and this will get me going on keeping the LED's from frying! Jun 17, 2013 at 21:30

Did you take into account the thermal resistance 1.7 °C/W in the datasheet?

The XLamp mK-r LeD operates at approximately 14 w of power when running at the chosen 1 A input current, and requires a heat sink to dissipate this thermal load.

• I've never worked with LED's that put out so much heat. Though I should have assumed with the high power dissipation as with all electronics, it would need a sink. I just didn't understand it would need one from looking at the original datasheet thanks for pointing out the design guide. Jun 17, 2013 at 21:25