Cree XT-E Xlamp problem :(

stomp

Newly Enlightened
Joined
Dec 11, 2014
Messages
3
Hi all :)
Need little help please!

Bought some Cree® XLamp® XT-E LEDs... From the datasheet the maximum forward current is 1.5A. I'm trying to use them @600mA (3.2V).

The LEDs are soldered on termal pad and I'm using LM2596 (http://www.ebay.com/itm/LM2596-DC-D...er-/400440730939?ssPageName=ADME:L:OU:US:3160) to drive them.

So the voltage is set to 3.2V (max is 3.4V) and the output short-circuit current to 600mA.

When LED is connected it shines for ~10s then starts to blink and finally degrades to glimmer... :sick2:
I connected a new LED to a programmable power supply with the same settings 3.2V/600mA - LED died much faster than with the LM2596 - 2s :mad:

So what I'm doing wrong?!?!
Sorry but I'm complete n00b in this :)
 

jason 77

Enlightened
Joined
Oct 2, 2008
Messages
518
Location
cali
Can you take some pictures of your set up? You say you have the LED soldered to a "thermal pad" what type of thermal pad are you using? High power LEDs like this need to be attached to a good heat sink. It is possible you are burning the LEDs out because they are overheating. How are you testing the voltages and currents to make sure they are at 3.2volts and 600mA?
 

stomp

Newly Enlightened
Joined
Dec 11, 2014
Messages
3

alpg88

Flashlight Enthusiast
Joined
Apr 19, 2005
Messages
5,343
you do not need to worry about voltage. only current, that board will adjust voltage automatically.

thermal pads alone are not enough for anything more than 100ma. if you do not have your star (termal pad) attached to aluminium heatsink, you simply killed leds with overheat.
i'm building xt-e light now, i run quads at 700ma (4 leds driven at 700ma on one 24mm star), the quad star screwed firmly to aluminum heatsink, no issues whatsoever.
 

AnAppleSnail

Flashlight Enthusiast
Joined
Aug 21, 2009
Messages
4,200
Location
South Hill, VA
It sounds like a heat problem. If you put the LED on those little stars (20mm stars, we call them) and applied a few watts of power, you heated the few grams of LED+star to a higher temperature than the LED is rated for. The star should be thermally attached to a substantial heatsink - About a palm-sized chunk of aluminum per 3 watts, or less if you use fans to actively cool them.
 

stomp

Newly Enlightened
Joined
Dec 11, 2014
Messages
3
Hmmm... I'll try with bigger heatsink then :)
Yesterday I ran @ 350mA with just 20mm star and let it shine for about 1h. The star wasn't hot but maybe it's different @600mA (I haven't touched it).
 

Anders Hoveland

Enlightened
Joined
Sep 1, 2012
Messages
858
If the LED is burning out after just 10 seconds, it is not a heat issue. The aluminum star has enough heat capacity that it would take more than 10 seconds for it to overheat.

The XLamp should be able to handle 700mA, if it cannot, I suspect there is something wrong with your power supply.

One big thing you may be overlooking is that your "constant current" power supply may only be able to drop the voltage down to a limited level. This may or may not actually be indicated on the labeling. In other words, it is only constant current within a certain voltage range. The solution to your problem is that you are going to have to run multiple LEDs at the same time in a series circuit.

With one of the cheap Chinese-made constant current power supplies, I have to run at least four LEDs in series or they will burn out. And even then, they still burn out if I plug the power supply into an outlet while the LEDs are still attached, there is obviously a power surge that they cannot handle. All this despite the fact that the LEDs are rated to handle 700mA and it is only a "350mA" rated power supply! But there are no problems when using six or more LEDs.

If the power supply states 600mA 12-48v, that means you will have to use at least 4 LEDs in series (if each LED is rated 3 volts), otherwise the current will be greater than 600mA, and chances are one or more of the LEDs will burn out. Some LED power supplies will explicitly state they are for a string of 20-30 LEDs. That means if you only use 18 LEDs there will be a burn out. Because the power supply cannot lower its voltage below a certain threshold, and therefore the current (mA) will be greater than intended.

This, I suspect, is most likely your problem.


Again, just to clarify, what matters is the amount of current going through an LED. Generally an LED can take any level of voltage so long as the current is limited. However, voltage is nevertheless still an important consideration when using power supplies. This is because a constant current power supply controls the current going through the LEDs by changing the voltage. The LEDs act as resistors, so the higher the voltage, the more current is able to flow through. If the power supply is not able to drop its voltage below a certain threshold, the current passing through the LED will be greater than what the constant current power supply is rated for. In other words, with most "constant current" power supplies the current is not actually limited at the source, rather it is a feedback loop active modulation through the circuit.

The reason it is the current passing through an LED that matters is because wattage is determined by current and resistance, so too much current means too much power/heat. Theoretically, there could be a thousand volts going through the LED, but the LED itself would only pick up a small fraction of the overall power travelling through it, so could do just fine. This is the phenomena of voltage drop / power drop. If the load in the circuit only absorbs a fraction of the voltage, it will only pick up a fraction of the power.
 
Last edited:

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
....
Again, just to clarify, what matters is the amount of current going through an LED. Generally an LED can take any level of voltage so long as the current is limited. However, voltage is nevertheless still an important consideration when using power supplies. This is because a constant current power supply controls the current going through the LEDs by changing the voltage. The LEDs act as resistors, so the higher the voltage, the more current is able to flow through. If the power supply is not able to drop its voltage below a certain threshold, the current passing through the LED will be greater than what the constant current power supply is rated for. In other words, with most "constant current" power supplies the current is not actually limited at the source, rather it is a feedback loop active modulation through the circuit.

The reason it is the current passing through an LED that matters is because wattage is determined by current and resistance, so too much current means too much power/heat. Theoretically, there could be a thousand volts going through the LED, but the LED itself would only pick up a small fraction of the overall power travelling through it, so could do just fine. This is the phenomena of voltage drop / power drop. If the load in the circuit only absorbs a fraction of the voltage, it will only pick up a fraction of the power.

I'm sorry, but this is just nowhere close to being correct. I don't mind people not knowing what they are talking about as long as they are aware of it. There are so many errors in this "clarification" that I must object.

Anders, I don't mind that you don't understand electronics, but please don't try to tell people that you do.
 

SemiMan

Banned
Joined
Jan 13, 2005
Messages
3,899
I am glad you posted that excerpt. I had not even made it through the whole post. It was good for a laugh :)
 

AnAppleSnail

Flashlight Enthusiast
Joined
Aug 21, 2009
Messages
4,200
Location
South Hill, VA
Again, just to clarify, what matters is the amount of current going through an LED. Generally an LED can take any level of voltage so long as the current is limited...The LEDs act as resistors, so the higher the voltage, the more current is able to flow through.

The first of these is false. The second is false, too, and more fundamental. To state that an LED "Acts as a resistor" means that it has some intrinsic behavior that Voltage = Current x Resistance. While you could take the Voltage drop of an LED at a particular current and multiply Volts x Amps, you would NOT have a resistance value for the LED. This is because the "resistance" value will change with respect to both voltage, current, temperature, and LED age. Therefore, LEDs do not act as a resistor because they do not follow Ohm's Law.

This leads us to the silliness of the first statement. If you apply any amount of potential (voltage) to an LED, there is a near-exponential relationship between Voltage and Current. Similarly, if you apply a given current to an LED, the voltage that will pass is dependent on this relationship, too. As with nearly all semi-conductors, this "Knee Bend" is very different from the linear behavior of a resistor. Here is some basic information on how diodes' voltage-current relationship behaves: EvilMadScientist on Zener Diodes. Check the Ideal Behavior chart, 4th or 5th image down.




Theoretically, there could be a thousand volts going through the LED, but the LED itself would only pick up a small fraction of the overall power travelling through it, so could do just fine.

This is false based on the above fundamentals of LED circuits. For example, the XT-E will pass about 600 mA at 3.2v. It cannot possibly pass 600mA at 50v - It will instead pass several dozen amps and fail immediately. It may fail to a dead short, in which case it will still emit light...:poof:


One thing is nearly-true, based on a naive model...
If the LED is burning out after just 10 seconds, it is not a heat issue. The aluminum star has enough heat capacity that it would take more than 10 seconds for it to overheat.
Wolfram Alpha Thermals In theory, each Joule raises the temperature of the LED+20mm star by about 0.67C. In theory, the circuit as specced would run at 2W, or about 1.3C per second temperature rise. I wouldn't expect even 20C to kill an LED on a star, suggesting very poor thermal contact between the LED and the star, or unusually high power.

So! There are measurements to take... Preferably very quickly, to find the root cause. Good luck!

Morning after edit: I should not CPF in the wee hours.
 
Last edited:

Latest posts

Top