Is LED damaged by cap. discharge?

[schnell]

Hello,

If i connect a 1 Amp LED straight to a 23u5F ceramic capacitor that's charged up to 26V, does the LED get damaged?

here is the LED datasheet

http://www.cree.com/products/pdf/XLamp7090XR-E.pdf

I can't see any I^2t data.

I need to know this since a customer may accidentally connect up a single LED to the output terminals of a certain LED driver which has the above capacitance at its output.

 

[IRstuff]

Doesn't the curve on page 7 tell you all you need to know?

TTFN

FAQ731-376

 

[schnell]

that curve does tell me that the current will be very high.......but the amount of energy is 7.94mJ (1/2CV^2).

The 23u5F will soon discharge and i'm not sure just how damaged the LED will be, and just how much shorter it's lifespan will have been reduced to?

 

[MacGyverS2000]

For (reduced) lifespan data, you need to talk with the manufacturer... no generic chart is going to tell you that. You need to look at the internal resistance of the cap and chip to determine actual current.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[itsmoked]

There is zero reason to expect that LED to take more than 2A without flash-bulbing.

The current out of a capacitor is considered to be infinite.

The reason you don't get infinite current is because of the resistance of the wiring between the cap and the load.

If we offer up 100mOhm which is ball-park, we have 26V/0.010 = 260A. POP!

It is general knowledge that an intermittent wire on the drive side always blows the string of driven LEDs. The driver promptly runs up to maximum output voltage as it tries to maintain its set current. This charges the output capacitor to full voltage. Then reconnecting the string discharges that cap thru the LEDs with extreme prejudice.



Keith Cress
kcress -

http://www.flaminsystems.com

 

[IRstuff]

Your OP asked if it would get damaged; how many systems do you know of that can tolerate 8 times the nominal operating voltage without massive protection?

If you know that it can happen, you need design a solution. Diode will not tolerate such abuse. Even if were to survive that indignity, it's life will probably be measured in the handfuls of hours, if at all.

TTFN

FAQ731-376

 

[schnell]

I now believe that the LED would die.

Therefore, i believe i now have absolutely no choice other than to use an "hysteretic" LED Driver such as AP8802.(?)

These do not require any capacitance on the output for the sake of feedback loop compensation.

Therefore, i need only use as much capacitance as i need to for EMC purposes.

http://www.diodes.com/datasheets/AP8802.pdf

The hysteretic IC has no oscillator and actually relys on the ripple in the current sense resistor to make it work.

 

[itsmoked]

Give it a shot. This may still not protect the LED string as an open will have the controller forcing full voltage output as the LEDs are connected. But, it might be a low enough current since it isn't capacitor backed.

If this is often having the string connected while energized you can always consider a shunt or a parallel resistor to the LEDs that always provides some load. It would drop your efficiency some and I wouldn't do it for a normal setup where the strings are not hooked up live.

Keith Cress
kcress -

http://www.flaminsystems.com