LED physics

[SSLtech]

A question about LED death...

Are they short-term power-sensitive (RMS) devices, or are they absolute peak-destroyed devices?

In other words, can I use them on 50Hz RMS AC, where the RMS is well within their rated value, but where the instantaneous peak current at the height of the cycle would cause them to die at DC?

...or do you derate?

Keith

 

[peranders]

LED's are like any other diode. They have a non-repetitive max peak current, a repetitive max current and a contineous current.

Are you planning to drive directly from mains? This can be done quite easily if you add a couple of parts.

 

[tmbg]

it's pretty common to try to squeeze brightness out of bright LEDs by running them at, say, 10% duty cycle, well above their continuous current rating. Some LEDs even have specifications as to max current and duty cycle ratings.

 

[SSLtech]

Actually it's a little more subtle... I want to drive directly from AC without a series resistor... (I'm not mad... -really!)

From the 6.3V secondary of a tube device, if I put 8 LEDs in a "series chain", with parallel pairs facing in opposite directions, I can avoid the use of a series resistor. However, as you probably know, the current draw accellerates rapidly approaching the forward-bias voltage at normal operating current. -Since the heater voltage will rise a little if all the tubes are removed, and the AC power can be a little higher at certain times of the day... then you factor in the AC waveform peak... I'm trying to figure out if they last a day will they basically last forever, or is there still a stress that is "doing its damage unseen"?

Keith

 

[clintrubber]

Actually it's a little more subtle... I want to drive directly from AC without a series resistor...

Can be done with a 1N400X anti-// to the LED and a 400V series-cap with a value giving the desired current.

For a Dutch LED the cap-value could be determined from:

(230 - 1.7)/10mA = 1/(2*pi*50*C)

so something like 150 nF.


Bye,

Peter

 

[peranders]

It's not wise to drive a LED with only a cap. Transients will(or can) kill it.

I have made a design for your purpose. Check my homepage in a few hours.

http://home.swipnet.se/~w-50719/hifi/mld01

If you are after driving a LED at 6.3 VAC directly, I wouldn't recommend it either.

 

[Mark Burnley]

Yes,

My homebrew solder-station has LED's run straight off the mains. When I was experimenting with the circuit, I did the equation to find the size of cap, and had the reverse 1N400x, but the green LED's kept popping (red okay- but the cap for the green needed to be larger for same brightness). A 330R series resistor limited the inrush current during the moment the cap charges and saved the LED. Been working great for ages now.

I find that some types of LED are more "forgiving" than others w.r.t abuse. Maybe you should hook up the chain to 6.3V and leave it on for a while, or variac it and monitor the current and forward voltage of LED pairs?

We all know you're trying to light your house up with the most LED's for a novelty Christmas bet with Scenaria :wink:

Mark

 

[synthetic]

I bought an LED pilot light that functions using straight AC line voltage. I'm still trying to figure out the circuit to make it a "power on" LED. This is the type of thing that's too basic to include in most schematics I've looked at. :roll:

 

[Gus]

Realy no easy way to do what you want Keith. Only thing that might work is a transformer with a very high resistance 6.3V winding and very low current output that would limit current .

You could use a CC device in the middle of a bridge. Make A bridge rec out of 4 leds make a string of how many LEDs you want and connect the string and a CC device to The +- of the bridge. The V drop will be two of the bridge and the sting connected to the + and -.

When the diodes needed go up this setup cuts down the diodes needed and because only 2 are not used on each AC swing the same ones are used and there for track better.

The normal resistor and LED setup is to give the LED something like a constant current the LED parameters moves with temp and current.

A fast low voltage drop CC device might be what you are looking for

 

[clintrubber]

Don't know what the 'environment' of that LED will be (w.r.t. looks).

If it's the old favorite way of styling with big meters & bakelite (hmm, why has this become the default assumption here... :wink: ), then why not a neon bulb + series resistor ?

 

[SSLtech]

No prob everyone... I got it all figgered out!

Lookin' pretty good, no apparent excessive stress on anything.

:thumb:

Keef

 

[peranders]

Have you checked my LED driver?
http://home.swipnet.se/~w-50719/hifi/mld01

Forget about the text at the moment (copied from an another project)

My design is made for long life and a similar design is used in ordinary time relays which are powered directly from the mains. You can also check my softstarter which is powered in the same way.

http://home.swipnet.se/~w-50719/hifi/sst01

 

[clintrubber]

No prob everyone... I got it all figgered out!

What did you end up with ?

 

[peranders]

A super bright LED, 1 mA + diode + resistor will also work fine.... without much heat. => 300 mW!