led directionality, how do I install them?

[jolly1423]

I tried to look this up but I can seem to work it well enough to get good search results. I'm going to work with led's for the first time and I've read a couple of things about "installing them the wrong way and ruining them". Am I interpreting that correctly? Do led's have a positive and negative? How do I ID which is which and which do I run the current into? I've noticed one lead is longer than the other, it seems there must be a reason for that that I don't know.

[R.G.]

LEDs are light emitting DIODES, and as such yes, they have a polarity.

All diodes have an anode and a cathode. The diode conducts current through it when the anode is more positive than the cathode by at least a certain minimum voltage. That minimum voltage is different for every type of diode material. Silicon diodes, for instance, always conduct when the voltage is the correct direction and bigger than 0.5V to 0.7V. You can in fact make a very good guess about what material the diode is made from by reading the forward conduction voltage.

In the reversed direction, diodes are intended not to conduct at all. How much reverse voltage they will stand up to depends on how they're made. Some silicon diodes break over in the reverse direction with as little as a few volts, some take over 1000V to reverse-break.

LEDs are specialized to emit light when they have current forced through them the correct way. They have a forward voltage of between 1.5 and 4V, most of them 1.8 to 3.0V.

For all modern LEDs without clipped leads, the longer lead is the positive one. Most LED packages, but not all, have a flat spot on the edge of the plastic package which denotes the cathode.

But the dead-simple way to figure out which way to install them is to test them. Take a 9V battery and a 4.7k resistor. Temporarily tack-solder the 4.7K to one lead, doesn't matter which one. Now touch the LED lead without a resistor and the free end of the resistor to the 9V. One way round, the LED will light up. One way it won't. The resistor keeps the reverse current from killing the LED when it's wrong.

[JKowalski]

EDIT: Nevermind RG added enough info  

[jolly1423]

Perfect! Thanks!

[PRR]

> "installing them the wrong way and ruining them"

Bosh.

As long as you HAVE a resistor or other current-limiting scheme, you won't "ruin" them by connecting backward.(*)

R.G.'s "touch and try" is utterly safe. I always just hook em up, if they don't light, I flip the leads. Never killed one. I often leave an LED hooked up backward for an hour when building a PC, don't get around to checking them right away, and never lost one yet.

I have fed LEDs raw AC through a resistor. They light half the time, flicker.

(*)Well, it is possible, but not easy.

If your design current is resistor-limited, very-very close to the LED maximum current, and your available voltage is high (dozens of volts), you can overheat an LED, on paper. But I've over-abused a few LEDs and my suspicion is that this would be very difficult. That the maximum specified forward current is nowhere near heat-limit (on little LEDs, I'm not talking the new super bright room-lights), so even at the higher 4V-7V reverse breakdown, they just sit there hot, not melting.

I think the one time I pushed the limit was feeding 110V AC to an LED. I figured I had to keep the current below 7mA (on a 20mA LED) so the reverse half-cycle didn't heat too much. That gizmo was still working (24/7) a decade later.

[blooze_man]

one other way to identify. if you look at the LED, you'll see what looks like a flag. the flag points to the positive lead.

[tiges_ tendres]

I always remember that having long legs is a positive attribute.  The long leg on a LED is the leg you attach the resistor to which you will be applying power (+).

Failing that,  I have flat battery (9 Volt) I use to test the LED.  I think if the battery is under 4 volts or so, it wont cook the LED with direct battery contact.

[JKowalski]

The diode function on my FLUKE multimeter is one way I like to test LEDs

It measures the forward voltage drop, so it outputs just enough voltage to barely turn on the LED, enough to make sure its working and double check the polarity (I have a batch of old mil-spec LEDs that are completely opposite - the cathode is the longer lead, and the anode has the flat side  )

Messes me up every time

[tiges_ tendres]

The diode function on my FLUKE multimeter is one way I like to test LEDs

It measures the forward voltage drop, so it outputs just enough voltage to barely turn on the LED, enough to make sure its working and double check the polarity (I have a batch of old mil-spec LEDs that are completely opposite - the cathode is the longer lead, and the anode has the flat side  )

Messes me up every time

Funny you mention that because I just bought some LED's that came with the same size legs.  There goes my theory!  I'm hoping these dont trip me up down the line! 

[JKowalski]

You should be okay if you take the long lead is anode approach most of the time, I think it's a standard for today and my LEDs were just that way because of their age. It's also very easy to immediately see.

Were your LEDs old stock as well? I wouldn't think any new manufacturers would risk going off standard for their products.



Blooze_man's method is a good one, the post markings are consistent with all LEDs I believe, because the semiconductor is always in the dome post and the anode wire comes in from the small post. Sometimes you get LED's that are hard to see through, and then... well.


Tiges, the dead battery approach is ingenious!   Can't source enough current to fry the LED, but enough voltage to turn it on somewhat no matter what the voltage drop of the diode is! Just got to make sure it cant go above a few mA of current. I suppose you could just check that before LED use with a multimeter and low ohm resistor. And everyones got dead batteries lying around

[tiges_ tendres]

You should be okay if you take the long lead is anode approach most of the time, I think it's a standard for today and my LEDs were just that way because of their age. It's also very easy to immediately see.

Were your LEDs old stock as well? I wouldn't think any new manufacturers would risk going off standard for their products.



Blooze_man's method is a good one, the post markings are consistent with all LEDs I believe, because the semiconductor is always in the dome post and the anode wire comes in from the small post. Sometimes you get LED's that are hard to see through, and then... well.


Tiges, the dead battery approach is ingenious!   Can't source enough current to fry the LED, but enough voltage to turn it on somewhat no matter what the voltage drop of the diode is! Just got to make sure it cant go above a few mA of current. I suppose you could just check that before LED use with a multimeter and low ohm resistor. And everyones got dead batteries lying around

I'm afraid I cant take the credit for it, I read it here.  Must have been some other genius!

Not sure if you were also referring to me with your comment about old LED's above, but as far as I know they are new stock, but they are on an ammo reel.  I would suppose that this would make machine operated assembly much easier if both legs were the same size.

[PRR]

> LED's that came with the same size legs.

One side of the "round" plastic should be flattened. That tell which pin is which, but I never remember.

Yes, one leg tends to have a small "flag". Also if you can see inside, one bigger leg hits the bottom of the chip, the other runs via a hair-wire to the top of the chip. Again, I never remember which is what.

> on an ammo reel

If you have a full reel, the packaging docs tell you if the parts are taped cathode-first or anode-first. There may be a standard. You program your automatic parts insertion robot to take the LEDs off the tape and rotate as needed.

Of course if you have a snippet, you can't know which way it came off the reel.

Back in the 1970s, I knew which-way. But since LEDs came down below a dollar apiece, I've mostly used the try-and-see method. It usually lights first or second try, which is easier than squinting and thinking. Most LEDs are "non-essential", in that the circuit will power-up fine without the LED. And usually no lethal voltage around. Then I just spread the legs and poke it at the connections, one way and the other way.

[head_spaz]

I robbed the battery out of an old motherboard, a coin cell, that I use to test LEDs.
Since it's 3 volts, no resistor is required, and it's thin enough that you don't have to bend the leads of the LED to test it.
Fast and simple, just the way I like it. Get some!

BTW... the current limiting resistor can go on EITHER leg, since it is wired in series. (providing your resistor is the non-polarized type.

[JKowalski]

(providing your resistor is the non-polarized type.

Nooo!

Don't confuse people! Who wants to bet that some topic will pop up with someone asking if they installed their resistors in the wrong direction?