Found some LEDs that SEEM to work in reverse

[GüdPedals]

So I've been building pedals for a few years now and encountered something I've never seen before.

It seems to be a batch of red LEDs that only work in reverse polarity. At first I thought maybe just the bulb itself was backwards, as it only worked with the flat side going to voltage rather than GND, but I looked closer and the internal structure seems to be arranged the same as every other LED I tested. But still, it only works if the side that SHOULD be the cathode is connected to voltage.

I thought I was losing my mind at first. Just wondering if anyone else has experienced this.


EDIT: Here's a pic of the red LED in question vs one of a normal white LED. Hopefully they're clear enough to see what I'm talking about:

[EBK]

We're going to need photographic evidence of this phenomenon, I think. 

[mozz]

Yes, i have seen this also. Whether the flat side was wrong, or the longer lead was wrong. I guess they are factory rejects but get sold anyway, maybe nobody caught the mistake and just shipped them out. Now i look inside if possible for anode cathode.

[GüdPedals]

I've added pics to my OP. It looks to me like the anode and cathode are indeed reversed on the inside. Unless I'm completely missing something.

[blackieNYC]

I have a few backwards LEDs, in terms of the long leg.  They're amber looking but glow red. Student tape machine parts.

[Rob Strand]

What is the voltage *across the LED* in the reverse state?

I posted a puzzle from Electronics Design some time back about getting negative collector voltages when a transistor's BE junction was reverse-biased.   According to Bob Pease when you breakdown a junction it gives off infrared, so you might end-up seeing a red glow if you break down the LED junction.

In another thread there was discussion on the reverse breakdown of an LED.   Back in the 70's or 80's I remember measuring low voltages like 5V or so but Jack Orman said he saw 20V or so.   When I checked some more modern LEDs I saw high voltages like Jack.  After some om'ing and stuff I managed to summon an ancient LED from the tombs and it showed a low breakdown voltage.

[GüdPedals]

What is the voltage *across the LED* in the reverse state?

1.86V

I would like to clarify that the LED does not work at all in the "correct" state. And it wasn't limited to a single unit unit. I bought 15 together and they're all like this.

[MikeA]

I regularly find red waterclear LEDs like this.  In the ones I've bought, the long leg is indeed the anode, but "fat/flat/cat" is false (flat side of the case and the larger/fatter internal terminal is anode, not cathode.)  Only red so far.  Good argument for testing every component before it's installed...  MikeA

[Rob Strand]

1.86V

I would like to clarify that the LED does not work at all in the "correct" state. And it wasn't limited to a single unit unit. I bought 15 together and they're all like this.

So it acts normally but it's like the little "chip" has been put in upside down - making connections reversed whatever way you look at it.
   

[ElectricDruid]

I guess if there's two ways to do something, sooner or later a batch is going to get done the wrong way around. That's just inevitable. And probably on a Friday afternoon...

[duck_arse]

from recent observation, it is my opinion that unless you buy your leds from the manufacturer, OR your supplier provides all the batch and QC documents relative to your parts, you can't say anvil is cathode, post is anode, short leg is [I never learnt this, or long leg .... ] or body flat indicates cathode.

if you buy from a surplus/disposals joint, they will very often disclaim 'high quality, known manuf, selling dirt cheap due to backwards lead ident' or similar, which is good for those particular parts.

if you buy from someone^* for whom leds are THE most interesting devices on earth, they can't help but test all their stuff to sort it for interestingnesses anyway, and they will note what they find in the stock they buy as surplus, run-outs, reject stock, etc, for orientation or useableness.

so, know your source, or get the info from your supplier, but then test everything before you use it anyway. in my opinion.

* as an example - https://www.ledsales.com.au

[amz-fx]

If you want to see the interior workings of an LED, then check this out:

http://www.muzique.com/news/closeup-of-leds/

regards, Jack

[Phend]

I've screwed this in (with help) both ways and it still lights up.

[anotherjim]

There are LED's that contain 2 colours in opposite polarity. Could they be selling rejects that had only one colour working?

[Rob Strand]

There are LED's that contain 2 colours in opposite polarity. Could they be selling rejects that had only one colour working?

You can also get flashing ones which do weird stuff if you power them like normal LEDs.   The OP's pics look normal.

[Radical CJ]

There are LED's that contain 2 colours in opposite polarity. Could they be selling rejects that had only one colour working?

You can also get flashing ones which do weird stuff if you power them like normal LEDs.   The OP's pics look normal.

Has anyone tried using those flashing ones as clipping diodes?

[EBK]

There are LED's that contain 2 colours in opposite polarity. Could they be selling rejects that had only one colour working?

You can also get flashing ones which do weird stuff if you power them like normal LEDs.   The OP's pics look normal.

Has anyone tried using those flashing ones as clipping diodes?

Does anyone have a datasheet that specifically shows how those types of LEDs work?

[Phend]

Kind of off and long winded:  But could these be used, maybe switched?

[anotherjim]

Kind of off and long winded:  But could these be used, maybe switched?

If you mean to have different flavours of LED for a clipper then yes. Would need to find the datasheet for one to check the Vf for the 3 colours. If the Vf's are too similar then the clipping flavours may not be different enough - although you can be sure that sombody has tried this already.

[Rob Strand]

Does anyone have a datasheet that specifically shows how those types of LEDs work?

I suspect they are all different.   It's just going to be some oscillator driving an LED.   

The ones I remember need a fixed voltage so you are just providing power a circuit.   You can see the chip for circuit inside the LED.  When you power them with a series resistor you can imagine the voltage across the device is high when the LED is off and low when the LED is on - so the circuit goes nuts.   There may be some which can work with some degree of series resistance.   It all boils down to the fine details of the design of the internal circuit.