We all know that on a PCB silkscreen, LED polarity can be indicated by a number of means. An "A" or a "K" or both, the package outline with flat, a full diode symbol or even just a "+" for the anode. It is widely touted on the internet that the internal structure of the LED is also a definite indicator, with the larger anvil shaped terminal being the cathode and sitting over the smaller one. That has always matched my experience over many years of working with them.

Well, apparently that last is an oddity because I have just had to repair boards carrying 3mm standard LEDs sourced from a "UK Unive-'RS'-al Supplier" where I found their internal terminal structure was the opposite way around. There are a mixture of colours on the board and the reds are reversed with all other green and yellow being correct to internet "opinion". Attached is a picture of my own LED with the cathode flat to the right and, though you can't see it, the shorter cathode leg matches that too. However, the cathode terminal is the smaller one with the anode on the overlying anvil shaped one as you can see.

I can certainly accept that the internal terminals are not the indicator to rely on but I have also seen LEDs without a flat and this would be the only indication of whether they are fitted correctly once the legs are trimmed down after soldering. Has anyone else seen this issue?

 

Guess I've answered my own question with a bit more Googling. Discussion re LED terminal design I have found reference to this back in 2013 in "another place". There seems to be no practical reason to have the cathode on the anvil, I had wondered about dissipation issues but apparently not. Terminal shape is definitely NOT the factor to use to check on polarity, it can be either even though most will be cathode -> anvil, anode -> post.

 

This is not so clear.

(I am not an expert on foundry practices concerning LEDs, so what follows is based on a somewhat-more-than-casual familiarity, not experience in the clean room. I do have colleagues who are genuine experts and I will be reaching out to one of them because I find this question interesting. When I get the chance, I will try to update this post with what they tell me.)

Here’s my understanding…

There are two sorts of LED dies: directional and side emitting. The side emitting dies are designed for backlighting applications to reduce hotspots. The directional dies are the sort in a typical indicator and emit perpendicular to the P and N layers.

In both cases, it is the active region where the P and N layers are alloyed, that emits photons. But in the directional LED the emitted photons are directed up to the top surface of the die by the internal structure. The cathode of the LED is almost always bonded to the substrate, which is the part that is bonded to the cup (or reflector cup) which is the part of the lead frame you are calling the anvil*.

The cup acts as a parabolic reflector partly determining the emission angle of the LED and so its apparent power. Since the internal structure of the die is designed to maximize top-surface (anode) emission, the reflector deals with photons that find another path—the directionality is not perfect.

While it is not physics that determines this directional characteristic, that is, you can design for directionality towards eother the anode or cathode, standardization and its benefits drive the almost universal practice of producing LEDs in the cathode-down configuration. I believe that there are cases such as yours where this convention is not followed but I have a suspicion.

It appears that the cathode is on the side of the LED case with the flat, is this true? Also, is the cathode the long lead?

*I had never heard this terminology before, and strangely it seems to have only one source, an illustration which appears over and over again copied from… copies?

 

https://eva.fing.edu.uy/file.php/585/materiales/HD/RedLedHP.pdf
TS AlGaAs is a flip-chip LED technology, die attached to the anode lead and wire bonded to the cathode lead.

It's a semi process thing that can reduce or eliminate post chip production die processing steps.
https://led.samsung.com/insights/insights-detail-2/

https://www.fagerhult.com/knowledge-hub/LED/clo-with-l100/leds-with-flip-chip-technology/
View attachment 1692907165978.png

 

TS AlGaAs is a flip-chip LED technology, die attached to the anode lead and wire bonded to the cathode lead.

I’d totally forgotten about the transparent substrate LEDs. They are quite unusual, though, in typical T packaging from what I have seen. They are usually used in multi-LED arrays for high density, high output emitters (usually white).

I suppose the TS’ LED could be an HP, but then it wouldn’t have a red lens. On the other hand, if the flat side lead is the cathode, there is something wrong because even if the cathode is bonded to the substrate, the indicator on the case should still be correct. I can’t be sure from the photo, so I am waiting to find out.

 

but I have also seen LEDs without a flat and this would be the only indication of whether they are fitted correctly once the legs are trimmed down after soldering.

You can use an old school continuity checker (2 AA batteries and a flashlight bulb) to verify polarity in-circuit or out.

Sorry about the poor focus. Took the picture in a hurry years ago and haven't felt like taking a better one...