How do self blinking LEDs work?

I'm wondering how those 5mm LEDs with internal blinking circuitry work.
If that circuitry were expanded to surface mount components, what would be involved?
Transistors I can understand, but can capacitors be involved at that tiny scale? Maybe it uses a thermal system?

Could you provide a link to the device you are describing?

Probably not thermal. Capacitors and resistors are possible but values are small. It may be possible to make a high frequency oscillator and a frequency divider.

Other more clever techniques are also used.

For example, a large FET may consist of N+1 tiny FETs (N could easily be 1000) in parallel. The LED current could be conducted through N FETs, with the 1 FET serving to provide a tiny charging current for controlling blinking.

What does your device use? Who knows? Perhaps if we had more information then we could find out, maybe not. Sometimes, these mechanisms are considered proprietary.

INTP:
I'm wondering how those 5mm LEDs with internal blinking circuitry work.
If that circuitry were expanded to surface mount components, what would be involved?
Transistors I can understand, but can capacitors be involved at that tiny scale? Maybe it uses a thermal system?

Resistors, capacitors and transistors (BJT or FET) can all be fabricated on most silicon processes. Capacitor
values are necessarily small, so for a long period blink its likely to be an oscillator and divider chain,
since small caps lead to higher frequencies of oscillation. Think a 555 plus a 74HC393. It might even
be a small microcontroller, whatever I.P. is available cheaply to the designed of the device I guess.

It won't be thermal I don't think!

Well i'm not sure what kind you have but on the upper scale the diode it's self is pretty much the same as any led but it has an integrated circuit with one nanofarad capicator and one nanofarad variable capicator

Even a microprocessor as small as a SOT23-6 package can have two oscillators built onto it. One can easily fit into an LED.

I don't have any, and there's likely only one method used.
Here's a link with a data sheet, just says IC is involved.
https://www.circuitspecialists.com/l56bhd.html?otaid=gpl&gclid=CPSDktq9rNQCFQWSaQodVlwCiA

I've only seen them as slow as 1.6ish Hz so that must be the technical limit.

I was hoping someone had intimate knowledge of what the IC entails. Maybe someone with these and an oscope could describe what it looks like in a circuit- normal LED behavior of constant voltage drop or perhaps an open circuit during off-times.

INTP:
I don’t have any, and there’s likely only one method used.
Here’s a link with a data sheet, just says IC is involved.
https://www.circuitspecialists.com/l56bhd.html?otaid=gpl&gclid=CPSDktq9rNQCFQWSaQodVlwCiA

I’ve only seen them as slow as 1.6ish Hz so that must be the technical limit.

How does that follow?

I was hoping someone had intimate knowledge of what the IC entails. Maybe someone with these and an oscope could describe what it looks like in a circuit- normal LED behavior of constant voltage drop or perhaps an open circuit during off-times.

There will be someone, working for a semiconductor company, probably in the Far East, probably under
a non-disclosure agreement. And remember every LED company will have their own take on this, its a highly
competitive market.

A couple of "candle flicker" LEDs have been reverse engineered, as described here.
The device consists of two chips, the actual LED and a fairly simple controller.

Lately, some solar powered garden lights have come on the market that have multicolor LEDs and controllers that change color and brightness in a long term pattern. Those have some simple microprocessor built in to the LED housing.

Thank you for that link, jr. Fascinating stuff. I guess 1980's tech is still beyond my grasp lol.