What is this component? Its part of a garden solar light - flame flicker effect

[paul_b_78]

I am struggling to identify this component. Obviously it is a SOT23-5, and I also came across a previous post similar where a YX2018/YX2019 was identified. But that was in regard to a SOT23-5 but 2 pins were joined.
Also, to throw a spanner in the works, the GND pin is in a different location to the YX2018/YX2019.
I know these lights are cheaply made from China, but I have quite a few of them scattered around my garden, and as a set they were quite expensive to replace.
They produce a nice flame flicker effect at night.

Notethe ?? are also unknown connections

[bingo600]

Guess

Something like a yx8183h in sot23-5

https://blog.csdn.net/weixin_44788830/article/details/130505835

[paul_b_78]

That looks perfect. I have been searching for a few days and you have located exactly what I have been for. Thank you
Now I just got to find somewhere to purchase said component, ideally not from that link you sent me, because I am not too sure that will cause me some bank issues...if you know what I mean

[paul_b_78]

I’ve just been looking closer at this component. The the YX8183 requires 3.7v supply, but this Solar light is running from a 1.2v 600mAh Ni-MH AAA single cell.

Plus the pin configuration is not entirely correct because the Ground is positioned in the wrong place…and according to my calculations, the Sol (+Ve) is connected to pin 1

[RJSV]

   Apparently, there is a more capable assembly, that includes a little vane, moved by magnetic coil, that will give some extra motion, to flicker in intensity by way of LED circuit but also sway side to side.
   Then there are much cheaper products containing simply the LED flickering, which I suspect were brought to market after the more complicated / more expensive had sort of ESTABLISHED some market demand.

   More postings, please see, the thread,
of rogeorge, 'Say a weekend Project'.
Much discussion there you might read.

[paul_b_78]

I am struggling to locate that Thread. Would you be able to share the link to that thread please.

Yes, after a little more investigation, I did wonder if these were cheaply made, and the place I purchased them from either had no idea just how cheap they actually are, or if they knew and just wanted to make a bit of a tidy profit.

[Ranayna]

That suspiciously looks like the circuit in the latest video of BigClive:


That also has just a SOT32-5 chip, an inductor, a capacitor and a couple of LEDs, and uses a NiMH battery.
The chip in Clive's video is a YX805. I don't have the time at the moment to verify the pinout, but the video should have all info.

That might mean the the flicker effect comes from the LEDs themselves.

[RJSV]

Right now the Rogeorge thread is the next item in list, under GENERAL subjects.

   There are two things, an LED that flickers by itself, and secondly, a little coil used to physically move a little vane that the LED is focussed into.  That second type is yet more realistic, likely a bit more expensive, due to the extra mechanism, coil and magnet, plus a wire pivot....

[bingo600]

The chip in Clive's video is a YX805.

I actually just received 60 of those from "Ali"
https://www.aliexpress.com/item/1005006016803680.html?

I have multiple garden lights where 90% stopped working after the first winter. ... 
Suspicion moisture ... bad sealing

Super simple inside ...
Inductor + Diode + YX805 + Liion batt

I tried to change the batt ... No luck.
Next up is the YX805 - 4-pin "in transistor house"

[RJSV]

The device I've been seeing is different than your pictures.   In regular steady lamp, (not flicker type) I see a very thin plastic package, with 4 pins, approx. 1/2 the thickness of your pict.   Your device looks like a regular IC package.
So I'm not sure, but thought was more than just a transistor internally.
The units I see, whether flicker or not, all have the little thin single in-line pack.

[RJSV]

...and YES, they aren't weather tolerant.   I've seen corrosion / rust at the LED leads, mainly because that's the most exposed, sticking out of case.   Corroded battery and terminals inside.

   EARWIGS TOO !   Many times my older units from last year, have stopped running, but provided nice cozy EARWIG HOME.   Ugg

[RJSV]

I just went and  verified;  that's the YX805 4 pin SIP.
   Obviously not just one transistor.

[Andy Watson]

Could the mystery chip be a standard single-cell boost convertor designed with a deliberately unstable feedback loop to give the flickering effect?

[RJSV]

   Thanks for the suggestion.  It's a good time to check on that for you, as all the components and tools are spread out in bench, ready to be worked on right now.
   Will let you know, after checking on that.

[RJSV]

   Andy Watson:
   I checked and pretty much convinced that the flicker is done right at the LED (right on the internal die, itself).
    Too difficult to check for printed part numbers;  my age advancing so fast, even a big magnifying glass, out in the sun, cannot read any printing, on the low profile black plastic 4-pin.   So I assume thats the commonly seen charge control and boost circuit, taking the 1.2 V nimh battery voltage up to about 2.4 V.

That flicker is ultimately done right at the LED.
The in-line package has to be more than just a transistor, as it does the oscillation, and the comparator that enables running the voltage boost oscillator (plus a low to mid pwr output transistor).
 
   So, that little 4-pin SIP is what I've seen virtually everywhere, in small Solar Lights.

By the way:   The oscillator mentioned runs at some hundreds of Khz and maybe a good candidate for operating a string of counter / click dividers.   Eight stages of binary counter gets that higher frequency down.... to some 509 hz to 2,000 hz (depends on original osc speed).

[Buriedcode]

One cool thing with those self-flickering LEDs - amplify a photo diode and play it through a speaker - I vaguely remember doing this years ago and the ones I had were playing happy birthday, but at a different frequency than greeting card sounders.

[RJSV]

Ah!   What comes to mind is the (clever) developer just grabbed the best AUDIO FILE to 'play' a pulsating signal through the output LED.
(Thus you get the Happy Birthday audio, when you tried using a speaker.
   That's the kind of ingenuity a person can't learn from books !

   By the way, did you try that audio with a PIEZO speaker, or little 1 " toy type.
(Regular 8 or 16 ohms maybe too much power required ?).
Thanks

[mikerj]

Ah!   What comes to mind is the (clever) developer just grabbed the best AUDIO FILE to 'play' a pulsating signal through the output LED.
(Thus you get the Happy Birthday audio, when you tried using a speaker.
   That's the kind of ingenuity a person can't learn from books !

   By the way, did you try that audio with a PIEZO speaker, or little 1 " toy type.
(Regular 8 or 16 ohms maybe too much power required ?).
Thanks

The IC is built into the LED itself, it would be a bit tricky to connect a speaker directly to it!

[Buriedcode]

Could put it in series with a speaker - the LED's current draw is mostly when the die is powered, so it'll have a pretty big difference between the LED on and off.  As to whether it'll be loud enough to hear, who knows.

[RJSV]

   I believe you meant to say;  that most of the current draw to the whole die is LED when LED is driven more....and so an XTRA LED in series with die will be perceived fluctuating, as the led on die fluctuates.

I can test that.   However, that's more voltage drop, so likely it's best to use a red colored LED for least drop...lest the circuit in the die might not get enough supply volts.
   Easy to test...it's the busy personal schedule that slows down test bench activities.