WS2813 LED Strip with ESP8266: First Diode Lighting Issue and Logic Level Concerns

I have an ESP8266 and a 5m WS2813 strip. All powered from a 20A 5V power supply. The first diode goes into such a strange state: it lights up either white/blue, once it paired properly after a reset. And now after plugging in it lights up red all the time. I know that the WS2813 should be operated with 5V logic, and the ESP gives 3.3V logic. However, it is probably the other diodes that should be the problem, but tests over longer distances show that the rest light up rather correctly.

Is the first diode faulty, or should I dig out some 3.3V -> 5V buffer. The question is whether it can be unidirectional to WS2813, or whether it has to be bidirectional.

And one more thing: the first examples use the ESP8266's DMA and connections only on GPIO3, which I did by connecting the DI inia. Whereas in t most of the examples, they say nothing about the CLOCK ini. So in my case it hangs around. There was only one example where both lines were defined (the example was on control from down pins). Any documentation on what to connect the CLOCK line from the WS2813 to?

Maybe it is the case that the first diode, ingests the data in 3.3V logic and if its driver receives it, it gets all tangled up. If, on the other hand, it recognises that the command doesn't apply to it, it lets it go on in 5V logic, which every subsequent diode already understands without error? I wonder if I am right?

WS2813 do not have a CLOCK line.

.

The examples are universal and talk about one pin. I was connecting DI, I was interpreting it as Data.
So the other one is Backup?
How do we connect it?
We clip together at the Bi and DI input ?

Added after 2 [minutes]: .

Oki, I found a drawing here that it is Backup:
https://www.superlightingled.com/sp110e-bluet...-rgb-addressable-led-strip-lights-p-1367.html

Regarding the 3.3V logic control of the ws2812 ribbon, someone suggested a cool idea:
https://www.youtube.com/watch?v=letCT3vq8V4

Comments under this video suggest that the first diode can regenerate the signal, which would confirm that the first diode under 3.3V control can light up strangely.

Added after 26 [minutes]:

And here someone propose another conversion with the help of the "first diode" https://hackaday.com/2017/01/20/cheating-at-5v-ws2812-control-to-use-a-3-3v-data-line/
But that requires cutting off the first diode from the rest of the strip so the "myk" with two diodes to feed the ESP seems cleverer.

@marcingebus Below is the wiring diagram for Wemos. The transistor acts as a unidirectional 5V/3V3 logic level converter and emitter secondary.
Source: https://github.com/arendst/Tasmota/wiki/WS2812B-and-WS2813

.

What library for handling WS are you using?

khoam wrote:

What library for WS support are you using?

.
So far I have fired up the ticker tape to see if it works. I used the library and examples from it "Neopixel by Makuna". I searched in the libraries: "ws2813".

Instead of combining, wouldn't it be better to give 74HCT1G08? One element smaller than the transistor + 2 resistors, certainly sharper edges especially rising.

You could possibly try the FastLED library. It supports ESP8266 and there are quite a few examples with different effects.
https://github.com/FastLED/FastLED

khoam wrote:

@marcingebus Below is the wiring diagram for Wemos. The transistor acts as a unidirectional 5V/3V3 logic level converter and emitter secondary.
Source: https://github.com/arendst/Tasmota/wiki/WS2812B-and-WS2813

.

What library for handling WS are you using?

.
Strange this schematic - when D4 is low - transistor Q1 conducts. When high - it also conducts (3.3V+0.7=4V, and the supply is 5V). Why not make a classic circuit on an NPN transistor with emitter to ground?

krzbor wrote:

Strange this schematic - when D4 is low - transistor Q1 conducts. When high - also conducts (3.3V+0.7=4V, and 5V supply)

The schematic around 4V is the high state for WS2813 supplied from 5V. On a single NPN(phase inverter) you would have to do some reworking of the ready made libraries .

ekrzychoooo wrote:

krzbor wrote:

Strange this schematic - when D4 is low - transistor Q1 conducts. When high - also conducts (3.3V+0.7=4V, and supply 5V)

The schematic around 4V is the high state for WS2813 supplied from 5V. On a single NPN(phase inverter) you would have to do a bit of reworking of the ready-made libraries .

As I wrote this, I also thought that someone did not want to improve the libraries. Otherwise this solution has no justification.