Quinled dig-next-2 RMT driver issu

[emmanuelborgetto-lgtm]

Using a 12V WS2811 RGB led strip, I get flicker when I get past 12 leds in my layout.
The color from a solid color effect is right for the first 7 leds, then it is messed up. White seems to work better (less flickering)

Sometimes, it lights more leds than set up, and these one are not changed when I change the color.
When I reduce the number of leds, the leds do not shut off, I need to turn off the driver (wich turns off the power via the relay).
When I shut off the hue or brightness in the effect the leds I did not set up will eventually shut off after some time, and the leds I set up shut off normally.

When I set up brightness to full, after it's on 0, more leds than needed light up. And with a random color...

My RMT log :
{"name":"RmtLed","type":"RmtLedDriver","role":"driver","enabled":true,"tickTimeUs":786,"classSize":776,"dynamicBytes":24579,"status":"driving 18 of 18 lights (54 channels)","severity":"status","controls":[{"name":"localBrightness","type":"uint8","value":255,"min":0,"max":255},{"name":"curve","type":"select","value":0,"options":["CIE 1931","gamma 2.2","gamma 2.8","linear"]},{"name":"lightPreset","type":"select","value":0,"options":["RGB","GRB","BGR","RGBW","GRBW","WRGB","Curtain GRB6","Lightbar RGBWYP","RGBCCT","IRGB","MH BeeEyes 15","MH BeTopper 32","MH 19x15W-24","MH Mini 10W 11"]},{"name":"whiteMode","type":"select","value":1,"options":["None","Min","Accurate"],"hidden":true},{"name":"presetRef","type":"text","value":"RGB","hidden":true},{"name":"start","type":"uint16","value":0,"min":0,"max":65535},{"name":"count","type":"uint16","value":65535,"min":0,"max":65535},{"name":"pins","type":"text","value":"2,4"},{"name":"ledsPerPin","type":"text","value":""},{"name":"loopbackTest","type":"bool","value":false,"advanced":true},{"name":"loopbackTxPin","type":"pin","value":-1,"min":-1,"max":39,"hidden":true},{"name":"loopbackRxPin","type":"pin","value":-1,"min":-1,"max":39,"hidden":true},{"name":"loopbackFrame","type":"bool","value":false,"hidden":true}]}

I made a video of the behaviour I see.
You can find it here (my personnal website) : https://illegalcommand.org/nextcloud/s/bTKbEATAwE9iJjb

[MoonModules]

I checked your video, made a screenshot of your leds:

Can you provide exact details what they are (link to website, or photo of its sticker), as they are not the traditional led strips, but "drops", I have them as well for led curtains and they have "particular behavior" sometimes.

[MoonModules]

WIP: adding led strip timings control, will be shipped later today

The four options are named by bit rate, not by chip, because chip names don't map cleanly onto timings: SK6812 and WS2812B decode identically, and "WS2811" exists in two different speeds.

800kHz WS2812B/SK6812 — the default, unchanged behaviour
400kHz WS2811 — try this one first
800kHz WS2811 fast — WS2811 strapped to high speed
custom — reveals t0hNs / t1hNs / periodNs for a strip matching no preset
Why 400 kHz is the first guess for your setup: 12V drops are almost always WS2811, and that part ships in a low-speed 400 kHz mode where every bit cell is twice as long. Fed 800 kHz timing, such a strip decodes the first few pixels and then reads noise, which looks exactly like your video: correct colour for the first 7, garbage after, white surviving best (all one-bits are the most forgiving pattern), and extra pixels lighting with colours that never update.

Switching is live and doesn't touch your pin configuration, so you can try the presets one after another without re-entering anything. Note a 400 kHz strip takes twice as long per frame, so max frame rate halves at a given pixel count.

If neither preset works, custom takes the numbers straight off a datasheet, and whatever values you find working are useful for us to know.

Still worth having the exact part if you can find a sticker or a link. Drops also run much longer data cable between pixels than a strip does, so signal integrity can produce the same symptom, and the part number is what separates the two.

[emmanuelborgetto-lgtm]

Here is the link I purchased it from : https://fr.aliexpress.com/item/1005008643187951.html

I know it may not be the most reliable leds, but that is the one wich fitted the estetics of the project.
I'll also attach photos of the sticker.

[ewowi]

Two fixes are now on main, addressing different halves of what you saw.

1. Strip timings (a setting, try this first). Your 12 V drops are almost certainly WS2811 in low-speed mode, where every bit cell is twice as long. Fed 800 kHz timing, such a strip decodes the first few pixels then reads noise, which matches your video: correct color for the first 7, garbage after, white surviving best, extra pixels in colors that never update.

The RMT driver has a new timing control: try 400kHz WS2811. Also there are 800kHz WS2811 fast and custom (which reveals t0hNs / t1hNs / periodNs for a strip matching no preset). Switching is live and does not touch your pin config. A 400 kHz strip halves the maximum frame rate at a given pixel count.

2. Flicker under WiFi (automatic, no setting). Separately, the classic ESP32's RMT has no DMA: it plays from a small on-chip block that an interrupt must refill every few dozen microseconds, and the stock driver services that at a priority any critical section can block, so a late refill re-clocks stale symbols into the strip. The refill now runs at interrupt level 5, above anything that can hold it. Flicker-free on our two boards, one a Dig-Next-2.

If 400 kHz does not fix it, the remaining suspect is signal integrity rather than firmware: drops run long cable between pixels, and a 3.3 V data line into 12 V pixels has little margin. Then custom takes numbers off a datasheet, and whatever works is useful for us to know.

Please report back either way, since we cannot reproduce your exact strip here. The sticker photos and video are what made the diagnosis possible.