QUERY · ISSUE

Neopixel "flicker"/noise only when Wifi is on

openby jomasnashopened 2025-04-14updated 2025-12-28

bugport-esp32

Port, board and/or hardware

esp32-C3 Mini

Matrix board like this https://www.hobbyelectronica.nl/wp-content/uploads/2021/11/8x8_WS2812_Led_matrix.jpg
with WS2812b Neopixels, all individual Neopixels have decoupling capacitors.

MicroPython version

MicroPython v1.24.1 on 2024-11-29; ESP32C3 module with ESP32C3

Reproduction

When Wifi is enabled and connected, a neopixel matrix (8x8) gives unexpected glitches.
Power and signal levels are checked and ok. (checked with an oscilloscope.)

This bug is the same as in https://github.com/micropython/micropython/issues/8161 however the fix will not work because the esp32-C3 Mini has only one core.

If this bug cannot be solved, I suggest adding a warning in the documentation that Neopixels on eps32 (single core types) cannot be used reliable when Wifi is on and used.

import time
from neopixel import NeoPixel  # Standard lib of MicroPython v1.24.1
from machine import Pin
import config as cfg
import network

wlan = network.WLAN(network.STA_IF)
wlan.active(False)
ap_if = network.WLAN(network.AP_IF)
ap_if.active(False)

if not wlan.active():
    wlan.active(True)
    time.sleep_ms(100)
if not wlan.isconnected():
    wlan.connect(cfg.ssid, cfg.passwd)
while not wlan.isconnected():
    time.sleep_ms(100)
    print('.', end='')

np = NeoPixel(Pin(4, Pin.OUT), 32)  # Use 4 rows of 8 pixels
np.fill((0,0,0))  # All pixels off
np.write()

np[0] = np[8] = (1,0,0)  # Red, lowest intensity
np[16] = np[24] = (0,1,0)    # Green, lowest intensity

while True:
    np.write()
    time.sleep_ms(20)

The configuration of the pixels of this board is as follows:

0  1 . . . . . 6 7
8  9 . . . . . . . 
16 etc.

So the example program above should give

R . . . . . . .
R . . . . . . .
G . . . . . . .
G . . . . . . .

Expected behaviour

First column (0) should be as above and steady

Observed behaviour

Only when Wifi is on and connected, glitches appear like for example:

R . . . . . . .
R . R . . . . .
G . G . . . . .
G . . . . . . .

or

R . . . . . . .
R . . . . . . .
G . G . . . . .
G . G . . . . .

The glitches always appear in column 2 (two pixels away from the first column)
and has the same color as the left one. I have not seen glitches in top row (0)

Additional Information

Conditions:

Power supply (both 5 and 3.3Volt) has 100nF ceramic parallel with 470uF capacitors.
Input signal of matrix module is within datasheet specs and look clean (Checked with oscilloscope).
No long wires.
No high currents (average current of the setup is below 50mA)

I checked the (total) time of the pulsetrain on the output of the esp with a oscilloscope, it shows 32 * 24 * 1.25uS = 960uS
This is as expected. However sometimes the pulsetrain is extended an extra 60uS.
This is probably when the glitches appear.

I also did the test with all 64 pixels (total time of pulstrain 1.92mS, column 0 -> RRRRGGGG) then also sometimes the extra 60 uS appear. Also parts of (bottom of) column 0 is copied to column 2 like above.

I have tested this with a few esp32-C3 and a few matrix boards. They all have the same behaviour.
I also tried using machine.bitstream but the same behavior shows up.

Code of Conduct

Yes, I agree

CANDIDATE · ISSUE

RMT + Neopixel + WiFi = Flickering

closedby scoatesopened 2022-01-11updated 2024-03-08

port-esp32

This is related to #8158 + #7985.

I am experiencing neopixel "flicker"/noise only when Wifi is on.

Some background on my setup (same as what I described in #7985):

a 16x16 RGB panel (256 LEDs)
connected directly to my LOLIN D32 Pro board on GND, USB (+5V), 25
I'm using the firmware that @jimno built and shared in #7985
no capacitors or level shifters (maybe this would help, but this setup seems to work fine in FastLED, so I'm doubtful)
I've powered it both through USB on my computer, and by unplugging the USB connection and putting +5V over the panel's connection, which feeds back into the "USB" pin as +5V
works fine with FastLED from what I can tell
the current amount of flickering (which is signalled with RMT, on this firmware, I assume), is much less than what I experienced in #7985
it seems to happen only in the "later" part of the panel—I haven't seen any flickering in the first 4 or 5 rows of 16 pixels
I've tried this on different ESP32 boards (a second LOLIN D32 Pro, and a different WROOM board); same results
I tried a different LED panel that appears to be the same hardware, and the same problem
I've taken a short video. It's a bit hard to see, but definitely visible. The 0 pixel is on the right and the 255 pixel is on the left.

Here's the code I'm using (with various timings; all similarly flickery, but only with wifi turned on and connected to a local WLAN in boot.py):

import machine
import neopixel


np = neopixel.NeoPixel(
    machine.Pin(CONFIG["DATA_PIN"]),
    CONFIG["NUM_PINS"],
    # timing=(350, 800, 800, 350),  # , timing=(400, 850, 800, 450)
)
while True:
    print("doing a round.")
    for p in range(CONFIG["NUM_PINS"]):
        np.fill((0, 0, 0))
        np[p] = (64, 0, 0)
        np.write()

in #7985, @odi6331 suggested:

I changed them on my setup to a timing=(350,800,800,350) and all the flickering was gone since then.

I tried this timing on my setup, and it worked fine with wifi off, but when I turned it on, the flickering came back.