Issues with simple Neopixel code

Toit

#Issues with simple Neopixel code

Thread channel in help

alex (he/him) 10/26/2025 07:52 PM

Using the following code, I get this when using jag monitor. Any ideas?

import gpio
import pixel_strip show PixelStrip

PIXELS ::= 1
main:
  pin := gpio.Pin 08
  pixels := PixelStrip.uart PIXELS --pin=pin

  r := ByteArray PIXELS
  g := ByteArray PIXELS
  b := ByteArray PIXELS

  r.fill 0x44
  g.fill 0x80
  b.fill 0xff

  pixels.output r g b

  sleep --ms=1

floitsch 10/26/2025 07:55 PM

Interesting.
10:1 is uart::create if I'm not wrong

floitsch 10/26/2025 07:55 PM

I don't see how and/or why that would take too long

floitsch 10/26/2025 07:55 PM

Is it reproducible?

floitsch 10/26/2025 07:57 PM

In the meantime, you can just use Pixelstrip.rmt. (for one pixel that should work perfectly fine)

floitsch 10/26/2025 08:08 PM

I can reproduce.

floitsch 10/26/2025 08:08 PM

And the rmt variant doesn't seem to work either.
Very strange.
I will debug.

alex (he/him) 10/26/2025 08:08 PM

rmt

puts it into deep-sleep

alex (he/him) 10/26/2025 08:09 PM

alexanderniebuhr@MacBookPro ~/D/P/c/toit [1]> jag monitor
✔ /dev/cu.usbmodem578E0248291
Starting serial monitor of port '/dev/cu.usbmodem578E0248291' ...
J�ESP-ROM:esp32c6-20220919
Build:Sep 19 2022
rst:0x1 (POWERON),boot:0x8 (SPI_FAST_FLASH_BOOT)
SPIWP:0xee
mode:DIO, clock div:2
load:0x40875720,len:0x9c
load:0x4086c110,len:0xb94
load:0x4086e610,len:0x2534
entry 0x4086c110
[toit] INFO: starting <v2.0.0-alpha.188>
[toit] DEBUG: clearing RTC memory: invalid checksum
[toit] INFO: running on ESP32C6 - revision 0.0
[toit] INFO: entering deep sleep without wakeup time

floitsch 10/26/2025 08:10 PM

How did you install the rmt version?

floitsch 10/26/2025 08:10 PM

For me it just doesn't do anything.

floitsch 10/26/2025 08:11 PM

could it be that the program finishes, and then nothing is left to do, and the system goes into deep-sleep?

alex (he/him) 10/26/2025 08:11 PM

Just installed it and then jag monitor

floitsch 10/26/2025 08:11 PM

installed it how?

alex (he/him) 10/26/2025 08:11 PM

Neopixel doesn't light up

alex (he/him) 10/26/2025 08:11 PM

/Users/alexanderniebuhr/.cache/jaguar/sdk/bin/toit compile --snapshot -o hello.snapshot hello.toit && /Users/alexanderniebuhr/.cache/jaguar/sdk/bin/toit tool firmware -e firmware.envelope container install -o hello.envelope hello hello.snapshot && /Users/alexanderniebuhr/.cache/jaguar/sdk/bin/toit tool firmware -e hello.envelope flash --port /dev/tty.usbmodem1401 --baud 921600

floitsch 10/26/2025 08:11 PM

Ok. That confirms my suspicion.

floitsch 10/26/2025 08:12 PM

Basically, your program finishes, and the system then decides that nothing is left to do and shuts down.

floitsch 10/26/2025 08:12 PM

If you increase the sleep, you will see that it doesn't go into deep sleep immediately.

floitsch 10/26/2025 08:12 PM

However, there is a bug with the rmt PixelStrip.rmt on the C6(edited)

alex (he/him) 10/26/2025 08:12 PM

Oh let me put a loop with logs, but I think that doesn't light up the pixel anyways

floitsch 10/26/2025 08:13 PM

You can work around it for now with the (deprecated) I2sPixelStrip:

import pixel_strip show I2sPixelStrip
...
main:
  ...
  pixels := I2sPixelStrip PIXELS --pin=pin

floitsch 10/26/2025 08:14 PM

I will debug the other ones.

alex (he/him) 10/26/2025 08:16 PM

I still don't get it to light up. But I'm using a strange waveshare dev board

floitsch 10/26/2025 08:17 PM

https://www.waveshare.com/esp32-c6-dev-kit-n8.htm
This one?

ESP32-C6 Microcontroller, WiFi 6 Development Board, 160MHz Single-c...

ESP32-C6 Microcontroller, WiFi 6 Development Board, 160MHz Single-core Processor, ESP32-C6-WROOM-1-NX Series Module, Supports USB And UART Development | ESP32-C6-DEV-KIT-N8

alex (he/him) 10/26/2025 08:18 PM

yeah

floitsch 10/26/2025 08:18 PM

pin 8 looks correct, and I manage to make mine light up. So that's a bit weird.

floitsch 10/26/2025 08:19 PM

So the I2sPixelStrip variant doesn't work for you?

alex (he/him) 10/26/2025 08:20 PM

No. But can be my board..

floitsch 10/26/2025 08:21 PM

Either way, the uart-deadlock is clearly a bug. And the RMT should also work.
I just tested on the normal ESP32 and there all three work.
Will now try with the s2, s3 and C3.

floitsch 10/26/2025 08:25 PM

S2 works.

floitsch 10/26/2025 08:30 PM

C3 only works with uart and I2s.

floitsch 10/26/2025 08:31 PM

I'm guessing RMT is currently broken on C3 and C6, and uart is broken on C6.

alex (he/him) 10/26/2025 08:42 PM

I'll get new prototype boards next week.. c3 & c6. But sounds like I have to switch to s3

floitsch 10/26/2025 08:44 PM

The c6 shouldn't be broken.
I'm already debugging the UART issue.

floitsch 10/26/2025 09:55 PM

I made a bit of progress, but no solution yet.
We are stuck in uart_ll_update, which waits for the peripheral to clear a bit.
I think the clock isn't active which is why we just keep looping there.
Will give up for today, but this is now a high priority.

36 messages in total