Problem with Micropython on ESP32-C3 Super Mini

[bcux]

I have a problem with Micropython (1.22.2) on ESP32-C3 Super Mini

from machine import Pin
import time
p=Pin(8,Pin.OUT)

while True:
    p.on()
    time.sleep(1)
    p.off()
    time.sleep(1)

The program runs without any problems when I enter it via the console (USB connection with terminal program or via rshell and repl).
However, if I save it via rshell as main.py in the ESP32-C3, it does not run: LED does not flash.
The version of Micropython does not matter. On ESP8266 or ESP32-WROOM it runs as main.py without any problems.

[robert-hh]

Are you sure that this code resides as main.py on the board? What do you get when typing the below line at REPL:

open("main.py").read()

[bcux]

Yes, I'm shure

/home/bcux> ls -la /pyboard/
   139 Jan  1 2000  boot.py
   125 Mar  6 17:51 main.py

I can edit it with nano (which is my editor for rshell) and:

/home/xc/esp> repl
Entering REPL. Use Control-X to exit.
>
MicroPython v1.22.2 on 2024-02-22; ESP32C3 module with ESP32C3
Type "help()" for more information.
>>> 
>>> open('main.py').read()
'from machine import Pin\nimport time\np=Pin(8,Pin.OUT)\n\nwhile True:\n    p.on()\n    time.sleep(1)\n    p.off()\n    time.sleep(1)\n'
>>> 

[robert-hh]

That's strange. Similar, on a ESP32C3, if I enter REPL with rshell and then push Ctrl-D, main.py is executed. What is the content of your boot.py file?

[bcux]

If I enter repl and press Ctrl-D, main.py starts, but it does not start, when the ESP32-C3 is powered up. The behavior is very different to ESP8266 and ESP32-WROOM where main.py is started when powering up.

boot.py is not changed, it is the same as after flashing the ESP

# This file is executed on every boot (including wake-boot from deepsleep)   
#import esp                                                                  
#esp.osdebug(None)                                                           
#import webrepl                                                              
#webrepl.start()  

[robert-hh]

There is a larger capacitor next to the 3V regulator, which is connected to 3V3 and GND. I assume that this ought to be the 10µ part. The Multimeter says 17µF. But that may be wrong, since the part is soldered.
The board does not have Reset (EN) exposed at the solder pads. The one next to the RST label is GPIO 4.

[davefes]

OK on RST! If the non-ground end on the RST button is closest to GPIO 4 it still might be "close" to the area in question.

Another, perhaps relevant question, where is system ground? On the shell of the USB connector?

[robert-hh]

The USB shell is connected to GND, and the GND pin is next to it.

[davefes]

Do all the regulator bypass caps return to that point by a copper ground plane or by thin tracks back to the USB connector? I am diving-into layout details that may not even be relevant here.

I would want to understand why the added cap fixed things as that might suggest a bigger problem or a better solution. Adding caps may fix one problem but could introduce other problems.

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwig5saVrOOEAxV12DgGHcrZAs0QFnoECBsQAQ&url=https%3A%2F%2Fwww.compuphase.com%2Felectronics%2FParallelCapacitors_Antiresonance.pdf&usg=AOvVaw1XyxXbtGMVCb8Ebh8CiymC&opi=89978449

[robert-hh]

This is a cheap board. 2 layers, as it seems without a ground plane. And the habit of having a black solder stop lacquer makes it hard to follow traces. So I stop here.

[bcux]

Yesterday i received 2 ESP32-C3 from a friend and his devices were fully functional. I'm now sure that my parts have a manufacturing defect.