Posts made by Paulskpt

I worked for weeks on an enhancement of an Arduino sketch that shows time data of seven different timezones in the world. At intervals the RTC is updated with a datetime stamp from an SNTP service server.
The past two weeks I was adding the functionality to send some data, like time sync time_t value to a Google Sheets spreadsheet. This all works OK.
The sketch ran a whole night, every 15 minutes sending data to a Google Apps Scripts script that adds the data to the spreadsheet.
In the sketch I have built the possibility to switch the display off (go to sleep) by using the command: "M5.Display.sleep()" and switch the display
on (awake) by using the command: "M5.Display.wakeup()". The switching off or switching on is done by holding an RFID card in front of the display. The sketch also has the possibility to switch the display on or off by touching the display (it is one or the other option. For now I use the RFID tag option). I used the Arduino IDE v2.3.3. for MS Windows 11. The Arduino IDE compile and build reports :

Sketch uses 1190657 bytes (90%) of program storage space. Maximum is 1310720 bytes.
Global variables use 53384 bytes (16%) of dynamic memory, leaving 274296 bytes for local variables. Maximum is 327680 bytes.
esptool.py v4.5.1
Serial port COM6
Connecting...
Chip is ESP32-S3 (revision v0.2)
Features: WiFi, BLE
Crystal is 40MHz
[...]
Wrote 1191024 bytes (721826 compressed) at 0x00010000 in 8.6 seconds (effective 1113.8 kbit/s)...

During the execution of the sketch calls to ESP.getFreeHeap(), result in values like: 265556 downto 264052.

After the display goes black, the sketch continues to execute fine, sending at the preset intervals its data to the Google Sheets spreadsheet.

Anyone that give me a clue to solve this abnormal behaviour?

I just uploaded a very short video impression to X. Here is the link: Post on X

I just uploaded a very short video impression to X. Here is the link: Post on X

FYI, I added functionality to make WiFi connection to a mobile phone. The WiFi credentials for the mobile phone are in the file secret.h. After reset the sketch will first try to establish WiFi contact with the mobile phone (when it's credentials are present in secret.h). If this fails, the sketch tries to establish WiFi connection with the fixed WiFi Access Point. In case of WiFi connection lost, the sketch will try to re-establish WiFi connection. The user will be informed on the OLED through brief messages.

Hi there, members. Last Saturday I received an ordered M5NanoC6. I ported the sketch that I created earlier for M5Dial and other devices as published in this forum. Amazing what fits in 4Mb or RAM!. The sketch is available on Github: repo.

@BoM_M5 I had the same problem today. In Arduino IDE v.2.3.3, BOARDS MANAGER, searched for esp32. Found "esp32" from Espressif Systems. It appeared that I had version "3.1.0 RC" installed. On Espressif's site on Github I saw that the latest stable version is: "3.0.5", so I downgraded to 3.0.5. Then I saw that the M5NanoC6 was available via: Tools -> Board -> esp32 -> M5NanoC6. After having selected this board, I saw also that: Tools -> CDC on Boot was "Enabled". I saw that there was here no more option "Disabled".
From this moment all Serial.print and Serial.println command worked as expected.

Hi members, yes, I was quite busy the past days creating projects.
This is a combination of the two recent projects: "M5Dial Timezones" and "M5AtomEcho ATOMECHOSPKR class to play simple beeps". In this combination the M5Dial sends a digital impulse (read: beep command), on pin 1 of its Grove Port B. The beep command is sent at the moment of a NTP Time Synchronization. In this moment a "TS" occurs every 15 minutes. The sketch running on the M5Atom Echo device is "listening" on its Grove Port pin 1 for a digital impulse. If such a "beep command" is received, or when the button on top of the M5Atom Echo is pressed, the Atom Echo will produce a double tone beep. At the same time the builtin RGB Led of the Atom Echo will first show Green and after the beeps have finished, the RGB Led will turn RED (again). See: repo.

Hi members, FYI I just created a new repo on Github. This repo contains a C++ the ATOMECHOSPKR class that I created and an Arduino sketch to test the new beep functionality. This sketch also contains functionality to use the builtin RGB Led of the Atom Echo.

I did not create the C++ class from scratch. I used the class ATOMSPK that M5Stack published on Github for the AtomSPK device.

See my: repo

Hi, FYI I created a new repo on Github:
M5Dial NTP local time

Hi, FYI I just created a new repo on Github. It is a port of the Atom Matrix Timezones repo.
See: repo

Hi members, FYI I created a new repo on Github.
See: repo.

Note: creating this project was only possible with great help from Microsoft CoPilot, AI assistant. Thank you Microsoft!.

Hi, I encountered the following:

In an Arduino sketch for my M5Stack Atom Matrix, I used the following code snippet,
that Microsoft CoPilot presented me to get the MAC address of the Atom Matrix:

uint64_t chipid = ESP.getEfuseMac(); // The chip ID is essentially the MAC address 
  Serial.printf( "\nESP32 Chip ID = %04X%08X\n", (uint16_t)(chipid>>32), (uint32_t)chipid);

The result of this code is: "ESP32 Chip ID = 90E08008B764".

Also

byte mac[6];
WiFi.macAddress(mac)

results in: 90E0808B764,

However, M5Stack M5Burner app reports the following MAC address: "64b70880e090".

Both values are a "mirror" of each other. How come?
Has it to do with "little endian" versus "big endian"?

Hi, FYI, I just created a new repo on Github. It is an Arduino sketch for the following M5Stack hardware setup:

• Atom Matrix;
• AtomPortABC;
• GROVE HUB;
• OLED unit;
• RTC unit.

The titel of the repo is my acronym for the used hardware.

See: repo

Hi there, FYI I just created a repo on Github of a M5Stack Cardputer with external RTC unit
See repo
PaulskPt @ Github

Hi there, FYI I just created a repo on Github of a M5Stack Cardputer with a M5Stack ENVII unit.
See repo
PaulskPt @ Github

Hi, FYI, just published on Github a repo called "M5Stack Cardputer syncd clock".
See: repo

@JonnyMac Hi Jonny, your post is quite old however I want to share with you (and anyone else) my experience.
I had the same "flash" or "flickering" while updating data on the M5Stack Core display. In my case update of elapsed seconds (every 10 seconds).
It was caused by the fact that quite a bit of redrawing took place in a same second. So I wrote the following lines to prevent redraw in the same second:

See especially the following set of lines of code:

 int elapsedSecsMod = elapsed_secs_t % 10;
 if (elapsedSecsMod > 0)
   elapsedSecs_shown = false;  // reset flag
 
 if (start || (!elapsedSecs_shown && elapsedSecsMod == 0)) {
   sprintf(charVal, "%6lu", elapsed_secs_t);
   disp_nr(120, charVal);    // Display elapsed seconds
   elapsedSecs_shown = true;  // set flag. Prevents value get redrawn many times in the same second
 }

Below the lines above in the total example:

unsigned long start_t;
unsigned long curr_t;
unsigned long elapsed_msecs_t;
unsigned long elapsed_secs_t;
unsigned long elapsed_mins_t;
unsigned long elapsed_mins_old_t = 0;
unsigned long interval_t;
int cur_h = 10;
int cur_v = 10;

void setup() {
  M5.begin();
  M5.Lcd.fillScreen(BLACK);
  Serial.begin(115200);

  start_t = millis();
  M5.Lcd.setCursor(cur_h, cur_v+120); 
  M5.Lcd.print("Elapsed Secs:");

}

void disp_nr(int v_incr, const char* sNr) {
  int cur_h_new = 0;
  
  M5.Lcd.setTextPadding(M5.Lcd.width() - 50);

  cur_h_new = 80;

  // NOTE: the 4th parameter to M5.Lcd.fillRect() must be at least 14 to cover all the height of a digit of M5.Lcd.setTextSize(2) !!!
  M5.Lcd.fillRect(cur_h_new, cur_v + v_incr, 100, 14, BLACK); // erase partial place for updating data

  if (strlen(sNr) > 0) {
    M5.Lcd.setTextSize(2);
    M5.Lcd.drawString(sNr, 100, cur_v + v_incr); 
    M5.Lcd.setTextSize(1);
  }
}

bool start = true;
bool elapsedSecs_shown = false;

void loop() {
  char  es2[] = "Elapsed Secs:  ";
  int le = 8;
  char charVal[le];
  curr_t = millis();
  elapsed_msecs_t = curr_t - start_t;
  elapsed_secs_t = (long) elapsed_msecs_t / 1000;
  elapsed_mins_t = (long) elapsed_secs_t / 60;

  int elapsedSecsMod = elapsed_secs_t % 10;  
  if (elapsedSecsMod > 0)
    elapsedSecs_shown = false;  // reset flag
  
  if (start || (!elapsedSecs_shown && elapsedSecsMod == 0)) {  // Every 10 seconds

    sprintf(charVal, "%6lu", elapsed_secs_t);
    disp_nr(120, charVal);   // Display elapsed seconds
    elapsedSecs_shown = true;  // set flag. Prevents value get redrawn many times in the same second
  }

  // [ more code here ]

  if (start)
    start = false;
}

Hello,

On Friday 2022-12-02 on a desktop PC with MS Windows 11 Pro, and runing the Arduino IDE v1.8.19, a sketch
for a M5Stack Core2 would not compile while this sketch built successfully on 2022-04-24.

To investigate the problem, I turned on 'File > Preferences > Show detailed messages during compilation'.

I discovered that the name of the folder of the M5Stack board package contained a '+' character (ASCII decimal value 43, hexadecimal value 2B).
The compiler reported that it could not find (the) file(s). (see below).

There are some sources that describe the limitations of which characters can be used in the Microsoft Filesystems.
For example:
a) Naming Files, Paths and Namespaces;
b) Wikipedia Filename.

As far as I understand it: for the current Microsoft filesystem, NTFS, the '+' character is not a 'reserved character'.
That is probably why there is no error generated when installing the M5Stack board package with the '+' in the name,
however source b above mentions limitations in FAT12, FAT16 and FAT32 file systems. See the text in the most right column,
the text I marked with an orange rectangle. Here the '+' character is listed as 'reserved character'. In other words: one cannot
use the '+' character in file or folder names for these file systems:
.

Despite a successful install of the M5Stack board package, the compiler was not able to find files because of
the '+' character in the folder name.

Fortunately I was able to install the package with version ID: '2.0.5-1.0'. Next I was able to delete the package with version ID: '2.0.5+1.0'.
After this correction the sketch compiled and built successfully (again)

Here the last part of the build session with the compile error:

The offending folder location and name was:
C:\Users<User>\AppData\Local\Arduino15\packages\m5stack\hardware\esp32\2.0.5+1.0

Today, when using the newest version of the Arduino IDE (version 2.0.3), I was asked to update a boards package from M5Stack.
While the version 2.0.5-1.0 was installed, the Arduino IDE suggested me to install the M5Stack version 2.0.5+1.0. Again the version
with the '+' sign in the version (and sure, the folder name too). With my previous experience I did not install the proposed update. (See the partial screenshot below):

Maybe I address the wrong persons with this post.
Perhaps I should report this problem to the maintainer(s) of the compiler used in the Arduino IDE.

Can I ask the moderators of this M5Stack forum to look into the matter and take corrective measures, if you think that is necessary?
Thank you.

The Espressif ESP32-S3-Box is a product that looks like some M5Stack products. I own quite a bunch of M5Stack products. When the Espressif ESP32-S3-Box (V 2.5) came out, I bought it. It was my intention to port my ongoing MSFS2020 GPSout project to the 'esp-box' (the (nick)name that Espressif uses for this device). It was until last weekend when through @Neradoc and @hathach, the maintainer of Adafruit tinyuf2 bootloader repo, on my request, had added the 'esp-box' to the list of boards compatible for the tinyuf2 bootloader. From that moment I was able to build a tinyuf2 bootloader firmware for the 'esp-box'. When that was done, I was able to copy the latest version of CircuitPython (V8.0.0-beta.2) to the device. Finally I was able to port my project to the 'esp-box'.

That is why I created this post here so you can read about the existance of my project. See my repo on Github.

Below some REPL output:

soft reboot

Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
code.py output:
]0;🐍Wi-Fi: off | code.py | 8.0.0-beta.2\
loop(): my_machine= "ESP32-S3-Box-2.5 with ESP32S3"

MSFS2020 GPS GPRMC data reception decoder script by @PaulskPt. 

Number of loops in this run: 99
........................

Start of loop 1
ck_uart(): nr of bytes=  120
ck_uart(): rcvd data: bytearray(b'$GPRMC,215807.00,A,5837.7935,N,13810.7771,W,164.0,319.8,161022,18.2,E*72\r\n$GPGGA,215807.00,5837.7935,N,13810.7771,W,1,05\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00')
ck_uart(): nr of bytes=  33
ck_uart(): rcvd data: bytearray(b',0.0,4266.7,M,0.0,M,0.0,0000*51\r\n810.7771,W,164.0,319.8,161022,18.2,E*72\r\n$GPGGA,215807.00,5837.7935,N,13810.7771,W,1,05\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00')
ck_uart(): nr of bytes=  153
ck_uart(): rcvd data: bytearray(b'$GPRMC,215811.00,A,5837.9408,N,13811.0171,W,164.0,319.6,161022,18.2,E*76\r\n$GPGGA,215811.00,5837.9408,N,13811.0171,W,1,05,0.0,4266.7,M,0.0,M,0.0,0000*5B\r\n\x00\x00\x00\x00\x00\x00\x00')
ck_uart(): * at -11, $GPRMC at 0, $GPGGA at 74, n4 =114
loop(): characters rcvd:  153
loop(): split_types() result = True
is_ac_stopped():  value of gs = 164.0, len(gs) = 5
is_ac_stopped():  value of v_gs = 164
loop(): handling msg nr: 01
lcd_pr_msgs(): Duration rx -> lcd: 241 mSecs
loop(): gc.mem_free()= 8148128
Waiting for serial com line to become available...
End of loop  1
........................

Note: because the 'terminalio.font' doesn't have a glyph for the 'degrees' character I used the caret '^' instead,
until I have added a font that can display the degrees character.

For some other microcontroller i have ports of the same project on Github.

PROBLEM SOLVED

At startup :
if flag 'power_lost' is True:
the external DS3231 is set with pre-determined datetime (adjust by hand)
next the built-in RTC is started with the datetime from the DS3231

if the flag 'power_lost is False:
the built-in RTC is started with the datetime from the DS3231

During the main loop displaying the datetime from the built-in RTC.

I put a short video in the images folder inside the dropbox folder (see the link in my post above).