M5Stack Paper I2C Doesn't work
-
I am trying to use the Paper and I can't get any of the ports to work. The code I am using in below. One strange thing in this code is if I comment out M5.begin() I see the internal i2C devices but if I include it I don't see any devices.
#include <M5EPD.h>
// I2C (internal) - SDA: 21, SCL: 22
// I2C (external - Port A) - SDA: 25, SCL: 32#define PortA_SDA 25 // Port A
#define PortA_SCL 32#define PortB_SDA 26 // Port B
#define PortB_SCL 33#define PortC_SDA 18 // Port C
#define PortC_SCL 19#define INTERNAL_SDA 21
#define INTERNAL_SCL 22void setup()
{
// M5.begin();
Serial.begin(115200);Wire.begin(INTERNAL_SDA, INTERNAL_SCL);
}void loop()
{
int address;
int error;/********** Internal ********/
Serial.println("Scanning internal I2C");
for(address = 1; address < 127; address++)
{
Wire.beginTransmission(address);
error = Wire.endTransmission();
if(error == 0)
{
Serial.printf("Internal Device Found %#04x\n",address);
}
delay(10);
}
Serial.println("Done Scanning internal I2C");
delay(500);/********** PortA ********/
Serial.println("Scanning external I2C (PortA)");
Wire1.begin(PortA_SDA, PortA_SCL);
for(address = 1; address < 127; address++)
{
Wire1.beginTransmission(address);
error = Wire1.endTransmission();
if(error == 0)
{
Serial.printf("PORT-A Device Found %#04x\n",address);
} else Serial.print("."); //Serial.printf("No Device Found PORT-A %#04x\n",address);
delay(10);
}
Serial.println("Done Scanning PortA I2C");
delay(500);/********** PortB ********/
Serial.println("Scanning external I2C (PortB)");
// M5.begin();
Wire1.begin(PortB_SDA, PortB_SCL);
for(address = 1; address < 127; address++)
{
Wire1.beginTransmission(address);
error = Wire1.endTransmission();
if(error == 0)
{
Serial.printf("PORT-B Device Found %#04x\n",address);
} else Serial.print("."); //Serial.printf("No Device Found PORT-B %#04x\n",address);
delay(10);
}
Serial.println("Done Scanning PortB I2C");
delay(500);/********** PortC ********/
Serial.println("Scanning external I2C (PortC)");
// M5.begin();
Wire1.begin(PortC_SDA, PortC_SCL);
for(address = 1; address < 127; address++)
{
Wire1.beginTransmission(address);
error = Wire1.endTransmission();
if(error == 0)
{
Serial.printf("PORT-C Device Found %#04x\n",address);
} else Serial.print("."); //Serial.printf("No Device Found PORT-C %#04x\n",address);
delay(10);
}
delay(5000);
} -
Hello @ispybadguys
calling
M5.begin()
already takes care ofSerial
andWire
. I suggest you only callM5.begin()
(which also initialises other M5Paper hardware) insetup()
.As for the I2C scanning on the three ports. The latest ESP32 Arduino framework doesn't allow to just call
Wire1.begin()
repeatedly with different GPIOs for SDA and SCL. If you look at the debug output you'll see this warning[ 41263][W][Wire.cpp:204] begin(): Bus already started in Master Mode.
every time you try that in theloop()
.So when the scanning is done for one port you need to call
Wire1.end()
before callingWire1.begin()
again.Thanks
Felix -
Thanks. I'll try that. When I call M5.begin(); I get the following behavior.
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OK
M5EPD initializing...OKcontinues forever.
-
If I comment out the internal I2C section so I only have the PORT-A scan I get the following:
Scanning external I2C (PortA)
..............................................................................................................................Done Scanning PortA I2CIt takes about one second for each address scanned. This is probably the default timeout.
Kurt
-
@ispybadguys What external devices have you tried connecting?
-
Hello @ispybadguys
are you saying that simply calling
M5.begin()
insetup()
, e.g.#include <M5EPD.h> void setup() { M5.begin(); } void loop() { }
produces
M5EPD initializing...OK
over and over again?Thanks
Felix -
Hello @ispybadguys
please find the I2C internal and all port scanner (which works for me) here.
Thanks
Felix -
Yes. I tested the very code you have here with just to include and M5.begin() and I get M5EPD initializing... followed by a 4 second pause and then the OK
-
Felix
Thanks for the code . For scanning the I2C. The code produces the following output:
M5EPD initializing...OK
I2C Scan - internal
.M5EPD initializing...OK
I2C Scan - internal
.M5EPD initializing...OK
I2C Scan - internal
.M5EPD initializing...OK -
Hello @ispybadguys
that is very weird - I am stumped.
Have you tried to only scan the ports by commenting these two lines:
// Serial.println("I2C Scan - internal"); // scanI2C(i2cBus_Internal.wire, i2cBus_Internal.sda, i2cBus_Internal.scl);
BTW: I also get a 4 seconds delay before
OK
is printed out - so that should be fine.Thanks
Felix -
Felix
Yes. I just tried that. I get the same behavior. I moved the script to a new Windows computer and installed the IDE there. I get
13:10:21.290 -> I2C Scan - Port A
13:10:21.290 -> .M5EPD initializing...OK
13:10:25.790 -> I2C Scan - Port AI can actually scan the internal ports correctly if I don't call M5.begin(); and I just do a Serial.begin() and the wire commands.
BTW the boards.txt has an error in it for the M5-Paper and you cannot upload code from the Arduino IDE. I fixed that and I can upload ok now.
Kurt
-
Hello @ispybadguys
the debug output suggests that as soon as the I2C port scan starts (first dot) your M5Paper seems to reboot.
- Have you tried a different / shorter USB cable? Maybe the issue is power related?
- Does the touch example (requires internal I2C) work?
Thanks
Felix -
Yes. That looks to be the case that it crashed after one or two Wire1 commands. Usually one period is printed but sometimes two. Also the internal I2C does not work is I call M5.begin but it does if I just call Wire.begin.
-
I'm having exactly the same issue. Have you figured out what causes it?
-
@Koiru I have not found a solution but I have found a workaround.
The culprit creating the constant reboots are the touchscreen, if I call the begin function without enabling the touchscreen I can get the unit to work instead of doing constant reboots.
To initialize the unit without touchscreen, use the following arguments:
M5.begin(false, true, true, true, true);