tuupola / micropython-mpu9250

MicroPython I2C driver for MPU9250 9-axis motion tracking device
MIT License
145 stars 46 forks source link

[Errno 19] ENODEV #13

Closed pidou46 closed 4 years ago

pidou46 commented 5 years ago

Hello,

I'm running into trouble trying "the simple test" code

My mcu : MicroPython v1.11-312-g22099ab88 on 2019-09-15; ESP32 module (spiram) with ESP32

Error: import test_mpu Traceback (most recent call last): File "<stdin>", line 1, in <module> File "test_mpu.py", line 7, in <module> File "mpu6500.py", line 97, in __init__ File "mpu6500.py", line 139, in whoami File "mpu6500.py", line 155, in _register_char OSError: [Errno 19] ENODEV

pidou46 commented 5 years ago

When trying to scan i2c bus I don't see anything

`>>> from machine import Pin,I2C

i2c=I2C(scl=Pin(22),sda=Pin(21)) i2c.scan() [] `

I have checked with a VL53L0X and I see it when I scan the bus:

`>>> from machine import Pin,I2C

i2c=I2C(scl=Pin(22),sda=Pin(21)) i2c.scan() [41]`

It sound like I have a bad hardware...

IHOXOHI commented 4 years ago

I have the same error (OSError: [Errno 19] ENODEV ) with a pyboard and lcd160, both from original manufacturing... For me, it looks like a problem with the library...

tuupola commented 4 years ago

Library works just fine. Make sure you have correct pins. If i2c.scan() does not find anything it means there is no properly connected i2c devices on those pins.

jasoncook-fr commented 4 years ago

Honestly I think Tuupola has got much more interesting things to deal with in this project (like working on the gyro calibration). @pidou64 take a closer look at the error message. If you are getting a result from the i2c.scan() then you can move on to the next step. Make sure that the correct address for your device is stated in the file mpu6500.py, line 88. The hexadecimal address here is 0x68 by default. My mpu9250 sensor address is decimal 105, or hexidecimal 0x69. I changed it in the code and all works now. If your sensor's address is 41 decimal, then you might try changing the address in the code to 0x29

IHOXOHI commented 4 years ago

Hi, Thanks for your help. I have found the origin of the problem which come from an other connection of dattalogging on a serial port. No problem without this connection. Sorry for alls the desagrements. This Library works fine with pyboard, arduino, ... I enjoy it. All the best, Have a fun.

elingeniero3 commented 4 years ago

Hi mates, I tried the example test code:

import utime
from machine import I2C, Pin
from mpu9250 import MPU9250

i2c = I2C(scl=Pin(22), sda=Pin(21))
sensor = MPU9250(i2c)

print("MPU9250 id: " + hex(sensor.whoami))

while True:
    print(sensor.acceleration)
    print(sensor.gyro)
    print(sensor.magnetic)
    print(sensor.temperature)

    utime.sleep_ms(1000)

but I have the sames issues:

>>> %Run -c $EDITOR_CONTENT Traceback (most recent call last): File "<stdin>", line 7, in <module> File "mpu9250.py", line 45, in __init__ File "mpu6500.py", line 97, in __init__ File "mpu6500.py", line 149, in whoami File "mpu6500.py", line 181, in _register_char OSError: [Errno 19] ENODEV

My mcu is:

MicroPython v1.13 on 2020-09-02; ESP32 module with ESP32

I cheked that the pins are correctly connected. Someone could help me please?

elingeniero3 commented 4 years ago

I've just solved the issue, I didn't correct right the device to the sensor

APMonitor commented 2 years ago

The pyboard PYBv1.1 internal accelerometer creates a conflict with the mpu9250. With pyb_Acc=True, there is the OSError 19: ENODEV. When pyb_Acc=False, there is no error.

[12, 76, 104]
Traceback (most recent call last):
  File "<stdin>", line 15, in <module>
  File "mpu9250.py", line 94, in whoami
  File "mpu6500.py", line 149, in whoami
  File "mpu6500.py", line 181, in _register_char
OSError: [Errno 19] ENODEV

Here is the code:

import utime
from machine import SoftI2C, Pin
from mpu9250 import MPU9250
import pyb

pyb_Acc = False

i2c = SoftI2C(scl=Pin('X9'), sda=Pin('X10'))
print(i2c.scan())
s = MPU9250(i2c)

print("MPU9250 id: " + hex(s.whoami))

if pyb_Acc:
    a = pyb.Accel()

while True:
    t = pyb.millis()  # get time
    print(s.acceleration)
    print(s.gyro)
    print(s.magnetic)
    print(s.temperature)
    if pyb_Acc:
        x, y, z = a.filtered_xyz()  # get acceleration data

    utime.sleep_ms(10)

I would like to be able to read both the MPU9250 and Internal accelerometer sensors. Any ideas on a way to avoid the conflict in address?

tuupola commented 2 years ago

How do you know there is a conflict? Does the internal accelerometer use the same address as the attachhed MPU9250?

APMonitor commented 2 years ago

The i2c.scan() command gives [12, 76, 104] for pins X9 and X10. Whenpyb_Acc = True in the above script, there is an error that the MPU9250 is not found with code ENODEV.

Line 181 is from the line self.i2c.readfrom_mem_into(self.address, register, buf) of mpu6500.py

    def _register_char(self, register, value=None, buf=bytearray(1)):
        if value is None:
            self.i2c.readfrom_mem_into(self.address, register, buf)
            return buf[0]

That is called from line 149 of mpu6500.py.

    @property
    def whoami(self):
        """ Value of the whoami register. """
        return self._register_char(_WHO_AM_I)

The _WHO_AM_I constant is 0x75 (hex) = 117 (decimal). I'm not sure why this register read is having trouble when the pyBoard accelerometer is also reading.

The micropython documentation for pyb.Accel() has a hardware note:

The accelerometer uses I2C bus 1 to communicate with the processor. Consequently when readings are being taken pins X9 and X10 should be unused (other than for I2C). Other devices using those pins, and which therefore cannot be used concurrently, are UART 1 and Timer 4 channels 1 and 2.

The MPU9250 is I2C so this shouldn't create a conflict unless both accelerometers have the same address. I'll keep digging into it to see if there is something in the pyb.Accel() function that is creating the conflict.

tuupola commented 2 years ago

Ok so i2c.scan() can see both MPU6500 (0x68) and AK8963 (0x0c). The internal MMA7660 is the 0x4c. However communicating with MPU6500 fails with ENODEV. This error means the device does not respond. You would get the same error if you remove the MPU9250 from the circuit.

Is the i2c.scan() result the same with both pyb_Acc = True and pyb_Acc = False?

APMonitor commented 2 years ago

The i2c.scan() is the same with both pyb_Acc (or mma7660) as True and False. The only error is when mma7660 and mpu9250 are both True.

import utime
from machine import SoftI2C, Pin
from mpu9250 import MPU9250
import pyb

mma7660 = True
mpu9250 = True

i2c = SoftI2C(scl=Pin('X9'), sda=Pin('X10'))
print(i2c.scan())

if mpu9250:
    s = MPU9250(i2c)
    print("MPU9250 id: " + hex(s.whoami))

if mma7660:
    a = pyb.Accel()

while True:
    t = pyb.millis()  # get time
    if mpu9250:
        print(s.acceleration)
        print(s.gyro)
        print(s.magnetic)
        print(s.temperature)
    if mma7660:
        x, y, z = a.filtered_xyz()  # get acceleration data
        print(x,y,z)
    utime.sleep_ms(10)

Here are the i2c addresses:

When the MPU9250 is removed, only the (0x4c) 76 address remains.

sriwahyni commented 2 years ago

i2c.scan() [104]

why?

APMonitor commented 2 years ago

@sriwahyni I'm not sure about the context of your question. When I have the MPU9250 connected, the result of i2c.scan() is [12,76,104] so it is seeing all of the addresses. The error occurs when mma7660 and mpu9250 are both True so it appears to be a software issue.