Neo Basic M4 I2C detect

The below code to detect I2C devices (MCP3424 ADC) on arduino like micro-controllers works perfectly on an Arduino Uno R3 and a Teensy 3.2, but when run on the M4 shows that all addresses have a device connected. Why doesn't this work on the Neo? Do I need to do some special setup on the Neo (hardware and/or software)? I can successfully read the MCP3424 from the M4 when connected to I2C2 on the Neo.

Thanks in advance.

The code is from Steve Marple's github MCP342X library.

Code:

#include <Wire.h>

void setup(void)
{
  Serial.begin(115200);
  Wire.begin();
  Serial.println("The MCP3422 and MCP3426 use I2C address 0x68, all other devices can be");
  Serial.println("configured to use any address in the range 0x68 - 0x6F (inclusive).");
  Serial.println("Be aware that the DS1307 uses address 0x68.");
  Serial.println();
  for (uint8_t add = 0X68; add < 0X6F; add++) {
    //Serial.print("Trying ");
    //Serial.println(add);
    Wire.requestFrom(add, (uint8_t)1);
    if (Wire.available()) {
      Serial.print("Found device at: 0x");
      Serial.println(add, HEX);
    }
  }
  Serial.println("Done");
}

void loop(void)
{
  ;
}

 

No, Wire1 did not work.

 

The Neo Doc under Arduino M4 Processor, Arduino differences says:

UDOO Neo M4 Processor has 2 I2C buses wired to the pinout headers and consequentially have 2 Wire objects:

• Wire: on Brick Connector
• Wire1: on SCL SDA pins

Be careful and double check when calling Wire.begin() or Wire1.begin() in sketches or libraries. Also check out to have I2C-2 disabled on A9 if you use Wire.

So to use the I2C on the M4 I have to:
- Remove I2C-2 from the bricks connector using the Device Tree Editor (the linux side will not have access to I2C-2)
- Reboot
- If I use SCL and SDA on the Arduino header then I should always use Wire1.
- If I want to access the bricks connector from the M4 then I use Wire.

Is this correct?
Can someone post example code for reading an I2C device from the M4?

Thanks