Problems with I2C & mma8652fc

[Martin K]

Hello,

we developed a custom board using the nRF52832. We are having problems communicating with the mma8652fc accelerometer using I2C. We also have a Rigado BMD-200 EVAL board (based on nRF51822), which has the same accelerometer on. With the Rigado board we can successfully communicate with the accelerometer, and are having no problems whatsoever.

Our custom board was soldered by a professional Pick and place company, so the pins are probably soldered ok. We have 6 boards and we are having problems with all of them, so it may be a problem on our side. I attached the accelerometer schematic at the bottom.

Did anyone had similar problems, or anyone has some pointers to guide is in the right direction? :)

I can provide more details if required.

Thanks!

[Michael Antwih]

Do you have an logic analyzer? Do you know if the bits are making it out to the bus? Common grounds? I know that is a silly question but I noticed you have GNDA instead of just GND.

[Michael Antwih]

Also when you step thru the code and you throwing any errors? In debugger mode you used be able to review the TWI/I2C register with either Keil or gdb.

[Martin K]

Sadly we don't have a logic analyzer. GNDA is just common GND, it's just marked this way in the schematic.

The debugger points us to some weird areas of code that are not being executed at all. I can post some code if it would help.

[Michael Antwih]

Make sure optimization is turned off and then set break point on the error handler function. Which IDE are you using?

[Ole Bauck]

Go buy yourself a logic analyzer or oscilloscope, or borrow some. If you are developing hardware this is mandatory to have in your inventory.

Determining if the bits are actually clocked out on the lines would help finding the issue.

As Michael said you can also go into the debugger and check what happens. It might be that a function is returning an error code and the device is resetting, see here.