nRF5340-DK & CCS811 sensor, doesn't reach main()

[raul7]

Hi,

I'm working with the CCS811 sensor example from the `v1.5.1/zephyr/samples/sensor` folder and I managed to compile it successfully after modifying the files in the `ccs811/boards` directory. When I debug the solution, the debugger enters in continued execution mode and never reaches the main() function. Every time I hit  "Pause", the code always pauses in the same place, as it is shown in the following screen capture:

I'm using the nRF Connect SDK V1.5.1 (tried with V1.5.0 also). In the `ccs811/boards` folder I have two files:

nrf5340dk_nrf5340_cpuapp.overlay:

&i2c1 {
        status = "okay";
        sda-pin = <0>;
        scl-pin = <1>;

	/* Sparkfun Environment Combo uses second I2C address */
        ccs811: ccs811@5b {
                compatible = "ams,ccs811";
                reg = <0x5b>;
                label = "CCS811";
                irq-gpios = <&gpio0 2 GPIO_ACTIVE_LOW>;
                wake-gpios = <&gpio0 5 GPIO_ACTIVE_LOW>;
                reset-gpios = <&gpio0 6 GPIO_ACTIVE_LOW>;
        };
};

And nrf5340dk_nrf5340_cpuapp.conf:

CONFIG_STDOUT_CONSOLE=y
CONFIG_SENSOR=y
CONFIG_CCS811=y
#CONFIG_CCS811_DRIVE_MODE_2=y
#CONFIG_CCS811_TRIGGER_GLOBAL_THREAD=y
#CONFIG_CCS811_TRIGGER_OWN_THREAD=y

Any hints about what could be the issue? Thanks in advance!

[Simon]

All the drivers will get initialized before main, so maybe a fault happened in one of the initialization fuctions, you could try to set some break points in ccs811_init() and see if that causes it to fail.

Maybe a hard fault happenend, could you try to set these in your prj.conf:

CONFIG_RESET_ON_FATAL_ERROR=n
CONFIG_LOG=y

Then you should be able to see more information about the hard fault. Open a serial terminal like Termite to see the log

This section might help as well:

NCS Tutorial-->Error handling-->Other tools and tips-->Hard faults

[raul7]

Hi Simon, thanks for your help!

Evidently, the execution hangs inside cs811_init(), concretely in the k_msleep() function:

#if DT_INST_NODE_HAS_PROP(0, wake_gpios)
	drv_data->wake_gpio = device_get_binding(DT_INST_GPIO_LABEL(0, wake_gpios));
	if (drv_data->wake_gpio == NULL) {
		LOG_ERR("Failed to get pointer to WAKE device: %s",
			DT_INST_GPIO_LABEL(0, wake_gpios));
		return -EINVAL;
	}

	/*
	 * Wakeup pin should be pulled low before initiating
	 * any I2C transfer.  If it has been tied to GND by
	 * default, skip this part.
	 */
	gpio_pin_configure(drv_data->wake_gpio, WAKE_PIN,
			   GPIO_OUTPUT_INACTIVE
			   | DT_INST_GPIO_FLAGS(0, wake_gpios));

	set_wake(drv_data, true);
	k_msleep(1);
#endif

If I comment k_msleep(1); the code goes on forward, but it gets stuck again in any subsequent k_msleep() instance. If I comment out every instance of k_msleep(), the code reaches the main() function.

Could it be an issue with the clock configuration? Any hint about what could be the cause?

I didn't get any hard fault errors in the serial terminal.

Thanks in advance!

EDIT:

Nevermind, I changed the following pins in my nrf5340dk_nrf5340_cpuapp.overlay:

sda-pin = <34>; // P1.02 (34)
scl-pin = <35>; // P1.03 (35)

irq-gpios = <&gpio0 36 GPIO_ACTIVE_LOW>; // P1.04 (36)

And now it runs fine without hanging in the k_msleep() instances. Perhaps, the sensor wasn't responding because of a conflict with the pins I was trying to use and that was causing the processor to hang in the css811_init() function? Not sure though...

Thanks!

[Simon]

Sorry for not seeing that initally. P0.00 and P0.01 are already occupied:  

https://infocenter.nordicsemi.com/index.jsp?topic=%2Fug_nrf5340_dk%2FUG%2Fdk%2Fconnector_if.html&cp=3_0_4_3_5

[raul7]

Now it does make sense, thanks Simon!