[TW#14489] i2c errors when using adc2

[jmorsap]

I'm new to esp32/esp-idf and am moving a large embedded project from an atmega328 in the arduino environment to an esp32 wroom32 using esp-idf.

I'm aware that adc2 is not supported and plan on eliminating it in the next design iteration, but I needed the functionality for our current design and so added the ADC2 driver according to ESP32DE's procedure found here #461

The adc2 works fine in its own program, as does I2C.

When I configure the two settings below, the I2C ceases to work and will not work again until a hardware reset.

    adc2_config_width(ADC_WIDTH_12Bit);
    adc2_config_channel_atten(ADC2_TEST_CHANNEL,ADC_ATTEN_11db);

Is there any way to undo these two lines of code or reset these registers after I finish using adc2 so that I can use the I2C again?

[CarlosGS]

Hi! I've used I2C + ADC2 without a problem (apart from espressif/arduino-esp32#349 ). What pins are you using for I2C?

…

On Tue, Aug 1, 2017 at 5:40 PM, jmorsap ***@***.***> wrote: I'm new to esp32/esp-idf and am moving a large embedded project from an atmega328 in the arduino environment to an esp32 wroom32 using esp-idf. I'm aware that adc2 is not supported and plan on eliminating it in the next design iteration, but I needed the functionality for our current design and so added the ADC2 driver according to ESP32DE's procedure found here #461 <#461> The adc2 works fine in its own program, as does I2C. When I configure the two settings below, the I2C ceases to work and will not work again until a hardware reset. adc2_config_width(ADC_WIDTH_12Bit); adc2_config_channel_atten(ADC2_TEST_CHANNEL,ADC_ATTEN_11db); Is there any way to undo these two lines of code or reset these registers after I finish using adc2 so that I can use the I2C again? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#869>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABj3qrzKd2pLLjQl952ztfOXLpGBJI8Uks5sT8WRgaJpZM4OqdqK> .

[jmorsap]

Hello CarlosGS, Thanks for the reply

I2C does not work with the following config

#define I2C_EXTERNAL_MASTER_SCL_IO    15    /*!< gpio number for I2C master clock */

#define I2C_EXTERNAL_MASTER_SDA_IO    4    /*!< gpio number for I2C master data  */

#define I2C_EXTERNAL_MASTER_NUM I2C_NUM_0   /*!< I2C port number for master dev */

#define I2C_EXTERNAL_MASTER_TX_BUF_DISABLE   0   /*!< I2C master do not need buffer */

#define I2C_EXTERNAL_MASTER_RX_BUF_DISABLE   0   /*!< I2C master do not need buffer */

#define I2C_EXTERNAL_MASTER_FREQ_HZ    100000     /*!< I2C master clock frequency */

Though it does appear to work on GPIO 27 and 14.

I have also found that

` adc2_config_channel_atten(ADC2_TEST_CHANNEL,ADC_ATTEN_11db);

`

is the problematic function, but have not been successful at pinpointing the exact piece of code in this function that is preventing me from using I2C on pins 15 and 4. I suspect that it is something in the rtc_module.c file.

I am using adc2_9.

I2C is initialized like so:

static void i2c_external_master_init()
{

	int i2c_master_port = I2C_EXTERNAL_MASTER_NUM; //here

	i2c_config_t conf;

	conf.mode = I2C_MODE_MASTER;

	conf.sda_io_num = I2C_EXTERNAL_MASTER_SDA_IO; //here

	conf.sda_pullup_en = GPIO_PULLUP_ENABLE;

	conf.scl_io_num = I2C_EXTERNAL_MASTER_SCL_IO; //here

	conf.scl_pullup_en = GPIO_PULLUP_ENABLE;

	conf.master.clk_speed = I2C_EXTERNAL_MASTER_FREQ_HZ;

	i2c_param_config(i2c_master_port, &conf); //here

	i2c_driver_install(i2c_master_port, conf.mode, //here

	I2C_EXTERNAL_MASTER_RX_BUF_DISABLE,

	I2C_EXTERNAL_MASTER_TX_BUF_DISABLE, 0);

}

Below is an example of how I am using I2C

static esp_err_t readhumidity(i2c_port_t i2c_num, uint8_t* data_h, uint8_t* data_l, float* rh)  //note, data_h and data_l are pointers to where the data is stored
{
	i2c_cmd_handle_t cmd = i2c_cmd_link_create(); //create 'instance' of i2c


	i2c_master_start(cmd);
	i2c_master_write_byte(cmd, SI7013_SENSOR_ADDR << 1 | WRITE_BIT, ACK_CHECK_EN);  //write two bytes to si7013 //note in axample this was taken from, the right most bit indicated a write or read command
	i2c_master_write_byte(cmd, SI7013_CMD_HSTART, ACK_CHECK_EN);
	i2c_master_stop(cmd);
	
	
	int ret = i2c_master_cmd_begin(i2c_num, cmd, 1000 / portTICK_RATE_MS); //this task sends all ques commands, the task will be clocked until all commands ahve been sent out. This is not thread safe, 
	
	i2c_cmd_link_delete(cmd); //delete link to free resources
	
	
	if (ret == ESP_FAIL) {  //if timeout (or some other error occurs, return a "fail"
		return ret;
	}
	
	
	vTaskDelay(30 / portTICK_RATE_MS); //wait 30ms. 
	

	cmd = i2c_cmd_link_create(); //create 'instance of i2c
	i2c_master_start(cmd);
	
	
	i2c_master_write_byte(cmd, SI7013_SENSOR_ADDR << 1 | READ_BIT, ACK_CHECK_EN);  //write a byte //note in axample this was taken from, the right most bit indicated a write or read command
	i2c_master_read_byte(cmd, data_h, ACK_VAL); //read two bytes
	i2c_master_read_byte(cmd, data_l, NACK_VAL);
	
	
	i2c_master_stop(cmd);
	ret = i2c_master_cmd_begin(i2c_num, cmd, 1000 / portTICK_RATE_MS);//sends all qued commands
	i2c_cmd_link_delete(cmd); //free resources back up

	unsigned int rawHumidity = (*data_h << 8) | *data_l;
	rawHumidity &= 0xFFFC; //Zero out the status bits but keep them in place
	//Given the raw humidity data, calculate the actual relative humidity
	float tempRH = rawHumidity / (float)65536; //2^16 = 65536
	*rh = -6 + (125 * tempRH); //From page 14
	
	if (ret == ESP_FAIL) {  //return fail if failed
		return ESP_FAIL;
	}
	return ESP_OK; //otherwise return "ESP_OK". data_h, data_l are stored in the appropriate locations (different than how I usually do these fxns)
}

Everything works well the first time, but as soon as I initialize ADC2, my I2C functions on pins 15 and 4 begin to store values of zero in the data but do not return any errors.

[negativekelvin]

ADC2_TEST_CHANNEL = ADC2_CHANNEL_9 ?

[jmorsap]

yes,

#define ADC2_TEST_CHANNEL (9)

[negativekelvin]

I think you should delete these 3 lines https://github.com/espressif/esp-idf/pull/466/files#diff-9934b2bd0fdda201ccf125abd90da487R662

[FayeY]

Hi @jmorsap , did the above suggestion help you resolve the issue?

[jmorsap]

Hey I apologize for not responding. We cut a trace on our boards and wired the signal to an ADC1 IO. Our new boards don't use ADC2 for anything, so I haven't tried this out yet. I'll let you know if I get a chance. Has anyone else tried this yet?