Re: [PATCH v2] iio: adc: ina2xx: Make calibration register value fixed

["Brüns, Stefan" <Stefan.Bruens@rwth-aachen.de>]

On Donnerstag, 7. Dezember 2017 13:02:47 CET Maciej Purski wrote:
> Calibration register is used for calculating current register in
> hardware according to datasheet:
> current = shunt_volt * calib_register / 2048 (ina 226)
> current = shunt_volt * calib_register / 4096 (ina 219)
> 
> Fix calib_register value to 2048 for ina226 and 4096 for ina 219 in
> order to avoid truncation error and provide best precision allowed
> by shunt_voltage measurement. Make current scale value follow changes
> of shunt_resistor from sysfs as calib_register value is now fixed.
> 
> Power_lsb value should also follow shunt_resistor changes as stated in
> datasheet:
> power_lsb = 25 * current_lsb (ina 226)
> power_lsb = 20 * current_lsb (ina 219)
> 
> This is a part of the patchset: https://lkml.org/lkml/2017/11/22/394
> 
> Signed-off-by: Maciej Purski <m.purski@samsung.com>
> 
> ---
> Changes in v2:
> - rebase on top of the latest next
> - remove a redundant variable - power_lsb_uW
> - fix comments
> ---
>  drivers/iio/adc/ina2xx-adc.c | 58
> +++++++++++++++++++++++--------------------- 1 file changed, 31
> insertions(+), 27 deletions(-)
> 
> diff --git a/drivers/iio/adc/ina2xx-adc.c b/drivers/iio/adc/ina2xx-adc.c
> index ddf8781..3e4972f 100644
> --- a/drivers/iio/adc/ina2xx-adc.c
> +++ b/drivers/iio/adc/ina2xx-adc.c
> @@ -124,11 +124,11 @@ enum ina2xx_ids { ina219, ina226 };
> 
>  struct ina2xx_config {
>  	u16 config_default;
> -	int calibration_factor;
> +	int calibration_value;
>  	int shunt_voltage_lsb;	/* nV */
>  	int bus_voltage_shift;	/* position of lsb */
>  	int bus_voltage_lsb;	/* uV */
> -	int power_lsb;		/* uW */
> +	int power_lsb_factor;
>  	enum ina2xx_ids chip_id;
>  };
> 
> @@ -138,6 +138,7 @@ struct ina2xx_chip_info {
>  	const struct ina2xx_config *config;
>  	struct mutex state_lock;
>  	unsigned int shunt_resistor_uohm;
> +	unsigned int current_lsb_uA;

I don't think you need current_lsb_uA, see below ...

>  	int avg;
>  	int int_time_vbus; /* Bus voltage integration time uS */
>  	int int_time_vshunt; /* Shunt voltage integration time uS */
> @@ -149,20 +150,20 @@ struct ina2xx_chip_info {
>  static const struct ina2xx_config ina2xx_config[] = {
>  	[ina219] = {
>  		.config_default = INA219_CONFIG_DEFAULT,
> -		.calibration_factor = 40960000,
> +		.calibration_value = 4096,
>  		.shunt_voltage_lsb = 10000,
>  		.bus_voltage_shift = INA219_BUS_VOLTAGE_SHIFT,
>  		.bus_voltage_lsb = 4000,
> -		.power_lsb = 20000,
> +		.power_lsb_factor = 20,
>  		.chip_id = ina219,
>  	},
>  	[ina226] = {
>  		.config_default = INA226_CONFIG_DEFAULT,
> -		.calibration_factor = 5120000,
> +		.calibration_value = 2048,
>  		.shunt_voltage_lsb = 2500,
>  		.bus_voltage_shift = 0,
>  		.bus_voltage_lsb = 1250,
> -		.power_lsb = 25000,
> +		.power_lsb_factor = 25,
>  		.chip_id = ina226,
>  	},
>  };
> @@ -229,14 +230,16 @@ static int ina2xx_read_raw(struct iio_dev *indio_dev,
> 
>  		case INA2XX_POWER:
>  			/* processed (mW) = raw*lsb (uW) / 1000 */
> -			*val = chip->config->power_lsb;
> +			*val = chip->config->power_lsb_factor *
> +			       chip->current_lsb_uA;
>  			*val2 = 1000;
>  			return IIO_VAL_FRACTIONAL;
> 
>  		case INA2XX_CURRENT:
> -			/* processed (mA) = raw (mA) */
> -			*val = 1;
> -			return IIO_VAL_INT;
> +			/* processed (mA) = raw*lsb (uA) / 1000 */
> +			*val = chip->current_lsb_uA;
> +			*val2 = 1000;

.. this is the only place where current_lsb_uA is used ...

> +			return IIO_VAL_FRACTIONAL;
>  		}
> 
>  	case IIO_CHAN_INFO_HARDWAREGAIN:
> @@ -541,28 +544,34 @@ static ssize_t ina2xx_allow_async_readout_store(struct
> device *dev, }
> 
>  /*
> - * Set current LSB to 1mA, shunt is in uOhms
> - * (equation 13 in datasheet). We hardcode a Current_LSB
> - * of 1.0 x10-3. The only remaining parameter is RShunt.
> - * There is no need to expose the CALIBRATION register
> - * to the user for now. But we need to reset this register
> - * if the user updates RShunt after driver init, e.g upon
> - * reading an EEPROM/Probe-type value.
> + * Calibration register is set to the best value, which eliminates
> + * truncation errors on calculating current register in hardware.
> + * According to datasheet (INA 226: eq. 3, INA219: eq. 4) the best values
> + * are 2048 for ina226 and 4096 for ina219. They are hardcoded as
> + * calibration_value.
>   */
>  static int ina2xx_set_calibration(struct ina2xx_chip_info *chip)
>  {
> -	u16 regval = DIV_ROUND_CLOSEST(chip->config->calibration_factor,
> -				   chip->shunt_resistor_uohm);
> -
> -	return regmap_write(chip->regmap, INA2XX_CALIBRATION, regval);
> +	return regmap_write(chip->regmap, INA2XX_CALIBRATION,
> +			    chip->config->calibration_value);
>  }
> 
> +/*
> + * In order to keep calibration register value fixed, the product
> + * of current_lsb and shunt_resistor should also be fixed and equal
> + * to shunt_voltage_lsb = 1 / shunt_div multiplied by 10^9 in order
> + * to keep the scale.
> + */
>  static int set_shunt_resistor(struct ina2xx_chip_info *chip, unsigned int
> val) {
> -	if (val <= 0 || val > chip->config->calibration_factor)
> +	unsigned int dividend = DIV_ROUND_CLOSEST(1000000000,
> +						  chip->config->shunt_div);
> +
> +	if (val <= 0 || val > dividend)
>  		return -EINVAL;
> 
>  	chip->shunt_resistor_uohm = val;
> +	chip->current_lsb_uA = DIV_ROUND_CLOSEST(dividend, val);

1st: shunt_div no longer exists, but shunt_voltage_lsb[nV], so this code does 
not even compile ...

This function does two things - range check and calculation of the 
curren_lsb_uA auxiliary variable.

Lets start with the second part:

read_raw returns the IIO_VAL_FRACTIONAL
val / val2     = current_lsb_uA / 1000             (1)


from set_shunt_resistor(...):

dividend       = 1e9 / shunt_div	                  (2)
current_lsb_uA = dividend / shunt_resistor_uohm    (3)

shunt_div no longer exists, but can be substituted:
shunt_div      = 1e6 / shunt_voltage_lsb           (4)

substituting (2), (3) and (4) in (1), we end up with

val / val2     = 1e9 / (shunt_div * shunt_resistor_uohm * 1000)

=> val / val2  = 1e6 / ((1e6 / shunt_voltage_lsb) * shunt_resistor_uohm)
=> val / val2  = shunt_voltage_lsb / shunt_resistor_uohm
=>  [mA]       =     [nV]          /     [uOhm]

The auxiliary variable is gone, as are the two DIV_ROUND_CLOSEST calls.


Part 2, range check:

lower bound is obviously 1 uOhm. The upper bound is only limited by
typeof(val2) and typeof(shunt_resistor_uohm), both are ints.

so the range check can be simplified to

if (val == 0 || val > INT_MAX)
	return -EINVAL

Kind regards,

Stefan