BME280 relative altitude not stable

[Myzhar]

Hi all,
I'm using a BME280 sensor to estimate the realtive altitude of an object, but the results are not really stable.

This is the formula I use:

float heightOutput_mm = (44330.77f * 1000.0f) * (1.f - pow((p1 / p0), 0.190294957f));

Where:

• p1 -> current pressure
• p2 -> reference pressure taken at ground level

The problem is that keeping the object STATIC at ground level I get a very highly changing relative altitude in the range [-0.5,+0.5] m

The sensor is configured in as suggested in the datasheet for INDOOR NAVIGATION:

	bme280_dev.settings.osr_h = BME280_OVERSAMPLING_1X;
	bme280_dev.settings.osr_p = BME280_OVERSAMPLING_16X;
	bme280_dev.settings.osr_t = BME280_OVERSAMPLING_2X;
	bme280_dev.settings.filter = BME280_FILTER_COEFF_16;
	bme280_dev.settings.standby_time = BME280_STANDBY_TIME_0_5_MS;

Any suggestion to inprove the stability of the readings?

Thank you
Walter
 

[shellywang]

The relative accuracy for the pressure part of the BME280 is 12pa -> 96cm. The highly changing relative altitude in the range [-0.5,+0.5] m is acceptable.

Could you please provide the ground truth and the height value you measured?

[Myzhar]

My ground truth is "zero" since I'm not measuring the altitude relatively to the sea level, but to the first value I get from the barometer (the sensor is steady).

This is the formula that I use to estimate the relative altitude:

float rel_height_m = ((pow((p0 / p1), 1.f/5.257f) - 1.f) * (temp + 273.15f)) / 0.0065f;

where:

• p0: reference pressure
• p1: current pressure
• temp: current sensor temperature

Attached a 30' plot of the relative altitude

[o_o]

Hi Myzhar,

Your plot suggests a testing time of of 30 minutes. Without a reference sensor, there is no way to real real pressure change (Building ventilation, doors/windows opening, weather, etc.) with sensor drift.

Do you have access to a reference pressure sensor?

o_o

[Myzhar]

I have not a pressure reference sensor... I search for one. Any suggestion?