Sensor sanity check

[baranyaib90]

Arduino board: Arduino Nano v3
Arduino IDE version (found in Arduino -> About Arduino menu): 2.3.2

Dear Maintainers!
I would like to ask for some improvement of this library to make it more robust in case of wiring issues.
Because of thermal expansion sometimes my wires get loose and I have contact issues. I know, this is my problem, I should have solder it, and so on.
But I would expect to have some sanity check functionality from the library side.

Given the following test scenario:

1. Running the code of: bme280test.ino

BME280 test
-- Default Test --

Temperature = 28.53 °C
Pressure = 1003.00 hPa
Approx. Altitude = 85.73 m
Humidity = 50.89 %

Temperature = 28.74 °C
Pressure = 1003.03 hPa
Approx. Altitude = 85.46 m
Humidity = 50.77 %

2. When I disconnect the sensor, I do not get NAN values (except Approx. Altitude), instead invalid values:

Temperature = 179.55 °C
Pressure = -160.23 hPa
Approx. Altitude = nan m
Humidity = 100.00 %

Temperature = 179.55 °C
Pressure = -160.23 hPa
Approx. Altitude = nan m
Humidity = 100.00 %

3. After reconnecting the sensor, finally there are NAN values:

Temperature = nan °C
Pressure = nan hPa
Approx. Altitude = nan m
Humidity = nan %

Thanks to NAN values, I can call the begin() function again and I can make the sensor work properly.
But I have no workaround for now to avoid the invalid values.
If there would be a possibility to check if the sensor is really connected and responding properly, that would be great!
For example:

1. sensorID() function could actually querry the sensor ID and not returning a cached value. I don't want to call begin() all the time before measurement.
2. There could be a isConnected() or similar function which checks if the sensor responds to a command with valid data.

Thank you in advance!

[LeunyEumel]

I am facing the same Issue - or have the same feature request

trying to run two bme280 with I2C over long cables frequently the sensor readings got stuck on:
Humidity: 100%
Temperature: 180.59°C
Pressure: -157.83
I assume this is due to the I2C erroring

I would prefer a cleaner way than to check my readings against above values to decide when to reset the sensors/the I2C

(I have seemingly solved this now with 1k pullups to 3V3 on SDA and SCL and going to slightly shorter cat8 cables.
I would feel more confident to deploy longterm though if there was a proper way to detect failed readings)

[LeunyEumel]

Update: the pullups only partially solved the problem.
Every 12 to 24h approximately the sensor still returns invalid readings.

Seems the Issue might be related to the Sensor itself:
https://community.particle.io/t/bme280-sensor-problem/49627/32

The proposed solution of providing 3v3 via a GPIO and power cycling whenever invalid readings has improved my situation.
I will report back after letting it run for some days