sensor: bme280 regulary fails to fetch samples

[visuphi]

Describe the bug

I have the following setting

• Zephyr 4.1.0
• nRF52840DK 3.0.2
• BME280 Pimoroni Module, connected to I2C on P0.27 and P0.27, VCC and GND.

The BME280 regulary fails to fetch samples, returning the error code -EAGAIN.
It looks like the function bme280_wait_until_ready in the bme280 driver regulary runs into a timeout.

*** Booting Zephyr OS build v4.1.0 ***
Found device "BME280", getting sensor data
[00:00:00.416,992] <err> BME280: Failed to fetch samples
BME280: sensor_read() failed: -11

To Reproduce

• connect bme280 module to to I2C0 on P0.27 and P0.27, VCC and GND.
• use the BME280 sample in samples/sensor/bme280
• add overlay file for nrf52840dk:

&i2c0 {
    status = "okay";

    bme280@76 {
        compatible = "bosch,bme280";
        reg = <0x76>;
        label = "BME280";
        status = "okay";
    };
};

• build and flash west build -b nrf52840dk/nrf52840 && west flash
• see logoutputs on serial

..or use my sample: https://github.com/visuphi/test_bme280_zephyr

Expected behavior

bme280 should be able to fetch sample without running regularly into timeout.

Impact

no samples fetched

Logs and console output

when using zephyr sample:

*** Booting Zephyr OS build v4.1.0 ***
Found device "BME280", getting sensor data
[00:00:00.416,992] <err> BME280: Failed to fetch samples
BME280: sensor_read() failed: -11

when using my sample, referenced above:

*** Booting Zephyr OS build v4.1.0 ***
[00:00:00.525,482] <err> main: sensor_sample_fetch() bme280 failed: -11
[00:00:00.525,543] <inf> main: temperature: 0.000000
[00:00:00.525,604] <inf> main: humidity: 0.000000
[00:00:00.525,665] <inf> main: pressure: 0.000000
[00:00:02.678,833] <err> main: sensor_sample_fetch() bme280 failed: -11
[00:00:02.678,924] <inf> main: temperature: 0.000000
[00:00:02.678,985] <inf> main: humidity: 0.000000
[00:00:02.679,046] <inf> main: pressure: 0.000000
[00:00:04.746,368] <inf> main: temperature: 23.740000
[00:00:04.746,520] <inf> main: humidity: 34.657226
[00:00:04.746,673] <inf> main: pressure: 96.534425
[00:00:06.759,521] <inf> main: temperature: 23.730000
[00:00:06.759,674] <inf> main: humidity: 34.634765
[00:00:06.759,826] <inf> main: pressure: 96.533644
[00:00:08.913,024] <err> main: sensor_sample_fetch() bme280 failed: -11
[00:00:08.913,146] <inf> main: temperature: 23.730000
[00:00:08.913,299] <inf> main: humidity: 34.634765
[00:00:08.913,452] <inf> main: pressure: 96.533644
[00:00:11.064,300] <inf> main: temperature: 23.720000
[00:00:11.064,453] <inf> main: humidity: 34.632812
[00:00:11.064,605] <inf> main: pressure: 96.535851
[00:00:13.077,423] <inf> main: temperature: 23.720000
[00:00:13.077,575] <inf> main: humidity: 34.667968
[00:00:13.077,728] <inf> main: pressure: 96.535695
[00:00:15.090,576] <inf> main: temperature: 23.710000
[00:00:15.090,728] <inf> main: humidity: 34.667968
[00:00:15.090,881] <inf> main: pressure: 96.534429
[00:00:17.103,729] <inf> main: temperature: 23.710000
[00:00:17.103,881] <inf> main: humidity: 34.689453
[00:00:19.257,232] <err> main: sensor_sample_fetch() bme280 failed: -11

Environment (please complete the following information):

• OS: macOS Sequoia 15.4.1
• Zephyr 4.1.0 with Zephyr SDK 0.17.0

Additional context

[vignesh-aerlync]

Hi @visuphi I’m able to reproduce the issue in both your code and the Zephyr sample. The problem occurs because BME280_MEASUREMENT_TIMEOUT_MS in bme.c is set too low.

*** Booting Zephyr OS build v4.1.0-3678-gaf77efb77e91 ***
Found device "BME280", getting sensor data
temp: 24.959991; press: 100.292968; humidity: 0.7812
temp: 24.949981; press: 100.292968; humidity: 0.7812
temp: 24.929992; press: 100.296875; humidity: 0.7812
temp: 24.929992; press: 100.296875; humidity: 0.7812
temp: 24.939971; press: 100.292968; humidity: 0.7812
temp: 24.939971; press: 100.292968; humidity: 0.7812
temp: 24.949981; press: 100.296875; humidity: 0.7812
temp: 24.949981; press: 100.296875; humidity: 0.7812
[00:00:09.195,709] BME280: Failed to fetch samples
BME280: sensor_read() failed: -11

[visuphi]

hi @vignesh-aerlync

Thank you for your response.
The error really occurs less if you increase the timeout. However, I have increased the timeout up to 450ms but the error still occurs in a few cases.

which value for BME280_MEASUREMENT_TIMEOUT_MS do you suggest?

[vignesh-aerlync]

Yeah, I increased the duration to 1500 and run a long test, but the issue still occurs after sometime. We need to find a way to handle timeout gracefully.

[ldomaigne]

I've observed this kind of behavior when some of the contacts to the BME280 sensor is loose. Have you already eliminated that possibility? Missed the -11 error code (EAGAIN); loose connection would give an -5 error (EIO). My bad🙄

[ldomaigne]

I could reproduce the issue on the NRF52840DK. My preliminaries investigations let to think that we have some regressions on that platform:

Zephyr	Platform	Test
LTS 3.7	nrf52840dk	✅
LTS 3.7	mcxn947	✅
4.1	nrf52840dk	❌
4.1	mcxn947	✅

Too early to draw any conclusions yet, however.

On a side note, I've observed some suspicious behaviors if a transient I/O errors is occurring on all tested platforms/versions. It's possibly an unrelated issue.

[ldomaigne]

On my test setup, If BME280 mode is set to forced (CONFIG_BME280_FORCED=y), then the problem disappears...
@visuphi : do you see the same?

Suspect that it correlates with commit 91e006.

[visuphi]

Hi @ldomaigne

I assume you mean CONFIG_BME280_MODE_FORCED=y :).

I can confirm that the problem disappears when using this config.

[ldomaigne]

The root cause is an initialization issue, which may cause to use BME_STANDBY_05MS and BME280_FILTER_OFF instead of the application's settings. commit 91e006 exposes this bug due to the 0.5ms standby timeout.