Barometers gone wild

[espruino-discuss2]

Posted at 2024-03-23 by @wminarik

I have an original Bangle.js 2 with the BMP280 pressure chip. It's running the current 2V21 firmware. I've recorded temperature and pressure logging (10 sec interval) using the 'Recorder' app.

A couple of years ago I cross calibrated the BMP280 against a DPS310 breakout; the BMP280 gave roughly similar pressures (offset by 0.2 hPa) and temperatures, with a bit more noise.

Now, however, the BMP280 gives very strange pressure readings, and seems to be extremely sensitive to temperature changes. Temperature readings seem to be OK. The watch hasn't been submerged, and both sensors were protected from stray breezes during logging.

I've attached the comparison plots: Temperature in blue, Pressure in orange, darker is the Bangle BMP280 and lighter is the DPS310. The drop in BMP280 recorded pressure with temperature is apparent in the full log. At the end, the sensors were lowered by about 6 meters. The DPS310 shows the small increase in pressure and is steady, but the BMP280 follows the temperature change.

Could this be a software algorithm issue in the device. driver? Or has the BMP280 membrane degraded?

---

Attachments:

• Bangle_BMP280a.png
  
• Bangle_BMP280b.png

[gfwilliams]

Posted at 2024-03-25 by @gfwilliams

I don't think the algorithm has changed on the device at all.

It's worth noting that since the sensor is inside the watch body, potentially other sources of heat within the watch (the CPU, backlight, battery charging) could affect readings...