I've been running some tests on my Forerunner 255 after noticing an anomalous behavior in elevation gain calculation.
Setup:
- Manual calibration performed
- Barometer set to Sensor Mode: AUTO
- The device was kept completely still throughout the entire observation
Theoretical background (from official manual)
According to the Garmin manual, AUTO mode uses both the altimeter and barometer based on detected movement. When stationary, barometric behavior is expected: pressure changes are treated as weather variations, elevation remains stable, and sea-level pressure varies accordingly.
To clarify the two values visible in the sensor screen:
- Ambient Pressure: raw value measured directly by the sensor
- Barometric Pressure: sea-level compensated value calculated by the software, based on the calibration elevation
Observed behavior
With the device stationary, ambient atmospheric pressure varies (normal weather variation). I would therefore expect:
- barometric pressure (at sea level) to vary accordingly
- elevation gain to remain stable at zero
Instead, I observe the opposite:
- barometric pressure remains constant
- elevation gain increases as ambient pressure changes
This behavior is consistent with the device operating in altimeter-only mode: it interprets the meteorological pressure variation as a vertical displacement, recalculates elevation upward, and the sea-level compensation remains artificially stable. This happens while stationary, which is exactly when AUTO mode should not behave this way.
Ruling out hardware failure
If this were a physical sensor issue, raw readings would be erratic and random. The firmware logic of AUTO mode would apply regardless. The fact that a systematic and consistent pattern is observed instead suggests a software problem, not a hardware one.
Hypothesis
I suspect the latest firmware update 27.09 introduced a regression in the movement detection algorithm used by AUTO mode, causing the device to behave like an altimeter even when stationary.
Has anyone else experienced this issue? Can anyone verify the behavior on a previous firmware version?

