drwilliams
New Member
- Joined
- Oct 20, 2018
- Messages
- 29
- Reaction score
- 10
- Country
- United Kingdom
From my own knowledge as a former embedded systems software engineer, I believe that the the G sensor is programmed to trigger an interrupt over an i2c bus at first power up, with a threshold value that is only marginally above the RMS noise level of the sensor chip. When the camera is asleep, that interrupt wakes it. Seems to me that the threshold is set in firmware (rather than nvram) and initialized very early on in camera boot - perhaps even in the boot loader itself. Which is presumably why it isn't trivial to make it user configurable. Also, by the time the interrupt handler is running, the G event has finished, so it has no way of knowing if it was real, or just sensor noise, after the fact.
I also believe that manufacturing variances in the G sensor will make some cameras 'better' that others (e.g. if they have a particularly good sensor with less noise), and I would be pretty sure that lower ambient temperatures will also have a positive effect (cooler should mean less sensor noise).
My camera is being replaced tomorrow. Will be keen to see if it behaves better, and I will let you know.
I also believe that manufacturing variances in the G sensor will make some cameras 'better' that others (e.g. if they have a particularly good sensor with less noise), and I would be pretty sure that lower ambient temperatures will also have a positive effect (cooler should mean less sensor noise).
My camera is being replaced tomorrow. Will be keen to see if it behaves better, and I will let you know.