On 24/05/16 09:57, jic23@xxxxxxxxxxxxxxxxxxxxx wrote: > On 24.05.2016 09:45, Linus Walleij wrote: >> On Mon, May 23, 2016 at 6:29 PM, Jonathan Cameron >> <jic23@xxxxxxxxxxxxxxxxxxxxx> wrote: >> >>> I forgot this part also has gain and offset trim registers. Will add them before >>> applying as that is easy enough and reduces the differences further. >>> >>> These exist on some other parts but we can add them any time. >> >> LIS3LV02DL also has x,y,z offset and gain trim registers, but when >> I read up on it, I understood that these are just factory-programmed >> registers. I.e. they are undocumented in the register list, but a table >> states them as "calibration" and "loaded at boot". >> >> So I always understood them as something the sensor loads at >> boot to calibrate itself, so you can ignore the contents, and they >> probably exist on the other sensors too, albeit undocumented as >> they have no use. >> >> I toyed with the idea of reading them and pusing the values to the >> entropy pool with add_device_randomness() as they are device-unique >> data. >> >> But if you have proper docs on them I guess maybe they are used >> differently on your sensor... >> >> Yours, >> Linus Walleij > As I understand it you can actually apply corrections to them (was certainly > how I read that years ago). Take into account that these early sensors were > incredibly noisy and perhaps more prone to shock damage than newer ones. > (my test sensor will vary by 10% or more whilst sat on a steady concrete > floor - even with huge levels of oversampling...) You can indeed apply corrections to them. It might be an entertaining attack vector if the machine gets compromised and someone is using it for free fall detection as you can make it read whatever you like ;) Anyhow, I have working code but need to tidy up before posting a V2 with support for this on this chip included. Can support other chips off datasheet but have no hardware. > > Back then we were using some sensors where if you dropped them on the floor > you had to send them back to the manufacturer for recalibration before they > were remotely useful again. > > Funnily enough, given we had them attached to people that happened all the > time :( > > ST certainly aren't believers in 'detailed' documentation. Always feels > like somewhat of a guessing game when implementing drivers for their > parts. > > J > -- > To unsubscribe from this list: send the line "unsubscribe linux-iio" in > the body of a message to majordomo@xxxxxxxxxxxxxxx > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-iio" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
