I think there is some "gain" to be had here....
Use pairs of MEMS sensors so any of the "G" bias effects cancel out.
Do a good job of vibration isolating the sensors....
Control the sensor temperature....
With GPS to correct the biases more than good enough to get to orbit.
Blueball and my autonomous Helicopter used mems gyros....
(So did Silver ball, but I never got that one stabilized so it might have
been a mems sensor issue)
The current best of the Mems gyros are an order of magnittude better than
what I used on Blue ball.
On Tue, Jan 19, 2016 at 2:34 PM, Uwe Klein <uwe@xxxxxxxxxxxxxxxxxxx> wrote:
Am 19.01.2016 um 20:35 schrieb Henry Vanderbilt:
My guess would be that much of the error within such a one-chip sensor
farm might not be random - rather, there might tend to be a systematic
bias, as the individual sensors all share similar errors due to sharing
the same overall imprecisions in the chip-fabrication environment.
Modern measurement stuff is all about being able to describe your errors.
directed, noise
Remember the effort put into linearizing sensors.
Its "egal" today.
you measure/describe it, place it in a table and apply the fix to
the harvested values.
( also the step from forex OCXO to TCXO clock sources )
uwe
