Unfortunately, the ADXRS budget went to higher taxes. Enjoy your new houses. Haven't had any problems with GPS aided heading. Wind skews the heading measurement, but even though sensed North is not true North, it's whatever heading is required to compensate for the wind. If it got to where wind exceeded the maximum horizontal velocity, it would have other problems.
When I mentioned about Ice Shaker IMU I was saying clearly about manual tumbling of the IMU on the table, which usually reveals obvious artifacts of mathematical motion model, which on the other hand shows excels in flight.
As a matter of fact, the vibration resistance is inversely proportional to sampling frequency.
The choice of sensors usually does the job of moving resonant frequencies up high.
However if you are playing with filtering out much lower frequencies say at 50Hz range that are above plane's dynamics threshold anyway, you might want to use analog filtering stage with sharp cutoff, not too high bandwidth for all sensors (observe research aviation IMUs going as high as ridiculous 7th order analog filters, sometimes as low cutoff as 8Hz, when RC filter is 1st order and has suboptimal frequency domain characteristics for phase delay). This however, contradicts the use of ADC oversampling (there is not much to oversample when the signal is completely denoised).
Very interesting...especially the bit about gyro drift compensation without the use of GPS. I was always under the impression that drift was somewhat random or varying in nature and not something that could be minimized so simply.
Among the things Dean talks about in this podcast are:
--His selection of the Parallax Propeller chip and how to work with it
--Dealing with centrifugal force in an IMU
--Gyro selection and how to do temperature compensation
--The ins and outs of getting an export license.
--The hidden charms of thermopiles (and a horrifying story of what a single raindrop once did to his plane!)
--What the heck "RC filtering" is and a simple equation to get right.
Comments
As a matter of fact, the vibration resistance is inversely proportional to sampling frequency.
The choice of sensors usually does the job of moving resonant frequencies up high.
However if you are playing with filtering out much lower frequencies say at 50Hz range that are above plane's dynamics threshold anyway, you might want to use analog filtering stage with sharp cutoff, not too high bandwidth for all sensors (observe research aviation IMUs going as high as ridiculous 7th order analog filters, sometimes as low cutoff as 8Hz, when RC filter is 1st order and has suboptimal frequency domain characteristics for phase delay). This however, contradicts the use of ADC oversampling (there is not much to oversample when the signal is completely denoised).
I've learnt a lot and I really enjoyed the way the discussion was steered.
Thank you guys for organizing it and thank you Dean for sharing your knowledge with us !
--His selection of the Parallax Propeller chip and how to work with it
--Dealing with centrifugal force in an IMU
--Gyro selection and how to do temperature compensation
--The ins and outs of getting an export license.
--The hidden charms of thermopiles (and a horrifying story of what a single raindrop once did to his plane!)
--What the heck "RC filtering" is and a simple equation to get right.