ihbarddx
Well-Known Member
So here are the products of an analysis of gtg738w’s datasets from the Hurricane 38 Rocket, as posted in the FlightSketch flight log. (Search the user name in the flight log at flightsketch.com and look at datasets labeled Hurricane 38)
The object was to compute tilt angles (angles with respect to vertical) from the fusion of X-axis accelerometer readings with barometric data, and to compare them with the (gyro?) tilt angles in the datasets. I do not understand why tilt angles are in the dataset. I assume it’s a new instrument they are testing.
The altimeter, which responds to pressure differences caused by gravity, serves as a vertical reference. The X-axis readings, which appear aligned with the long axis of the rocket in these flights (as would make sense with a gyro), serve as information about extra-gravitational accelerations parallel with the velocity vector. (That assumes stability. i.e.; small angles of attack – the rocket must be following its nose). The result is a basis for two-dimensional trajectory analysis – even without a gyro.
Barometric data were adjusted for temperature. (i.e.; all altitudes were multiplied by [GroundTempInKelvins/288.15] )
Accelerometer data were adjusted for launch angle (i.e.; All values were multiplied by [g*sin(LaunchAngle)/AverageOfPreflightReadings] )… although the launch angle was only 3.6 degrees off-vertical.
The moment of launch was judged by the accelerometer, and not by the time column. The spreadsheet was modified to accommodate sampling intervals of varying sizes.
==
The russet curves (labeled Tilt) are the values from the instrument. The blue curves (labeled WGX) are the values from fused inertial and barometric data.
In the F50-9 flight, the FlightSketch data are noisy on the way up. The blue curve is smoother, although smoother doesn’t necessarily mean more accurate. An altimeter is likely to be less responsive than a gyro (if that is what was used), and smoothing may be the result. The ascending curves intersect – hard to judge correspondence. The whole thing is a mess.
Correspondence among the curves from other flights is rather good until somewhat past nose-over (900). Note that the plateaus represent parachute descents, where the rocket isn’t following its nose, and all bets are off. The plateau regions comprise most of the graphs; whereas the important parts are the ascents (to the left), so the effect is deceptive.
This sort of analysis can be done with any flight computer that has an altimeter and at least a single-axis accelerometer (e.g.; standard FS hardware). The tilt angles make possible broader 2D analysis. I’ve posted some of this stuff before, but it was nice to have tilt angle data to compare. Thought it more or less postable.
Looking forward to FlightSketch hardware becoming available.
Also looking forward to collecting such data from my new Blue Raven.
The object was to compute tilt angles (angles with respect to vertical) from the fusion of X-axis accelerometer readings with barometric data, and to compare them with the (gyro?) tilt angles in the datasets. I do not understand why tilt angles are in the dataset. I assume it’s a new instrument they are testing.
The altimeter, which responds to pressure differences caused by gravity, serves as a vertical reference. The X-axis readings, which appear aligned with the long axis of the rocket in these flights (as would make sense with a gyro), serve as information about extra-gravitational accelerations parallel with the velocity vector. (That assumes stability. i.e.; small angles of attack – the rocket must be following its nose). The result is a basis for two-dimensional trajectory analysis – even without a gyro.
Barometric data were adjusted for temperature. (i.e.; all altitudes were multiplied by [GroundTempInKelvins/288.15] )
Accelerometer data were adjusted for launch angle (i.e.; All values were multiplied by [g*sin(LaunchAngle)/AverageOfPreflightReadings] )… although the launch angle was only 3.6 degrees off-vertical.
The moment of launch was judged by the accelerometer, and not by the time column. The spreadsheet was modified to accommodate sampling intervals of varying sizes.
==
The russet curves (labeled Tilt) are the values from the instrument. The blue curves (labeled WGX) are the values from fused inertial and barometric data.
In the F50-9 flight, the FlightSketch data are noisy on the way up. The blue curve is smoother, although smoother doesn’t necessarily mean more accurate. An altimeter is likely to be less responsive than a gyro (if that is what was used), and smoothing may be the result. The ascending curves intersect – hard to judge correspondence. The whole thing is a mess.
Correspondence among the curves from other flights is rather good until somewhat past nose-over (900). Note that the plateaus represent parachute descents, where the rocket isn’t following its nose, and all bets are off. The plateau regions comprise most of the graphs; whereas the important parts are the ascents (to the left), so the effect is deceptive.
This sort of analysis can be done with any flight computer that has an altimeter and at least a single-axis accelerometer (e.g.; standard FS hardware). The tilt angles make possible broader 2D analysis. I’ve posted some of this stuff before, but it was nice to have tilt angle data to compare. Thought it more or less postable.
Looking forward to FlightSketch hardware becoming available.
Also looking forward to collecting such data from my new Blue Raven.
