I'm relatively new to quads myself but I don't think from a logical perspective that would work.
here's how I figure it....
If the board reads data ( changes in degrees ) from a central position, then no matter which way it tilts, the system will give the same opposite effect to counter that movement. Its like a sea-saw - sure you can have 2 small kids at one end and one fat kid on the other, but a non equal distribution of weight without the same amount of weight to compensate I think would cause the motor ( on the side closest to the board ) to work harder to compensate. - then there's the central pivotal point to consider. for "trim"
I believe the closer to center the board is placed, the easier the trim would be to set and maintain.
Hope Im on the right track here..
I would like to bump this one!!!
Can someone in the know chime in?
I am a little conflicted on this topic, and have not yet received my APM to do any testing. Just running the idea through my head, being further from center would result in a greater impact to some measurements, and not others. accel vs gyro... So one could move away from center in X and/or Y axis to increase accel sensor readings, as the same angle change would result in the sensor traveling an increased distance in the same amount of time. But to any benefit?
Or maybe I'm nuts and a complete clown?
Typically you want to select a point on a robot to be the center of its coordinate system. When dealing with quadcopters and airplanes it makes sense to select the COG. However that doesn't mean you have to place all of the sensors at the COG, you could place them all over the robot, but then you have to transform (rotate, and translate) them to the point about which you want to make your decisions.
An interesting story that pertains to this is the Space Shuttle Approach and Landing Test program. The pilots in the shuttle were placed very close to the center of rotation in the shuttle, and this created a .5s delay in their ability to sense pitching of the aircraft. This almost ended catastrophically when on the last of 5 test flights the pilots nearly flubbed the landing because they put in delayed controls ( as well as several other factors...). Luckily they were epic pilots and were able to save the landing, but it's a great example of how being at the center of rotation can be bad. Of course this depends on your sensor, the inner ear is epic, but not the greatest rotational sensor.
The thing you have to balance here might be the saturation or max of the sensor, for example you might only be able to sense up to 30rad/s^2 with the gyro, but it's actually doing 35rad/s^2 then you might want to move it farther from the center of rotation so that it can sense the true rotational acceleration... I think that makes sense, someone double check me on that.
I'm not sure I follow.
If the angle changes 10° it does not matter where on the chassis this is, it is still 10°. Assuming the time taken is the same then further from center requires greater accel readings, correct? So you will saturate the accelerometer sooner by moving away from center, correct?
I have no basis, but would not expect the accel sensor to be at it limits during any part of normal flight, at least with the types of flight typically discussed here. Anyone know off hand what typical accel range would be?(both sensor and measured)
I agree Sam! Only the accs will have inconsistant readings. For the gyros, dosent matter at all.
My guess is that the accs output information for lat/long/alt only, while roll/pitch/yaw infos are given by the gyros. This means that if the board is assembled far from the COG, during a roll or pitch, the acc might read that the aircraft is going up or down.
Have no idea about their range... Would be for sure interesting to know.
my bad. you guys are right about those gyros.