some redundancy for added safety and more compactness....
Actually, there isn't any redundancy. Notice that in this Y configuration there is no servo on one of the arms. That means that for each pair of motors the torque must cancel. If a motor goes out then the controls have no way of correcting the yawing without creating a pitch or roll.
Just to be clear this isn't a problem with Parallax's Y6, but all flying systems of this type.
Actually, there isn't any redundancy. Notice that in this Y configuration there is no servo on one of the arms. That means that for each pair of motors the torque must cancel. If a motor goes out then the controls have no way of correcting the yawing without creating a pitch or roll.
Just to be clear this isn't a problem with Parallax's Y6, but all flying systems of this type.
I guess Parallax should change the wording then:
This 'double-stack' of the motors gives you added payload capabilities, and makes your system redundant in case of a sudden motor loss, making the platform safer to fly.
Which begs the question, can the HoverFly board compensate for a one motor outage?
How about this: if a motor goes out the control system takes any motor on any other arm that rotates counter to what the failed motor turns at. That way, you still have an even number of motors and, in theory, could still fly. But I doubt that the system supports that as is.
How about this: if a motor goes out the control system takes any motor on any other arm that rotates counter to what the failed motor turns at. That way, you still have an even number of motors and, in theory, could still fly. But I doubt that the system supports that as is.
Wouldn't that need some form of feedback from the motors? How would the control board know that a motor failed. (Just thinking out load).
I was thinking more along the lines of the input from the gyro. If a motor failed, the gyro would compensate.
It is really hard to say by looking at the picture but it looks like each set of motors is acting as one (same handed prop reversed and opposite each other)?
If that is the case wouldn't the failed motor set produce less adverse yaw and less lift?
Since there is less adverse yaw then yaw shouldn't be an issue to correct. Since there is less lift the other motors just compensate and it should still be controllable.
This looks cool but I have what is probably a dumb question. Why is this better than a hex-copter? Is it because with fewer arms it is easier to avoid having one of the arms end up in the picture if you mount an camera on the copter?
If that is the case wouldn't the failed motor set produce less adverse yaw and less lift?
No. Between each pair on an arm the props rotate in different directions, and this cancels out the yaw. One goes out then that produces a net yawing torque on the arm, which spins the whole craft. The other motors can't compensate because then they produce too much lift, which flips the craft.
Since there is less adverse yaw then yaw shouldn't be an issue to correct. Since there is less lift the other motors just compensate and it should still be controllable.
It's a system of two equations to remain in a stable level flight: balancing the yaws, and balancing the thrusts. You have to make sure that you take both into account.
Okay, if each set of motors are counter-rotating then I can't think about it anymore my brain hurts. Reading discussions on just how redundant Y6 configurations was interesting and produced further confusion. I digress lol.
I think redundancy is the correct term. You have four degrees of freedom (pitch, roll, yaw, and overall thrust) and six motors. Assuming the roll axis lies along the arm for motors 1T and 1B, here are the equations:
thrust = 1T + 1B + 2T + 2B + 3T + 3B
pitch = 2T + 2B + 3T + 3B - 1T - 1B
roll = 2T + 2B - 3T - 3B
yaw = 1T + 2T + 3T - 1B - 2B - 3B
Suppose motor 2B stops, then the equations are:
thrust = 1T + 1B + 2T + 3T + 3B
pitch = 2T + 3T + 3B - 1T - 1B
roll = 2T - 3T - 3B
yaw = 1T + 2T + 3T - 1B - 3B
Solving for 1T, 1B, 2T, 3T, and 3B in terms of desired pitch, roll, yaw, and thrust is still overspecified (i.e. 4 independent equations and 5 unknowns). Holding T1 constant, the solution is:
1B = (thrust - 2*1T - pitch)/2
3B = (-yaw + 2*1T + pitch)/2
2T = (thrust + 2*roll + pitch)/4
3T = (2*yaw + thrust - 4*1T - 2*roll - pitch)/4
So, yes, you can still control the 6-motor Y configuration with a damaged motor.
I do believe that it can fly with one motor out - as long as it is not too heavily loaded. When one motor in a pair fails the other motor must turn faster to keep level. That solves the problem of keeping level, providing there is reserve power available. As far as yaw goes that can be handled by the other two pairs of motors.
This is so easy to test, just disconnect a motor and go fly it.
Another way to look at. Any single pair of motors can handle yaw. Since a pair contains counter rotating propellers one can speed up and one slow down to affect yaw while maintaing the same amount of lift.
So, arm 1 has lost the top motor, the bottom motor is running at 100%, arm 2 has each motor at 50% and arm 3 has the bottom motor at 0% and the top motor at 100%. The torques from 1 and 3 cancel out and all arms produce equal thrust.
Publison
Posts: 12,366
Comments
http://www.parallax.com/product/80100
Actually, there isn't any redundancy. Notice that in this Y configuration there is no servo on one of the arms. That means that for each pair of motors the torque must cancel. If a motor goes out then the controls have no way of correcting the yawing without creating a pitch or roll.
Just to be clear this isn't a problem with Parallax's Y6, but all flying systems of this type.
I guess Parallax should change the wording then:
Which begs the question, can the HoverFly board compensate for a one motor outage?
Wouldn't that need some form of feedback from the motors? How would the control board know that a motor failed. (Just thinking out load).
I was thinking more along the lines of the input from the gyro. If a motor failed, the gyro would compensate.
Haven't thought the whole thing out just yet.
If that is the case wouldn't the failed motor set produce less adverse yaw and less lift?
Since there is less adverse yaw then yaw shouldn't be an issue to correct. Since there is less lift the other motors just compensate and it should still be controllable.
[video=youtube_share;S0L58SMryDs]
No. Between each pair on an arm the props rotate in different directions, and this cancels out the yaw. One goes out then that produces a net yawing torque on the arm, which spins the whole craft. The other motors can't compensate because then they produce too much lift, which flips the craft.
It's a system of two equations to remain in a stable level flight: balancing the yaws, and balancing the thrusts. You have to make sure that you take both into account.
I wrote something about this here: http://robotics.stackexchange.com/questions/1562/if-you-can-create-pure-yaw-motion-with-a-quadcoptor-then-why-wont-this-work-wit/1563#1563
pitch = 2T + 2B + 3T + 3B - 1T - 1B
roll = 2T + 2B - 3T - 3B
yaw = 1T + 2T + 3T - 1B - 2B - 3B
Suppose motor 2B stops, then the equations are:
pitch = 2T + 3T + 3B - 1T - 1B
roll = 2T - 3T - 3B
yaw = 1T + 2T + 3T - 1B - 3B
Solving for 1T, 1B, 2T, 3T, and 3B in terms of desired pitch, roll, yaw, and thrust is still overspecified (i.e. 4 independent equations and 5 unknowns). Holding T1 constant, the solution is:
3B = (-yaw + 2*1T + pitch)/2
2T = (thrust + 2*roll + pitch)/4
3T = (2*yaw + thrust - 4*1T - 2*roll - pitch)/4
So, yes, you can still control the 6-motor Y configuration with a damaged motor.
-Phil
This is so easy to test, just disconnect a motor and go fly it.
So, arm 1 has lost the top motor, the bottom motor is running at 100%, arm 2 has each motor at 50% and arm 3 has the bottom motor at 0% and the top motor at 100%. The torques from 1 and 3 cancel out and all arms produce equal thrust.