Huh? Magic relieves you of the laws of physics when the smaller wing goes first?
Brett
Section IV, subsection 37, paragraph (A),
Unusual Planforms, of
International Standards for Model Aviation Aerodynamic Analysis clearly states:
"Aircraft that do not conform to the usual monoplane layout, with a single main lifting surface ahead of a single, smaller stabilizing surface, may be analyzed using magical thinking, as long as you can make it sound convincing. This rule specifically applies in the case of biplanes and other multiplane lifting surfaces, flying wings, 'flying lawnmowers', asymmetrical aircraft,
and canards." (my emphasis)
There. Now you know.
The gliding angle is about 15°, even with full up (elevator 45" down). That convinces me that the canard is stalled (not the wing!).
As long as the canard is in the propeller slipstream, it is effective.
So the prop is ahead of the stabilizer. Could you
please post some pictures?
Just because the shallowest glide angle is steep doesn't mean the canard is stalled -- it just means that it's generating insufficient lift for the conditions. Have you tried moving the CG back? I assume this is going to be an interesting balancing act, as having the elevator so close behind the prop is going to really crank up the sensitivity when the motor is running.
The strange motor sound during level flight is caused by ny motor speed regulating system. The accelleration sensor based system tries to keep the line tension (more corruptly: the centripetal acceleration) at around 3g. If too low, the speed is increased a little. Next measured after 8 ms, If too high, the speed is decreased with the same amount. Tim, this is an integrating regulation, correct?
If the increment is constant then it's sorta-kinda integrating control. True integrating control would have you calculate the throttle as:
<throttle now> = <last throttle> + <integrator gain>(<target accelerometer> - <accelerometer reading>)
If you have what I think you have it'll adjust slower than it could for large speed errors, and it'll always oscillate around the correct value instead of settling in to a constant value. For a large enough increment the constant seeking of the throttle will cause the system to consume more power than if it settled out to a fairly constant value, but that effect may be washed out by the necessary throttle changes in response to maneuvering.
Given the various lags in the system, I expect that you could get snappier control with a proportional-integral controller, but if you do that then tuning will almost certainly be harder than with integrator-only control (and based on the tuning parameters that Igor exposes on his system, I suspect that's not what he does).