We’re using a YANGDA FW320 VTOL to conduct (almost) daily flights. After each flight, we do some basic log analysis to make sure there are no hidden red flags and that the drone’s performance is more or less consistent (similar throttle % to cruise, similar airspeeds, similar battery consumption etc).
I observed that (for reasons I do not know) the drone is appearing to use more and more up elevator percentage on average to maintain level cruising flight while in fixed wing mode.
Please note that a month passed between the test flight by the manufacturer and our first 2 flights, and that they were conducted in different locations (China , West Africa). All the flights after the manufacturer’s test flight were conducted in the same location, in more or less the same weather conditions, and within a 2 week period.
We’ve already tested the Center of Gravity, and it appears to be just fine (almost level, with a bit of leaning towards the nose, as normal). Both elevator (VTAIL) servos appear to be working perfectly fine, and their trim values are relatively normal (approx. 1100 min, 1450 mid, 1800 max). I cannot think of any reasons for this type of elevator behaviour other than CG or servo issues, which both seem to not be the case.
Has anyone here experienced similar issues? What are the potential causes of this behaviour / what can be done about it? Thanks!
There is not much to a fixed wing in reality, if you are using increased up elevator on each flight to maintain level flight at the same cruising speed then only causes could be a progressive forward moving c of g, or a wing install incidence that’s changing, or a servo mount/linkage/servo arm or control surface horn that is moving during the progressive flights/storage.
The flight controller just moves the servo to the position required to maintain attitude and altitude, what you’re seeing is a larger and larger up elevator position with each flight. If it continues you may well run out of elevator authority and crash
This is how the average cruise speed (airspeed) has been changing over the course of these flights. It looks quite normal / constant to me.
23.69 m/s
24.4 m/s
23.32 m/s
24.33 m/s
23.14 m/s
Based on this information, would you suggest inspecting the elevator movement / servo linkages there?
Do you think running an AUTOTUNE would help, or is this not a PID tune issue? @tridge
And what could cause a progressive shifting towards the nose of the cg? @Scott_Nunan
(This is just an idea - could the fact that this is a VTAIL configuration, meaning the same servos are used for yaw and pitch, somehow be skewing these observed values?)
I don’t see anything other than a payload or slipping battery etc, but you have checked the c of g and found it correct.
I’m not suggesting your problem is c of g related I was just listing all the things that would contribute to more up elevator being required on successive flights at the same cruise speed.
It’s almost like the control surface is not any more deflected in later flights but just the servo out percentage is increasing so that would suggest something moving or slipping in the servo/linkage/horn etc.
have you checked physically what a 67% up elevator position actually looks like on the ground?
If the back of the ruddervators are only marginally above centre then your surfaces are not at 67% up at all and the increase is being absorbed somewhere in the servo/linkage etc, if the surface truly is rising 2/3 of the throw then it has to be c of g or something changing in the geometry of the wing/tail etc
As long as you’re looking at elevator position being common to both servos and not just a rudder command where the other servo is getting progressively more down while the first is getting more up then you are talking elevator and it shouldn’t be at all related to the mixed function.
I calculated these in the following way: Max up elevator endpoint (e.g. 1930 uS) - Center point of servo (e.g. 1530 uS) = 400 uS. Therefore, 30% of 400 uS = 120 uS, so the signal that should be sent to the servo to mimick 30% elevator up = 120 uS + 1530 uS = 1650 uS,
Are there any other possible physical biases? Maybe an aileron slipping on its horn causing some lift/drag artifacts? Is it possible the motor has shifted slightly,
producing less (or negative) lift?
Your analysis seems to have isolated the issue to the elevator itself, but it never hurts to be thorough.
The 30% up appears to be showing some down deflection, as if zero deflection would occur at about 40-45% up elevator. This could be the reason for the offset you see in recent flights.
Why isnt zero at zero? There aren’t any transmitter trims being used are there?
Looks like a linkage is winding its way out or something?
What does the airframe manufacturer have to say about what you’re experiencing?
i dont currently have the tools to look at your bin logs, so i apologize for something that may be obviously answered by them…
.
i am assuming that these elevator servo variations are significant enough to indicate a problem. If they are just a normal variation, never mind. it’d be interesting to hear if they ever move back towards the less positive direction.
.
Air density: less air density = less lift. this usually requires more speed with the same control deflections or, inversely, same airspeed with more deflection (possibly indicated by pitch angle). are you using an airspeed sensor, or just GPS airspeed estimate? i think arduplane accounts for air density with an airspeed sensor (or maybe that’s just an AS sensor’s operational physics?), but i don’t think GPS based estimates account/adjust for it (don’t know for sure). in other words, i think arduplane will increase ground speed in a higher density altitude situation if using an AS sensor, but not if using GPS estimates.
…ps: i see the plane comes with an AS sensor. hopefully it’s working correctly. maybe you could crosscheck it with the GPS groundspeed variations, but wind might confuse that effort.
do the logs show that pitch attitude is the same? in other words, is the elevator deflection difference causing an attitude difference?
… more positive elevator should result in a more positive (nose-up) attitude. if it doesn’t, that may suggest a variation in the elevator controls (like slipping/changing linkages/horns/arms), elevator incidence (shifting more positive), wing incidence (relative to elevator incidence - shifting more negative). are those empennage connection tubes secure and always resulting in the same distance from tail-to-wing (shorter distance = more up-elevator)? unless, of course, the vtail has some rudder/elevator relationship that i’m not aware of.
.
Is it possible that any of those VTOL rotor blades are causing aerodynamic deflection in forward flight?
.
per yuri’s comment about potential aileron problems: are your other servos showing variations (ailerons)?
.
is throttle position for cruise following the elevator servo trends?
… check it.
.
Is height being maintained?
… If same height is maintained, same speed, same attitude (pitch), same throttle position, then elevator changes don’t even make sense - then you really need ask why changes in servo control are resulting in elevator at the same position.
.
and another thing… hopefully you are averaging from just level flight at cruise speed. turning, loitering, maneuvers and landing will, obviously, increase the elevator’s average. forgive me for stating the obvious. oh, wait! will the elevator deflection change with “rudder” changes? how’s that ‘servo’ looking?
