Tricopter Tail Servo Jitter on Naze32, please help.

[x0054]

I decided to swap my APM board for a Naze32 (afroMini) to see what the difference would be. The APM flys just fine on this tri, so the problem isn't with the frame. With the Naze I am getting a lot of servo jitter. Even wile sitting on the table, twisting the model causes the servo to go into uncontrollable jitter. While flying, the servo get's into a jitter that's getting faster and faster, and faster, uncontrollably. I tried playing with PIDs with no luck. The only thing that helps is getting the P all the way to 1, but that makes the yaw very twitchy, and the jitter is still there, stressing out the servo a lot.

I know on other flight controllers there is a low pass filter for servo input, but I can not find the same on Naze32. Any ideas?

If you are flying a smaller tri with naze, could you also tell me your PIDs. Thanks!

- Bogdan

 

[x0054]

Correction, the APM USED to fly just fine on this Tri. About 2 weeks ago I crashed the Tri, and broke one of the arms. Today I finally got replacement parts in the mail, and put everything back together. I took the opportunity to lighten the craft by shortening wires and swapping for smaller ESCs (still more than adequate). It a bit lighter now, but the rest is the same.

While constructing it, I installed the Naze32 rather then APM, to see how it performs, and got a lot of that jitter I spoke of earlier. After no luck in getting the Naze to fly, I decided to switch to APM. With APM I get the same thing, tail wag. The difference is, the APM is slower in terms of it's control loop time, so what looked on the Naze as a jitter, on the APM looks like a very pronounced wag. After a fast yaw, the tri starts to oscillate left and right. With APM it settles after 3-5 seconds. With Naze it would build more and more, until it would loose control and crash.

The strange thing is, before the crash 2 weeks ago, the tri would fly grate, no problems at all. The crash was my fault I rebuilt it largely identical to before. Same frame, same motors, new ESC's, but they work fine, same props, same everything. And same APM flight controller with same settings as from 2 weeks ago.

I swapped out the Servo twice, trying 3 different servos that I know work well. No luck. Looks like the problem is somewhere with the frame, not the FC. Naze32 amplified the problem, but looks like it did not cause it.

I am a bit at a loss as to what else it could be. Any ideas?

 

[Anand]

I have similar problems, i have a HK x900 (its somewhat of a cheap aluminium RC explorer shape) with DT750's. I have flown it before with the KK2.0 but had some problems with it (later it turned out to be a faulty throttle gimbal on my 9XR) Yesterday i got 2 Naze32 controllers in the mail and installed one in the Tri, after some testing hooked up to the baseflight app it looked like the jitter was giving the gyro's and Acc's a hard time. Still a went out for a maiden, took off and about after hovering for 5 secs at 1 meter it did a backflip and landed upside down in the grass. (tailservo part bent, 1 prop and 1-2 motors damaged)
Maybe the weight of the (shifting)DT750 on the tailservo was too much, or not enough stiffness in the tailboom.

 

[x0054]

Anand, take a look to make sure that you have the naze32 orientation and servo direction setup correctly. I had that problem too, went out to fly, 5 seconds and it was face down on the ground. But that was a simple enough problem to solve. Now I am trying to figure out the problem with the jitter.

As for my problem, I think I figured at least part of it out. I lightened my design, concentrating all of the weight in the center of the craft. This resulted in the rear tail having a lot less mass, making it very yaw happy. This is why the APM would not fly right after the rebuild. I tried adjusting the PIDs, but nothing changed. Today I figured out why. Something in the APM went haywire, and it was not saving any of my PID changes. Eventually today it completely went insane! I had it armed, pushed up the throttle and nothing happened. So I throttled down, placed the controller on the floor, and looked over to my laptop. As I was looking away the tri came to life, took off, flew up 5-6 feet, and then fell down. After that I figured out that the APM was somehow exacting all my RC commands 0.5-2 seconds after receiving them. God only know how that's even possible.

I ended up reflashing the firmware and resetting all the settings. After that the tri started flying a lot better! The PIDs still need to be adjusted, because it yaws way too crazy, but at least now it actually saves and reacts to the PID settings. Just need to replace the broken servo (got damaged in the many crashes), and tune it out.

As for the Naze, I have no idea what the issue is. That thing makes the tri way too twitchy. If anyone flies the Naze on HJ-Y3 frame, or for that matter, the x900 frame, please post your PIDs, it would really help.

Thank you.