Naze32 FPV BNV problems. Please help.

[Bleumunkie]

Originally posted in FPV, but since the problem is not specific to FPV, but the Naze32 - I figured it could go here too..

If I broke a rule, let me know and I'll delete, just trying to get help as soon as I can.

I have built a BNV for FPV flight. Plan was to use a Naze32 for flight stabilization and OSD support. Maybe RTH if the firmware ever improves, but that is outside this post.

Here is the problem:

If i test the ailerons they respond normally - quick and to the extent of their travel, returning to neutral when the stick is released. However - If I have the throttle anywhere above 0, my left aileron will start dropping even though it is just sitting on my table. The ailerons react to the right stick being moved, but stick out of neutral spot when the stick is released. Dropping the throttle back to zero immediatly brings the ailerons back to neutral (as if they where out of place and you plugged in a battery).

If i put the Naze32 in passthru mode, I get about 1/2 the travel and slightly slower response, but it is otherwise unaffected by throttle position.

Video attached demonstrating problem. Please help!!!


Stats:

BNV completely to spec - built out of corrugated plastic

Naze32 Acro rev 5 w/ Cleanflight.
..... turning the board 90 degrees for wiring convenience.
..... Aux 1 arms the board
..... Aux 2 handles switching between passthru / acro / horizon modes.

any other questions I will answer, but unaware of any other factors ahead of time.

 

[HawkMan]

I don't know much about the Naze as a fixed wing flight stabilization, but does it have a setting for countering motor torque ?

 

[makattack]

My guess is that the software is doing that because that's not a great test. What do you see in cf configurater when you have it powered, connected and level? Does it show the craft as level? If not, it might be trying to level it if you are in angle/horizon mode. You didn't state what mode you were in or if you calibrated the accelerometers in a level surface.

Also, in a stabilized mode, when you give stick input and the craft doesn't respond, the software probably will probably drive the surfaces to max limits until the sensors responds. I test fixed wing flight controllers (using apm) by tilting the craft on roll pitch and yaw to see that the control surfaces do the opposite to counteract the movement. Eg. Roll left should cause there flight controllers to use right aileron. For tx input testing, I just check the standard inputs map correctly. The real testing comes in the air. I always start in manual mode (I yes it's called pass through in cf) get it really high and switch to a stabilized mode to test them.

 

[Bleumunkie]

I don't know much about the Naze as a fixed wing flight stabilization, but does it have a setting for countering motor torque ?

not that I am aware of.

 

[Bleumunkie]

My guess is that the software is doing that because that's not a great test. What do you see in cf configurater when you have it powered, connected and level? Does it show the craft as level? If not, it might be trying to level it if you are in angle/horizon mode. You didn't state what mode you were in or if you calibrated the accelerometers in a level surface.

Also, in a stabilized mode, when you give stick input and the craft doesn't respond, the software probably will probably drive the surfaces to max limits until the sensors responds. I test fixed wing flight controllers (using apm) by tilting the craft on roll pitch and yaw to see that the control surfaces do the opposite to counteract the movement. Eg. Roll left should cause there flight controllers to use right aileron. For tx input testing, I just check the standard inputs map correctly. The real testing comes in the air. I always start in manual mode (I yes it's called pass through in cf) get it really high and switch to a stabilized mode to test them.

Cleanflight configurator does show the craft as level, and the accelerometers are calibrated. This is troubleshooting after a maiden where it would fly like absolute crap (if I could get it in the air) and multiple barely controlled crashes from it behaving erratically on control inputs.

I will probably try absolutely backup Naze32 board to see if it may be HW related, but it doubt it. Maybe tomorrow tho- UFC fight tonight.

 

[makattack]

Does it fly the same way in pass-through/manual mode? Unfortunately with problems like this, flight data logs can really be useful. With APM, I would example the logs for vibrations and to compare "RC Input" vs. "Servo output" and "desired" roll/pitch/yaw vs. "actual/measured" roll/pitch/yaw.

Unfortunately, unless you have a naze32 with onboard memory or the logger peripheral, then you might have difficulty doing that.

If this happens when you run the motor, I suspect vibration is the issue... how is the board mounted? Is the motor balanced? Maybe it's the PID's? I'm not familiar enough with the PID values in cleanflight for fixed-wing, but with APM, one of the most difficult tasks is doing the PID tuning -- so much so that there's an autotune feature that still requires a lot of work (flying with lots of max limit roll/pitches).