The unit on my 6hvu flies perfect, but the one on my warp 360 slowly pitches the nose up (back elevator) while hovering. I have to hold forward pressure on elevator to keep the disk level. It acts just like my radio has a few clicks of back elevator trim, but it does not. I checked that first thing.

The head is correctly set up. My swash was leveled using a tool, so it's nearly perfect. Servos are correctly centered via subtrim tab. 0, 6, and 12 degree pitch set with rc logger gauge. So I'm pretty sure the issue doesn't lie in the head setup.

I have noticed the swash sometimes creeps backward on the bench as well.

Anyone have any thoughts on what the problem is? vibrations? Bad sensor? Bad servo?

It is hard to tell you where exactly is the problem. But if you can, please check in the diagnostics tab (immediately after the flight) if there is no deviation on the elevator channel more than approx. +/-3%.
Also you can check if there is "0" in the Sensor part - first digit from the left side is elevator axis.

What is your value of Stick Deadband?

Can you measure vibrations in Elevator axis with rudder blades on?

_________________
Spirit System developer

Stick deadband is 10. Vibes where around 20. The input channels for aileron and elevator read a steady 0% on diagnostic tab. No flickering at all.

However the senor reading are strange. The first time I hooked it up, all 3 sensor readings where at 0, but when I wiggled the cyclic the left senor slowly moved up to around 80 and flickered +- 2.

I could move the left reading back down by pushing forward elevator, but it would then crawl back up to 80ish again when I released the stick. The center sensor returned to zero no matter what I did.

I then power cycled the unit and reconnected to settings app. This time both the left and center sensor would always return to zero after wiggling the cyclic.

After many power cycles I managed to reproduce the issue with the climbing left sensor reading a couple more times, but unfortunately I haven't figured out exactly what makes it happen. When it did happen, the initial value was a steady zero. It didn't start changing until I touched the elevator channel, at which point it would climb up to 80 and stay there.

I'm a software developer, so I'm very familiar with needing a solid and repeatable repro, but I can't figure it out. Sorry.

Thanks for good info.

The best condition to reproduce the issue is to wait 5-10 minutes and then to move with elevator stick.
Aileron and Elevator "sensor" outputs in the diagnostics tab should return slowly back to 0 always.
If not, then
1) something is affecting the sensor because of exceptional noise in the voltage - what BEC/ESC do you use?
2) sensor is damaged
3) do you use a fan?

_________________
Spirit System developer

I am using the internal bec in my HobbyWing Platinum Pro V3 50A. No fan.

Additional data using the suggested 10 minute wait:

First power up of the day:

• Turned on radio, sat aside on bench after verify throttle hold
• attached programming cable between PC and FBL
• powered up ESC
• connected settings program
• observed following sensor readings on diagnostics tab: 5, -1, 0
• observed following sensor readings after 10 minutes: 12, -3, 0
• moved elevator stick on radio, sensor readings rapidly moved to: 80 +- 2, -3, 0
• moved aileron stick, sensor readings moved to: 80 +-2, -81, 0
• senor readings after 5 more minutes: 80 +-2, -81, 0
• this is the first time I have observed issue on aileron. Had been elevator only until now.

2nd test

• power cycled fbl unit, reconnected software
• initial sensor readings: 0, 0, 0
• sensor readings after 15 min: 0, 0, 0
• wiggle both ele and ail, sensor readings move with input, then return to 0, 0, 0.

3rd test

• same results as 2nd test above

Seems to be associated with the initial start up. I'm going to leave the ESC unpowered for a while so the caps bleed down, etc. and try again. Will post results later.

Thank you very much for the info.

What is your BEC voltage? Can you see bold values in elevator channel?

_________________
Spirit System developer

So I tested again after leaving ESC/BEC disconnected for 45min or so.

For this set of tests I power cycled the FBL by pulling the throttle signal wire while leaving the ESC/BEC powered. So the BEC was running on initial boot up for all 3 tests.

The 1st test was similar to test 1 above. Sensors did not start at 0, I didn't record the actual values, but they were not the same as before. Wiggling cyclic causes sensors to run up to 80, -80, 0.

2nd test started out looking good, sensors at 0,0,0 and returned to 0 after wiggling cyclic. However after a few min the values where 3, -1, 0. After wiggling stick they again went up to 80, -80, 0.

3rd test was showed completely normal behavior. No value creep at rest and always returned to 0,0,0.

BEC is configured for 7.4v. Measured with multimeter on throttle signal wire fluctuating between 7.53 and 7.55. Power cycled ESC a few times and measured same voltage every time.

Input channel values turn bold about half way between 3% and 4% on both pos and neg directions.

I did some more testing and the incorrect behavior is definitely related to how long the unit is without power.

If I leave the unit powered down for a few min it shows the incorrect behavior after power is returned. It is also not necessary to wait the 5-10 minutes to see the bad behavior. It starts immediately.

If the power is returned quickly the unit seems to behave correctly.

This was consistent across many tests. I didn't try to nail down exactly how long it needed to be without power to reproduce the problem, but it was only a couple minutes.

It was also the same if I power cycled the ESC/BEC or if I cycled the FBL only via pulling the throttle wire. So it doesn't seem to be related to how long the ESC is off.

It seems as if the internal regulator is getting very hot and because of very big temperature change sensor is drifting. If you powercycle the unit quickly, then drift is compensated.
Do you use Spektrum satellite?
If it is possible to change the voltage to 5-6V and to connect normal receiver, it should be better, but definitively not solution.

So there are possibilities that Spektrum satellite (if used) is drawing too much current or we have to replace regulator in the unit.

_________________
Spirit System developer