Troubleshooting Auto-Pilot Issues

[lrobichaux]

I’m in the process of troubleshooting autopilot drift, jerkiness and overall unsatisfactory performance smoothly holding course, altitude and glide path. This morning I noticed that the G3X shows that there is no software installed on the pitch and roll controls. Have no clue if this is normal, or if it’s an oversight at CubCrafters. Any thoughts or suggestions would be appreciated.

 

[Cactus Charlie]

The blank entries for roll and pitch usually mean the servo switch was off. Servos must be powered for their status to be read.

With avionics bus on, then servos on, the PFD mode annunciator should show briefly PFT (pre-fight) which is the servo test before AP engagement is allowed.

It would not be possible to have AP engagement if no software was loaded in the servos. Servo software is updated with each G3X software update as long as the servos are powered.

 

[lrobichaux]

That’s the answer. Appreciate it.

 

[Cactus Charlie]

The autopilot performance questions were not answered. However, a couple of points-

Procedures for tuning the autopilot are included in the G3X Touch installation manual. All these adjustments are available in experimental aircraft but are likely only available to the system installer in a type certificated aircraft.

In turbulent air a Carbon Cub is not going to fly like a Cirrus which has a higher wing loading and much higher cruise speed. The first test of AP performance should perhaps be to compare what it does with what the pilot would have to do in the same conditions. (In my FX-3 the stick roll activity in turbulent air was far greater than when I was hand flying. I reduced the roll gain.)

 

[Neal]

@lrobichaux while flying or on the ground, and you don't have to be on the config side, you can adjust your roll and pitch rates. I can't remember what mine are set to but I had to adjust them. You can control the aggressiveness of the responses. That's probably the first step to adjust.

 

[Neal]

@lrobichaux my apologies. It was brought to my attention I used wrong terminology. I didn't reference a manual nor am I at my plane to get screenshots. I tried looking it up in the manuals and didn't find what I was after so I gave up. But I'm told the correct terminology that you can adjust are roll servo gain and pitch servo gain.

Anyone is welcome to share screenshots of their settings to share what they're using as it may help.

 

[lrobichaux]

Thanks all. When I get the plane back from annual I'm going to go spend some quality time tweaking the AP settings.

 

[Cubonaut875]

Here are some photos of my autopilot settings:

 

[lrobichaux]

Very helpful! I’ll start with those and tweak from there.

 

[Cactus Charlie]

Very helpful! I’ll start with those and tweak from there.

Please record the current settings before making any changes.

If I were tuning the AFCS of an NXCub I'd start with the settings used by CubCrafters and only change those parameters that flight test, or ground test, showed to give improvement. Optimum NX settings may not be the same as optimum FX settings.

Servo authority can be set on the ground and clutch breakout should be evaluated with stick centered and at all extremes of travel. I found the pitch breakout force was much higher with stick well forward than with stick centered. I 'd like to be able to take control if the AP decides to stuff the nose down.

In my FX-3 I have only changed pitch servo max torque, roll servo max torque, roll gain, and min airspeed limit.

 

[lrobichaux]

clutch breakout should be evaluated with stick centered and at all extremes of travel.

What exactly is clutch breakout? Is this the amount of manual force required to overpower the servos?

 

[Cactus Charlie]

What exactly is clutch breakout? Is this the amount of manual force required to overpower the servos?

Yes, but Garmin expresses this as a percentage of maximum possible servo authority. The torque at which the clutch slips is expressed as a percentage of the torque available if the clutch did not slip.

 

[Cubonaut875]

I found with the factory settings the stick bounced all over. Less gain helped that quite a bit

 

[Neal]

I found with the factory settings the stick bounced all over. Less gain helped that quite a bit

Same, it was far too aggressive in roll for me.

 

[Cubonaut875]

I thought the autopilot shouldn't correct more than I would. I might have it dailed back too much but it seems to work better in kinda bumpy air. I still turn it off completely when it starts jumping about and yawing

 

[Cactus Charlie]

I still turn it off completely when it starts jumping about and yawing

Any aileron input will cause some adverse yaw. The G series ailerons may reduce adverse yaw but they don't eliminate it.

I reduced the roll activity but still assist the AP with rudder for turns to a new heading. The aircraft needs a yaw damper.

 

[Dollyartasia8800]

I thought the autopilot shouldn't correct more than I would. I might have it dailed back too much but it seems to work better in kinda bumpy air. I still turn it off completely when it starts jumping about and yawing

It can feel lethargic in smooth air but strangely joyful with a little turbulence if the gains are cranked way back, as I have also experienced. When it starts hunting, I usually adjust them a couple of times before giving up and hand-flying.

 

[Cactus Charlie]

I had reduced roll gain to 0.8 which worked quite well in cruise flight with light turbulence. However, I found lateral control on RNAV/LPV approaches was sloppy. I have increased roll gain to 1.0.

The good thing is that it is adjustable in flight.