What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Autopilot Pitch runaway GSA28

mciaglia

Well Known Member
Started the autopilot config per the G3X manual. At the point of turning on the AP and adjusting the heading to confirm stick movement. As soon as the AP is engaged the pitch servo goes to a full nose up position (stick comes full back).
Stein built the panel
Software is most current
G3X
GMC507
On configuration page all servos are “green check”
Tried switching the servo - same problem
Reversing the servo in setting caused just the opposite...stick goes full forward
Trim test works on the trim motors

What next?
 
Turn on the FD (flight director) and see what it's commanding, the AP follows the FD. I assume you did the pitch/roll calibrations?
 
Started the autopilot config per the G3X manual. At the point of turning on the AP and adjusting the heading to confirm stick movement. As soon as the AP is engaged the pitch servo goes to a full nose up position (stick comes full back).
Stein built the panel
Software is most current
G3X
GMC507
On configuration page all servos are “green check”
Tried switching the servo - same problem
Reversing the servo in setting caused just the opposite...stick goes full forward
Trim test works on the trim motors

What next?

Hello Mark,

There is no such thing as a pitch (or roll) servo runaway with a Garmin autopilot, so no worries there.

What you are seeing is just the normal limitations to engaging the autopilot on the ground where the autopilot has no ability to lift/lower the nose to correct for small differences between selected pitch and measured pitch (like it does when in the air).

To test the pitch axis, simply press the AP button to engage the autopilot in pitch/roll modes, then immediately reach up and roll the pitch wheel down on the GMC 507 to move the flight director symbol up to 5 degs above the horizon. The autopilot will slowly ease the stick back in an attempt to raise the aircraft nose to match the flight director position. Since it can't actually lift the nose on the ground, it will continue to ease the stick back until it is against the stop if you let it.

Once you have confirmed proper stick movement to climb, reach up and roll the pitch wheel up on GMC 507 to lower the flight director symbol to 5 degs below the horizon. The autopilot will slowly ease the stick forward in an attempt to lower the aircraft nose to match the flight director position.

Test over.

Steve
 
Last edited:
Unfortunately it doesn's follow the flight director inputs.
It does in roll. I select heading mode, turn the heading bug and the stick slowly follows as if its turning the plane. Unfortunately no matter what is done with the pitch wheel the pitch servo does not respond.
 
Unfortunately it doesn's follow the flight director inputs.
It does in roll. I select heading mode, turn the heading bug and the stick slowly follows as if its turning the plane. Unfortunately no matter what is done with the pitch wheel the pitch servo does not respond.

Mark,

What do you have your pitch gains and max torque set to?

Are you using PIT mode?

Are you sure you are giving it time to slowly reverse direction when you move the flight director to the opposite direction? It can take a while, especially if you have let it run to the stops in one direction.

A short video might let us see something to explain what you are seeing.

Steve
 
Ap

I assume that you did the servo wiring test prior to the autopilot test?

Do you have the auto trim disabled for the autopilot test?
 
To test the pitch axis, simply press the AP button to engage the autopilot in pitch/roll modes, then immediately reach up and roll the pitch wheel down on the GMC 507 to move the flight director symbol up to 5 degs above the horizon. The autopilot will slowly ease the stick back in an attempt to raise the aircraft nose to match the flight director position. Since it can't actually lift the nose on the ground, it will continue to ease the stick back until it is against the stop if you let it.
Steve

WHen you press the AP button, the System should go into pitch hold (tries to hold the current attitude) and therefore shouldn't be commanding elev movement. The face that the OPs servo rapidly goes to full control indicates that something is amiss. After you confirm that the config and calibration is correct, I would call Stein or Garmin for guidance. Something doesn't sound right.
 
I went through the entire configuration steps this AM once again per the G3X manual.
Prior to this I checked the DSub that Stein wired for the pitch servo and everything appeared correct.
Same exact results.

I have a call into both Stein and G3Expert.

I also provided my data log of the entire morning going through the steps.
A fellow RV10 pilot walked through the steps with me as well and was utterly confused with the actions he saw.
 
was the AP already spooled up, mine does that when I turn it on and engage it.... because I wasn't turning it on with the avionics and letting it spool up. Just a thought maybe your a lot smarter than me.
 
Longez IS the expert!

If anyone can help....its longez Steve....he was at the sharp point of the spear before retirement .....or like you said...call Stein.
 
If anyone can help....its longez Steve....he was at the sharp point of the spear before retirement .....or like you said...call Stein.

Hi Bob,

This is very kind of you, but I don't think I did much good here.

While it is true that there was nothing wrong with the servos, Mark was doing everything right and it was Mark collecting data and working with the real pros at Garmin that figured out what was happening with this specific installation.

Thanks,
Steve
 
Mark,

I may not be a G3X expert, but I have installed it in my own RV-9a that I am still building and I have been an avionics tech for 30 years. If you still need some help with it I am nearby and also a member of EAA302. I powered mine up today and confirmed that the pitch motor did not run until I moved the wheel up or down and followed the FD command bars.

Jay
 
I had the same setup and the same problem. I eventually just turned the auto trim off.

I decided I liked it off anyway. It will keep you more involved in airspeed changes and gives a trim up or down message as soon as it starts to get a little out of trim.
 
I had the same setup and the same problem. I eventually just turned the auto trim off.

I decided I liked it off anyway. It will keep you more involved in airspeed changes and gives a trim up or down message as soon as it starts to get a little out of trim.

Can’t imagine spending all that money on equipment and then turn a feature off instead of troubleshooting it and figuring out what is really wrong.
 
Basics

I am only 10000 electrons away, so cant see what you see, but I agree it is worth fixing.

First make sure no wires are shorted to ground that should not be. Then I would make sure the wiring is per Garmin. Then make sure the trim works in the correct direction. Make sure the autopilot steers in the correct direction, then switch the trim wires around that go from the servo to the trim motor and see if that helps.

Gotta be something wrong somewhere.
 
Last edited:
Gremlin found

With the help of the amazing garmin team and Joe Waltz my EAA302 president and fellow RV10 builder/pilot we figured it out

We went through several test scenarios and it always was still presents. We tried disconnecting the GMC to see if that had an issue. Nope.

I traced every wire from the servo to the canbus to the gmc and everything was correct

I sent a data log to garmin and they said it seemed fine, but wanted the engineers to dig deep into it

I reset the software to the previous version. No help

Redid the new software... no help

Then we tried something. I turned off my primary AHAR’s and forced the backup G5 to control everything...IT WORKED!

But when I spoke to garmin they didn’t see any abnormal code from the GSU. So I fortunately had a fellow builder with a brand new GSU25 AHARs and I installed it ....it worked perfectly .

Garmin was great and my new gsu is shipping out. They are extremely interested to see mine and dissect this bug since they reportedly have never heard of this. Odd thing was , the AHAR’s calibrated fine and I flew 3 uneventful flights with normal pitch, airspeed and altitude information. I wonder if I had two GSU25’s if one of them would have seen this error.

Bottom line is...we are now flying a computer. Be diligent in your setup and troubleshooting.
 
Great news on the troubleshooting and support. I was surprised by the auto-trim runaway because I know Garmin has the system programmed to disconnect trim servo if it runs more than a couple of seconds. That was done precisely to nip runaway trim in the bud. To test this feature try pressing and holding the electric trim control switch continuously. After a few seconds the switch ceases to work when the system software makes the assumption that the trim command is coming from a continuous short in the harness instead of the pilot. The system expects small period "blips" of trim from the pilot, not continuous engagement of the button. I found this out when a panicked pilot held down the trim button with a death grip and wouldn't let go of it due to the lack of the response he was looking for (counter-intuitive, especially when the mind is under duress).
 
...
Bottom line is...we are now flying a computer. Be diligent in your setup and troubleshooting.
As someone that's been developing and working with computers for a long time, this is scary. My car is "drive by wire" and at any moment it could do something that kills me, but at least I know in my RV-8 I can power it all off and override anything dumb it does, as long as I recognize it before it's too late. I hope we hear what happened to this GSU and that we learn from it. Glad it happened on the ground, not in the soup.
 
GSU 25 Exchange

With the help of the amazing garmin team and Joe Waltz my EAA302 president and fellow RV10 builder/pilot we figured it out

We went through several test scenarios and it always was still presents. We tried disconnecting the GMC to see if that had an issue. Nope.

I traced every wire from the servo to the canbus to the gmc and everything was correct

I sent a data log to garmin and they said it seemed fine, but wanted the engineers to dig deep into it

I reset the software to the previous version. No help

Redid the new software... no help

Then we tried something. I turned off my primary AHAR’s and forced the backup G5 to control everything...IT WORKED!

But when I spoke to garmin they didn’t see any abnormal code from the GSU. So I fortunately had a fellow builder with a brand new GSU25 AHARs and I installed it ....it worked perfectly .

Garmin was great and my new gsu is shipping out. They are extremely interested to see mine and dissect this bug since they reportedly have never heard of this. Odd thing was , the AHAR’s calibrated fine and I flew 3 uneventful flights with normal pitch, airspeed and altitude information. I wonder if I had two GSU25’s if one of them would have seen this error.

Bottom line is...we are now flying a computer. Be diligent in your setup and troubleshooting.

Thank you for your kind words, and patience while we worked through this one.

We took a look at all the data and photo's you provided to us and the problem is well understood. This issue only affects the on-ground autopilot test and there is no risk of the autopilot or airspeed indicator behaving abnormally in-flight as a result.

Due to the non-linear response of airspeed to pitot pressure, airspeed is very inaccurate while stationary. Each GSU 25 will indicate slightly different airspeeds while stationary due to subtle differences in components, but are individually calibrated to a high degree of accuracy before leaving the factory to ensure proper indications at operating speeds. Per the datalogs you provided, the stationary airspeed reported by your GSU 25C appeared to be within normal tolerances, but on the higher end. We identified a corner case in which a higher airspeed value reported while stationary, prevents the on-ground autopilot test from being performed normally. We will take a look at the unit when we receive it to ensure nothing is amiss.

Thanks,

Justin
 
Tim Motor Run Time

Great news on the troubleshooting and support. I was surprised by the auto-trim runaway because I know Garmin has the system programmed to disconnect trim servo if it runs more than a couple of seconds. That was done precisely to nip runaway trim in the bud. To test this feature try pressing and holding the electric trim control switch continuously. After a few seconds the switch ceases to work when the system software makes the assumption that the trim command is coming from a continuous short in the harness instead of the pilot. The system expects small period "blips" of trim from the pilot, not continuous engagement of the button. I found this out when a panicked pilot held down the trim button with a death grip and wouldn't let go of it due to the lack of the response he was looking for (counter-intuitive, especially when the mind is under duress).

In this case, the system was commanding movement of the elevator, rather than the elevator trim tab, so this was not a runaway trim event. Your description of the trim system is accurate however, and you can actually program the MOTOR RUN TIME LIMIT setting in configuration mode to customize the maximum continuous run time of the trim motor in your system.

Thanks,

Justin
 
Back
Top