I have a GRT autopilot in my RV-8. I don't test it mechanically on the ground or before flight. I just leave it off, and then activate it when I'm over 1000ft AGL. This gives me time to disconnect it if it misbehaves, which has never happened. I know I can overpower it - I've tested that in the air. My servos have a safety feature that would allow me to overpower them even if they fully locked up - it's a type of "shear pin." If the servos are working normally, testing overpowering them won't cause any damage to mine - you need to check with Garmin to see about their servos. I'm pretty sure that they won't be damaged by overpowering them on the ground or in the air. I commend you for looking into this and making sure your students and instructors understand the AP. If nothing else, it might provide something positive to the tragic death of the pilot and her father that you mentioned.

We recently installed a GFC500 in our Cherokee 235 and had the EXACT same question. We came to the same conclusion: - test aileron - test elevator - press the AP disconnect button, don't try to manually override the servos

Great video! Thanks Meredith

First i will qualify my comments with the provision that i am not an AME and truly you should talk with your AME or avionics shop for final guidance. I have however built several Homebuilt aircraft and recently installed a full Garmin system and Autopilot in my RV-10. Here's what i thing are a few answers to some of your questions: 1. Garmin doesn't recommend overpowering the autopilot servos in flight because if you have autotrim installed the A/P will adjust the trim to remove the force of your inputs from the controls. It cannot tell if the loads it is seeing are from air loads on the control surface or from the pilot. Should you disengage the autopilot with these abnormal forces trimmed into the aircraft you will end up with a potentially seriously out of trim airplane. In a slightly different lead up that is believed to be the cause of the crash of the Beech Debonair flown by KZfaqr Tennessee Fly girl. 2. The autopilot servos have adjustable clutch breakout (or strength) settings. If you find the forces required to overpower these forces too great you can have your AME or avionics shop go into the configuration menu and adjust these forces. They are adjustable from 10 inch pounds of torque at the servo all the way up to the maximum of 60 inch pounds. Most likely the lowest end would not be enough to overcome aerodynamic forces in flight so the torque will be set somewhere above that number. Perhaps your shop set the forces set at their maximum or a very high number. I've never rigged a 172 but i would suspect that the torque settings would be somewhere around 30% of the maximum. Again check with your AME or avionics shop. 3. Because of the built in clutch you cannot break the servo by overpowering it on the ground. the clutch is designed to disengage before anything breaks. If you overpower it in flight be very aware of my comments in (1) above. 4. Garmin doesn't publish the forces at the controls as depending on the configuration and geometry of the aircraft those forces would be different for every aircraft. They do publish specifications on their servos that they can put out between 10 and 60 inch pounds of torque but that is all. 5. If you feel you need absolute numbers for your aircraft i suggest perhaps you work with your AME or avionics shop to measure these. They will most likely be different for each individual aircraft. Also keep in mind that in your 172's case your static 'UP' force will feel significantly higher than the 'DOWN' as with the 172's elevator naturally sits down so you will have to overcome it's weight naturally to hold it up on top of the servo force. I hope this answers your questions and i understand how you are confused because there is truly little guidance out there on these items. Keep in mind i am an airplane builder and neither a certified avionics technician or a licensed AME and take the above advice for what you paid for it.

These questions reflect deeply examining and understanding the documentation! I’m yet to fly with an autopilot so I very much appreciate seeing this system from your more experienced and in depth understanding. In addition to “pushing the red button” I would think that one would want to practice pulling the breaker in case the button doesn’t work as expected. Is Garmin OKWith doing that for real on an installed, running system?

interesting video

I’ve seen airplanes more being moved on ramps as you describe it

As an engineer myself, I am sure Garmin engineers know the answers to your questions. Have you reached out to them?