What's new
Van's Air Force

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

AutoPilot Scary Incident

mfleming

mfleming

Well Known Member
Patron
I'm posting this because it surprised me on how easy and incipient it was. These RV are slippery...I know this and you know this but when it happens it can happen without warning.

Here's a link to a flight I made this morning. Headed to the WAAAM Museum for their annual Fly-in. Level at 8500', cruising at 163 KTAS and on the autopilot. The smoke has been bad in Oregon and on this flight it was OK until it wasn't. Decided to descend from 8500' to 6500' and do a 180. The autopilot was commanded to descend at 500'/m. After telling center that I was reversing course, I switched from NAV mode to heading mode and punched in 080°. The AP started a turn to the right and continued to descend. I'm flying VFR so my eyes are outside, not much of a horizon and I didn't notice that the bank angle was increasing and the autopilot was struggling to follow the flight director. Just as I noticed the upset, the airspeed warning started annunciating. Autopilot off, wings level and raise the nose...already too late, Vne busted!

Move ahead on the timeline until about 07:43 that's when I commanded the AP to descend. At about 07:45:50 I command the AP to fly 080°. It's all over about 07:48

EDIT: here's a better timeline.
  1. 07:43:56 commanded the AP to descend from 8500' to 6500' at 500ft/min
  2. 07:45:50 Commanded the AP to turn to 080° - switched from NAV mode to HDG mode
  3. 07:46:31 The AP losses the ability to follow the Flight Director. The AP is still engaged.
You can see the AP status in the top right of the display.
 
Last edited:
That is scary!
My only suggestion is to be religious about your scan beyond just traffic. Yes, head outside, but don’t forget about scanning your instruments. A panel scan might have given you a clue in time to question what’s up.
Look on the positive side, you passed a flutter test ;) .
No harm no foul. Now time to dig into what’s up with that auto pilot!
 
mfleming said:
I'm posting this because it surprised me on how easy and incipient it was. These RV are slippery...I know this and you know this but when it happens it can happen without warning.

Here's a link to a flight I made this morning. Headed to the WAAAM Museum for their annual Fly-in. Level at 8500', cruising at 163 KTAS and on the autopilot. The smoke has been bad in Oregon and on this flight it was OK until it wasn't. Decided to descend from 8500' to 6500' and do a 180. The autopilot was commanded to descend at 500'/m. After telling center that I was reversing course, I switched from NAV mode to heading mode and punched in 080°. The AP started a turn to the right and continued to descend. I'm fly VFR so my eyes are outside, not much of a horizon and I didn't notice that the bank angle was increasing and the autopilot was struggling to follow the flight director. Just as I noticed the upset, the airspeed warning started annunciating. Autopilot off, wings level and raise the nose...already too late, Vne busted!

Move ahead on the timeline until about 07:43 that's when I commanded the AP to descend. At about 07:45:50 I command the AP to fly 080°. It's all over about 07:48
Click to expand...
Good catch.
Have you set Max and minimum speed in your A/P settings? Most A/P have the max and minimum airspeed but it seems something serious has gone wrong in your A/P as at some point you had nearly 4000 fpm descend rate.
 
Last edited:
in addition to airspeed limits there are bank limits to set. at a minimum, it certainly looks like you need to go through the A P tuning procedure that Dynon publishes. If you can‘t pass that there is a problem with the installation or AP.
 
RVDan said:
in addition to airspeed limits there are bank limits to set. at a minimum, it certainly looks like you need to go through the A P tuning procedure that Dynon publishes. If you can‘t pass that there is a problem with the installation or AP.
Click to expand...
I'll keep the EFIS & Autopilot brand under my hat until they have a chance to look at the data log...
 
JonJay said:
Look on the positive side, you passed a flutter test ;) .
Click to expand...
Actually he didn’t, but that's a discussion for another thread. Time to consider an RV-8 rudder though.

‘ArlingtonRV' and I flew to Forks (S18) for lunch today from Arlington (KAWO) and the smoke was bad on that route too with low vis and no horizon.

Remember to pull the power off in addition to rolling level and pitching up (remember the items for unusual attitude recovery).

Glad it all worked out for you. Thanks for the report.
 
Last edited:
RV8JD said:
Actually he didn’t, but that's a discussion for another thread. Time to consider an RV-8 rudder though.

‘ArlingtonRV' and I flew to Forks (S18) for lunch today from Arlington (KAWO) and the smoke was bad on that route with low vis and no horizon.

Next time remember to pull the power off in addition to pitching up. Glad it all worked out for you. Thanks for the report.
Click to expand...
I have a -8 rudder on my -7.
Good tip about pulling the power, it might have saved a couple knots. The autopilot definitely went fubar but I should have reduced power and slowed down as soon as I decided to reverse course. Lots of lessons learned on this one.
 
mfleming said:
I have a -8 rudder on my -7.
Good tip about pulling the power, it might have saved a couple knots. The autopilot definitely went fubar but I should have reduced power and slowed down as soon as I decided to reverse course. Lots of lessons learned on this one.
Click to expand...
What was the pitch and roll trim doing when the excursion occurred?

Thx.
 
Last edited:
Capt said:
The time to scan the instruments more is when you command a change to the A/P to confirm it’s doing what you expect, this we do in heavy metal, this I do in my own RV👍
Click to expand...
I watched it start the turn and descent but unfortunately started looking outside again. I’ve recently been training to complete a IPC and if on instruments, would have caught this early.
 
RV8JD said:
What was the pitch and roll trim doing when the excursion occurred?

Thx.
Click to expand...
No roll trim. Pitch trim is controlled by the autopilot when engaged. I’ll have to look at the data log and see if trim changes are captured.
 
Another clue that’s obvious to most is the increase in noise, wind noise as the ROD & AS go wild! Good lesson for all here to remain alert, in all senses👍
Whilst A/P’s have been a God send for flying in high work load situations it has had a negative effect on all of us, it’s dumbIng down our skills and situation awareness, this event shows that perfectly!
 
Last edited:
I'm sorry, but that is a long time not to monitor your instruments regardless of what else is happening. You also gave the AP a high work load, which you failed to monitor or even set up (reducing power) to ensure the safety of your flight.
Like all modern electronics- it works perfectly right up to the time it doesn't! You have to never let your guard down - infact when you are not physically flying is the time to ramp up your vigilance!
 
mfleming said:
I'm posting this because it surprised me on how easy and incipient it was. These RV are slippery...I know this and you know this but when it happens it can happen without warning.
Click to expand...

Thank you for sharing this Michael. I've never flown with an autopilot but I'm building an airplane that will have one (Garmin GMC-507). Your post has been super helpful.
 
I’ll add another caveat to AP/auto-trim operation: never apply stick inputs with AP engaged, the auto trim system is always trying to remove stick force and it doesn’t know the difference between pilot input or normal trim changes. As an example if you pull back on the stick the auto trim will start trimming down, can get your attention pretty quick when you let go, and the faster you’re going the more dramatic the effect.
 
Last edited:
larryMar said:
I'm sorry, but that is a long time not to monitor your instruments regardless of what else is happening. You also gave the AP a high work load, which you failed to monitor or even set up (reducing power) to ensure the safety of your flight.
Like all modern electronics- it works perfectly right up to the time it doesn't! You have to never let your guard down - infact when you are not physically flying is the time to ramp up your vigilance!
Click to expand...
Nothing to be sorry about. You’re absolutely right, the time between the autopilot losing control until I disconnected it was 15 sec. That’s too long. After thinking about this some more, I should have gone on instruments as soon as I made the commands to reverse course. Looking outside was doing little to help due to the stratified smoke layers giving false horizons.
 
Capt said:
Another clue that’s obvious to most is the increase in noise, wind noise as the ROD & AS go wild! Good lesson for all here to remain alert, in all senses👍
Whilst A/P’s have been a God send for flying in high work load situations it has had a negative effect on all of us, it’s dumbIng down our skills and situation awareness, this event shows that perfectly!
Click to expand...
The really scary part is there was no noticeable increase in wind noise 😳 just a number on the EFIS.
It must be the slick nature of the RV plus the noise canceling headsets. I can definitely see how if you were disoriented in the clouds things could get out of hand in a hurry 🤯
 
I was testing my AP's 180 turn around function and had the AP turn and descend rapidly. I've been given some things to try after reporting it to the manufacturer, but have not yet taken the time to implement them. I'm so happy that I tested this before I needed the feature, since a few weeks after the problem I had a situation where had I used it, I would not be typing this. If you have an AP and have not tested features like this in VMC with a safety pilot, please consider it.
 
rv8ch said:
I was testing my AP's 180 turn around function and had the AP turn and descend rapidly. I've been given some things to try after reporting it to the manufacturer, but have not yet taken the time to implement them. I'm so happy that I tested this before I needed the feature, since a few weeks after the problem I had a situation where had I used it, I would not be typing this. If you have an AP and have not tested features like this in VMC with a safety pilot, please consider it.
Click to expand...
It will be interesting to see what the manufacturer has to say about my data. My servos were replaced less that 6 months ago under warranty.
 
Walt said:
Yes, they will keep applying maximum torque trying to regain control. Basically being overpowered by the aircraft though.
Click to expand...
On the dynon, I can definitely hear the slip and if I wasn't ready for it, it could be disconcerting. When doing the 180, if I don't pitch the trim up the Dynon autopilot will let me know and it will slip
 
mfleming said:
The really scary part is there was no noticeable increase in wind noise 😳 just a number on the EFIS.
It must be the slick nature of the RV plus the noise canceling headsets. I can definitely see how if you were disoriented in the clouds things could get out of hand in a hurry 🤯
Click to expand...
I have the fixed pitch propeller and if I pick up the speed up like this the propeller will go over 2700 RPM and I will hear it and feel the extra vibration. But you have the constant speed and the RPM is the same and why you didn't hear the change.
 
Holy Cow, glad you're okay! Good on you for catching that before it got even more out of control.

I'm going to be an outlier here and say I wouldn't have reduced power in that situation either. Right up until it came off the tracks, the IAS was below 160 and the TAS was around 180.

When it went rogue it quickly rolled to 40 degrees which is well above a bank that would correlate to a standard rate and then it dropped the nose to about -10 degrees, which resulted at one point in about a 4,000 fpm descent. Pulling power back a few inches isn't going to save you from that and pulling it back below a level that will sustain level flight is just an accident waiting to happen.

Most modern autopilots are just trying to make the airplane bars match what the flight director is displaying. Your flight director seemed to be representing pitch and bank appropriately, but at 14:46:16 the autopilot completely lost track of the flight director command bars. Also at 14:46:46 when the heading blew through the heading bug it was supposed to capture, you can see that Nav mode was momentarily armed and then went away.

Again, really glad you came through this okay. pleased keep us posted on what you find out.
 
Desert Rat said:
Holy Cow, glad you're okay! Good on you for catching that before it got even more out of control.

I'm going to be an outlier here and say I wouldn't have reduced power in that situation either. Right up until it came off the tracks, the IAS was below 160 and the TAS was around 180.

When it went rogue it quickly rolled to 40 degrees which is well above a bank that would correlate to a standard rate and then it dropped the nose to about -10 degrees, which resulted at one point in about a 4,000 fpm descent. Pulling power back a few inches isn't going to save you from that and pulling it back below a level that will sustain level flight is just an accident waiting to happen.

Most modern autopilots are just trying to make the airplane bars match what the flight director is displaying. Your flight director seemed to be representing pitch and bank appropriately, but at 14:46:16 the autopilot completely lost track of the flight director command bars. Also at 14:46:46 when the heading blew through the heading bug it was supposed to capture, you can see that Nav mode was momentarily armed and then went away.

Again, really glad you came through this okay. pleased keep us posted on what you find out.I've been flying with a CFII these last few weeks to get instrument current. We did some upset trading last week and the fix for a nose low turn was to level the wings and pull up...no reduction of power.
Click to expand...
I really think the smart thing to have done is gone to my default slow down speed of 20" MP and 2350 RPM before descending and turning. This puts me just above maneuvering speed with plenty of room to Vne and stall.
EDIT: My recent instrument upset training for my IPC for a nose low turn was - level the wings while raising the nose. No power adjustments.
 
A friend of mine never seems to pull power to descend and goes well into the yellow almost every time. I pull power to descend every time. This story is exactly why I’m not in a hurry to get on the ground. Take your time. Stick to your checklists and flows, fly conservatively.

Descending to land, descending for meteorological phenomena, etc. is not the time to be speeding up. Great example of how something unexpected can get us out of sync with the aircraft.

Glad we’re all able to learn from this without another member of the GA community eating soil! Thanks for sharing.
 
Thinking back on the flight in question, I remembered that there was a minor AP excursion earlier in the flight. At about 07:22:45 with the AP on Nav, the AP loses the flight director and slowly wanders to the right. I noticed this and, I think, just bumped the stick to the left ( you can see the RV jerk to the left and the AP recaptured the course... Something's not right in AP land.
 
It looks like the roll servo torque settings are too low. The A/P wasn’t able to roll out of the turn. The pitch increased causing the turn to tighten and enter a spiral dive.
Time to revisit the autopilot setting.

regards Peter
 
kiwipete said:
It looks like the roll servo torque settings are too low. The A/P wasn’t able to roll out of the turn. The pitch increased causing the turn to tighten and enter a spiral dive.
Time to revisit the autopilot setting.

regards Peter
Click to expand...
Yep, I'm getting ready to send a email to the manufacturer with all the setting info.
Roll servo torque is set to 15% and the pitch at 30%. The roll does seem really low. Not sure why I set it there but I vaguely remember following some internet advice 🫣
 
mfleming said:
Yep, I'm getting ready to send a email to the manufacturer with all the setting info.
Roll servo torque is set to 15% and the pitch at 30%. The roll does seem really low. Not sure why I set it there but I vaguely remember following some internet advice 🫣
Click to expand...
Vans recommendation for Garmin pitch and roll torque for a 14 is 50%
 
Recommend settings are are good starting point but it’s worthwhile going through the setup procedure step by step as every aircraft is going to be slightly different, depending upon control friction, even the amount of trailing edge bend radius will effect control forces Etc.
 
mfleming said:
Yep, I'm getting ready to send a email to the manufacturer with all the setting info.
Roll servo torque is set to 15% and the pitch at 30%. The roll does seem really low. Not sure why I set it there but I vaguely remember following some internet advice 🫣
Click to expand...
I run roll servo at 30%, 15 is not enough for roll, 30% is usually enough on the 6/7/8 for pitch assuming nice free floating elevators.
I’ve installed many (including my own) with these settings and never heard of or had any issues.
You could bump the pitch to 40 if you want it to have a bit more authority.
I also run the roll servo arm in the middle hole which increases torque and gives full throw on the servo arm.
 
Last edited:
Walt said:
I run roll servo at 30%, 15 is not enough for roll, 30% is usually enough on the 6/7/8 for pitch assuming nice free floating elevators.
I’ve installed many (including my own) with these settings and never heard any issues.
Click to expand...
That sound good Walt. I was surprised to see the roll setting so low when I looked into it.
 
kiwipete said:
Recommend settings are are good starting point but it’s worthwhile going through the setup procedure step by step as every aircraft is going to be slightly different, depending upon control friction, even the amount of trailing edge bend radius will effect control forces Etc.
Click to expand...
Most surface TE are too fat causing overly sensitive controls.
 
I notice on the digital flight record that during the decent through the max decent rate and the max airspeed the power remained at 75% +/-. In a decent power has to came back or Ne will arrive quickly.
That is presumably what may have happened to TNFlygirl a few months ago in her bonanza.
 
Do you have an autotrim module?

besides the graveyard spin bank angle excursion, the AP seemed unable to hold the set 500fpm on several occasions, especially with a bit of bank. i'm also in the "not enough servo torque" both for roll as well as pitch camp (and possibly no automatic pitch trim?)...
as others have said, at least reduce the power a little bit. one additional effect that i feel is very noticeable is that whatever throttle position you have at 8500, if left untouched, due to the increase in static pressure while descending, the engine will automatically produce even more power. making matters worse...

another key function in such a scenario is that whenever the autopilot is "losing it", there must be an audible alert/callout with the autopilot disconnect! was there any warning or only once you manually disconnected the autopilot? or not at all?
 
flyvans.com said:
Do you have an autotrim module?

besides the graveyard spin bank angle excursion, the AP seemed unable to hold the set 500fpm on several occasions, especially with a bit of bank. i'm also in the "not enough servo torque" both for roll as well as pitch camp (and possibly no automatic pitch trim?)...
as others have said, at least reduce the power a little bit. one additional effect that i feel is very noticeable is that whatever throttle position you have at 8500, if left untouched, due to the increase in static pressure while descending, the engine will automatically produce even more power. making matters worse...

another key function in such a scenario is that whenever the autopilot is "losing it", there must be an audible alert/callout with the autopilot disconnect! was there any warning or only once you manually disconnected the autopilot? or not at all?
Click to expand...
Yep, reducing power is the right call.
The autopilot never disconnected, so no tone or warning.
 
Last edited:
flyvans.com said:
another key function in such a scenario is that whenever the autopilot is "losing it", there must be an audible alert/callout with the autopilot disconnect! was there any warning or only once you manually disconnected the autopilot? or not at all?
Click to expand...
Did it disconnect? In any event, note that if you have a 305 or 307 with a G5, then there is NO audible alert/callout when the autopilot disconnects. I consider this a major safety disadvantage of that equipment combination. I’ve had my autopilot take a nap on me several times, and the little flashing A on the G5 is the only warning. It’s very easy to miss. I looked into upgrading to a 507, but the upgrade path was “buy a 507 at full price.” 🤣
 
Dugaru said:
Did it disconnect? In any event, note that if you have a 305 or 307 with a G5, then there is NO audible alert/callout when the autopilot disconnects. I consider this a major safety disadvantage of that equipment combination. I’ve had my autopilot take a nap on me several times, and the little flashing A on the G5 is the only warning. It’s very easy to miss. I looked into upgrading to a 507, but the upgrade path was “buy a 507 at full price.”
Click to expand...
Edit: I was wrong and Dugaru was right. The GMC 307 controller doesn't support audio annunciations so the system can't provide a disconnect sound with a stand-alone 307/G5 install. So the balance of this post can be ignored.

(Edit: I just saw you were specifically referring to a G5-only installation, which I realized only after typing the below message. However, the G5 with autopilot flight manual 7-7 specifically states you should also have a disconnect tone . In fact I occasionally fly a Cherokee with a G-5 only installation with GFC that has the autopilot disconnect audio chime. But to be fair type certificated aircraft only use the 507)...

I have the older GMC 307 on my GFC system in my RV-8 and I get the very distinct chime audio sequence tone when I disconnect the autopilot. Whether the on-stick disconnect button, on the face of the 307 or if I hit the trim switch. Same audio tone as the 507. Which makes sense when we remember that the GMC you see on the panel is just a control panel of knobs and switches to make the autopilot easier to use. The actual autopilot brain is in the "smart servos" and can be fully controlled by a G5 or GDU from internal software menus, which is why the GMC (MC = mode controller) is an option, at lease on EAB aircraft. The smart servo requests the fight deck send the tone through a discrete channel to the audio panel. The GMC isn't in that loop. The only significant difference between the 307 and 507 is that the 507 has CAN Bus connectivity whereas the 307 uses a direct RS-232 connection to the flight deck. So the GMC is kind of out of the loop with respect the the CAN Bus architecture of the G3X system. In real world flying the pilot would not notice any difference between the functionality of the 307 and 507 nor have any idea of which data bus the mode controller is using. So I would suggest you check your autopilot configuration settings for audio alerting. Also make sure the G3X alerting system isn't muted. An make sure the audio has been wired correctly from the GDU to the audio panel. Upgrading the 307 to the 507 would be a lot of money an effort since they are not plug and play and would require wiring changes due to the difference in data bus type used. In the end your current lack of autopilot disconnect annunciation would not change. Neither would your autopilot experience in flight.
 
Last edited:
jliltd said:
Edit: I just saw you were specifically referring to a G5-only installation, which I realized only after typing the below message. However, the G5 with autopilot flight manual 7-7 specifically states you should also have a disconnect tone even with a G5 combine with a GMC 305/307 mode controller. In fact I occasionally fly a Cherokee with a G-5 only installation with GFC that has the autopilot disconnect audio chime. But to be fair type certificated aircraft only use the 507)...
Click to expand...
I was told by Garmin that no disconnect tone is available with a G5/307 combination. Perhaps something has changed? I will investigate!
 
An "yellow arc" article for reviewing....
 

Attachments

  • The Yellow Arc - Aviation Safety.pdf
    1.9 MB · Views: 37
Dugaru said:
I was told by Garmin that no disconnect tone is available with a G5/307 combination. Perhaps something has changed? I will investigate!
Click to expand...

Not sure how that would be possible with that combo since neither of them have any audio outputs.
 
I find it most likely that the servo max torque value prevented a recovery, and the autotrim either wasn't fast enough, or was already at max. Please do report back once you know!
 
It's important to understand that an autopilot is a sort of PID-controller.
Some combination of set up settings may cause the output signal to be unstable, the regulator fails to regulate.
Therefore it's important to follow the guidelines in set up procedures.
As Capt has said, it is important to check what the autopilot is doing after a mode change. If you change two setting more or less at the same time, the workload for the autopilot increases and it will be more difficult to see what is going on.
Before I change from NAV to HDG, I push the left knob on the GMC507. This will synchronize the controller with present heading of the aircraft.
Then I use the left knob to set the new heading.
It's my experience from industry regulators that abrupt changes sometimes may upset the controller output signal.

Good luck
https://en.m.wikipedia.org/wiki/Proportional–integral–derivative_controller
 
Dugaru said:
I was told by Garmin that no disconnect tone is available with a G5/307 combination. Perhaps something has changed? I will investigate!
Click to expand...
You are correct. My experience with stand-alone G5 and GFC 500 has been with a GMC 507. My experience with G3X and GFC 500 has been with both. Since the pinout of the 307 doesn't have audio circuits and the pinout of the GMC 507 does, that would explain it. The reason the GMU 307 works in my RV-8 is because it is integrated with the G3X Touch. I learned something.
 
You must log in or register to reply here.
Back
Top