Autopilot initiates a spiral dive

[Bert Pieke 7,585]

52 minutes ago, andrecarli737 said:

Solution for me -> Setting a deadzone on the aileron axis so it wont activate while AP is on.

It could indeed be that my Yoke has drifted out of calibration  🙂

I have re-calibrated and set a 5% deadzone for ailerons and elevators.

I might have expected the AP to switch off, if the controllers are out of whack.. but it stays engaged as the airplane banks and then dives..

[andrecarli737 177]

2 minutes ago, Bert Pieke said:

It could indeed be that my Yoke has drifted out of calibration  🙂

I have re-calibrated and set a 5% deadzone for ailerons and elevators.

I might have expected the AP to switch off, if the controllers are out of whack.. but it stays engaged as the airplane banks and then dives..

I came to a conclusion that it is a bug. Elevator is not affected with manual activations, but ailerons are until you get no input.

[leprechaunlive 1,859]

On 10/1/2020 at 9:43 PM, Mike S KPDX said:

Bert, try this.  Please turn off (via slider) airliner and general traffic densities.  Then try your complete flight.  I have stated elsewhere on these forums that there is an issue here that can be easily reproduced.  This has to do with 'ghost' manipulation of the controls/switches.  It seems that if you are within a certain radius (either in the air or on the ground) of a same or similar AI controlled type, those AI actions also effect your aircraft.  So you may be overflying an airport that has these types of planes.  I can sit in the TBM at a midsized airport, turn on the battery (this seems to activate the AI operations) and then turn it off.  If I wait about 5-15 minutes, multiple switches will suddenly start flipping on/off or changing in random order, including the AP.  Maybe I screwed something up in my install, but this is definitely a reproducible error on my set up.

That is my thought at least, would love to know if this effects others.

 

That sounds mad haha. I seen you stated that a few times, i fly without AI so cant reproduce, but if you'r theory is true, thats mental 😄

[jabloomf1230 2,445]

Just when you think it's safe to go into the water... I tried flying the Caravan for a fly flights and guess what? Spiral dive into the ground. The bug apparently isn't specific to any user aircraft, but I've never seen it happen on a flight where I didn't use the autopilot. I don't remember this happening so often with earlier MSFS releases, but maybe that was because the autopilot was so broken, it was impossible to use.  There are posts about the "death spiral" autopilot on reddit and the official forums, so I doubt that it's user error.

[desbean 357]

40 minutes ago, jabloomf1230 said:

Just when you think it's safe to go into the water... I tried flying the Caravan for a fly flights and guess what? Spiral dive into the ground. The bug apparently isn't specific to any user aircraft, but I've never seen it happen on a flight where I didn't use the autopilot. I don't remember this happening so often with earlier MSFS releases, but maybe that was because the autopilot was so broken, it was impossible to use.  There are posts about the "death spiral" autopilot on reddit and the official forums, so I doubt that it's user error.

I have issues with the Caravan AP so I try to get to a safe altitude before engaging, but while on a stable well trimmed manual climb, just barely moving the stick to the left or right to correct drift, and the nose drops dramatically. You expect to have to adjust trim some during a turn, but not a big dive from a slight adjustment. So yes, issues are there besides AP.

Edited October 3, 2020 by desbean

[Toneal2 6]

When I have used Active Pause, although the aircraft stops flying, the IAS runs down to zero so when try to start flying again the inevitable stall and crash.

I am new to glass cockpits so I have problems should I want to do things in the air without being able to pause the flight. Meantime I enjoy the sim, find (for my old brain) a welcome challenge, trying to build my hand flown experience (last night from Anda to Florø along the fjords at low level to the inevitable dire landing).

Edited October 3, 2020 by Toneal2
Added a bit.

[Moria15 950]

11 hours ago, Bert Pieke said:

It could indeed be that my Yoke has drifted out of calibration  🙂

I have re-calibrated and set a 5% deadzone for ailerons and elevators.

I might have expected the AP to switch off, if the controllers are out of whack.. but it stays engaged as the airplane banks and then dives..

I doubt this is anything to do with it, but I have seen this once since the last patch in my 50+ hours of flying since 1.9.3...  and managed to create it a couple of times.

It was in the cessna 172  and I was setting up for an approach to a visual runway, so was slowing down andgetting ready for flaps...   I let the speed get a touch to low.. around 50 knots  and I saw the A/P trying to compensate by hammering the nose trim up to maintain height...   of course that made it worse and the speed dropped further...   next thing I was into a stall and dropping to the right like a rock...

When I was trying to recreate it...  I let the system stay on A/P at a speed of about 60  and then flicked off A/P..  by this time the trim is pretty high on the nose, and flicking off the A/P suddenly results in a straight nose up to a stall to the right.

I have found there is a distinct chance of this happening if you let the speed drop below about 75 in the 172 with any form of flaps...

Now whether this is relevant to your deathspiral or not, I have no idea I am afraid but as I say it is the only time I have succumbed to this since the patch. 

Graham

Edited October 3, 2020 by Moria15

[Muklum 28]

14 hours ago, Mike S KPDX said:

During active pause, all aircraft functions continue except movement.  This includes flight velocity vectors the plane was experiencing just prior to the pause.  So, if the plane had any variation from exact level flight (not really doable in FS), then that continues.  So, for example if the nose was pitching down at a rate of 1 foot per minute, and you active pause for 15 minutes, your rate will be 15 feet per minute on the unpause.  So extrapolate that if you are climbing at 50 feet per minute, you are now at 750 FPM and possibly stalled.

On top of this, input device deflections follow the same rule, this is why is it advised to a full 'swipe' (deflection) of all control axis prior to the unpause.

This would agree with my observations, and confirms my feeling that Active Pause is of no interest to me.

[jabloomf1230 2,445]

2 hours ago, Moria15 said:

When I was trying to recreate it...  I let the system stay on A/P at a speed of about 60  and then flicked off A/P..  by this time the trim is pretty high on the nose, and flicking off the A/P suddenly results in a straight nose up to a stall to the right.

Thanks. This is about 90% of my death spirals also. 

[GACSavannah 227]

On 10/1/2020 at 1:15 PM, RobJC said:

I saw something similar. I was on AP and decided to use active pause. When i took it off spiral dive. It could be a feature. 🙂

I have experienced this one as well.

[eslader 1,791]

Got the spiral dive again this morning in the Longitude. It was not a pitch up issue because I didn't have alt hold on, because of the stupid porpoising that the Longitude does at high altitudes. I active-paused to go do something for my wife, maybe 10 minutes. When I came back, it immediately went into a spiral dive. I turned AP off and it stabilized. AP back on, immediate sharp left roll into a dive. Went into external view and the ailerons were fully deflected, so it's not a stall, it's being commanded by the AP. Even if I turned nav hold off, it still yanked the ailerons full left when the AP was engaged.

[Bert Pieke 7,585]

I wonder if switching between cockpit view and external view has anything to do with this..

Still not sure what causes it.. ☹️

[eslader 1,791]

I don't think so. I never switch to external view unless something is going wrong that I can't explain. I only switched *because* it initiated the spiral dive, to see if the control surfaces were operating or if it was some weird air problem.