Auto pilot pitches down

pgtuck

Member
Joined
Feb 3, 2020
Messages
37
New install of dual skyview touch with AP panel and knob panel. Center stack is garmin gtr 200, GMA245 and, GNC355. AP panel is configured with auto trim for pitch only. When aircraft is straight and level autopilot is engaged roll hold and pitch hold. Within 10 to 20 seconds the pitch slip warning comes on the screen. Autopilot pitches nose down to about 400ft/min. A couple seconds later the same thing. Additional nose pitch down. Now vertical speed, shows about 800 ft/min down. This nose pitching down will continue without ever stopping. During this activity the skyview screen starts asking for trim nose down. Flight returns to normal after cancelling auto pilot.
confirmed wiring is correct on all connections. Three times. Returned pitch servo to Dynon for Inspection. No issues. Swapped out pitch and roll servos. Issue continues. Nose pitches down in simplified and expert modes. AP panel was swapped out with another AP panel from known aircraft in which it performed perfectly! Same issue. Roll servo hold course pretty well. I do geat a lot roll slip notices but. The aircraft stays perfect on roll course.
Dynon currently has flight data down load for review. I’m stumped ! Dynon’s help is good but scarce. Usually takes a couple days just to get a question asked. Any ideas would be greatly appreciated. Thanks in advance!
 

DBRV10

Active Member
Joined
Jun 15, 2008
Messages
925
Location
Brisbane, Qld. Australia
I assume that you did the TRIM CALIBRATION.

When it asks you to trim Nose Up that means the trim tab must go down, and when it asks for trim Nose Down that means the tab itself must come up. It is possible that your system is back to front due to being wrongfully calibrated.

I would suggest you do this again just to be sure. Plenty of people get this wrong because the trim tab works in the opposite direction as it is used to deflect the elevator.

Outside of that there is a bug in the trim sensing for pitch, which has been around for almost 10 years. I had been working on this with the engineers at Dynon, but due to its random nature we could never replicate it and if it could not be replicated on the bench then it was near impossible to find. Last year, about this time in fact, Lawrence had an intern working for him and he tasked them to try a bunch of things to replicate my bug. By then they had a couple of others also reporting the same issue, and by shear luck they managed to do it. Once they replicated the issue they quickly narrowed the bit of code that was randomly behind it and fixed it. I have been flying the beta test software since September last year with zero complaints and a much improved vertical nav performance. This is incorporated in the next software release which is due soon.

But please check your calibration again.
 

pgtuck

Member
Joined
Feb 3, 2020
Messages
37
I assume that you did the TRIM CALIBRATION.

When it asks you to trim Nose Up that means the trim tab must go down, and when it asks for trim Nose Down that means the tab itself must come up. It is possible that your system is back to front due to being wrongfully calibrated.

I would suggest you do this again just to be sure. Plenty of people get this wrong because the trim tab works in the opposite direction as it is used to deflect the elevator.

Outside of that there is a bug in the trim sensing for pitch, which has been around for almost 10 years. I had been working on this with the engineers at Dynon, but due to its random nature we could never replicate it and if it could not be replicated on the bench then it was near impossible to find. Last year, about this time in fact, Lawrence had an intern working for him and he tasked them to try a bunch of things to replicate my bug. By then they had a couple of others also reporting the same issue, and by shear luck they managed to do it. Once they replicated the issue they quickly narrowed the bit of code that was randomly behind it and fixed it. I have been flying the beta test software since September last year with zero complaints and a much improved vertical nav performance. This is incorporated in the next software release which is due soon.

But please check your calibration again.
Calibrations for Trim servo and pitch, roll servos checked and confirmed operating by the book.
 

pgtuck

Member
Joined
Feb 3, 2020
Messages
37
When nose pitches down I wanted the pitch trim indicator on skyview . It does not move. I have never seen the auto trim actually move.
 

pgtuck

Member
Joined
Feb 3, 2020
Messages
37
Another interesting fact. These same servos were in the same plane with my old panel. Servos were connected to D10A and AP 74 panel. Servos were sent back to Dynon for software updates to match the skyview system.
 

pgtuck

Member
Joined
Feb 3, 2020
Messages
37
Should have said “ watched the trim indicator on skyview. It never moved”
 

DBRV10

Active Member
Joined
Jun 15, 2008
Messages
925
Location
Brisbane, Qld. Australia
Should have said “ watched the trim indicator on skyview. It never moved”

Exactly what were you watching? the Trim position indicator like this one?
1593082723939.png



Or this one?
1593082619920.png
 

pgtuck

Member
Joined
Feb 3, 2020
Messages
37
Watching trim position indicator. Auto pilot setting was TRK hold and ALT hold. AP was calling for trim nose down while AP was pitching the nose down!
 

kellym

I love flying!
Joined
Sep 29, 2013
Messages
272
Another interesting fact. These same servos were in the same plane with my old panel. Servos were connected to D10A and AP 74 panel. Servos were sent back to Dynon for software updates to match the skyview system.
I think you may be looking in the wrong place. First thing I would try is turning OFF the auto-trim. See if the system then outputs to the screen a call for trim change, when the plane is trimmed. Also, check that your trim indicator shows neutral when your elevator trim is flush with the elevator.
My two suspects are the A/P panel and the trim motor itself. (trim motor provides trim position information to the Skyview).
Something is telling the system it is out of trim, when it is not. Servo motors do not do that, they just exert a force when commanded to do so.
 

pgtuck

Member
Joined
Feb 3, 2020
Messages
37
Trim indicator shows neutral and trim surface is indeed level with elevator. With auto trim off autopilot will still show trim required. It starts with pitch slip then goes to trim nose down. The nose pitching down is still there. Autopilot panel was removed and replaced with a Known working panel. Aircraft still had nose down pitching.
 

pgtuck

Member
Joined
Feb 3, 2020
Messages
37
More interesting info. I set the auto pilot up for roll heading hold. Pitch is turned off. I use trim button to trim aircraft for a specific altitude. aircraft flys great. Occasional roll slip but quickly goes away. Altitude holds +/-20 ft. Turn the altitude hold on and, within 10 -15 seconds autopilot starts pitching nose down again.
 

DBRV10

Active Member
Joined
Jun 15, 2008
Messages
925
Location
Brisbane, Qld. Australia
The Trim indicator will not show such small movements.

You are suffering the bug I think.

Send me an email to davidbrown ' at' advancedpilot.com

I will help you out a bit. I need to know your aircraft type also. This will be a temporary fix before the new software release.
 

jkturner55

New Member
Joined
Sep 30, 2019
Messages
4
I have the same bug and could sure use you help as well. My autopilot exhibits the exact same characteristics. I have tried everything that I can think of and then some! It still wants to pitch down when engaged. I have had this installed for a few years now and have probably put 20 hours of flight time into trying to fix this issue. Any help would be much appreciated. I'm questioning my decision to remove my Century 2000 AP which worked fine. I know this has to be something that can be resolved (since so many are out there working well), I just haven't found the solution for mine yet. The aircraft is a Wheeler Express CT.
 

DBRV10

Active Member
Joined
Jun 15, 2008
Messages
925
Location
Brisbane, Qld. Australia
Here is what I wrote in a recent email.

First of all, you need to set the pitch servo up on the second inner most hole on the arm (middle hole) and it MUST have the travel limiting bracket installed. I have seen an RV6 where the servo can go over centre. That could be very bad. This gives good resolution for the position feedback. Bottom line is you need to have no risk of over centring and have the inner most hole possible. Usually that is the middle hole. (Unless it is a capstan)

Try these settings
Torque 100
Sensitivity 16-17
Pitch gain 1.6
Alt Gain 0.9
Pull rate 2.0
VSI Gain 1.5
G error gain 1.2
G error limit 0.7

Also your auto trim needs sensible numbers You want 100% speed below say 100 knots, and no more than say 35-40% speed above 130 knots.

See how you go with that. If you still have pitch down issues you also have the elusive and rare AP pitch bug which is exterminated in the next software release. I have been flying the BETA TEST version which also has better and tuneable GPS and ILS vertical parameters since September last year. It's awesome. We just need to wait for V16 to be released :)
 

jkturner55

New Member
Joined
Sep 30, 2019
Messages
4
Thanks for the response. My servo attachment is currently setup on the middle hole like you describe. I'll try these settings and see how this works out.
Thanks!
 

jkturner55

New Member
Joined
Sep 30, 2019
Messages
4
David, who did you speak with at DYNON regarding the beta test on the software upgrade? I would certainly like to do the same. I need to get the altitude hold to work. I was one of the original beta test planes for the Avidyne IFD 540 so I’m familiar with the process, any help would be much appreciated.
 

DBRV10

Active Member
Joined
Jun 15, 2008
Messages
925
Location
Brisbane, Qld. Australia
David, who did you speak with at DYNON regarding the beta test on the software upgrade? I would certainly like to do the same. I need to get the altitude hold to work. I was one of the original beta test planes for the Avidyne IFD 540 so I’m familiar with the process, any help would be much appreciated.
Unfortunately the engineer who solved the bug and did the AP development, has moved into a new aerospace consulting job, after almost 20 years at Dynon. So I cant refer you to him. Maybe email Don Jones at the support email address and ask.

I can only hope that the new release will be out shortly, and if OSH was on this year it would have been timed for that. It really is time for the next version. Maybe the spying eyes of a Dynon employee might comment, but they do keep their cards close to their chest.
 
Top