Servo calibration failed

Eugr

Member
Joined
Sep 12, 2020
Messages
57
I’m finally close to wrapping up the installation in my Bonanza, but I couldn’t get past servo calibration.
When I look in the network settings, I can see all 3 servos in the list of the devices with active status.

Then when I start calibration, it says that 3 servos are found:

D03AED84-F471-4E5A-A86A-09A2A7EAB35C.jpeg


Then I follow the steps but it fails at the very end:

0AFBB45F-B3A5-4470-B863-48F818668536.jpeg


I set up V-speeds, tail number and performed zero airspeed calibration before calibrating servos.

I held rudder pedals in place when calibrating pitch and roll and yoke centered when calibrating yaw. However, since Bonanza has aileron-rudder interconnect, I wasn’t able to hold it absolutely still, so other servos probably moved a small amount when calibrating.

Any ideas?
 

Eugr

Member
Joined
Sep 12, 2020
Messages
57
I also realized that the my autopilot servo power was wired to the avionics bus, not master. Does it make any difference? Is it better to keep the servos on master bus like on the diagram? When trying to calibrate, I had avionics master on and the servos were visible on SkyView network.
 

DBRV10

Active Member
Joined
Jun 15, 2008
Messages
925
Location
Brisbane, Qld. Australia
Perhaps do not hold the rudder tight, due to the inter-connect it might be loading up the servo in its test.

Did you get any page saying travel over X degrees do you have a capstan installed? If s what did you answer?
 

Eugr

Member
Joined
Sep 12, 2020
Messages
57
Perhaps do not hold the rudder tight, due to the inter-connect it might be loading up the servo in its test.

Did you get any page saying travel over X degrees do you have a capstan installed? If s what did you answer?

I held it pretty tight, but it’s impossible to prevent any movement, other than locking rudder mechanically. When calibrating yaw I even put the control lock in place, but it didn’t help.

I didn’t get any messages about travel over X degrees, that would indicate successful calibration. Once I followed all steps it went directly to the screen that said the calibration failed (see screenshot in my original post).

I guess next step would be checking the shear pins (but I’d be surprised to see it broken on a brand new servo, plus I believe we inspected all of them before installing).

Is there any way to see some raw data from the servos, similar to the sensor page? Otherwise the only way to detect which servo is faulty would be disconnecting them one by one from the hub and running the calibration again. Like disconnect yaw first, then run with only pitch and roll. If that fails too, disconnect pitch and run with roll only (although I’m not sure the calibration will succeed with just one servo connected as Dynon doesn’t support roll-only configurations).
 

DBRV10

Active Member
Joined
Jun 15, 2008
Messages
925
Location
Brisbane, Qld. Australia
I am strongly suggesting not to hold the rudder tight at all. I have never done a Bo installation but the interconnect could be the problem. It might be loading the servo up when it is not expected.

Centre the controls and keep your feet off the pedals.
 

airguy

Active Member
Joined
Nov 10, 2008
Messages
921
Location
Gods Country - west Texas
Agreed - from your first photo, see the instructions on the screen - "ensure free range of motion" - that means hands and feet inside the car until the ride has come to a complete stop. Do nothing to interfere with the flight controls - see if they do their thing correctly.
 

Eugr

Member
Joined
Sep 12, 2020
Messages
57
Well, Bonanza specific instructions have these two notes:

NOTE: Hold the rudder pedals centered in yaw during the pitch and roll portion of the calibration or the calibration will fail. This prevents the yaw servo from moving when performing the pitch and roll servo portion of the calibration.
NOTE: Hold the control yoke centered in pitch and roll during the yaw portion of the calibration or the calibration will fail. This prevents the pitch and roll servo from moving when performing the yaw servo portion of the calibration.

That makes sense, as if you don’t keep rudder centered, it will move with roll inputs due to bungee interconnect.
 

Eugr

Member
Joined
Sep 12, 2020
Messages
57
Agreed - from your first photo, see the instructions on the screen - "ensure free range of motion" - that means hands and feet inside the car until the ride has come to a complete stop. Do nothing to interfere with the flight controls - see if they do their thing correctly.

Ensure free range of motion means that nothing is blocking the controls during calibration. This is not the test phase, the servos do not engage while calibrating. You just move controls as guided.
 

Eugr

Member
Joined
Sep 12, 2020
Messages
57
An update: I came up with the way to find the faulty servo.

I disconnected pitch and roll, reconfigured network and and did yaw only calibration. It worked.
Then I connected roll and pitch, but left yaw disconnected. Reconfigured network and ran with two servos. Failed.
The I disconnected pitch, but connected yaw instead. Reconfigured network and ran calibration using yaw as pitch input. Calibration was successful.

So the culprit is pitch servo. The shear screw looks intact, so I guess I’ll have to get a replacement servo from Dynon.
 

Dynon

Dynon Staff
Staff member
Joined
Jan 14, 2013
Messages
14,217
Location
Woodinville, WA
So just for reference, to help understand the "programming", what the servo detect/cal routine is doing is detecting which servo is which, and which direction is which for each. And that's it. So, for each requested movement, it's looking for one axis to move significantly, and for all the others to not. If more than one servo moves, or if none are detected as having moved, then the calibration routine will fail.
 

Eugr

Member
Joined
Sep 12, 2020
Messages
57
So just for reference, to help understand the "programming", what the servo detect/cal routine is doing is detecting which servo is which, and which direction is which for each. And that's it. So, for each requested movement, it's looking for one axis to move significantly, and for all the others to not. If more than one servo moves, or if none are detected as having moved, then the calibration routine will fail.

Yep, that‘s what I figured too. Still would be extremely useful to have a debug mode similar to sensor raw data where it just shows position/force on servo #1, #2, #3. This way if one servo is faulty it would be easier to detect.
 

Eugr

Member
Joined
Sep 12, 2020
Messages
57
So, I’ve got the replacement pitch servo from Dynon, and after we swapped the old one for the new one everything worked as expected and autopilot passed calibration and test. So the Dynon portion is done, just need to address a couple of unrelated issues, do a new W&B and paperwork. Getting there!
 
Top