RV-9 with Dynon AP servo beta test

[Dynon]

Michael,
We are aware of the increased leverage the arm has as it approaches perpendicular to the pushrod and will be taking this into account.

Bill,
We never really tried hard to prevent engagement on the ground. An install should not have an issue hitting the stops, so our only concern was the AP turning itself off before you were off the ground, not when you were taxiing. Airspeed achieves this. We use the same logic in our D10A/D100 products that do not always have GPS data.

 

[aviationone]

The roll servo arm in the "reversed" position, jamming the controls.

The bent control attach plate.

Are you missing the two AN960 washers one under the head of the bolt and the other one under the control arm bearing. That nut looks like it may have bottomed out on the shoulder of the bolt. Hard to tell from the pictures but looks like over three threads.

 

[TThurston]

Pictures are important!

If the bracket is important, please include it in the drawing. Telling me to do something but not including it in the picture that shows what to do seems like you don't really mean it. Remember, "A picture is worth a thousand words."

Bill,
Thanks for bringing this to our attention. Clearly we're concerned to hear about any failure like this.

As has been identified, this is why we recommend the over-center limiting bracket on every install. Even in an install that appears that it can not go over-center, the bracket is a great backup to unexpected failures. While it isn't in the drawing, it is in the text requirements for the install and included with every servo we ship.

Our mechanical team at Dynon is looking over the install and doing some of our own tests to see if we can re-create this issue using only the strength of the servo (about 4 lb/ft of torque).

A few bits of misunderstanding to clear up.

1) This was not in a test mode. You can engage the AP on the ground, by design. At zero airspeed, the AP can be engaged. This allows you to still use the AP if your pitot system fails and goes to zero. So when you engaged on the ground, you engaged it in a standard mode and it attempted to roll the plane to the heading, track, or course you had selected. The AP will disengage when the airspeed hits 20 knots, so you can't take off with it on accidentally.

There is a self-test mode which can only be accessed via menus in the system, not in any kind of normal operating mode. This does not cause the controls to travel to the limits.

2) The shear screw does not exist to limit the force of the servo. If it did, you could break the shear screw with your knee when you reached over to grab your lunch. The shear screw is stronger than the servo motor and is only there to protect from a geartrain failure. In normal operation, the servo motor itself is limited in torque, and can be overridden without breaking the shear screw. Just grab the stick and override it. So it is not unexpected that the shear screw didn't break.

 

[Toobuilder]

Michael,
We are aware of the increased leverage the arm has as it approaches perpendicular to the pushrod and will be taking this into account...

I suspected as much, but your post specifying a finite force value gave me pause. Thanks for closing the loop.

 

[krw5927]

As has been identified, this is why we recommend the over-center limiting bracket on every install. Even in an install that appears that it can not go over-center, the bracket is a great backup to unexpected failures. While it isn't in the drawing, it is in the text requirements for the install and included with every servo we ship.

I see that the documentation for the roll servo contains the words, "We recommend the use of the optional Range of Motion Limiting Bracket..."

However the very first sentence in the Range of Motion Limiting Bracket instructions has the word "optional" in bold text. I took this to mean that there were no possible adverse conditions that would occur if I omitted the bracket, and I suspect many other builders did the same. Perhaps the word "optional" should be changed to "recommended" for consistency with the servo installation document. Not that I expect anyone to be responsible for my choice to not install them, but I probably would have installed the brackets had the documentation said "recommended".

As it is, I'm unable to locate the ones that came with my servos and will be calling Dynon on Monday to inquire about purchasing replacements. And as always, I am ever thankful for the outstanding Dynon customer service, as I know they will thoroughly look into this potential issue and get back to us with their findings.

 

[Low Pass]

After looking, thinking and evaluating and you feel the travel limiting bracket is necessary, put it on. You want a parachute, put it on. Etc.

 

[flightlogic]

Kurt. glad you are going to put the limit brackets on... from Dynon.
They do a great deal of behind the scenes work to make things work safely.
Your post, the photos... the discussion will also lead to further awarness to the RV community at large. (and other airframes, I suspect) You have done a service for people you will never meet.
Good on ya'.

 

[N941WR]

Are you missing the two AN960 washers one under the head of the bolt and the other one under the control arm bearing. That nut looks like it may have bottomed out on the shoulder of the bolt. Hard to tell from the pictures but looks like over three threads.

No, the purpose of the 960 washers is to keeo the rod end from falling off, if it fails. That said, there would be no harm in adding them.

Go back and read the original post, the servo slammed the controls a few times, bending the plate and allowing the servo arm to go over center, jamming the controls. This happened on the ground, not in the air.

 

[N941WR]

...
Bill,
...
We use the same logic in our D10A/D100 products that do not always have GPS data.

Even better, change the code to say, "if IAS=0 and GPS_AS=0 THEN Don't engage AP". On a D100 W/O a GPS airspeed, it should be zero and the AP will only engage when either the indicated OR the GPS airspeed is not zero.

 

[aviationone]

No, the purpose of the 960 washers is to keeo the rod end from falling off, if it fails. That said, there would be no harm in adding them.

I was talking about the 960 washers not the 970 washer. Not saying that was the issue but just looking at the install photo it does not match the Dynon install diagram. The photo looks like the nut may have been bottomed out. But hard to tell from photos. It was just an FYI
960 washers circled in black below.

 

[N941WR]

Sorry Tom, my mistake. The small washers are in there, just difficult to see.