Jul 1, 2016 – AutoPilots, Open Doors, and other drama

From earlier posts, you know I am going down a path of learning to do supervised maintenance on my own airplane. Recently, that has included removing and re-installing the pitch trim servo for my auto-pilot. That work necessarily includes learning how to remove and replace my beautiful interior without diminishing it. All of which develops a new respect for what an airplane A&P is able to do routinely.

DOING YOUR OWN MAINTENANCE comes with the benefit of better understanding your airplane and it’s failure modes, so they tell me. That certainly seems to be the case after this maintenance experience. The symptoms I experienced when the servo failed are now in context, and a theory has emerged as to the cause of the servo failure in the first place.

Additionally, during the post-maintenance test flight, I experienced serious unexpected consequences when an improperly secured door opened during a test flight. I always secure that door myself, but had ceded that responsibility to a knowledgable person in the right seat. He just wasn’t familiar with the precise method required to ensure that this particular door was secure, and I didn’t reach across to back him up.

Finally, I was reminded of a silent failure mode that could seriously bite me under some circumstances. There is no annunciator associated with my A/P, so it can stop working without you knowing about it.

Experience is a series of non-fatal mistakes – author unknown.

20160617 PitchTrimServoMY PITCH TRIM SERVO FAILED recently. Earlier this year I did an Angel flight down to Charlotte and back. That was a challenging flight in that I flew to Norfolk for the pick-up in IMC, knowing that the destination was below minimums and would be slow to improve. I explained to the passengers – a young mother and her son – that we’d have to wait until I could be certain we’d be able to get them into our destination airport. The risk to all of us was more of convenience – flying all the way there, only to have to return to the airport we’d left after being bounced around for 3 hours.

The weather started creeping up towards the minimums we needed, and I decided to launch on the trend and keep an eye on the progress as we went. We were flying along fat, dumb, and happy at 8000′ and between layers – watching the XM Weather to ensure continued improvement. The plan was to turn around at the first sign of worsening weather, and continue if I thought we could get in. I was very well focused on playing the weather game – diverting around cells and planning new alternates – when suddenly the nose tucked down 3 degrees and lifted us out of our seats! Not good!

I instinctively clicked off the auto-pilot and re-trimmed for level flight while we continued on our way. My experience to date told me two things: 1) this problem might be intermittent and you don’t throw money at problems that won’t likely be reproduced on the ground; and 2) something in the auto-pilot might have actually failed since it has given me years of solid service and I was probably due.

I knew my passengers were familiar with light airplanes from talking with them thus far, so I discussed with them the possibility of doing some testing along the way. Mom was very cool about it, after I explained what I expected might happen. The little boy in the back wanted more zero G’s anyway, so he was in! I reset and re-engaged the auto-pilot.

The flight continued with it on, and I even flew a fully automated approach to minimums in considerable wind with it engaged. The problem did not recur, though I was on a hair-trigger watching it for the next few flights. Eventually I stopped thinking about it and relegated the event to an anomaly.

In late April I was flying the airplane out of Easton for S37 – Smoketown airport to get some warranty work done on my paint job. You’ll remember a previous post regarding a warped aileron that made me an expert at getting a light twin in and our of this skinny, obstacle encrusted airport. Leveling off at 3000′ on the way up there, I turned on my auto-pilot and called for flight following. Just as Potomac Approach called me back, the nose pitched down aggressively – probably 5 degrees this time. I was lighter on this flight and felt it more this time. This was starting to piss me off!

With that on my mind, when I picked up the airplane after the paint work was completed, I stopped in at Delaware Coastal (KGED) and the Delaware Aviation Museum to talk to my friends about it. These guys got me hooked up with an expert on autopilots down in South Carolina who, based on their recommendation, spend time on the phone telling me how to test and isolate the problem. I built a test flight plan based on what he had told me, and Matt and I would fly that a day or two later (see the post on that flight published recently). Getting direct support from these folks was invaluable.

It took three tries for AutoPilots Central to get the rebuild done correctly. Approximately three weeks later the overhauled unit was re-installed and testing begun. Initially we found, Matt and I, that the trim system now had too much friction in it. The cable tensions were correct now, but may not have been when the servo was taken out. I’d have to put some force into moving the trim now, and knew something wasn’t right. Looking at the entire system, Matt found that two of the pitch trim pulleys had frozen in place and flat spots had been being rubbed into them by the pitch trim cable. Glue used by the interior installers was applied too liberally in that area, and had gotten into those two pulleys.

I freed those pulleys up, and added them to the list for replacement at annual. I believe that the flat spots on these frozen pulleys contributed to the symptoms I encountered in flight. The theory is that the pitch trim cable system was bound by these pulleys, and it took a larger force to move it past that flat spot. This circumstance may also have contributed to the premature failure of the servo.

Happily, the servo was put back in and Matt and I went for a test flight with the interior left off until we were sure. Out of Easton we headed over the Delaware Coastal for a few approaches to ensure the trim servo was working well in all modes of flight. Level at 3000′, our speed built towards 170 kts when the door popped open. I’d never experienced that in this airplane, but expected it to be a non-event based on all I’d experienced and read about previously.

I HAVE READ ABOUT DOORS POPPING OPEN on a Bonanza pilot in particular, where the event was so distracting that a fatal accident resulted. Armed with this information, I was ready when doors popped open while I was giving flight instruction in Cessnas, Pipers, and even in a Cirrus SR20. It can happen even to the most diligent, but by now I knew to continue flying the airplane and don’t distract yourself trying to close the door. It is a non-event in many aircraft if you stay focused. Nothing more than wind and noise to have to contend with.

HAVING THE DOOR OF MY TWIN COMANCHE POP OPEN IN FLIGHT is like nothing I’ve encountered in these other airplanes. When it popped open, the airplane began to noticeably shake, and the yoke shuddered in my hands. Matt was trying to close the door, but I stopped him because I didn’t think it possible; and also didn’t want him to bend any metal trying.

The shaking was a significant concern, so I dropped our plan to fly an approach from the waterloo VOR and set up for a left base entry to the longer runway 4. I reduced power through ranges down to landing speeds – wanting to be sure the airplane would keep flying with all the shaking going on. I was also looking for a speed that would reduce the shuddering impact on my airframe.

Nothing seemed to make the vibrations any better, but at least the airplane flew through all speeds well. I could land normally, but needed to get on the ground ASAP. I changed our plan again and opted for a straight in to runway 28. Matt opened the storm window on his side, and used the leverage to position the door so as to minimize the vibration. He could see something that I could not – the tail was visibly shaking in the turbulence caused by the open door, and he was concerned. I was too busy learning to fly like this.

We landed just fine and the shaking finally stopped. I reviewed with Matt how to ensure the door was secure for the next go, and we taxied back – none the worse for wear. Were this to occur while flying solo. this situation would be more difficult to manage. Flying solo and in IMC I would have to consider as an emergency. You’d be justified in getting priority handling to get back on the ground. I can say that with certainty that I would have declared an emergency were this to have occurred in IMC.

FINALLY, I WAS REMINDED THAT MY AUTO-PILOT HAS NO ANNUNCIATORS ASSOCIATED WITH IT. During the last post-maintenance test flight, I allowed the airplane to trim itself via the auto-pilot while I set a new airspeed. Once the speed stabilized, I would turn off the entire auto-pilot to ensure that the trim servo was doing it’s job. If the airplane pitched up or down, the pitch trim servo stopped short of a neutral setting and those pulleys might be causing a problem now. For this test though, the airplane was rock steady after A/P disconnect, and didn’t change pitch at all.

I remembered that the airplane is so well rigged and so stable that I once had trimmed for cruise flight and simply forgotten to turn on the auto-pilot. I didn’t discover that the ship was flying itself until I monitored a turn in our plotted course and watched while the airplane didn’t respond to the turn. The problem, I discovered, was that I’d been distracted from ever turning it on! Altitude and course had been rock steady on it’s own.

It got me thinking that a circuit breaker could trip, disconnecting the autopilot silently. I’d never know the airplane was flying itself unless I was looking for it. I need to add an annunciator of some sort when I do the ADS-B work.

Fly safe – never stop learning.

Never stop flying – even when the auto-pilot has it.