LOBO Training Devours the Swiss Cheese

[Dan O'Brien]

I had an incident report on Avweb. Here's what happened. It's a lesson in Swiss cheese, and how good training can devour it.

1. Got up and saw the weather indicated I need an instrument plan to get from Leesburg to Milwaukee (KJYO to KMWC). (Swiss cheese hole number 1). Filed. Cool. Two weeks prior I had flow a bunch of approaches in a simulator. Ready.

2. Everything was normal. Beautiful 1700 foot ceiling (LOL), and roughly 5000 foot tops. No ice. Happy to fly instruments whenever there is no ice. Otherwise, NOT!

3. Door, closed and latched. Nice over-center link click. Inflated the door seal, did whole checklist, all normal. Everything was normal. What a great airplane. I'm flying when I want to see my daughter in Milwaukee, heading back to the west coast. What a country!

3. I had just completed a condition inspection about four flights before. Everything was generously lubed among the other things we do a condition inspection. Door operated like the smoothest lever action rifle ever, way smoother than before. (I'm not a gun guy, but my uncles from WI are and brag about the action in their deer rivals.) The action was that way for flights from Redmond OR to Truckee CA, Truckee to Casper WY, Casper to Cumberland WI, Cumberland to Leesburg VA. And the door latched properly every time. Different tactile feel than before, due to lubing, but I was like "hey, this is great.". (Swiss cheese hole number 2).

4. Plane's running beautifully with new Tempest fine wires, timing set right, higher fuel flow than before (I wanted that), all great. Normal taxi, normal run-up, normal everything.

5. Normal takeoff. Turned to first heading, hit soup about 1700 on the way to first assigned altitude of 3000. I stayed off my autopilot because I like hand flying practice and wanted to do it through the soup to 5000. (Swiss Cheese hole number 3, maybe). ATC vectored me around a bit, and then instructed me to turn on course and climb to 10,000, which was my planned cruise til I got to Lake Michigan, where I would climb for glide distance over the lake. I turned on course to destination and continued climbing, and then, suddenly, the door was gone. (Swiss cheese hole number 3)

6. I'm like "WTF, am I in the twilight zone?" Well, skip all that Dan, and yes, you are in the twlighlight zone and need to deal with it. I declared an emergency. They asked if I wanted to return to Leesburg, and I said yes (now wish I had gone outside the Washington TFR). They gave me a vector. I was pretty calm. After a few seconds, I realized the plane was flying OK and responding to controls, so all's good, right? I was squarely in the soup, but I have good instruments, a good autopilot, and a half decent old school scan (though I'm on glass), and just did a bunch of practice approaches. All will be fine. And then ...

7. So far we have three elements Swiss cheers holes: IMC, my choice to hand fly, and a door failing to latch (more on that later) and departing the airplane. Here's the fourth hold in the Swiss cheese. I take off at a maximum of 25.5 inches MP, which near sea level is less than full power. Theory is that it keeps cylinder head pressures on 10:1 cylinders at or below their max with normal 8.5:1 cyllinders. Sounds good to me. But maybe about three times over 465 hours flown in my ES I didn't have the friction lock set right and the throttle slid out a bit (seems to happen more when it's partially out to begin with.) I always noticed and corrected it pretty quickly in VMC, but this was IMC and...and...and...it was rather NOISY in the cockpit (could barely hear ATC), and I was rather busy after the door exited stage left (actually stage right, more on that later), busy declaring, talking to ATC, trying to load a return airport and instrument approach, trying to set up my autopilot (I realized discretion was the better part of valor), and getting vectored around a bit. I got my autopilot set up on some vector, and then realized that it was climbing me steeper and steeper and steeper, LOL. No, don't laugh Dan, it's getting too steep. Before it was too late, I disconnected the autopilot and then realized that my manifold pressure had fallen to ~15 inches because the throttle slid back (I suppose it could have come back when the door departed, I don't know). I didn't got toooooo close to stall---I caught the AOA indicator in my eye when I finally realized what was happening, and lowered the nose and added power in time. All still in the soup. In any event, my throttle sliding back was a fourth hole in the Swiss cheese.

8. After that I was was still calm but a little rattled (four holes in Swiss cheese, one that coulda got me), and then the fifth hole came, maybe a more minor hole. Even when I hit direct to JYO, the first airport in the flight plan, the Garmin wanted to load an approach to the destination airport. I realized that that's what it was doing and know rationally that that's what it does, but when the sh_t is hitting the fan and you're in the soup and not on auto pilot realizing you need to remove the destination and go direct to the departure airport before loading an approach takes more time than you want. I removed the destination, added JYO, then loaded and activated the approach. They were still vectoring me, and for reasons I don't quite grasp as I type, I overshot the course a bit and got a little low, but not radically so. I woulda flunked a check ride. But I got that straightened out by the final approach fix.

9. The rest was pretty routine. Broke out around 1700 AGL, so not a low approach at all, and did one of my better landings with wind making the cockpit noisy.

10. The plane handled well with the door gone. I seemed to need more rudder at times (can't recall it all), but the plane handled surprisingly (to me) normally.

11. Swiss cheese: (1) IMC; (2) decision to hand fly when the autopilot maybe could have done better and eased workload; (3) door departs; (4) throttle friction lock not set righ and airspeed is lost during climb; (5) fumbling with entering an approach to the departure airport because it wasn't set up ahead of time and time was wasted recalling how Garmin works and fixing it. I guess I can take some credit for managing to fly the airplane during it all, although in retrospect, once I figured out that the plane was handling OK, I probably should have climbed above tops, got a vector, stayed on it, and calmly set up the approach. Yeah, that's what I should have done rather than allowing myself to get vectored before I was ready.

12. So what happened? The evidence is very clear. The door was found by some cows, LOL (see attached photo). The door was closed and "latched," but the hooks missed the pins and latched on the door frame on the fuselage. This is obvious from the fact the the door was latched (confirmed upon finding it) and it took a bite out of the glass as it departed (see attached photo). I guarantee I heard a nice over-center link click and thought it was latched before inflating the door seal. That said, I'm gonna take the heat---it was ultimately my responsibility to make SURE it was closed and latched ON THE PINs. I didn't successfully do that. But, to give myself a little break and not beat myself up too much, it is possible that it would have been quite difficult to distinguish between latching on the pins and latching on the door frame skin. I do think this is a flaw in the design, as did the FAA person who inspected the evidence post incident. It should not be possible for the door to latch anywhere other than on the pin. It's worth a discussion here on how to fix that. Glue a piece of delrin or similar that prevents the hooks from getting between the pins and the door frame skin? I dunno.

13. I stayed remarkably calm, and I attribute that mostly to (i) age, and (ii) LOBO training (and I'll also give a shout-out to Will Wobbe and Don Kaye, who trained me in my Mooney and continue to provide insight whenever I ask). Training and practice make you calmer when the sh_t hits the fan. I suppose it helped that flew several approaches in a simulator a few weeks before the incident.

14. An incident like this should be mostly a yawner if you're properly trained and do what you were trained to do. People at Leesburg congratulated me for getting it down, but I looked at my track and remembered the airspeed issue, over-shooting, and being too low on parts of the approach and say this is a C+ at best, maybe more like a C-. But that's a lot better than an F. LIVE, and LEARN.

15. Regarding my decision to stay off the autopilot for the departure. This incident makes me question doing that in IMC. That practice is good, and the blue button is there if there's trouble, but the workload is eased a lot when the autopilot is on. Curious what you all think. Having it off allowed me to figure out how the airplane was handling. But having it on would have made things easier, except for that friction lock hole in the Swiss cheese.

BTW, for those who don't know, the "Swiss cheese" "theory" in accidents is that sometimes holes in the Swiss cheese line up in ways that amount to a chain of events that lead to accidents. I think it's a nice way to think about how accidents happen, as they often involve multiple factors that line up a particular way. I think I've identified the main Swiss cheese elements here.

[J.C. Peterson]

Dan,

I’m glad to hear you’re safe and the airplane is (mostly) in one piece. Well done handling an incident like this.

The photos you mentioned don’t seem to be attached. Can you post them? I’m sure we’re all curious how everything looks.

Thank you.

[Dan O'Brien]

Here are a few photos. It's clear the hooks latched onto the door frame and ripped out part of it in flight. The latches were closed when the door was found, so it's clear what happened. The pic with the cows is kinda priceless. "Moooooooo! What's this? Moooooo".

A part of the Swiss cheese I meant to mention in my long post but didn't is that the door had just been nicely lubed in the condition inspection and was operating smooth as the smoothest lever action rival on the planet. The tactile feel was different, which might have contributed to not noticing failure to latch, I don't know. That said, I flew from Redmond to Trukee, Truckee to Casper, Casper to Cumberland, Cumberland to Leesburg -- all the way across the country -- after the condition inspection, with the door latching each time. It felt the same each time -- very smooth -- and it felt the same when departing Leesburg. But for some reason, the hooks didn't grab the pins, for the first time in ~500 hours. The grabbed the door frame.

My view of this is that either it should not be possible for the hooks to grab the door frame (e.g., glue a piece of delrin or similar to prevent latching over the frame),, or there should be a indicator light saying whether the door is closed and latched, maybe a micro switch that shows green when the door is closed and latched and shows red when it isn't (or something like that). Because I believe the door grabbed the door frame tightly enough that it was hard to tell that it wasn't latched. I'm not sure about that though. It's hard to remember how hard I pushed against it to check whether it was properly latched. One would think there would be noticeable play if it didn't grab the pins, but I'm not so sure about that. I might have had to break the darn door on the ground to detect the play!

Soon enough the door will be in your neck of the woods. I'm having Ryan Wolstenholme make a mold around it and ship it to Steve Lorentzen, and they will then rebuild it. Surprisingly (to me), the hinges didn't tear out--they are apparently very robustly installed. Instead, the inner skin broke in the upper third as the outer skin delaminated in the upper part and departed the airplane.

At first I was a bit shocked. Then I was sad. Then I was ticked off. Now I'm just longing for my airplane back. Four stages of grief? I dunno.

Happy and safe flying,
Dan

[Rock Barchfeld]

Hey Dan,

I do not know how the latches on an unpressurized ES compare to the latches on a pressurized IV-PT, so this could very well be an apples to oranges comparison, the latches do look very similar, however. In your post, you stated: "My view of this is that either it should not be possible for the hooks to grab the door frame..."
Not wanting a repeat of your scenario, I went to the airport today and looked at my door. Quite honestly, on mine, there is just no way to physically get the hooks past the pins. If the hooks on all 7 of my latches are not fully open, they will hit the top of the pin and the door will not close. Also, no matter what I tried to get the hooks past the pins and onto the frame, there was just nowhere near the clearance required to do so.
Like I said, this might not be relevant to your post, but I thought maybe some others might try the same thing and post their results.
Warning, thread creep. As the pressurized models have two separate latching mechanisms, has anybody ever heard what the minium number of latches it would take to secure the door under pressurized conditions. Engineers tend to design in redundancy, kit designers more so to make up for less than optimum building technique. Just curious if anybody ever heard this topic discussed?

Rock

[Tom Sullivan]

Dan,

Great job handling a very clear emergency !!!!

Is that door the same as the IVP? I bought a "Fast Build Door" when I was building mine and still have the original skins for my door.

I was offered a ridiculous price for it years ago and declined it, as having the spare was worth more than what the guy offered me (and I'm sure he spent more getting it repaired than what the parts would have saved him in labor).

Let me know if you have any interest.

Tom Sullivan

[Dan O'Brien]

I understand that the IVP has 8 latches. Bob Pastusek told me that there is no way to latch it around the door frame—the hooks either close on the pins, or they don’t close, as you are saying. However, my door, which basically copies the non-pressurized IV door, only has 2 latches at the bottom, not at 8 points all around the door, and apparently, it was able to hook around the door frame. Unless the latch wasn’t closed, was held in by the door seal, and then closed when it departed. Could that have happened? My recollection is that it felt like it closed and latched. But memories can deceive. Is it possible that the evidence (the photos, the fact the latch was closed when the door was found) could be consistent with the door latch not being closed? I am puzzling about this, as you can imagine.


Sent from my iPhone using Tapatalk

[J.C. Peterson]

Dan,

The door frames I’m accustomed to seeing on IV-Ps have a bigger cutout for the latch arm to go through. As Bob P said, there is no way for the latch to grab the frame itself, but that it partly due to the interference the upper latches would have unless the door was fully down. On your door frame I could definitely see how that could happen though. It might not hurt to take a good look at enlarging the frame openings toward the outboard side. You may not be able to do much without the door seal losing area to rest on, however.

[Dan O'Brien]

Couldn’t one glue a piece of delrin or similar to cover the space above the pin and below the frame so that if the hooks miss the pin there is no way the hook could grab the door frame? Then if the hooks missed the pins, it seems either the latch wouldn’t close at all, or if it did, it would be grabbing anything so it would be easy to tell that the door wasn’t latched by pushing out on it. Thoughts?


Sent from my iPhone using Tapatalk

[Dan O'Brien]

Dan,

Great job handling a very clear emergency !!!!

Is that door the same as the IVP? I bought a "Fast Build Door" when I was building mine and still have the original skins for my door.

I was offered a ridiculous price for it years ago and declined it, as having the spare was worth more than what the guy offered me (and I'm sure he spent more getting it repaired than what the parts would have saved him in labor).

Let me know if you have any interest.

Tom Sullivan

Tom,

Several people had told me that there wasn’t a fast build door available. But I did not contact you. If I had, I would have bought your door in a heartbeat, flown out of Leesburg with it, and brought it home and finished it. Since I didn’t have one available, I made arrangements to have someone go out and make a template in order to rebuild the existing door. How much easier would it be to scrap that idea and just buy and finish the fast build door? I don’t know what is already done on the fast build door and what is involved in fitting and finishing it.

I would give you a very fair price. You’d make a return on your investment.


Sent from my iPhone using Tapatalk

[Tom Sullivan]

You misunderstood what I said. I BOUGHT a Fast Build Door while I was building, several years after buying the kit. I have the door skin and inner support pieces left over from the original kit that you stated you would have to mold up, which these pieces would still save some considerable time.