Sunday, November 02, 2008

Sorry, Airplane

Have you ever scolded your dog for excessive barking, and then discovered that there was actually an intruder on your property? Or maybe goaded an apparently lazy horse before discovering a problem with the harness, or something caught in its hoof? This is the feeling I have for this poor hardworking airplane. It tried to tell us. We should have believed it instead of maintenance. But we didn't recognize what it was trying to say.

My coworker reports that the airplane ran perfectly for the rest of the day's work, with no surging or abnormal indications whatsoever. She said she was ready to believe I had been hallucinating and then she left to take the airplane south for scheduled maintenance.

Halfway to the maintenance facility, she experienced the same indications as I had. Maintenance determined that that engine was only operating on five cylinders. The poor thing. A four cylinder engine that is down to three cylinders shakes the whole airplane and tells the pilot with certainty that you had better land me soon. This was so subtle we were all willing to believe it was just a gauge. Thanks, engine, for doing so well for those last twenty hours despite your handicap. I promise to give you an extra ration of oil or something (how do you reward an engine?) when you get back.

I think of my engines a bit as if they were horses and I've just figured out why. It's not that we think of them in terms of the number of horses inside that cowling--thought I'm sure that starts the thought processes down the right path--it's because in the wintertime I put engine tents on them. The engine tents are made of the same sturdy quilted material as horse blankets. You walk up to the side of the engine, making sure it's reasonably clean first, and place the blanket high on the withers, um I mean the front of the cowling, pulling it back and securing it under the belly with velcro or other straps. You smooth it back, make sure that it's even and secure, and I can never, at this point, not pat it gently and speak a few words of praise.

My engines work hard for me, and I want to take good care of them. I still have to be careful that they don't kick or bite me though.

I didn't get a chance to say sorry to that airplane, or find out in more detail the nature of the cylinder problem, because while it was in maintenance I was moved on to the next job on another airplane, in another province.


Garrett said...

Most automobiles built in the last 15 years would have diagnosed this fault and given you a warning including the suspected cylinder and a snapshot of the engine parameters at the time of the fault. For a fraction the cost of an aircraft engine, you can get this technology plus a shiny new automobile to test it with.

If you buy a shiny new composite airplane, it will probably come with a set of engine controls that would be refinements of controls known to a 1930's flight deck. If you pay enough money, it might come with a fancy digital engine monitoring system with more data presented than some test cells. You then dedicate a substantial part of your busy single pilot brain to managing the engine with these controls. Single pilot IFR in light aircraft apparently isn't hard enough without additional LCDs to interpret.

Competing hypotheses for this technology gap:
-engine development is expensive, even without certification. GA production volume simply makes it uneconomic without regulatory intervention. The increased maintenance and operating costs resulting from the luddite solution are partially external to the engine manufacturers, and partially in their favor.
-Light aircraft types are actually anoraks who get a kick out of the mixture/EGT game, and rue the day that some boring microcontroller takes away their task.

Anonymous said...

Continuing from what garrett wrote: I recall a Porsche engine available on the Mooney several years ago that used a single power lever - using a similar computer box to the FADEC on modern jet engines (Full Authority Digital Electronic Control - if memory serves).

Here's a detailed article about GA FADEC.

Lynn said...

I find computerized engine controls on airplanes scary. I'm not a luddite--I've worked with computers for the past 35 years. But I've also had a few cars where the engine suddenly died because the engine computer wore out. It was a real annoyance stalling out in the middle of a two-lane bridge during rush hour--but nearly as much of an annoyance as having your airplane engine suddenly stop during a takeoff would be.

Maybe they build 'em better for airplanes. I certainly hope so.


Sarah said...

Very nice couple of posts, enjoyed the view of the relationship between pilot & plane. It is hard to pass by a spinner & not pat it on the nose in thanks for pulling you faithfully.

My nomination for quote of the month from Aviatrix:

You'd be amazed how often airplanes make you wonder if you're hallucinating.

And most often, it seems to be "what is *that* noise? ..." Or, more enjoyably, "wow,look at that view..."

Garrett said...

Lynn, most automobiles are barely maintained, let alone routinely inspected or closely regulated. Engine controls in automobiles aren't inspected until they fail, which is usually appropriate risk management given the context.

In the last 5000 hours of driving, I've had two faults result in a loss of power, both the same fault with a part that had a simple design error. Not a bad record considering the vastly lower bars of cost, regulation, inspection, safety over performance, etc. FWIW, the part that failed would be protected by system redudancy in any aircraft was an ignition coil.

I'm actually all for mechanical solutions in life critical systems. The Therac-25 disaster is a good case study in why I feel this way. (not to mention a good study of all sorts of bad design practices not limited to mechanical v. electronic)

However, there is a tremendous capability and workload gap between light aircraft and heavy ones, that may make the challenges present worthwhile. Pilots are the most common cause of accidents, after all.

Building an electronic engine control system as reliable as the current state of light aircraft art isn't trivial. I recognize that. Some early efforts have been pretty disturbing. See events like the double engine failure of a DA42 on takeoff last year. Poor design, poor oversight by the regulators. One of the specific design faults (poor handling of hardware resets) is something that a modern automobile would have operated through without the slightest drama.

One of the annoying facts of life is that doing something later doesn't guarantee you'll learn anything from those who did it earlier.

Sorry all for the long post, and I agree with Sarah that this pair of posts is great.

A Squared said...

I recall a Porsche engine available on the Mooney several years ago that used a single power lever -

You recall correctly. You many not recall that compared to same model mooney with a conventional engine, the mooney PFM had:

Less useful load (even with a lightweight prop and other weight savings modificatons on the PFM, the Porshe engine installation was significantly heavier)

Lower crueise speed.

Higher fuel consupmtion

Much higher purchase cost.

The Mooney PFM is often cited as an examplke of the luddites in aviation refusing to embrace new technology.

The thing is, Luddite or not, I'd be more than happy to do a little mixture and prop knob twiddling in exchange for being able to carry a heavier load, faster, at a lower operating cost and a lower initial acquisition price.

dpierce said...

Pertaining to above comments, it's also worth noting that many systems in modern cars aren't designed to be inspected until they fail.

With a lack of quick disconnects, few easily accessible components, and a plethora of mating surfaces that are designed to be prepared and left alone for years, you can easily increase your car's MTBF by over inspecting it. The computer is there to tell you that your normally sealed system needs to be cracked open.

A good compare - and - contrast is a race car, which is designed to be easily accessible and frequently torn down. Granted, many race cars are loaded with embedded systems, but that's more for performance than reliability.

Anonymous said...

The kind of reliability airliners get out of jet FADECs demonstrates that computerized controls for aircraft engines can be made extremely reliable.

The real problem for the GA sector is that the cost of making a FADEC as reliable as a manually-operated engine is beyond what GA operators are able/willing to pay. Quadruple redundancy and software designed for life critical real time control does not come cheap.

Nicolas said...

A former military buddy of mine told me a related story about a poor engine and a smart-a** mechanic.

While in flight one of my buddies engines misfires. So he writes this up saying that the "engine is missing." When he gets the plane back he finds noted in the mechanic's report that he opened the cowling and found that the engine wasn't "missing".

A joke to be sure but those hardworking aircraft engines deserve better.