Hacker News Reader

Comment by MSeven | Hacker News Reader

Just last week i heard a story from an older friend about a fighter jet crash investigation he was involved in, where the cause basically was that an engineer wired down some bolts (as required) but trimmed the ends of the wires without recovering the trimmed part. That little piece eventually found its way into the main stick mechanics, causing enough resistance at the wrong moment, preventing the pilot from recovering from a tight dive and killing him.

View on HN

This is one of the reasons we never trimmed the ends of cut wire when we tied bolts down. We just twisted them and bent them inward. No excess cuts needed, and no tiny bits of FOD that you have to dig out.

What is FOD in this context?

Foreign Object Debris

We always called it "Foreign Object Damage" but yeah, "Debris" would technically be more correct.

Gotta keep the FOD out of the aircraft and the flight lines!

Both are kind of correct. If it's laying where it shouldn't be it's debris, but if it's plowed through the windshield it's damage.

This is really interesting to me, how are the crash investigators able to pinpoint the exact cause of failure in such circumstances? I'd imagine a small piece of wire would be just another miniscule piece of debris at a crash site, and unlikely to still be in the location that caused the crash.

Starts with data logging, everything is logged. Then you inspect that area and look for anomalies in the metals in that area. At 5Gs, a little piece of wire becomes a razor blade and leaves evidence. A lot of odd nicks in an avionics bay will lead to testing, interviews and more testing. A worker has no reason to hide details, it would be criminal, so the interviews look at all the possible anomalies and then narrows down to the root cause.

Really amazing work, read NTSB crash reports and they outline the methods used

Thanks, yes, I've heard that NTSB crash reports are mind blowing in their attention to detail. A follow on question: how do they ensure people honestly answer questions, is the person not inviting personal liability by answering?

In my experience there has always been a _deep_ culture of mutual respect between industrial aviation players and the NTSB.

Everyone knows that the investigative body is there to help figure out what went wrong and to provide guidance on how to avoid similar mistakes in the future. It’s not about assigning blame, and there’s an understanding that if it were to become about that then the industry would be worse-off as a whole.

(Slightly off-topic, but this is precisely why Tesla’s reaction to the NTSB investigation in their Model X crash was so concerning. Any player that defects here runs the risk of poisoning the process for everyone else.)

Thank you both for the answers!

Look at the original metal fatigue investigations that lead to round windows.

They pressurized and depressurized an airplane (inside of a water tank) 50,000 times until it failed.

I am really impressed that we are able to fly safely at all if a tiny little piece of debris like this can cause plane crashes. Debris is a fact of life. Things fall apart, small pieces come off whatever substance or piece they were attached to. Are maintenance crews constantly just taking aircraft apart and vacuuming every surface they find?

Military aircraft are designed to be unstable, so that they can perform fast, odd combat maneuvers. The flight computers are the only think keeping those things flying.

Civilian aircraft are designed to be super stable.

And aerodynamic stability would not have prevented the accident described in grandparent.

This isn't necessarily true. Without stability you can't perform certain maneuvers. And the large aircraft are super stable.

That said, we take FOD and tool control very seriously. And the aircraft had 30 / 60 / 90 /365 day inspections that varied in detail

Do you think the pilot had any idea what happened or did he simply die never knowing?

Almost certainly the latter. :-(

I think at the speeds we are talking about here, it's just an "the plane is not reacting to control input" and nothing else. Probably not enough time to think about what exactly went wrong (and will kill you in seconds).

Its lack of relevance to the matter at hand not withstanding, I admit to be a bit bewildered by your question and thus feel the need to ask: what makes you think the pilot could possibly have any idea that trimmed parts from a bolt is what caused the stick resistance, or that he would spend the time thinking about that instead of recovering the aircraft in the few minutes he had at the most ...

He or she would spend time mentally running through the avionics system to think of what to trouble shoot. Experienced pilots know how their craft works just like you know how to debug a failing drive and not waste time looking at your network cable.

My neighbor is a pilot and I asked him completely off hand how to troubleshoot non responsive flaps and he went into detail about the hydraulics system, where the pumps are and the steps he would take to troubleshoot. This was off the top of his head without a checklist.

Pilots have loads of checklists, many of them committed to memory. If there's a problem, there's a checklist for it and following the checklist should lead you to a solution or at least tell you what is wrong. The downside is that checklists can't account for every possible thing going wrong. A little piece of wire jamming a hydraulic system is going to be a hard fix. Even if you have the tools and access to the component, do you have time while the plane is nose down?

Planes have to be predictable, they can have massive failure states but the solution must be repeatable. Unknown failures, regardless of the severity, are a major problem and should always be tracked and investigated. A small unknown failure could be a symptom or precursor to a much larger problem.

Everyone hates paperwork but it's what saves lives.

You’re actually not supposed to memorize checklists, it leads to complacency.

For those emergencies that demand quick action, commercial pilots do have to memorize the relevant checklists.

How is it possible to make that determination after the crash ...?

Reply on news.ycombinator.com