[16s]

Software issues are common in jet fighter aviation. Here are two little known, but potentially deadly issues:

* 1986 F-16 Inversion when flying over the equator - Flying the F-16 in simulation over the equator, the computer got confused and instantly flipped the plane over, killing the pilot [in simulation]. And since it can fly forever upside down, it would do so until it ran out of fuel. Note: some claim this is a myth and to my knowledge, it has never been verified, but I tend to believe it could have happened. Source - http://catless.ncl.ac.uk/Risks/3.44.html

* 2007 F-22 Complete systems failure over International Date Line - At the international date line, all systems dumped and when I say all systems, I mean all systems, their navigation, part of their communications, their fuel systems. They were—they could have been in real trouble. They were with their fuel tankers. The tankers – they tried to reset their systems, couldn’t get them reset. The tankers brought them back to Hawaii by visual, line of sight flight. Source - http://www.defenseindustrydaily.com/f22-squadron-shot-down-b...

[anovikov]

There have been a very similar but actually happening lock-up of all systems of F-16 and same with soviet Su-24 when flying below zero altitude. Designers haven't looked at the map to find out there are plenty of places with ground much below sea level, and it happened for Israeli F-16 near Dead Sea, and for soviet Su-24 near Caspian Sea, when they tried low-level flight.

[MichaelGG]

What kind of codepath causes systems to die when crossing the date line? Shouldn't all times be in UTC, and certainly any change of timezone is a display-only issue? What would be the point of using local timezones for any "systems" type stuff?

[ktizo]

My guess would be that the problem isn't time, but rather the code dealing with the position of the plane in longitude as it crosses over. Perhaps it spat out an unexpected negative, or even something undefined.

[don_draper]

I would think more than half the code for these types of systems is exception handling.

[petitmiam]

My lecturer for 'Software Testing' at uni used to bring up that first issue. He was a total ball-breaker with assignments and exams, but you knew it was for a good reason.

[majormajor]

The first one, if it's real, confuses me: how does the plane turning upside down kill the pilot?

[noonespecial]

Fighter planes are very different than any other craft "driven" by people. They are so fast and so strong that if you give the stick a good yank while flying one, the forces involved will kill you (unless of course good fly-by-wire software knows your limits and intervenes).

Its not at all far fetched to assume if the plane suddenly rolled at its maximum rate that it would pull the pilot's spine apart.

[WildUtah]

Which points to the real problem with the F35 and all next generation fighter planes.

It's not software, relationships, rivalries, cost, or maintenance. It's obsolescence.

By the time there's a full wing of F-35s in regular service, there will be drones capable of routinely shooting them down. The advantage of not needing to coddle a pilot and having all possible G-forces available will make even cheap drones distributed by corrupt Chinese or Russian backchannels superior to the best fighter jets in the world within ten years.

And a whole wing of such drones will soon come cheaper than a single F35.

Obama should do the same to the F-35 as he did to the F-22 and cancel it before it wastes more money.

[noonespecial]

Such drones have existed for some time now. They are colloquially known as "surface to air missiles".

[InclinedPlane]

The logistics of aerial domination are pretty much the same whether you're talking drones or piloted craft. If you are the attacker you use ECM and missile/bomb bombardment to take out the radar/SAM stations. Meanwhile, you try to destroy as much of the enemy's airplanes on the ground as possible. If you have superiority in the air you take advantage of it, destroying every enemy thing that flies whether it is piloted by a crew or not.

The biggest advantage that drones have is a tactical one. And perhaps a manufacturing/logistical one if they can be produced in greater numbers, though at present that is not the case. Tactics only apply if there is a logistical balance sufficient to allow for a roughly even battlefield.

[Daniel_Newby]

The biggest advantage of drones is you get to reuse the pilot. The second biggest advantage is you get to use a team instead of one pilot.

[_b8r0]

To expand, I think you may be (intentionally or unintentionally) referring to loitering munitions. These are munitions (not just SAMs but other bits of kit) that are launched in theatre with no specific target set on the ground, then reconfigured in-flight after relaying information back to the ground.

Traditional SAMs aren't quite the same thing and serve different purposes.

[salem]

Rather than replacing manned planes, once software and AI catches up, it's more likely that a huge numbers of semi-automonous drones would get pared up with F-22s and F-35s as wingmen. It would solve the bandwidth and pilot shortage problems, and reduce the risk that a few anti-SAT missiles take out the entire drone fleet. Oh, and that helmet and the sensor fusion software is probably going to be a key stepping stone to making that happen.

[Symmetry]

We already have AWACS that are more or less designed as control planes, no need to keep the manned fighter planes around.

[salem]

If I'm in denied airspace, I'd rather be in an F-35

[greedo]

The problem with relying on drones is threefold; Humans are excellent computers, the Mark 1 eyeball has some advantages over other optical systems, and drones rely on a passive electronic environment. Jamming GPS and other control frequencies would be the first course of action should we attempt to use drones on a serious opponent.

[Tuna-Fish]

> Humans are excellent computers

Humans are very slow. Humans are also squishy, fragile, heavy, and can't take G-forces very well. (Really, at all.)

> the Mark 1 eyeball has some advantages over other optical systems

These are rapidly going away with machine vision research.

> and drones rely on a passive electronic environment

Only ones with a human pilot. The air-to-air drones that will be deployed at the tail end of this decade will be completely autonomous after they receive their orders.

There is ridiculous amount of progress in a-to-a drones at the moment. Very nearly everyone who used to build/design fighters is now concentrating on drones. There will not be a seventh-generation manned fighter, and the sixth-generation ones will be outclassed very rapidly.

[greedo]

Humans are far superior at many of the tasks that require more than just following an algorithm. And the Mark 1 may not have Steve Austin abilities, but compared to scanning terrain with a Sniper/Lantirn pod, it is unbeatable. Talk to some of the guys who fly the Predators out at Creech, and they'll gladly tell you the unclassified shortcomings.

And autonomous air to air drones still need to be able to accept instructions/commands while in route, during combat, etc. The idea that we'll simply unleash 100 drones and create a killzone isn't something that's going to fly...

The reason all the companies are building and designing drones is because that what the governments/militaries are paying for. Private industry simply can't afford to do much R&D that isn't already contracted for by the DOD etc.

Finally, drones are easy pickings for manned aircraft. For permissive environments like the 'Stans and Iraq, sure. Up against someone with a real AD network, forget about it.

[jakeonthemove]

I suppose they could replace the pilot in the F35 in the future? I still think the F22 would've been more than enough for now and F35 is a waste of money, though...

[hastur]

That's rather naive. You conveniently forget the fact that before you shoot down an F-35, first you have to spot it.

Not to mention that building a drone with capabilities similar to a manned fighter is a very hard problem and you can be quite sure that US's rivals (China, Russia) are decades away from fielding one.

[Daniel_Newby]

The F-35's "secret sauce" is sensor fusion and human-machine interface. If their innovations actually work, they can easily be rolled into the next generation drones. In the meantime they can be used to fight dirty little wars.

[rbanffy]

> Its not at all far fetched to assume if the plane suddenly rolled at its maximum rate that it would pull the pilot's spine apart.

The pilot sits very close to the roll axis of the plane. I find it hard to believe just rolling very fast would seriously hurt the pilot unless the plane could roll insanely fast. A pilot can really hurt him/herself by doing severe pitch and yaw changes, but I don't believe the g-loads (and the g gradient) during roll would be that high.

[DrStalker]

An F-16 can turn fast enough to make the pilot black out of suffer severe injuries; one of the things the software does is restrict G-forces to acceptable levels. As potentially terrible as this bug was I've never seen any indication of where in the development life cycle it happened; it could have happened early on when it wasn't meant to be working yet (i.e.: working enough to fly, but with no handling of edge conditions worked on) or it could have happened in final testing of a release candidate.