Video of B-2 Stealth Bomber Crash

A reader tips Murdoc off to this:


  1. Well, I’m glad I didn’t wreck the billion dollar aircraft! At least the pilots got out OK, and in fairness it may very well have been hardware failure of some sort. Any idea when it happened, as I haven’t seen anything in the news recently about losing a B2.

  2. It looked like FOD to me. Or else the engine just exploded for some other reason (manufacturing defect, fatigue, etc.) Something seems to come flying violently out of the wing just as the aircraft pulls up, and it pitches up and they lose control. It could a loss of hydraulic pressure too. Certainly some kind of mechanical failure… Interesting nosewheel arrangement. Flanker, I think it was about six months ago. IIRC Murdoc had a post on it.

  3. I read that moisture in a sensor caused an erroneous reading, making the aircraft pull up violently. It looks to me from the video that the wing tip hits the ground, causing the crash.

  4. KTLA is right. This is a problem that maintenance people learned about back in 2006 but the jackasses never formally recorded the vulnerability or how to fix it. So some people knew about it and some didn’t this time around.

  5. TY for the time frame, T thought maybe it was a while back. At least we have hundreds of them in inventory and can easily make up the loss! 🙂

  6. I linked to an article about the results of the investigation here. Obviously in a stealth airplane you can’t use an angle of attack vane like non-stealthy airplanes use, so apparently they use a series of static taps on the leading edge. The data doesn’t make a lot of sense when water freezes in them. Confuse the computer in a fly-by-wire airplane and you’re in for a wild ride. I was glad to see those ejection seats leave the airplane about the time the left wing dragged the ground.

  7. There is more information available from the USAF accident investigation website, including another video and some digital recreations of the event. I’m glad those airmen taking that second video were not involved in the crash.

  8. Hopefully the USAF will consider changing the flight control laws to pay more attention to the inertial data feedback. Frankly, I’ve never seen a control loop that used angle of attack feeback to adjust attitude like that. My preferred approach is to let the pilot fly the airplane and provide them the feedback on what the air around them is doing. Of course, this isn’t your average airplane, so maybe the configuration requires that the control system react because the pilot is too slow. Even so, it should not be allowed to go to crazy attitudes like that at slow speeds. They didn’t have the thrust or energy to maintain that attitude and then did not appear to have the control authority to get the nose to come down. Even if they do need the flight control system to react quickly, the system should transfer control back to the pilots after the initial reaction to a transient event. The accident report says:

    MP1 [the pilot] initiated the takeoff with a normal stick force, attempting to establish a standard pitch attitude and climb-rate. As the MA main gear lifted off the runway, the miscalculated negative AOA was interpreted as a dangerously nose-low attitude by the FCS which therefore commanded ‘full nose=up’ to the flight controls. The result was an abrupt 1.6 times the force of gravity (G) pitch-up (0030:50z) and a 30 degree nose-high pitch angle. As the MA lifted off at 0030:15z, all air data was lost, causing the FCS to substitute inertially derived valuse for air data computations and the MA to revert to default flight control gain settings. With loss of air data, AOA, AOS, airspeed and altitude indications, flight control response, as well as longitudinal and directional stability of the aircraft, were degraded. MP1 attempted to recover from the low altitude, low airspeed condition, but was not successful. MP2 [the copilot] initiated the ejection sequence at 0031:06z just after the left wingtip made contact with the ground, successfully ejecting MP2 first, followed by MP1.

    So the computer reverted to inertial feedback once it realized all of its air data was screwed up, but that backup mode was insufficient to allow the pilots to fly their way out of the crappy attitude it had put them in. Then it became a classic example of running out of altitude, airspeed, and ideas! Very weird. I’ve never seen a control system set attitude solely by angle of attack. From what I’ve seen here, I’d say that’s not a good idea. As I stated earlier, the system should at least use inertial feeback to set some reasonable pitch limits. Gravity can be a bitch, especially in a big airplane.

  9. Hmm, in retrospect, I have heard of pitch control algorithms using angle of attack rate feedback. I guess even though I was seeing alpha, I was thinking theta (alpha is angle of attack, theta is pitch attitude). Whenever they draw those pictures, the air is always coming from the left parallel to the ground. Even so, this is a real eye opener regarding what can happen with angle of attack feedback when the data goes bad. I wonder what the advantage is to using angle of attack instead of pitch? Those controls guys are real bad about asking pilots for their preferences and not ever evaluating performance given the various control law approaches.

  10. What was with that cameraman? Was he drunk? He could hardly keep the bomber in the picture. If he’s trying to record a billion-dollar airplane’s performance, he should be showing us the billion-dollar airplane.

  11. On the website I provided the link to, that was identified as ‘security camera’ footage. It appears that the person operating the camera was doing it remotely. Who knows what kind of controls he had? I talked to a friend of mine who is a professor who works with things like fly-by-wire control laws. He said that angle of attack (aoa) feedback is common. He said it provides an unstable aircraft with the ‘feel’ of a stable airplane. This is understandable, because if the pilot cranks in 3 degrees aoa with the stick and then slows down, the airplane will start to nose over to maintain that aoa. If it used an inertial reference and was commanded to a 3 degrees nose high attitude and the airplane slowed the elevators would continue to deflect upward until they had no more travel left, only then would the airplane depart the commanded attitude. That departure could be very abrupt if it coincided with wing stall. Still, I’m thinking that perhaps some combination of aoa and inertial attitude should be considered in these control laws. The idea that the airplane would fly itself into an uncommanded attitude like that B-2 did is something that is not acceptable. Then once the airplane reverted to inertial feeback mode either the airplane did not have the control authority to get the nose down and build speed, or perhaps the pilot was trying to land it. He seemed to be on the back side of the power curve if that’s what he was trying to do. Once that wingtip started dragging, it was all over.