Again, not an expert, the biggest thing I flew had 4 seats, but…
Off the top of my head - If the load came loose and went backward causing a stall (due to climb angle) the aircraft would have stayed nose up.
Instead, it look like a classic:
-stall
-One wing dips
-recover to level and overcompensate a bit the other side
-dropping nose down
(How do you do that if the weight is shifted into the back behind recommended CG?)
-actually appears to be starting stall recovery, pulling up, but runs out of altitude too soon.
So I’ll suggest scenario #5, rare but not impossible - momentary distration or loss of power leading to stall, and failure to get the nose down in time?
Another point, in the Air France thread here a year ago, someone mentioned that large aircraft shake and rock violently when they stall - it’s not a clean experience; so a stall could conceivably lead to a wing-down situation, but not so significantly that it leads to a classic roll-right-over-onto-the-back spin… ? Although that’s almost how it looks like they were headed… Did they compensate for that very quickly with differential thrust maybe? Black box will tell…
I recall an accident report in the Transport Canada newsletters many years ago. It described a flying boat that the crew negelected to pump the bilge before takeoff. It had accumulated enough water that the takeoff angle caused it all to run to the back - at which point it went directly nose up before falling to earth. That’s kind of what I’d expect from a cargo shift down the center axis of the aircraft.
The nose only drops once the aircraft has banked far to the right. Weather cocking from the tail as the aircraft slips sideways towards the ground will cause the nose to drop like this regardless of how aft the C of G is.
I don’t think it’s possible to tell anything other than it stalled. Could be aft C of G, shifting C of G, incorrect trim setting, mishandling, etc.
Something about elevator trim that those used to smaller aircraft might not realise is that on some jets, including the B747, the pitch trim system moves the entire tail-plane. If the aircraft is badly miss-trimmed, full down elevator may not be enough to get the nose down, even if the aircraft is loaded within limits. The only way to get enough elevator authority in this case is to re-trim. This is unlike a conventional system with elevator trim tabs. Elevator trim tabs change the neutral point of the elevator but don’t change the effective range of elevator movement. If you are strong enough you can physically over-power an elevator trim tab system, this doesn’t work with a moveable tail-plane.
Water will travel all the way aft but a load of vehicles are only going to be able move slightly aft, even if they all come loose. It doesn’t take the whole load sitting in the tail to cause this crash, it only needs for the load to be too far aft for the take-off trim position of tail plane to not allow enough elevator authority to lower the nose. A load shift could cause a crash where the aircraft still follows normal stall aerodynamics, or it could be more unusual. The fact that it looks like a normal stall doesn’t necessarily rule anything out.
It is very common that there is a very high drag condition (more than gear down) as the gear transitions between full up and full down. Sometimes this is due to gear doors that close once the gear is down and have to open again to allow it to retract. Other times it is just interaction between the partially extended gear and the rest of the air frame.
Finally, flight testing may check stall recovery characteristics with the gear up, and with the gear down, but probably not everywhere in between.
Also, while reducing drag is not a bad thing, it probably isn’t what will save your bacon. Drag increases with airspeed, so when you are going too slow to generate 1 G of lift at max AOA, you can be pretty dirty and still have enough power for decent acceleration.
Wing spoilers are another matter, though. In addition to adding drag, they reduce lift.
I actually didn’t know that the -400 had that. (That’s why it ended up as a footnote.) But I thought I remembered that an earlier version, the -SP, did; and I thought it was in the vertical stabilizer. While googling to confirm that, I found the info for the -400 first.
I hate to think that mistake was made, but if it was, maybe it gives the best chance of solving this. Can’t bring the guys back; best we can hope for at this point is to find out why it happened so it doesn’t happen again. If there was a problem with the straps or the tie-down procedure, will the evidence have survived the fire?
There was an accident at a hotel in Kansas City about 30 years ago. The lobby was a tall atrium, and there were two walkways suspended from the ceiling. The way it was designed, there were metal rods coming down from the ceiling, and each walkway was supposed to be supported by those rods. The way it got built, the lower walkway was hung from the upper, putting twice the load on its supports. They failed. I wonder if something like that may have happened here; lots of redundant straps, but too many attached to a single point until it failed.
There was an Alaska Airlines flight that crashed off of Port Hueneme (‘why-nee-mee’, for non-locals) in 2000. The accident aircraft was a McDonnell-Douglas MD-83, and the crash was caused by a worn elevator jackscrew. The MD-83 was developed from the DC-9. AFIAK, the DC-9 was a new design that was not based on the DC-8. In looking up the crash of DC-8 N8079U, the probable cause was ‘A loss of pitch control resulting from the disconnection of the right elevator control tab. The disconnection was caused by the failure to properly secure and inspect the attachment bolt.’ For a moment, I was wondering if DC-series jets might share a problem with their elevators – especially since a DC-8 and a DC-9-derived MD-83 crashed in the same year; but the crashes were had different causes.
If lowering the nose would have worked I guarantee you they would have done so. The plane has a stick shaker and stall alarms screaming at the pilot to lower the nose.
Stall recovery is the most fundamental skill of flying required of beginners.
Close, but not quite right. Your description suggests that the engineering was sound.
The way it was “designed” (as shown in the drawings) was physically impossible to produce: There was a threaded section in the middle of the support rod that had nuts to support the upper walkway. There was no way to get those nuts over the un-threaded length of the rod. Just cutting those threads into the middle of the rod would have required either an impossibly long lathe or construction of special tooling.
It wasn’t that the support rods were overloaded, as they had the same tension as designed, but the box-beam of the upper walkway where it’s support rod attached carried twice the design load…it had to transfer the load of both walkways to the upper support rod, instead of the rod carrying it directly.
A different fix would have involved a long coupling nut under the upper walkway, still allowing a two piece support rod to be threaded only at the ends. That probably would have been fine, as the upper walkway beams would not have had to transfer the weight of the lower walkway to the upper support rods.
Sadly, the advent of computerized drafting has made this sort of error more common instead of less. Manual drafting used to be taught as part of an industrial arts curriculum where the students actually had to build from what they drew. Now I see all sorts of silly stuff that was done just because it was easy in Solid Works or ProE. Driving a CAD program takes a different interest and skill set from being a gear-head it seems. I’d much rather have slow drawers with mechanical sense.
Just wanted to add something about fuel in the stab. My current airplane (MD-11) also has fuel tanks in the horizontal stabilizer. These tanks are never full on takeoff.
The stab fuel is used as an in-flight CG adjuster - an aft CG at cruise is more fuel-efficient, so at cruise the airplane pumps fuel from the wings into the tail to get the CG as far aft as possible. On descent the fuel is moved back into the wings.
Obviously the pumps could fail and fuel could be trapped back in the tail, so we have procedures for that. But these only involve approach and landing. The airplane should never be fueled for takeoff and have fuel in the tail. If fuel does get into the tail the airplane will pump it into the wings as soon as the engines are started. I imagine the 747-400 is similar - there would have to be multiple failures (including the crew not noticing) for them to take off with fuel in the tail.
Everything I see here (and this is with 20 years of flying cargo for the military and the airlines) is a load shift on rotate.
The thing about aviation is that just because it worked the last time, doesn’t mean it will work the next time. The airplane in front of you made it on the arrival without deviating? No guarantees for you.
So it is with this - the tiedowns on one of those MRAPs might have been compromised, and were able to survive one takeoff. But those same tiedowns, on takeoff number two? That is when they decide to give up.
I hope we can find the cause of this accident. That is the only way to prevent something like this from happening again.
I’m not close to your level of knowledge on planes but I was under the impression the wet tail 747-400 carried fuel off the runway. I know it’s a range extender so I was making the assumption: From wiki: “The horizontal tail was also redesigned to fit a 3,300 US gallons (12,000 l) fuel tank, resulting in a 350 nautical miles (650 km) range increase.”
I was wondering if too many straps were attached to the same attachment point if that would do it. If someone was looking over the load they would see dozens and dozens of straps but might not see the tree within the forest. I’ve only been in a 747-400 once and really didn’t focus on the floor setup.
Like I said, I’m not a 747-400 guy, but I would assume that fuel capacity in the tail would be movable from other areas (ie wings).
If they can load up that tail with fuel prior to takeoff…damn. I’ll have to ask some of my buddies over at UPS about that, because that seems…extreme. Not unreasonable, depending on your load, but just…one step more toward the failure scenario.
As far as the straps go - yes, exactly. The forest for the trees.
FWIW- I checked with someone I know who flies the 747-400 regarding the stab fuel. He said that the stab fuel is typically disabled on conversions, which this plane was. (Retired from passenger service and converted to a freighter.)
Reading this thread made me aware of and appreciate the job of loadmaster.
On civil flights, there is someone with the same responsibility, right? Even straight suitcase flights?
Under what circumstances are loadmasters required to be on the flights? I can’t see a luggage-handler supervisor (ie, not a loadmaster) on every flight.
This is very true of most things involving risk. Feynman went into it when he was addressing congress, and less so in the written challenger report. There is a human tendency to assume that when a risky action has a good outcome, that the risk was initially overstated. “See, everything turned out fine, you were needlessly worried!”
Many years ago I saw a shuttle launch. I was able to do this because the scheduled launch was delayed several weeks until I was in Florida. Reason for the delay - NASA was looking into the problem that ice had come loose from the tank and damaged the shuttle tiles. But in the end, they concluded the problem was not significant. This was about 1998.
I fly both freighter and passenger jets. In both cases there is someone who designates how the load (including passengers) is carried. The freight/baggage component of this is sent to the loaders who load the aircraft accordingly. The load supervisor then signs the load sheet stating the aircraft has been loaded correctly.
The captain is ultimately responsible for the loading of the aircraft but as the physical task has been delegated to other people, all I can do is check the paperwork is correct. On the dedicated freighters we can check that the correct cans have been loaded onto the freight deck in the correct order but there is no way to check that each can has the right load inside it.
There is no requirement for us to carry anyone other than the pilots unless we are carrying horses in which case we have horse handlers to pat their noses and make soothing noises.
Normalisation of deviance is closely related to this. Here is a YouTube videoof a presentation touching on the normalization of deviance and how it contributed to the shuttle disasters.
