It’s ok, I’m pretty sure I spelt it wrong myself when I first changed it.
True, but one of the key causes of the famous BAe146 windscreen blowout (where the pilot was half dragged out of the cockpit at speed and altitude, but survived) was a maintenance supervisor signing off on their own poor-quality work. Not intentionally on their part - they just used bolts that were slightly the wrong size thanks to a series of coincidences that were fairly improbable in themselves), and there were no checks to catch the error after the fact. Hopefully a story that all aviation maintenance people take to heart.
I watch one of those air disaster shows (I can’t remember the name at the moment) which details various real-life air accidents. To be sure a fair number are cause by faulty maintenance.
For instance the DC-10 crash in Chicago in 1979 where an engine just fell off the plane was caused by maintenance crews using a shortcut when re-attaching the engine to the wing. The official procedure was deemed too long so they took a shortcut and that was the result.
Once the FAA figured out what happened they found this was a common practice and more than a few other jets were found to have a similar problem…just lucky their engines had not fallen off yet.
The history of aviation is loaded with failures but, to their credit, they are usually relentless in figuring out what happened and then working to see that problem does not happen again and that includes aircraft maintenance procedures. As a result commercial air travel is incredibly safe these days.
Still, the pilots cannot know all the maintenance details that are done to their planes. They simply have to hope the maintenance staff did their jobs correctly. The pilots will certainly do a cursory check to see nothing really stands-out but can’t be expected to do more.
With apologies for resurrecting this hijack, the UK’s Air Accident Investigation Branch (AAIB) has just released their report of an incident in which a passenger flight nearly failed to rotate on take-off because it was loaded to be outside of CoG limits. This was due to a late change of aircraft not being communicated effectively. Fortunately, in this case the commander applied TOGA thrust in time to lift the plane off the ground and the flight was completed safely. Could very easily have been a much worse outcome.
Here is a link to the report (from which you can click through to the more detailed pdf version if you wish): AAIB investigation to Airbus A321-231, G-WUKG - GOV.UK
Perhaps the most concerning aspect is that although the problem was investigated by the crew during the flight and the abnormal passenger distribution noted, the crew concluded that the root cause was incorrect stabiliser trim, and the plane landed (with no apparent abnormalities) with the CoG still outside prescribed limits. I don’t know how reasonable it is to criticise the pilots for not recognising (in consultation with the airline) the real cause of the problem during the flight. And I suppose it’s hard to quantify how much closer to the edge of the safety envelope the landing was as a result of the CoG being out. It’s just frustrating (with hindsight, of course) that it could have been so easily remedied in the air, by asking some passengers to move aftwards. But given the previous discussion about aircraft ‘needing’ to be ‘nose-heavy’ to fly, I guess it makes sense not to move the CoG aftwards unless you’re sure that’s the correct solution - is it fair to say it’s more dangerous/harder to fly a tail-heavy plane than a (too) nose-heavy one? Or to put it another way, if you had to, would you prefer to fly a plane with the CoG 20% too far forward than one with it 10% too far aft? I guess that might depend on the aircraft type also?
Good find! This does deserve some commentary. Bad show all around.
Today’s a very busy day for me, so no time to respond in depth.
No problem - I never ‘expect’ a response at all, never mind a detailed one, and certainly not on any schedule - but if you (or anyone else) can discuss the questions raised when when you can, all the better of course. I’ll just drop in a tag for @Richard_Pearse in case he feels so inclined to rejoin this old thread (I think he flies A320s but I may be mistaken).
I receive a weekly email from the AAIB with notifications about their investigations, I believe anyone who is interested can sign up for this free service: Get emails from GOV.UK - GOV.UK
I don’t fly anything at the moment, but the A320 and A321 were my most recent steeds. Unfortunately I was not senior enough at my company to avoid the fallout from COVID-19. The good news is that my place on the company list is preserved and I will be sucked back up into the fold sometime in the future.
I can’t offer much comment on the incident you’ve linked to other than to say the company seems to have a serious weakness in its system where the people responsible for distributing the passenger load weren’t in direct contact with the people producing the load sheet. As for the crew not working out the solution in flight, as you note, it is much easier with hindsight. This kind of fuzzy issue can be pretty confusing all round particularly if you haven’t seen or heard of it before. That said, I would hope a very careful look at the load sheet would be an automatic response from a crew in this situation and confirmation of actual load distribution verses the paperwork should follow. Easy to say that sitting in front of the computer though.
As the A320 is fly-by-wire, it should feel the same to the pilot no matter how it is trimmed. In practice I don’t think this is quite true as the A320 and A231 variants have different handling characteristics that don’t seem to be able to be programmed out.
For a normal mechanical aircraft, a forward CofG would be easier to fly provided you still have enough control authority available. I doubt that would have figured into their thinking though.
Thanks for the reply, sorry you’re grounded at the moment.
No, this was very much a theoretical question - but thinking it through, I guess a plane that has enough control authority to take off, despite being nose-heavy, is going to be similarly OK to land. Unless you were unable to quickly pitch up as necessary.
A related question - is it fair to say that a plane that is too nose-heavy to take off is more dangerous than one that is too tail-heavy? My thinking here is that in the former situation, you’re not going to discover the problem until past VR (as in the particular case under discussion), by which time if it is impossible for the plane to take off, you’re basically guaranteed to be ploughing into the first obstacle beyond the end of the runway. Whereas with the latter, presumably you might feel the plane trying to rotate too early and still have time to abort the takeoff (if you correctly recognise the problem and can do that before V1). But failing that, I guess the likely outcome is an early rotation with insufficient airspeed followed by a catastrophic stall at 100ft agl or so. Or are pilots taught to counter this with enough downward pitch to keep the plane stable? I suppose that would be the instinctive response, the question would be whether there is time to abort the take off or if you get in the air tail-heavy and then have to figure out how to try and land. If you can even keep the thing under control at all.
TLDR: is it possible for a too nose-heavy plane to be able to take off, but then have serious problems landing? Conversely, if a too tail-heavy plane gets in the air, how likely is a controlled landing from there?
Continuing with on Richard’s excellent work and addressing only the linked post, not @Dead_Cat’s follow-up just above.
IMO the AAIB really dropped the ball on the investigation. Which is not uncommon on almost-accidents regardless of which governments’ authority we’re looking at.
The load sheet shown in their Figure 1 contains passenger counts by compartment. Numbers that were/are pure fantasy. Whether those came from some human keystroking or from some interface between software systems, they are fantasy.
There totally needs to be root cause analysis for how somebody/something just assumed = guessed there were 42 pax in Cabin D when the truth was zero pax. And the other 3 numbers had to be fantasy also because if correct, they’d have totalled 42 more people.
This opens in my mind the possibility that that airline always just assumes the load is distributed a certain way, with no input whatever from the field on what it really is. If so, that’s criminally negligent. And results in load CG errors of unknown magnitude on every single flight. Including those being flown today, 9 months after the event.
Enough of investigations, back to flying.
Failure to rotate is a really crappy corner to find yourself in, with no time to think. Choosing to go max thrust was a wise move. In fact the incremental pitch-up moment from the underwing engines may have been the factor that let them lift off. Conversely, in an MD-80, adding thrust would push the nose down and make rotation harder.
The other thing to do is simply to un-rotate, and let the speed build while waiting until the last moment before you run out of runway, then rotate again. Whether it’s CG or insufficient nose-up pitch trim, faster will be better. Though for a truly gross nose-down trim error faster is worse, not better. The takeoff configuration warning system is supposed to alert immediately when you throttle up for takeoff if the trim is implausible = set too nose down for any takeoff. But there’s still a gray area that’s nose down enough to prevent this takeoff, but still nose-up enough to be correct for some takeoffs. A screwed up excess forward CG is exactly that situation.
Cranking in some up-trim at the same time as the extra power is also a reasonable thing to do. Not procedural, but there isn’t a formal procedure for “it won’t fly” on any jet I’ve flown. At some point the engineers sit back and a la Maverick & Goose just ask that we “do some pilot shit” to resolve save the situation. The hard part is you have no idea how much is enough and by the time the nose reacts, you’ve probably overdone it. But at least you’ve avoided crashing for a few more seconds. On to the next problem.
As to the cruise-phase post mortem …
Figuring out what went wrong is hard both because you’re still busy doing the flying, and because you don’t have much evidence. Depending on the details of your carrier, talking to HQ may be easy or hard. And may be either on a private line or being broadcast in the clear for anybody on freq to hear. Sometimes airing dirty laundry in public is undesirable before you know quite how dirty that laundry really is.
You can’t know what trim setting was actually used. You know what the paperwork says that you were supposed to have set in. But did you dial in 4.5 when it was supposed to be 5.4 and both of you fell for the transposition error and didn’t catch it during your taxi & pre-takeoff checklists? Did you deploy one notch too little flaps, which manifests the same way? Again your paperwork preserves what you were supposed to have done, but there’s no ability in flight to recall the aircraft’s data logs to know what you really did.
For sure asking the FA’s if there’s anything unusual about the passenger distribution in the cabin would have been an excellent question to ask and, if asked, would have clearly pointed out the fantasy load planning inputs I mentioned above. At which point you’d know you had a CG problem, not a flaps or trim problem. Depending on what tools are available inflight they maybe could have run new numbers with their actual load and gotten the big TILT light.
At a minimum they could have redistributed the actual pax according to the fantasy load plan and then they’d know they were distributed within a normal balance. It would be very unwise to just start moving folks willy nilly.
Depending on the airline and route, cargo is another issue. Sometimes forward-biased pax are on purpose to offset aft-biased cargo. Or vice versa. IOW, in the absence of any cargo oddities, redistributing pax evenly throughout the cabins is a decent rule of thumb. But in some rare corner cases can make the problem worse, not better. Or swap out of range one way for out of range the other. No guarantees.
My reading of it is that this is about as far as they got before they were approaching their destination. I guess their priority, quite rightly, was to rule out any problems with the controls of the aircraft first, and having determined that wasn’t the case, they moved on to other possibilities but essentially ran out of time (I believe the total flight time in this case is only a couple of hours). Possibly with a longer flight they could have corrected the load sheet and then moved some passengers.
ETA: I imagine the landing was a bit more tense than usual - the pilots were presumably on high alert for any unusual behaviour from the plane, I guess they figured they had to land sometime and with control inputs appearing normal, approaching the airport as usual was the right decision.
Ref this post only.
Excellent questions and excellent thinking; you’ve found the nub(s) of the two extremes. Slight CG /trim errors are manageable in either direction. A bad enough case of nose heavy means, as you say, no chance to fly and you run off the end of the runway at speed with a hull loss and some fatalities a near certainty. A bad enough case of tail heavy results in a lift off, uncontrollable pitch up to stall and fall to Earth on or just past the airport boundary. That’s a hull loss and 100% fatality event. The Fine Air crash you posted upthread is very typical of that sort.
The in-between cases come down to luck & skill. How bad is the load imbalance vs. how Yeagertastic is the crew vs. how good/bad are the weather, runway length, nearby terrain, and other anciillary factors?
As you say, tail heavy gives warning signs earlier in the takeoff process and should therefore be more recoverable. The earlier the signs appear, the worse the post-liftoff situation will be; so any sign of early self-rotation well below vR is a huge warning flag to STOP!!!.
There’s lots of non-linearities in play here, so I don’t think a pilot could validly say “I’d rather have 10% too far aft than 10% too far forward.” The engineers might be able to compute the equal risk percentages each way, but they sure don’t share it with us. Nor would that be decision making info for us if they did. You set what you’re given and deal with the consequences as they occur; be they benign or horrific.
As to takeoff versus landing.
I’ll use figure 1 of the AAIB report as an example. Pretend that was their actual loading and all is well. There are two points called out, CG at takeoff and CG at ZFW. Those represent the CG with all the fuel and with no fuel. For airliners the assumption is neither the cargo nor pax shifts meaningfully during flight, so fuel consumption is the only dynamic variable. In military W&B they also deal with enroute cargo or armament deployment, and aerial refueling as additional moving parts to the inflight calcs.
Back to airliners … If we were to repeatedly recalculate this spreadsheet in real time as the flight progresses we’d see the CG start at the takeoff point then as fuel is burned the CG would slide more or less straight-line towards the ZFW point. If they stayed airborne up to fuel exhaustion the CG would be sitting on the ZFW point. For a flight such as this, London to Prague, I’ll WAG they were planning to burn about 3/4ths of the fuel load, so I’d WAG the landing CG would be about 3/4ths of the way from the takeoff point towards the ZFW point.
In that diagram weight is vertical, with heavier upwards and lighter downwards. CG is expressed as percent of MAC, with nose heavier to the left and tail heavier to the right. The punchline is that during a normal flight in a swept wing airplane, the CG shifts towards more tail heavy. So as a rule of thumb, if you got airborne after a nose-heavy incident, the landing will be doable. Even for the gross misloading accurately portrayed in Fig 2, the situation was marginally better at landing than at takeoff. We can’t really say how those sloping lines of equal %MAC would look splayed that far out off-range to the left. But my intuition is the CG would still be better at landing than it was at takeoff.
Conversely, for a tail-heavy situation that you safely negotiated after takeoff, the more fuel you burn the worse the situation will be at landing. That can be compensated for by any/all of a) landing sooner = closer to the takeoff airport, b) landing at a higher than normal speed, c) using less flaps.
The CG shift effect from fuel burn is usually not huge on a short-range flight. The longer the flight, the greater fraction of your total takeoff weight is fuel, and the bigger the shift is. Looking at the Fig 1 example again and pretending that was the accurate load sheet, we see the takeoff CG is 27.1 and the ZFW CG is 30.1 Which implies the landing CG in Prague will be about 29.1, maybe 29.5. A shift of 2+%. At those weights the CG range is roughly 15% to 30%, a range of 15%.
So our shift is about 2 parts in 15, or only about 13% of the available CG range. Which implies that there are some loadings that are safe to takeoff with but would shift outside the range by landing. But the majority of actual loadings that start in-range at takeoff will still be in-range at landing.
Going back to DC-3, the tail is not lifted just for visibility. It changes the aerodynamic characteristics of the plane. It has to be lifted up to safely take off.
And if there is thrust vectoring backwards on a tail dragger it’s possible for it to neutralize the lift of the elevator. That was the case of the ME-262. It was originally designed as a tail dragger. The test pilot had to hit the brakes to bring up the tail to get it off the ground. They changed it to a tricycle gear for the final production design.
Is that specific to the DC3? I haven’t had any trouble taking off from the three-point attitude in other tail-draggers. The DC-3 would probably appreciate having the tail fin and rudder in clean air though.
It might be. I got that from a guy on the net with some good instruction videos. He owns and flies DC-3’s. I’ll try to hunt down the specific video of him talking about DC-3 crashes.
After watching short take off and landing competitions it looks like they can do whatever they want with a variety of small aircraft.
I will retract the DC-3 tail statement. I’ve seen a number of videos saying that is the case but none I want to quote.