View Full Version : NW emergency landing-ANC

12th Oct 2002, 14:54
The following link discusses a NW 747-400 currently on the ground in ANC after experiencing a full left deflection of the lower rudder in flight.


12th Oct 2002, 15:41

13th Oct 2002, 08:13
In both cases it seems that the redundant rudder design (and a professinal crew!) has saved the day. Maybe Boeing could retrofit this concept to the B737s or is it an engineering/financial nightmare?

14th Oct 2002, 07:46
Forgive me if it's been said before but it does appear that Boeing does have an inherant problem with their rudders, (hornets nest approaching) Is/was it an engineering design flaw or a get ahead of Airbus at any cost problem.

14th Oct 2002, 10:22
If you look at the more recent Belgique posts on this thread (http://www.pprune.org/forums/showthread.php?s=&threadid=2617&goto=lastpost), you might reflect that a split rudder setup affords much more systemic redundancy than that of the A300.

On that a/c the two yaw damps work through a single dual-acting concentric valve that operates the same singular (3rd of three) rudder actuator on a single panel rudder. Now just reflect on what could happen if that actuator's valve was to stick intermittently - i.e. let's say if it were to be deflected beyond normal operating limits during a wake encounter (followed quickly by yet another). The 737 rudder actuator has a similar singular valve - and it has a sticky failure mode. AA587 had a yaw damper fault on the ground (pre-start at JFK) that needed re-setting. Was it actually a disregarded alert that that valve was into an unsmooth operating condition (likely, because the pre-start BITE check checks it over its operating range).

Why might it stick intermittently? Well those actuators spend 99.999% of their lives working within a very limited range. That possibly means, for an hydraulic component, that it works smoothly and happily within that range - but equally that, once driven outside that range, (by say, a wake encounter) the likelihood of it sticking is much greater. In fact, let's use the term stiction - inferring a momentary (rather than an ongoing) condition. But it's a condition that may be enough to intellectually challenge the yaw damp computer's fair expectation of "intervention begets prompt result"..... or, in its programming, further action is warranted.

Come with me a little further into the realm of supposition and reflect how a mutually supportive two yaw damper system might respond to any such sticking valve. Would not the second yaw damper kick in? If the fault then cleared momentarily might not the primary YD then (assuming it hadn't tripped) reclaim it's primary responsibilities and intervene (perhaps not altogether appropriately). Is it possible to have such a "War of the Yaw Dampers"? Hopefully the NTSB Hearings on the 28th will address this issue. But my personal feeling is that having a two yaw-damper system operating through a singular dual-acting valve that drives a single rudder panel is not much flight-control redundancy at all. Boeing's similar malfunctions would always seem to be capable of a happier ending than the Airbus (A300 at least).

This isn't the whole story - as Belgique has also proposed a mechanism that explains why such events as a stall, CAT or a wake encounter might initiate a further complication involving the air-data computers (remembering here that prior events have involved fluctuating airspeed indications and been "resolved" by ADC swap-outs). What's common to stalls, CAT and wake encounters? Sharp pressure discontinuities. How might that find its way from ADC to rudder? Well to interim answer that query with a question….. What, according to the NTSB, has been the most annoying aspect of their investigation? They’ve found that despite prior moaning, the DFDR’s sampling rate has remained highly “filtered” i.e. much intermediate data is unrecorded – to the extent that a full L/R rudder deflection at 39 deg/sec could be missed). Why would that be the case?

Well this very original Airbus design’s systems (FCS, ADC, yaw damp computers etc) data-throughput had proven to be much too high and the DFDR is logging downstream of the filtration that’s required in order to suppress any distracting “jitter” on the crew’s displays. So what are we getting at here? Simply that a system that’s capable of moving very fast, once also connected to a system that’s being suppressed or subjected to highly fluxed inputs (the ADC say) is suddenly capable of tripping over its own feet. Throw in a rudder limiter and a pilot’s pedal with a high breakout force, both of which are dependent upon what the air data computer is instantaneously saying is appropriate airspeed-wise ( but whilst in an area of high pressure discontinuities and spikes…..) – well the inference is that it might be like inviting an epileptic on ecstasy to polka.

14th Oct 2002, 14:44
>>Forgive me if it's been said before but it does appear that Boeing does have an inherant problem with their rudders, (hornets nest approaching) Is/was it an engineering design flaw or a get ahead of Airbus at any cost problem.<<

Well, Airbus has the low cost feature of having the vertical stab come off in event of a rudder anomaly as demoed at JFK last year. Like a lot of us, I've flown both brands. The Boeings generally go higher, farther, faster, last longer and cost more.