Elevator Disconnection
 

It may not be common knowledge that many aircraft types have separate control systems for each side of the elevator. The sides are connected by a clutch which, in the event of a jam, disconnects and allows safe flight from either control column. If the sides do disconnect, pilots must take great care not to apply opposing control inputs. Doing so will subject the tail to asymmetric loads for which it is not designed.
On 26 January 2016 a BAe ATP (nee Avro 748) sustained separation between its two elevator systems, possibly caused by opposing inputs by the two pilots. See https://www.gov.uk/aaib-reports/aaib...bae-atp-g-buur.
On 20 February 2014 opposite inputs by Captain and FO of an ATR 72 caused the left and right elevator control systems to uncouple from each other. Independent and opposite movement of the left and right elevators subjected the tail to asymmetric loads well beyond design requirements. The aircraft was seriously damaged but the damage went unnoticed. See https://www.atsb.gov.au/publications...r/ao-2014-032/ and http://www.pprune.org/australia-new-...albury-12.html
ATR issued an AOM warning of operational and maintenance hazards whenever elevators become uncoupled - essentially avoid opposing inputs and check for structural damage after any elevator disconnection.
Disconnect systems on the ATP and ATR are significantly different in detail but the same hazards arise. If elevator disconnects are more common than generally appreciated, ATR’s warnings should be disseminated for all types with disconnect systems.

Opposing elevator inputs after a split will also put a twisting load on the stab which could cause it to detach. I seem to remember an E110 many years ago losing a stab with obvious catastrophic results.

ozaub

It may not be common knowledge but it's a great advertisement for Airbus fly-by-wire

The design intent is to overcome the jammed elevator case (FOD, icing. . .). I think some/most transport aircraft have a manual disconnect to uncouple L from R.

do really believe that there is so little design margin against these loads that the stab will detach?

And if one side of the elevator jams up or down, is it dangerous to try to offset it with the other side after the split? Even if the jammed side is at neutral, presumably you'll need relatively large deflections on the other side to get reasonable control authority.

As SLF, though, I wonder what the procedure is for flying with split controls (absent a jam). I'm guessing that one side is normally left at neutral, and both pilots don't need to constantly try to synch inputs?

I doubt this is as tender a structure as we are treating it here. It has been verified in flight test, and I'd be surprised if any prudent limitations are not documented.

In the incident quoted by ozaub, the horizontal and vertical stabilisers needed to be replaced before further flight due to the significant damage caused by the asymmetric loads. The manufacturer found "the ultimate load was exceeded by about 47 per cent, and the exceedance lasted about 0.125 second". It is indeed fortunate the exceedance only lasted 0.125 seconds, otherwise the stabiliser might well have detached from the aircraft.

Elevator asymmetry protection and limitation are all part of the hydro mechanical control mechanisms for 737, 757, and 767 aircraft.

Although rare, prior to taxi, it was possible for the DC-9 series to have a control column lockout caused by elevator asymmetry, due mainly to wind gusts, which could move one surface without the corresponding surface following thanks to the free-floating elevator tab design. This aircraft did not have elevator asymmetry protection, yet to the best of my knowledge, there have been no incidents or accidents relating to a failure in this regard.

Well that's pretty obvious, isn't it, ozaub? The idea is the jammed side be left alone and the aeroplane controlled with the serviceable side, not have the pilots fighting each other with their prongs. Instead of casting doubts on the system, you'd be better off concentrating on the CRM that resulted in this fight. You do fly/have flown a type with split elevators, haven't you?

I hope it as least common knowledge for those who fly these types. The various disconnect systems tend to have associated control jam memory items after all :hmm:.

@lomapaseo

I was thinking of this https://aviation-safety.net/database...?id=19841206-0

Had a complete, rock solid, lock up of both elevators once. Quite alarming as the column felt like it was set in concrete. It wouldn't move one millimetre, on either side. Aircraft doing big phugoids with the AP in....

The a/c had a splitter but we managed using the trimmer and eventually it all came free at about 6,000' when the ice, which was causing the jam, melted.

Diverted, as we were due to land at LCY, which seemed ambitious, but didn't tell ATC much and the pax, nothing.

Not much that anyone could do to help really.

kind of missing the point, on the ATP incident, the AP stuck in and the controls split due to the pilots pulling on the column to overcome the AP, force of more than 50 lbs causes a "breakout" and the Standby control system activates, they did this to get rid of the stuck AP.

Aircraft is completely flyable with asymmetric Elevators, it was designed this way as a secondary control system

This probably isn't realistic, but I guess if you had an unrecognized pitch disconnection at high speed and somehow one pilot kept pushing while the other kept pulling, you could end up with large elevator deflections but no change in pitch. Loading on the horizontal stabilizer might be as much from the deflections themselves as from the fact they were in opposite directions.

Chu Chu, this would result in an aileron-like effect and, consequently, a roll, would it not?

On the DH8D, a disconnected elevator (the selections leading to this are indeed memory items initially) will result in a maximum speed of 170 to reduce the bending moment on the fin.

I guess it would cause a roll. My thinking (for what it's worth as SLF) is that if the pilots were making opposing inputs, the feedback mechanism (change in pitch and, I guess, g-loading) that would ordinarily stop pilots from making full elevator deflection in high-speed flight wouldn't be there. And I suspect for many aircraft full deflection at high speed would risk damage to the stabilizer, even if the controls weren't split.

I hadn't thought about the roll, but I have my doubts about how quickly that would alert the pilots to what was happening. (If it's not clear, I'm not saying it's likely that the pilots would make large opposing inputs, or that if they did they would persist for long, just that the roll doesn't seem like it would immediately clarify the situation.)

Chu Chu, be practical. The other pilot is sitting less than a metre away. They talk. They see. One pilot flies, the other watches.