AAIB special bulletin ~ Loganair

[parkfell]

Special Bulletin S1/2015 Saab 2000, G-LGNO

Well worth a read.

[DaveReidUK]

Well worth a read.But not to the extent of providing a link?

http://www.aaib.gov.uk/cms_resources.cfm?file=/S1-2015%20G-LGNO%20v2.pdf

[lomapaseo]

Seems to be the same old lesson learned between the crew and the perceived condition of their automatics.

Is it on or is it off?

[jack11111]

Very simple system...no THS or elevator trim tabs. The control column always tells you where the elevator is.

[Capn Bloggs]

Very simple system...no THS or elevator trim tabs. The control column always tells you where the elevator is.
Apparently not:

Pitch trim consists of elevator movement without associated control column movement.

Fly by wire with an AP that will stay in at Vmo + 80 ?? Great system...

[jack11111]

Capn Bloggs, Reread page 4...Right you are!

[snchater]

I have been told by an industry insider that CVRs on CAT were randomly selected for playback as part of a HP study.

One of the most frequent recording heard was " what's it doing now?".

Increased automation has significant benefits but the system limitations must be fully understood by the crew.

[NorthSouth]

Scary biscuits. I wonder if it was a factor that the captain only had 143 hours on type and looks like he'd only been flying it for about four months?

[Basil]

Spot of remedial training to tick the box but neither of them will ever make that mistake again.

" what's it doing now?".
Yup! Heard that a few times on them new-fangled hairyplanes.

[Nock187]

Reading the report it seems more basic in that the pilot may have expected his control column inputs to disconnect the autopilot as per more sophisticated APs.

It was implied in the report that how to disconnect the autopilot may not have been fully known by some pilots/operators. Remedial actions included the operator revise their SOP to explicitly state AP disconnect methods and to incorporate it into their Type Conversion and Recurrent training.

It's interesting though, how would most pilots instinctively expect an AP to function when say, after having converting from a type with a more sophisticated AP to a less sophisticated AP or vice-versa?

Further, the design sounds a bit contrary.....it has sophisticated FBW, no mechanical trim-tab elevator control system but mated to an AP with only basic logic/functionality that ignores most of the flight control inputs the pilot is making?

[Basil]

Nock187, Yes, doesn't matter what the book says, I'd still press the disconnect button.
I guess we should remember they'd just had a lightning strike which I recollect as VERY loud (and Bas is ex artillery) which could have caused a little disorientation leading ol' skip to think he'd already pressed the disconnect.
Not excusing, merely looking for reasons - there's always a reason.

[parkfell]

This Saab 340B AAIB report was published in September 2006. (09/2006)

Icing conditions with the AP engaged. Again some initial confusion as to what was happening. An experienced Captain, with an FO who was previously a licensed engineer on type. Probably helped the situation?
As part of the AAIB involvement, an Inspector spent a day on the Saab jumpseat. He was then flying a fews days a month with Flybe on the Q400.

[Nock187]

Thanks Basil, quite possibly, although initial reading of the report that may not quite fit? Admittedly the initial bulletin doesn't state either way.

The sequence of events was:


Approach with AP On with Heading Select and Altitude Tracking modes selected
Decision was made to discontinue, and the pilot then manually turned the aircraft using control inputs. So he was manually making control inputs before the lightning strike, whilst the AP was still engaged. (I'm interested though, on a type like the Saab 2000, if the approach was discontinued and needed manual control inputs like a heading change, would a crew disconnect the AP first or not? Maybe he thought the manual control inputs had auto-disconnected the AP at this stage in the flight?)
As rolling out, lightning strike and pilot pulled back in a climb in response and (tried) to manually add in more up-trim to boot. Initially he "won" against the AP, but it progressively added more and more down-trim to get back to its altitude setting


The very first bit of the subsequent advice published by Saab, in response to the incident, to operators is interesting too, as it implied that some operators/pilots may not know that manual control inputs or main trim switch inputs do not disengage the AP on this type. That in turn implies that some pilots may "instinctively" think that manual control inputs disable the AP and they don't necessarily have to hit the Disconnect button?

"Manual control inputs will not cause the autopilot to disengage and the main trim switches are disabled when the autopilot is engaged. Consequently, operation
of the main pitch trim switches will not have any effect on aircraft trim nor cause the autopilot to disengage.

Disengaging the autopilot is normally done by pushing the disconnect button on either control wheel..."

[Nock187]

Ah, my apologies, I need to read reports properly!

You're right, the bulletin doesn't say (and my error was saying the report stating that pilot made manual control inputs to turn away to a new heading after discontinuing the approach.....re-reading it actually doesn't say the pilot used the control column to initiate roll/heading change).

Further re-reading it, it implies it is more in line with what Basil said, they probably left it on AP and used it to change heading after discontinuing approach....this is implied by the event straight after the aircraft just completed the roll out and the almost-immediately-after lightning strike with the commander announcing that "he had control". This means he probably didn't have his hands on the control column prior to the lightning strike but he was effectively announcing he now had his hands down to it.

[parkfell]

This was a disorientating experience, and would result in a significant reduction in spatial awareness.
You can discuss these events during recurrent training, but until you actually experience it you don't fully appreciate just what impact it can have on your mental processes at the time.
Sleepless nights are not uncommon, and visits to the "Trick Cyclist" would be perfectly understandable.

The Captain was PF, and following the event he said " I have control ". This implies reversion to manual flying? So did the AP "cavalry charge" alert sound due to the strike, but the AP stayed engaged?
Why did he not push the AP disconnect button on the control wheel? The effects of mild shock?
The AP operation of the 2000 is not the same as the Saab340 which I guess was the crews previous type? On the 340, operating the pitch trim switch WILL disconnect the AP. This is not the case on the 2000. You can see how confusion might arise.

[Jwscud]

I do have to wonder why the system was designed to allow pitch trim with the AP engaged.

On every type I've flown, manual pitch trim with the AP engaged disconnects the autopilot immediately. Indeed on one, using the pitch trim to disengage the autopilot was the most common method used by pilots (the other was the MCP button) as the master cutout switch on the yoke also took out the YD, which was typically not taken out until minimums.

[haejangkuk]

Wow, what a scary incident? Tsk tsk tsk! How did this incident escaped notice by aviation forums? Hmmmmmmm.

The A300 Supervisory override feature with A/P engaged and THS auto-trimming was the cause of a couple of fatal accidents.:(

[Naali]

Would the Go Around buttons in the power levers have solved the problem? Haven`t flown 2G,but in 340,if i remember right,GA buttons disconnected the a/p and Eadi moustache reverted to G/A pitch guidance,so it was hand-flying then on. Correct me if i`m wrong.

[parkfell]

The lightening strike triggered the AP disconnect "cavalry charge" to sound.
They believed that the AP had disconnected ~ that was their mind set ~ that was the fundamental issue ~ until the penny finally dropped.

Learn from the mistakes of others, as you won't live long enough to make them all yourself.

[stilton]

Well, apart from anything else, 330 Kias in a Saab ?



They built a strong aircraft !

[Smilin_Ed]

1. I continue to be puzzled as to why they were still in autopilot at that time. Nothing concentrates the mind on keeping situation awareness than manual flying. That's what they should have been doing, not playing "video games" by driving around pushing buttons.

2. Even more puzzling is why the AP doesn't disconnect when the trim switches/buttons are activated. If you need to manually trim the AP isn't doing its job properly and should be dumped. At the end of the report, I fully expected a recommendation that the AP be automatically disconnected when trim was activated or control forces above a certain level were experienced but was disappointed.

3. The first thing they should have done when the plane did not fly right was to push the autopilot disconnect switch(es). :ugh:

[Oakape]

I do have to wonder why the system was designed to allow pitch trim with the AP engaged.

The system does not allow pitch trim with the AP engaged. The manual pitch trim switches are inhibited with the AP engaged, so they don't actually do anything when the AP is engaged. That is why they will not disconnect the AP as well.

[Self Loading Freight]

The Daily Record has a pretty breathless (http://www.dailyrecord.co.uk/news/scottish-news/hero-airline-captain-saves-scots-5495288#rlabs=1) 'hero pilot wrestles with the controls' report now, which is basically a rehash of the AAIB narrative.

The AAIB report doesn't say that the cavalry charge annunciator sounded - was that info from other sources? Also, I've always thought that a lightning strike on an aircraft isn't that serious a technical event. Was there another reason for the mayday? (You don't need a very active imagination to see that getting hit at night, at 2000' over the North Sea just as you've broken off an approach, in a small aircraft and a big winter storm, may dispose one to expedite being somewhere else.)

I've done a fair bit of paxing into Scottish islands and know how exhilarating the weather can be - doesn't Saxa Vord still hold the wind speed record, achieved just before the anemometer blew away? I would not have liked to have been up front that night, that's for sure.

[Gertrude the Wombat]

One of the most frequent recording heard was " what's it doing now?".
Ah! - so maybe the voice recognition feature of the system ought to be configured to recognise that phrase and ...

...

...

... hmm, yes, tricky, init - perhaps your choice of:

(1) cut all the automatics, so that the pilots are just flying an aeroplane and there is no longer any scope for confusion

(2) disable all manual controls and fly the aircraft completely automatically, because the pilots have obviously lost the plot and can't be trusted to mess with it (maybe enable the controls again once the pilots have explained to the computer's satisfaction that they really are back in the loop).

[RAT 5]

There has been much chat & comment during various different threads on here that all lead to the same conclusion: why did the crew not disconnect the automatics, stabilise the a/c and then determine the true status of the a/c? The thought is often that they were scared to do so, or were uncertain how to control the a/c manually. It could also be they did not detect an automatics malfunction in the first place. (see the thread about Unreliable airspeed) Thinking back to my years of recurrent training sessions, where there is a 3 year cycle so all systems are covered, there was very little manual flying of a broken a/c. (I do not consider a few moments of unusual attitude upset recovery to be manual flight training). I do not remember any LOFT or Reccurency training sessions which involved automatic runaways or subtle misbehaviours. Yet this is what we see in the real world as leading to many major incidents & accidents. We train and experience many failures that are quite rare and not too difficult to handle, but the really dangerous and difficult to handle scenarios are not included.
Food for thought? When writing recurrency exercises I often included scenarios of recent accidents that had been reported. The crews knew of their background and some of the why's and wherefore's. They learnt a lot from the experiences. That was back in there day when the chief pilot & chief instructor had more flexibility. Those scenarios ticked some of the mandatory systems boxes for recurrent training, but there was more meat on the bones. That's what I find missing these days; no meat on the bones.

[w14]

Aircraft Accident Report AAR 2/2016 published today.

https://assets.publishing.service.gov.uk/media/57ce96af40f0b65264000010/AAR_2-2016_G-LGNO.pdf

"Five Safety Recommendations are made relating to the design of the autopilot system and the certification requirements for autopilot systems."

[Nil further]

Not a great advert for CRM or training within Loganair. ......Although is it the same everywhere ? Emirates looks like botched GA multiple EZY flaps selected instead of gear or vis versa . Asiana can't fly visual approach Turkish 3 pilots can't monitor speed .
Air Asia land at Melbourne instead of KL.

Are we now seeing the " children of the magenta line" coming home to roost ? What will it be like in another few years when the last of the old school are gone ?

[HeartyMeatballs]

Oh dear god. If I have a dollar for every time "children of the magenta line" mentioned here I would be minted.

The problem is two fold. You now have airlines who 'encourage' pilots to hand fly but also encourage the highest appropriate level of automation during an approach. If you stuff up the airline is not going to congratulate you for fine tuning your hand flying skills. It is going to come after you. Some even get twitchy about a visual side step and will put in the ILS and fly said ILS if offered the side step.

You've also got the social media generation. For them the job is just an excuse to stick themselves on Twitter or Instagram. Take one of our yung'uns: literally every time you see him he's taking a selfie. People are doing the job for the social status and bragging rights and have no interest or passion for the job. These go on to be young captains and can be prima donnas and think they're gods gift where in fact they're more donner kebab when it comes to their command skills.

Unfortunately there's very little than you can do about those with poor CRM skills. As long as they pass their sims they will continue.

[Nil further]

Not sure what your argument is there Hearty , are you saying that the concept of "children of the magenta line" is not relevant to the above noted incidents ?

You may have a better definition of the phenomenon but it seems to me the clearest and easiest way to describe automation dependency .

have another dollar !

[Chronus]

Extract from AAIB report below:

"After reaching 4,000 ft amsl, the aircraft pitched
nose-down to a minimum of 19° and the applicable maximum operating speed (VMO) was
exceeded by 80 KIAS with a peak descent rate of 9,500 ft/min."

The bit about minimum 19 degree nose down is somewhat puzzling. I would have thought it is max that matters.

We all do of course express concerns about failure to disengage the A/P. Here are some clues.

Low time on type, yes, but what type(s) before.
Trigger lightning. This is a particularly nasty phenomenon, ask any North Sea chopper pilot. A trigger lightning strike forced down a chopper to ditch in the North Sea in January
1995 and was a causal factor in another fatal accident that occurred in July 2002. It has been suggested that it is the aircraft that triggers it off. There is no prior warning of it n`or of its presence. More info on this at https://www.linkedin.com/pulse/triggered-lightning-phenomenon-puts-fear-helicopter-kevin-forbes
I can well imagine the effect this sort of event may have had on the crew.

[HeliComparator]

Extract from AAIB report below:

...

The bit about minimum 19 degree nose down is somewhat puzzling. I would have thought it is max that matters.




Nose down is expressed as a negative number, so for example the minimum of -19 and -18 is -19.


Anyway, helicopter pilot here, highly automated type. It is easy to ridicule the pilots for not noticing that the AP was still engaged but that is the sort of mistake that should cause you a bit of confusion, not the sort of mistake that should nearly result in the aircraft spearing into the ocean at 9500'/min rate of descent well over max speed having held the control column right back for the duration. In other words, whilst the pilots have some culpability, most of the culpability rests with the system designers and certifiers. Which I think is what the AAIB are trying to say.

[The Ancient Geek]

The main point in the report is that when the pilot is clearly trying to fly manually by moving the control column or the trim switches the autopilot should get out of the way and let the pilot fly the aircraft. The Saab 2000 is the only aircraft which does not do this and pilots are conditioned by their experience on any other aircraft to expect this to happen.
This is a basic design flaw which creates a nasty trap for the unwary. When the going gets tough we all rely on our experience which in this case is dangerous.

[Tinstaafl]

The report also mentioned that there are two commonly used a/p response to manual input - control column *or* trim switch - and that is for the a/p to disconnect. This type doesn't do that.

The captain said he was trying to trim as well as pitch up. In all the other types he flown, including the Captain's recent previous, the a/p would have disconnected with those same inputs. Doesn't matter which input - column or trim - because the pilot was doing both.