ATSB report just published on A320 throttle asymmetry incident
 

The report:

Investigation: AO-2012-022 - Inadvertent thrust lever asymmetry during the take-off roll involving an Airbus A320, VH-JQX

It is well written and well worth close study. It took a year to publish which is surprising for such a relatively simple incident. Seems during the takeoff roll a small asymmetry between the two thrust levers was noted. The left seat pilot was under training for a command while the right seat occupant was his instructor. Without going into details here (see the ATSB full report), the captain under training heard what he thought was the word "Rotate" from the instructor (PNF) and commenced rotation before realising the airspeed was still 20 knots below the correct VR. The aircraft became airborne but the pilot was able to pick up speed by reducing the climb attitude.

Rotating on the call from the PNF of "Rotate" is something often seen in simulator training even though the airspeed indicator of the PF shows well below the correct VR. It goes to show that the word "Rotate" in a firm voice can very much sound like a command, rather than just a support or advisory call by the PNF. Of course, in real life and no significant airspeed difference between the two main ASI's, "Rotate" is invariably ties in with what the PF expects to see on his ASI.

In the old days, the calls were simply "V1.....VR". Boeing changed "VR" to the new term of "ROTATE" and while that has been in use for decades I wonder how many unreported incidents have ocurred where "Rotate" called in a firm voice has caused momentary confusion if an ASI defect is present and the tendency is to respond to a firm voice (in this case "ROTATE") rather than on a correct bugged airspeed.

But there is a powerful tendency to start rotating at the word "Rotate" depending how forcefully the word is annunciated and despite the fact the airspeed on the side of the PF has not yet reached the correct VR. In other words, because things happen so fast approaching VR, it is a human factor problem where it is all too easy to act on the PNF's support call rather than rotate on the bugged VR on the PF's ASI.

In the simulator, this is most likely to occur when the instructor introduces an airspeed error between the two main ASI's with the PNF calling "rotate" on his own ASI despite the airspeed of the PF being nowhere near VR. Invariably we observe the PF then start pulling back on the control column in momentary confusion even though his ASI may be nowhere near his bugged airspeed. This is where it is good airmanship to be aware of the expected ground speed at VR as a double check.

ATSB missed a good opportunity to emphasise the need for the PF to plan on rotating on his own ASI bugged VR and treat the PNF support call of "Rotate" as purely advisory. if the PF reaches his own bugged VR speed but fails to get a support call from the PNF for whatever reason, he simply rotates (a quick glance at the ground speed indication is wise at that point). There should be no confusion and the problem sorted out at a safe altitude.

Every so often you com across some gold on pprune.

This is just such an occasion.

It certainly makes me think.

Pay peanuts........:yuk:

How would you propose to observe that groundspeed?

For me, I scan the ASI along with temps and centreline until I am crosseyed by which time it must be time to haul back. Non standard calls during the roll are extremely difficult for just this reason.

By the way, tell me about these throttles? What sort of jet engine uses carburettors?;)

[QUOTEHow would you propose to observe that groundspeed?
QUOTE]

If you don't know where to find the ground speed you need refresher in the sim:ok:

If you have not heard of autothrottles ditto above :=

A37575

Right on the button. Three cheers to that man!

Thanks for illustrating my point. If the PF has time available to look at the GS on the panel then he or she could just as easily look at the much more relevant airspeed on their side. Scanning groundspeed for rotate point is about as useful as scanning groundspeed at 100' agl on final.

Yes there is an autothrottle system sometimes also referred to as autothrust but the levers in the cockpit/flightdeck/bridge/control room/reading room/pointy end are most definately called thrust levers or power levers by the jet manufacturers I am familiar with.

Anachronism anyone?

A37575 - I think you need to consider the psychology of the process. In essence, the word "rotate" could be changed to "Vr" or "banana's" - it wouldn't matter. Any call has the potential to become a classically conditioned response. What you are arguing (correctly) is for the PF to ensure he does not fall for this conditioning process, and actually cross-check against some independent system. Changing the words isn't enough. If you look at the Boeing Takeoff manoeuvre in the QRH it simply has V1 ...VR Rotate, and offers no guidance or hint of a crosscheck. The FCTM goes into elaborate details of tail clearances - again anything but the need to crosscheck ASI's.

This is a huge advantage that the HUD offers, your own ASI is right there in your field of vision, together with the tail strike warning, is worth its weight in gold.

A37575 – Pprune at it's very best. There is no icon for 'Hats off' but consider mine doffed.

It's an old debate now, but years ago a couple of overseas companies I had the pleasure of working with used a "80 knots" callout, the idea was that it provided, well before any critical speed an ASI, engine power check and a warning of the approaching V1 speed range. I liked it and found it helpful. The GS check is also very useful, particularly when the speed tape and GS readout are on screen, but the early cross check built 'time' into equation. I still do it – in my head – just a quick look to make sure the power is balanced, the GS about what I'd expect and a glance at the standby, takes longer to write it out than it takes to do. Not advocating - just my slightly battered two bob's worth.

This report is interesting as it describes one of those 'sneaky' little holes in the cheese which can distract and create havoc; experience, training and discipline greatly benefit on those occasions, particularly at the pay attention end of a take off roll. All's well that ends well.

We used 80 knots as a check. I thought it was in general use.

For the love of God, this was in Australia you guys are talking about. The pilots here are supposedly God's gifts. It just cannot be!

Rotating at 20 kts below Vr? Unless it is a contaminated runway with the correct take off speed computation, there is a very small V1/Vr spread. As such, it was likely that rotation was accomplished before the V1 call. I do not recall take off V speed callouts on the A320 reverting to previous practice by certain operators of F-27 whereby the only callout was " rotate " when V1=Vr!

Just as a matter of interest, what inputs are used to calculate ground speed and will it still read correctly if the ASI is in error?

T/L issues
 

I guess after a minor 'lever assymetry' issue on take off ( and by the way as the detentes wear out it gets even easier to not quite have them both/ all four in the slot, thus generating the assym ecam and requiring a minor amount of fiddling on the odd occassion ). The next thing we might see in an ATSB report is the use of FLEX/MCT when a flex temp is missed out, generating the THRUST NOT SET ecam as well....ending in an RTO or similar perhaps?

Both of these very minor issues IMHO should 'not result in anything more than either pushing up the levers to TOGA ( actions for the latter ECAM actually tell you what to do) or a bit of adjustment . Bearing in mind the actual thrust set will be TLA determined ,normally checked at 80 kts ( yet with no acknowledgement of the 'THRUST SET ' PNF call by the PF - which seems strange to me , but is AB standard now.) .. Then '100 kts ' as the hi / low speed RTO divider and x check etc ( or 80 kts on a Boeing if I remember)

As far as rotating 20 kts early, yes, perhaps a pavlov dog reaction , but hopefully just an ease off of the pitch rate would have prevented any tail strike as with both engines operating the jet would pretty quickly be there anyway.

To be honest , if they did not bend the jet or generate a QAR event which they had to answer to the safety department for, it is hardly worth even being a reportable incident in my view... Simply a debrief item to learn from ...well done for the crew coming forward if they did so voluntarily as it has generated some good discussion.

But this is Australia And we do like to get wrapped around the axles for trivial issues...Go work overseas for a while and see how bad things have to get to hit the papers!!!

This is not that different to other incidents where the gear up was called for and the flaps were raised instead. It all boils down to the age old communication problem of what was said being different to what was percieved. Mandating use of the interphone I think would help mitigate that problem. A jet cockpit is still a noisy work environment and as both pilots are looking out the front, the other cues that help a person decipher what is being said by the other person are not present. At least using the interphone what is being said is heard in the headset and not a distorted bunch of words coming from the other side of the cockpit.

I would imagine that the Training Captain got a nasty shock when the nose started coming up in response to his "Go..TOGA" command.

Especially if that call comes from a Training Captain in the RHS!

Interesting. Are there any jet operators in Australia that do not use headsets & intercom as standard below transition? Maybe not mandated but always used?

Both looking forward? Not sure I agree. As PF I am trying to keep the beast on the centreline, scan speed and temps, when the FO is the handling pilot the same as well as being ready to take-off for an RTO.

Headsets always on below cruise altitude.

Interphone is optional.

Most guys don't use it, however overseas operations can require both speakers on ears and interphone on.

Geragau, I agree with you completely. If one does not receive the V1 call but only the call to " rotate ", something is definitely not right...cause for pause, query and a quick crosschecking. For these two highly experienced pilots to be caught in this incident is highly troubling.

Agreed but where is most of your focus?

The rest of the statement puts the first part in context.

A headset without using the interphone actually hinders intercockpit communication. One ear is covered and the other is not so its not surprising that in this instance the PF misinterpreted what was said by the PNF.

Doesn't matter how experienced you are as an F/O, when you are doing Command training on a new type you are on a steep learning curve and mistakes will occur. The Training Captain can't anticipate every unexpected response if he hasn't experienced it previously.

As a mere PPL two things come into my head; would they not have calculated the proposed take off roll, and would this not be able to be gauged by the amount of runway used up, and, would it not be unheard of to glance at the ASI before committing to becoming airborne?

Read the report 10 times. I'm not looking forward to my A319 endo! http://www.smilies.our-local.co.uk/i.../scratchhd.gif

I think more plain language from training captain was in order, and/or just push up the throttles to TOGA himself. "Thrust asymmetry, selecting TOGA" or similar.

"Go"? one syllable, easy to mishear?

Good learning experience nevertheless, even for us non-computer types. Good on the crew for reporting it. :D

Why airlines dick around with manufacturers recommended procedures and SOP's is beyond me.

The ATSB show again that they cant do any sort of analysis or have any expertise. This is just a regurgitation of an internal Jetstar report it would seem.

OEMs build them - they don't operate them!
 

Hey Rudder,

What are the "manufacturers recommended procedures and SOP's" that apply in this situation? :hmm: :hmm: :hmm:

Then this trainee captain IS DEFINITELY NOT READY for command! As has been pointed out, rotating without crosschecking is a no no, especially if there was no prior V1 call.

Ah, ah, ahhhhhh bullish!t he's not :cool:

How did this make it to report stage? Bro in the right seat shoves the throttles full stick, tells bro in the left seat, lower the nose a bit bro, climb, climb, climb. Don't do that again bro! Sorry bro....

Scrubba,

I don't know what Rudder has in mind but being an Airbus trained TRE/TRI my view is that the SOP's call for a thrust set call by 80 its. If it isn't then a slow speed rejected takeoff would have been more appropriate and required by the SOP's otherwise why have the call in the first place.

From what I read, the Pilot under training was not with the aircraft even at this early stage and the training captain sitting there and hoping this would resolve itself by directions probably is a bit passive. While clearly the Training Captain knew that the thrust was at least forward of the detent and had at least the minimum power expected for this particular takeoff, the guy with his hands on the thrust levers had no idea.

While I clearly think that this takeoff should have been rejected. Another course of action given that the training captain was situationally aware of the power, was as Capt Bloggs indicates and that was to push the lever forward himself to TOGA ASAP and take over.

I have no idea how anyone can get Rotate out of GO TOGA but it does indicate that probably the guy under training was still trying to in fact figure out what was going on and was well behind the aircraft and this is the guy pulling the aircraft into the air!!

Bloggs you will love this aircraft. You just have to give up on Boeing and come across to the dark side!! (Sniggering aside they are really great)

Never trust an A320 NEVER :ugh:

Anti skid on, take off distance is not calculated. Speeds are calculated to fit in with all the factors. Estimating distances is pure guesswork based on familiarity with the runway and conditions. But yes, checking the ASI should very much be part of the scan.

"Go toga" and "Rotate" probably sound similar in the fog of war.

Maybe so, and impossible to argue against that.

However, why didn't the cognisant lack of a V1 call raise a hair on the back of the PF's neck? Shouldn't the PF have a chronological feel for when the call should be coming from the inertia of the aircraft? 130-150 kts is a fair split and if you "thought you'd missed the call" wouldn't you also be thinking it's a bit early and double check your speed tape?

I assume the speed tape scan was part of the PF's scan? The report doesn't mention the 100kt call though either....

PS. I'm giving the ATSB the benefit of the doubt for the lack of detail in this report due to the fact that the CVR would have been overwritten...

With the 737 there is a 80 knot call and with Airbus there is a 100 knot call. The principle as far as I know is a check that both ASI's are within a bull's roar of each other. In other words guarding against a gross error. A wise man will know the wind velocity before the start of the take off run and by association the expected ground speed read-out at the 80 or 100 knot call for the aircraft type.

As a general observation it is probable that an experienced simulator instructor will see more events or incidents in one week while training various ranks of pilots than most line pilots would see in 20 years. So the opinions following are the result of watching years of "events" with erroneous airspeed during type rating and general simulator training.

It takes less than one second to glance at the ground speed indicator during the take off roll at the support call of 80/100 knots. That is of course if you know exactly where the ground speed indicator is located on the instrument panel.

You can never know if a hidden defect is present giving erroneous airspeed. It could be on the take off roll which is critical - or some other time. Indeed, a reminder of this fact may be found in the B737 FCTM where there is a statement "ground speed information is available from the FMC and on the instrument displays (as installed). These indications can be used as a cross-check."

A year or so ago ATSB published an incident report to an A330 on take off at Brisbane with the F/O as PF. I don't recall the exact details but I think there was an insect stuck in the captain's ASI which had the effect of causing his ASI to under-read by a significant amount. Passing 100 knots the F/O expected a support call of 100 from the captain but didn't get one. He said nothing assuming the captain had been distracted and forgot to make the call. Approaching VR on the copilots side and as the copilot was preparing to rotate, the captain called "100 knots". There was immediate confusion as you can imagine. After a very brief moment the captain took control and made high speed abort. Back at the gate the fusible plugs let go as the brakes were hot.

In the Boeing FCTM there is a statement under the general heading of "Callouts" and says "The PM makes callouts based upon instrument indications or observations for the appropriate condition. The PF should verify the condition/location from the flight instruments and acknowledge. If the PM does not make the required callout, the PF should make it."

That is why it is good airmanship for the PF to call something like "95 knots my side" if he didn't get the call of "80 Knots" from the PM. The high speed of the abort by the Brisbane A320 might have been avoided if the PM (the F/O in this case) had called "110 my side" during the take off run instead of saying nothing. While 100 knots was the SOP call here by the time the PF has reacted to no support call by the PM the aircraft would likely have been 10 knots faster - hence the theoretical "110 my side" call by the PF. That would have alerted the captain to a potential problem and he would have to decide to take over and abort or tell the PF to keep going especially as an erroneous airspeed defect is not normally a reason to make a high speed abort.

This then leads us to the value of the ground speed glance where in this case the PF could have quickly noted his ground speed as he passed 100 knots. Having already had in his mind what the ground speed should read approximately as he saw his own IAS going through 100 knots he would be in a good position to know his own ASI was good in event of an erroneous airspeed rare event. In fact there was an erroneous airspeed event on the captain's side - so these things can happen albeit at an unexpected time and very rarely. But that is what good airmanship is all about.

In the simulator an instructor may select a 25 knot airspeed over-read or under-read on one of the two pilots ASI's. Example follows: Captain as PF has the over-reading error but doesn't know about it of course as he starts the take off run. His ASI comes off the stops early and soon rockets past 80 knots. The PM will not know of course. The captain doesn't receive the expected support call from the PM of "80 knots" (Boeing). He quickly calls
"90 my side". The astonished PM mentally thinks "Bull****" as his ASI is around 55-60 knots. He says something out aloud about airspeed disagreement and by this time the aircraft is rapidly accelerating through 100 knots real airspeed.

The captain realises there is an erroneous airspeed problem but there is no time to glance at the standby ASI and evaluate the umpire. Now he could have picked up the problem earlier if he was in the habit of glancing at his ground speed as he passed 80 knots IAS. With his over-reading airspeed he would have been a bit shocked to see his ground speed was something like 50 knots. Providing as a good captain he knew what the W/V was before the start of the roll, he knew the airspeed of 80+ and the GS of 50+ didn't compute and it would then be up to him to continue or stop. He could rotate at the scheduled VR speed on his ground speed reading (allowing for wind component) and at a safe altitude sort out the problem in the air.

This technique taught in the simulator is invaluable and once the crew get into the habit of ground speed awareness because of simulator practice then it becomes second nature for all future take offs. Honestly it is no big deal. But it is a real big deal if a premature action is taken to make a high speed stop simply because of a perceived airspeed difference between the two main ASI's in the middle of the takeof run.

One Pprune contributor in this thread scorned use of the ground speed indication on very short final as useless. In fact those pilots who have flown into some of the atoll runways like Nauru, Tarawa, Truk, Pohnpei and the like will remember the unreliability of local anemometers and at night the windsocks may be unlit. In fact years ago, we had the situation at one island airport the flight service operator announced the wind as calm by simply looking at the forecast wind. His anemometer was u/s and some vandals had shot out the windsock lights. We were on final approach at night at 5 miles and kept a wary eye on the ground speed which showed a 15 knot tailwind from 10,000 ft down. "Wind calm - runway clear" said the local in the flight service tower. That didn't tie in with on course on the VASIS rate of descent. At 500 ft the ground speed was now 10 knots tailwind. The runway was just 5600 ft long and no way were we going to land with any TW so we went around and landed from the opposite direction. It was good to see a 10 knot HW component on that runway.

Morale of that story? Be fully aware of the ground speed indications that may alert you to unforecast or actual tailwinds. Nothing wrong with a quick glance at the ground speed at 200 ft on short final. Especially with varying winds. Similarly on the take off run be aware of the expected ground speed indication as the PM makes his IAS support call for the aircraft type. This could prevent an unwarranted high speed abort because of perceived airspeed reading difference.
Please note: The above comments are personal opinion only. Arguments against should preferably be based on technical disagreements - not playing the man!

Great stuff Centaurus.

Just as a matter of interest, what inputs are used to calculate ground speed and will it still read correctly if the ASI is in error?

RTO for thrust mismatch
 

Thanks d_concord,

I thought the TLA mismatch was surprisingly small to cause this problem and even more surprised that the ATSB made no comment about the apparent sensitivity. For the non-Airbus people, I was also disappointed that there was no discussion of what the available solutions were - I'm guessing that repositioning the TLs is not an option, but I'm also surprised that it is not an option up to a certain point in the TO sequence.

I'm unclear why you seem to think it warranted an RTO. Presumably there is some cross-check of engine parameters to confirm that at least flex thrust is being generated - so if it is available, why not continue? Can normal autothrust be reinstated airborne at a safe altitude? Is manual thrust setting difficult or are automatic features disabled with the A/T?

FWIW, I am particularly glad that it was reported, investigated and published. I thought that a lot more could have been drawn from the event and the lead up to it, as well as the company SOPs, preparation of training captains, etc and what reviews took place post-event.

For those who have never suffered a loss of SA by a distraction at a critical time while under the duress of training, don't be so quick to judge - I've seen some very good pilots do some very unexpected things in similar circumstances and everybody was much better for understanding the experience.

Some very good points Centaurus. I cannot disagree. However you said...
That would be me, for the record what I said was...
My bolding this time around. Your point about "trusting' ground reports of wind is one I take to work everyday. Looking for GS for me is not the solution or really required, simply glancing at the wind vector arrow does the trick :ok: GS is only part of the story ie is it a quartering tailwind vs headwind etc.

I still feel that scanning GS on the TO roll is problematic, partly due to update rates on the particular box, small as that may be in the scheme of things but mainly due to the small window between "power set" and "V1" to scan all the other useful stuff. Anyway, each to their own.

For those unfamiliar with airbus, the GS comes from the GPS through the inertial reference system. It's displayed at the top of the Nav display.

Hi Scrubba,

The thrust TLA can be just a millimetre out of the detent to give this indication on the FMA. It will give this whether one thrust lever is forward of the detent or behind as well. The Levers are just a slide switch so position is everything.

What happened here is not unusual. I have seen other do it and have done it myself. And as you say, in all the cases I have seen or done myself you just put the lever instantly back into to the detent and presto now you have the power you programmed. I have never heard of anyone to just sit there with the problem until now.

This is from the Airbus QRH for the PNF

BELOW 80 kt:
N1 (EPR)CHECK
THRUST SET ANNOUNCE
PFD/ENG PARAMETERS SCAN

You will note it says below 80Kts not a specific speed.

This all takes place almost instantly after the levers have been put into the correct position and the FMA will indicate what has been commanded. This is followed by the PNF confirming that the thrust is set from the engine parameters. This will all happen well before 80 Kts. That is what is unusual here. The reality is the thrust set call normally takes place well before 80 kts. The airbus call is "thrust set" not "80kts Thrust set" or "80 kts thrust not set"

The PNF called "80 Kts. Thrust Not Set" (maybe another issue is why wait till then!)

Its very clear that this was now a non normal take off .

A slow speed RTO is a non event in this case. (That's not the case with a slow speed RTO with an engine out but that is another issue )

To be honest, I just don't think the PF understands the FBW characteristics of the aircraft. If he did he would have instantly known the cause and the remedy.

It is interesting that the findings did not see the need to look at or make any comment on any deficiency in understanding or the need to improve the initial conversion training onto the aircraft. Seems as though they did not even look at the ground course or training syllabus of the organisation that did the training. Given the PF had only 120 hours on type this incident started there.

As Rudder indicated you really have to wonder about the ATSB and the level of experience, skill and analysis.











A very interesting event that has two distinct aspects. First is the actions of the PNF in the 80-120kt region, second the actions of the PF in rotating @130kts. I think the actions of the PF were dealt with above.

But IMHO the action that started this incident was the non standard call "Thrust Not Set" when it appears that it was. That lead to the PF's confusion.

To follow on from d_concord's comments:

Immediately after the Thrust Levers are moved into their takeoff detent, MAN FLX 62 (or TOGA etc) appears on the screen. But the engine isn't at the Thrust Set stage yet. It also takes longer for the IAEs to get to power than the CFMs.

The required N1 (CFM) or EPR (IAE) is displayed on the screen just to the right of the engine N1 or EPR. Only when these two are matched can you call "Thrust Set". This is to be done before (not always at) 80 knots.


Now back to this incident: From the little information available, it would appear that the thrust was indeed matched between that commanded by the TLA and that required (it should have exceeded it, but we aren't told this :rolleyes:) . As such the correct call would have been "Thrust Set" even if there was a MAN THR displayed on the FMA. in this case, the FMA has nothing to do with the Thrust required, which is the purpose of the thrust set call.

The correct call (in addition and either before or after the Thrust Set call as noticed) should have been reading the FMA - "Manual Thrust".

Then if was all still going to custard which is the subject of another debate, Go - TOGA was the correct action (as the thrust was indeed set).


Hopefully this thread can remain a healthy discussion here with appropriate respect to the pilots involved about an event from which we all can learn.

Apples to apples.
 

A valid point, well worth some time and consideration. The slippery feel of 2012 report offers little 'meat' compared to the 'food for thought' of 2006 report and very neatly demonstrates the differences we are consistently seeing in ATSB reports.

Airbus 2006.

Airbus 2012.

There is in the GA section a report on an engine failure, at night of a single engine training exercise. The report has a smug, almost patronising feel to it and provides little of value. It arrives at the end of a long string of similar offerings. If you can bear it, have a look at the Air Rescue, Air North, Pel Air and Canley Vale efforts, they are enlightening.

Oh, almost forgot, overheard at the after cricket BBQ – and this is strictly "off the wind"; but it seems some bright spark at CASA has outlawed the "80 knot" cross check call for "light" jets and similar. Remove all traces from SOP or evil will be visited on your house. Only BBQ fuelled hot air at this stage – but I do intend to get to the bottom of this rumour; if true then someone better tell Boeing and a couple other not very well informed manufacturers.

C/S

I agree with you overall. While it is clear from the report that thrust was probably set. The fact is that neither pilot indicated they knew what thrust was set. The PF had no idea what was going on and the PNF (and training captain) called it as not being set. Smart move would be to pull it up from low speed. I don't agree with your go call at all "in this case".

There is lots not to like about this incident, most of which has been covered above. Another thing to look at is the role of the training captain. A simple advice to bring the lever back to the detent would have fixed this from the second the FMA came up . Thats what training people would normally do.........! There were passengers on this aircraft. The concept of lets see if he works it out is probably not a great one.

Let me expand on the intent of my penultimate paragraph which i admit isn't clear, probably due to watching tv at the same time...

IF the thrust had been determined to be at least what was commanded then there is nothing inherently wrong with the GO command, particularly if the MAN THR was noticed over 100kts.

I still feel had the Airbus SOP calls been followed it is likely that this would never have happened.


Further criticisms of the nuances and actions (as to whether he should have left it or not) in the role of the Training Captain I'll leave to those of you more qualified than I.