Pilots and ATC both at fault in TCAS incident

[armchairpilot94116]

Taipei Times - archives (http://www.taipeitimes.com/News/taiwan/archives/2008/08/16/2003420504)

Pilot, controller to blame for air incident, ASC says
FASTEN YOUR SEATBELATS: An investigator said that passengers on an FAT flight experienced something like a rollercoaster ride when their plane took evasive action
By Meggie Lu
STAFF REPORTER
Saturday, Aug 16, 2008, Page 2
The Aviation Safety Council (ASC) yesterday concluded a 21-month investigation of a midair incident over South Korea between Taiwanese and Thai passenger aircraft, and said that both the Taiwanese pilot and the South Korean air traffic controller were at fault.
The incident occurred on Nov. 16, 2006, when Far Eastern Air Transport (FAT) flight EF306, which was flying to South Korea, made an abrupt midair dive in order to avoid another passenger plane, Thai Airways (TA) flight TG659.

The maneuver resulted in 20 passengers and crew suffering injuries, said the ASC’s director of accident investigation, James Fang (方粵強).

Passengers would have been able to avoid injury had they been wearing their seat belts at the time of the plane’s sudden descent, Fang said.

“The near-collision was the result of a chain of events,” said the director of the ASC’s investigation laboratory, Michael Guan.

TOO CLOSE

As the FAT and the TA aircraft were too close to each other, the traffic advisory (TA) signal on both planes was triggered, Guan said.

At about this time, the controller at Incheon International Airport made several vital mistakes.

Initially he gave flight EF306 contradictory instructions and then referred to the plane as “EF308,” Guan said.

“The controller later said that because he was busy trying to find a plane that had disappeared off his radar screen, and since he was the only person tracking both incidents, his hands were tied,” he said.

Without clarifying the instruction, and with the auto-pilot turned off in response to the TA signal, the FAT pilot, upon seeing the resolution advisory alarm triggered, failed to obey the traffic collision advisory system (TCAS) on his plane and put the aircraft into a nosedive, Guan said.

“Making a decision based solely on visual data is against protocol … the pilot later claimed that he had seen a shadow above him, which was why he descended, however later analysis confirmed that there was no object near his plane; he should have followed TCAS advice and gradually descended,” he said.

Though the pilot only dipped the plane for four seconds, the aircraft plummeted a total of 2,400 feet (730m) in 17 seconds because of the acceleration, he said.

To the passengers inside the plane, the experience was similar to riding a rollercoaster, he said.

To avoid future aviation safety mishaps, the ASC has formulated several guidelines.

STAFF LEVELS

The council suggested that Incheon International Airport increase staff levels at its air traffic control center, that pilots as well as controllers use professional language when communicating and that the public be made aware of the importance of wearing seat belts during flights.

“In many air accidents we found that those passengers who were injured almost always did not have their seat belts on, whereas those who wore seatbelts had a much better chance of coming out uninjured,” he said.

[punkalouver]

It seems to be a belief among more than a few pilots that an RA requires an extremely aggressive action to avoid a collision. It does not. If you get an RA to descend like this pilot did, do a basically normal although purposeful descent placing the VSI needle in the range where you are being directed to do so. It is a 1/4 g manouver. In other words if you were at 1g like normal, the manouver for descent will initially have you at 3/4g. People should not be floating in air.

[misd-agin]

Ham fisted response caused the problem, not the TCAS alert or the correct response to the alert.

[PJ2]

It seems to be a belief among more than a few pilots that an RA requires an extremely aggressive action to avoid a collision.
This link to the TCAS (http://www.pprune.org/TCAS%20II,%20Version%207) II Version 7 document provides detailed information on this and other system design characteristics as well as expected crew responses.

Some carriers have used their FDA Programs intelligently by actually doing studies on crew TCAS responses. The data on crew response is enlightening to say the least. No.1 error was responding incorrectly to negative RA's, (don't climb nnnn fpm, don't descend nnnn fpm, maintain nnnn fpm etc). Interpreting the "fly to" commands ("maintain vert spd" especially if one was in the red area) was next and the largest "contributing" factor was leaving the flight directors ON (Airbus fleet types). Most noted errors in correct responses was "too aggressive" - overshooting altitudes, exceeding vertical speed requirements and by virtue of that, exceeding 'g' levels required.

So it isn't only ignoring ATC instructions to climb or descend that is the issue in TCAS responses. Crews are mis-interpreting commands, not turning f/d's off on Airbus and responding too aggressively.

[Magnet77]

TCAS Resolution Advisory (RA)

Our Company’s procedures for a TCAS RA are identical to Airbus’s procedures to my knowledge, and that is:

AUTOPILOT –-------------- OFF
BOTH FLIGHT DIRECTORS -- OFF
VERTICAL SPEED –--------- ADJUST TO THE GREEN (promptly & smoothly)

Note: Avoid excessive maneuvers ....

Apparently the TCAS systems are designed to give the pilot(s) 5 seconds to respond to a primary RA (Resolution Advisory) which is carried out only in the vertical plane (eg no turns). In the unlikely event of a secondary TCAS RA occurs (you’re diving or climbing into the path of a third aircraft) while carrying out the first RA maneuver (nightmare scenario) the pilots have 2 1/5 seconds to respond correctly to the secondary RA.

5 seconds is a fairly long time so there is no need to rush it, but also there is no time to sit on your hands either. Just a normal disconnect of the A/P and F/Ds - OFF then push or pull gently in the direction that TCAS is directing. TCAS events should be part of every simulator training and license renewal, but unfortunately they are not. If they were, then line pilots would not get startled at an occasional real world RA occurrence and possibly overreact.

Granted, if an RA occurs and at that same instant you fly under a dark cloud that shadows the flight deck and the unexpected combination sends a chill up your backside, thinking another aircraft is descending onto the flight deck from above, you might over-react and shove a bit too fast and too hard on the yoke or side stick. That’s a natural human adrenalin response to fear invoking situations. (P.S. Pilots are human too.)


The quick abrupt response of the Taiwanese pilot would have been correct and would have saved lives if there really was an aircraft present and descending onto his flight deck from above, but the identical same pilot actions are easily criticized by the Media, Management, Safety Review Boards, etc, if there was no other aircraft; just the perception of one.

In my early days of flying, I climbed out of a small uncontrolled airport with no radar coverage and well before TCAS was invented. During climb passing 3500’ a sudden shadow above the aircraft caused me to look up abruptly through the eyebrow windows to see the underbelly of another aircraft a few meters above us. The other aircraft had flown overhead the airport without seeing us taking off below because we were below his line of sight over the glare shield. We did not see him because he was above and behind us.



Only my quick abrupt response of pushing the nose down and pulling off power avoided a mid-air collision, but all onboard floated up against their seat belts with the negative g's produced. The two aircraft could only have missed one another by a few feet. And yes everyone thankfully had their seat belts fastened so no one got hurt.

It’s easy to throw stones at pilots that have to make split second decisions during a “oh-sh*t” moment, but that’s why we get paid the big bucks (or the little bucks as the case may be). Our lives and the lives of the crew and passengers depend upon the pilot’s quick reflexes tempered with experience and training.

Be that as it may, if no one had gotten hurt in the TCAS RA event executed exactly as the Taiwanese pilot referred in the primary article did, then the ASC investigation, if any at all, would have centered on the Inchon ATC under-staffing, or inattention due to overloading in his sector, or lack of procedures, or lack of supervisory oversight, that contributed to the circumstances that produced two aircraft at the same level and then giving contradictory ATC instructions using the wrong call-sign numbers.

Seatbelts fastened make a heck of a lot of difference to outcomes under such unexpected circumstances.

I hasten to add that the Cabin Crew working the cabin and galleys would have still been vulnerable to injury even if all passengers had been belted in, and therefore, such abrupt maneuvers must be avoided as much as possible in airline operations.

That’s another side of the story to consider. Cheers, Magnet77

[Centaurus]

AUTOPILOT –-------------- OFF
BOTH FLIGHT DIRECTORS -- OFF
VERTICAL SPEED –--------- ADJUST TO THE GREEN (promptly & smoothly

I don't know about the Airbus types but why the urgency to switch off the FD? It takes a finite time to click off the autopilot and get cracking on the avoidance and at least 2-3 seconds for the PF to lift a hand to switch off both FD. And even if you request the PNF to switch off both FD's that too takes a couple of seconds - and all for what? With only five seconds to act, switching off the FD is the least priority. No one in his right mind is going to follow FD gyrations during a TCAS RA and therefore the FD indications are simply disregarded until the danger is passed, then appropriate FD re-programming can take place. Perhaps the order of action should be autopilot off, vertical speed adjust to the green and finally both FD switches off.

[DeeTee]

Ah, fair question....ON the Airbus (319 anyway), the FD are intrinsically linked to the autothrottle. So assuming a high rate climb, and then a Descend TCAS RA, if you ignored the FD on an Airbus, disconnected the AP and pushed down the FD would still be commanding a climb, the autothrust would follow the FD and command CLB thrust and you would dive away from the RA with climb thrust set, the Thrust would Lock to climb (so even closing the thrust levers would not close the thrust). So, there would be a massive risk of an overspeed (despite the aircraft attempting to protect itself it would still overspeed). The only get out woud be to both disconnect the A/T and close them, this is not a normal Airbus manouver as the natural inclination is to leave the A/T on. So with Airbus, switching the FD off removes the link between FD and A/T, so diving away and closing the thrust would work normally. Does that make sense? (I refer only to my explanation and not to why the Airbus works this way :ok:)

[Magnet77]

Quote from Centaurus #6
“I don't know about the Airbus types but why the urgency to switch off the FD? It takes a finite time to click off the autopilot and get cracking on the avoidance and at least 2-3 seconds for the PF to lift a hand to switch off both FD.”

Great question and observation Centaurus. From your comments, you are not familiar with the Airbus models so I will give a few general tidbits for you to consider, having flown both manufacturers and currently flying the A330.

On Boeing aircraft, generally, when you disconnect the Autopilot (AP), you must also disconnect the Auto Thrust (A/T) and should likewise turn the Flight Directors (FDs) OFF. A lot of pilots simply look through the FDs on the Boeing and ‘fly the airplane.’ So, when manually flying the Boeing models (disconnecting the AP), the thrust should also be disconnected (manual) as well. However, on Airbus models and by Airbus procedure, the A/T is left engaged even with the AP disconnected. The A/T does a great job of looking after the thrust and speed for the pilot.

As pointed out above by DeeTee, on Airbus, the thrust settings will follow the FDs if left engaged during a TCAS maneuver and create all manner of havoc, so the FDs are turned off procedurally along with the AP OFF.

The way I wrote would lead you to the conclusion that you drew, but actually the whole TCAS RA procedure happens simultaneously and takes only a few seconds at best. Your point about the time it takes to turn off the FDs is not an issue because the PF is doing it all simultaneously with disconnecting the AP and beginning the pitch movement as the PNF is turning off both FDs, (unless the PF instinctively reaches up and turns his own FD off).

At any rate, this Airbus procedure did not figure into the incident at Inchon, so the pilot apparently followed the Airbus procedure, albeit too abruptly for those in the cabin without seatbelts attached.

I hope this answers your question. Regards, Magnet77

[Shark Slayer]

I've always understood that you have 5 seconds to respond to an RA and 2.5 seconds to the reversal of that manoeuvre.

Remember, commercial air transport aircraft only certified to +2.5/-1.0 G Clean.

[Capn Bloggs]

I was told in a SIM once that in response to a TCAS RA, the ROC/D relates to the Mach No as follows: At M0.7, a 1° change in pitch will result in a 700fpm vertical speed change (up or down). It works.

I am not suggesting that one looks particularly at the Mach to decide how much to pull or push when an RA occurs, but this does illustrate just how little a pitch change is required to satisfy an RA.

Secondly, the comments on the aircraft systems and in particular the Airbus FDs are interesting. In the 717, the FD bars automatically disappear when an RA occurs and pop back on when the RA goes away. I wonder why that hasn't been put into the AB (with concurrent suitable auto-throttle mode change)?

[Blues&twos]

As a controls engineer I was wondering the same as Capn Bloggs about automatic disconnection of AT and FDs under these circumstances. Are there any situations in which you'd want the AT and FDs left ON during an RA (given their apparent 'interference' with your avoidance actions)?

[His dudeness]

We get a special FD symbol which is red and as soon as you follow it, it turns green. So you keep the airplane in the green and everythings fine.

But that is just a Cessna (C680 Sovereign)

[SNS3Guppy]

It seems to be a belief among more than a few pilots that an RA requires an extremely aggressive action to avoid a collision. It does not.


Not necessarily. Fly what's commanded, and it may be an agressive maneuver. What's required really depends on the conflict.

I've always understood that you have 5 seconds to respond to an RA and 2.5 seconds to the reversal of that manoeuvre.


In a perfect world, that might be true. In the real world, it's not necessarily true.

[Capn Bloggs]

it may be an agressive maneuver.
Depends on your definition of aggressive. Positivey but smoothly easing the stick back or forward a few degrees at the most is all that is required (at least in my experience in real RAs and in the SIM). I would definitely not class following an RA as requiring aggressive handling.

We get a special FD symbol which is red and as soon as you follow it, it turns green. So you keep the airplane in the green and everythings fine.

But that is just a Cessna (C680 Sovereign)
Goes to show what progressive companies can do. I appreciate that airliners have more-strict certification requirements, but having to turn off your FD (Airbus) when following an RA? Is it really that hard to modify the software to be more user-friendly? Although I will say that Flight Intl wrote a few months back that Airbus is looking at the autopilot following RAs automatically. That would be good.

[ATC Watcher]

Airbus is looking at the autopilot following RAs automatically. That would be good.
Indeed, I am now also convinced that this is the solution and that it would be a good thing. This would prevent encounters like Ueberlingen to happen again ( and it probably will one day as the number of similar incidents continue to rise ) But there seem to be tremendous resistance from the pilot community.
When I suggested this during an IFALPA ATS Committee meeting some 3 years ago, there was not a single member around to support the idea, most were absolutely against in fact.

But when we look at all the serious TCAS incidents post Ueberlingen , the vast majority occurred in the upper airspace , and all without exception would have been solved should the RAs had been linked to the AP.

[Jet Fuel Addict]

My guess is the pilot community would feel they become obsolete.

On the other hand if someone was flying his plane by hand (lower down) this could cause confusion as what to do...?

Best thing for the moment now IMO is to just keep pilots trained and informed!

[Capt Pit Bull]

Indeed, I am now also convinced that this is the solution and that it would be a good thing.

Have to disagree with this.

What we need is the current recommended practices to be correctly trained on a world wide basis. Collectively, the industry can't be bothered to go to the (very modest) expense of doing so. As a result we have 'one size fits all' dogma about how TCAS should be used, rather than a proper understanding of its strengths and weaknesses, in a significant proportion of the industry. So we have dumb incidents happening, which people think can be solved by yet more automation.

The existing equipment and training recommendations are fit for purpose. We just need to use them.

pb

[Blues&twos]

Sorry Capt Pit Bull, I don't understand, why wouldn't automation solve the problem?

(I'm making the assumption that we're talking about two TCAS equipped a/c. I realise that if one isn't suitably equipped or isn't using the equipment, then it's effectively useless).

[punkalouver]

[QUOTE=SNS3Guppy;4336348]Not necessarily. Fly what's commanded, and it may be an agressive maneuver. What's required really depends on the conflict.


Necessarily..... for an initial RA, if you fly what is commanded. Some seem to believe that a strong high G pullup or low G pushover is required. All they are doing his risking injury to pax.

From a Honeywell manual about their TCAS operation...

"A typical TCAS II maneuver requires crew response to the initial RA within 5 seconds and G-forces of ± .25 G.This force is similar to that experienced when initiating an enroute climb or descent."


On page 24 of this link.

http://www.sisadminov.net/tcas/docs/cas81_pilotguide.pdf

[LowVizRTO]

... Airbus is looking at the autopilot following RAs automatically. That would be good.Most of us pilots have somewhat of an aversion to being relegated to 'the most forward window seated passenger', especially during a life threatening event such as a TCAS RA.

I guess you could always disconnect the autopilot if it was leading you up the garden path, but I personally would prefer to retain the control and take decisive action. Shades of the old "Spam in a can" aversion of the astronauts in "The Right Stuff" of yore, refusing to fly in a capsule without a window and no pilot controls. The engineers lost that battle and good thing they did. The moon landing in July 1969 would have ended in disaster if Neil Armstrong had not taken over from the computers and landed the craft safely.

Computers are not afraid of dieing - pilots are. That gives the pilot's survival instinct the winning hand in almost all situations. A pilot's desire to live has gotten aviation to where it is today, albeit with a few pilot errors along the way. For the most part, pilots are the Captains of the most safe form of transportation known to man. There are no other forms of transportation even close to the safety record of commercial aviation.

Just increase the mandated pilot training for TCAS events during simulators and leave the autopilot out of the equation. That's my two bob worth. Cheers

[ATC Watcher]

Low Vis RTO , in the absolute, I fully agree with you: humans are , and by far, superiors to computers as far as decision making is concerned. . However for monitoring processes, computers are, also by far, superiors to humans. I hope you can agree on this.

TCAS is basically for 80% a monitoring system ( checking the altitude reports of the Mode C and S transponders around you, coupled with a basic azimuth detection ). The remaining 20% is an algorithm issuing TAs and RAs when the system detects a collision risk.
This 20% is the problem as it mixes computer generated logic ( an RA) needed to be followed immediately with a human analytic mind (pilot) decision/action .This does not work well .

After many years (more than 25 to be precise)of testing and evaluation , the current version we have (7.0) dictates that to be most effective, initial RAs have to be followed within 5 sec and corrective RAs within 2,5 sec. This leaves very little time to analyze the situation and decide to follow it or not.

The many TCAS incidents I saw up to now, indicate that those pilots, when confronted with a sudden computer “ advice “ ( as an RA is in the current definition) trend to evaluate the options, try to make visual acquisition, take into account the last ATC clearance received, etc.. and , in those incidents, decide to follow their own logic , even if it means disregarding the RA or, worse, if it is opposed to the TCAS RA issued.
The pattern is recurrent, even after Ueberlingen. Until a couple of years ago, I also believed that the problem was lack of ACAS understanding and foremost lack of Training , and I pushed, like everyone else, for more training and for unified,, mandatory simplified guidelines and procedures.
Now, years later , the percentage of TCAS incidents involving pilots not following or acting against RAs has not changed much, it is still extremely high. Therefore I also, like many of my colleagues, now start to believe that coupling the ACAS to the AP is probably the best solution.Especially in the upper airspace ( let's say above 20.000 ft)

At this altitude , all the AP will probably do in most cases, is make an altitude change on 4 to 500 ft max with 0,25 G , but most importantly in the correct sense . That alone will prevent a metal to metal encounter, which is afer all what we expect this system to do.

I fear that we sadly will have to wait until another Ueberlingen type collision before the minds are ready to accept the AP coupling solution. But that is often the case in many other industries as well .
Ah, before you ask, I am not, even remotely, involved with Airbus . I just think they are correct is their thinking on that issue.

[LowVizRTO]

The many TCAS incidents I saw up to now, indicate that those pilots, when confronted with a sudden computer “ advice “ ( as an RA is in the current definition) trend to evaluate the options, try to make visual acquisition, take into account the last ATC clearance received, etc.. and , in those incidents, decide to follow their own logic , even if it means disregarding the RA or, worse, if it is opposed to the TCAS RA issued.

ATC Watcher - All good and valid comments, and I do agree with your analysis of monitoring vs. analysis. In all my experience (41 years and nearly 26,000 hours) I cannot imagine a pilot having to think about what to do when faced with a real world TCAS RA. All you have time for safely is to disconnect the AP and fly the computed commands by the TCAS. Why anyone with adequate training, understanding and regular simulator TCAS experience would take any other action is beyond me.

The 'swimming in cold syrup' when an emergency occurs is a symptom of the lack of recency training. Well trained and recent pilots have no problems with what to do, and do it correctly time after time. A few do not. If it weren't for the passengers and cabin crew that they take with them, it's not a bad thing to cleanse the gene pool of such deep but self-effacing thinkers. Because of the few, must we assume all pilots are incapable of functioning properly and hand the reins over to computers and automation?

On account of the lowest common denominator among us pilots (eg the poorly trained ones), you may well be right about the AP being part of the solution in the future. But, having said that, the same sort of individuals, with certainty, will disconnect the autopilot during a TCAS RA while saying something akin to, "what is it doing now?" Try as one may, "You will never make the cockpit idiot proof".

With computerization and automation carried to their logical conclusions, we will have fully automated computer controlled takeoff, climb, cruise, TCAS avoidance, descent and landing, all without the need for a pilot. And ATC could uplink active CPDLC commands and fly the bugger from their comfortable swivel chairs on the ground like the various air forces do with their drones.

We have a name for such aircraft. They're called ICBM s (Inter-Continental Ballistic Missiles). I wouldn't allow my family to passenger on one, nor are you likely to allow your family on such an aircraft. There has to be a human pilot at the controls for safety. Safety will never be 100% as much as we would like it to be. That's not the world we live in.

Having run completely off the road and into the field of extremes, I rest my case for now. All this for a simple TCAS RA and the mere mention of an autopilot? LOL. Cheers. ;>)

[modeselectorignition]

Hi All
Airbus procedure to switch BOTH FDs OFF is to revert the AutoThrust to SPEED mode. SPEED mode? Well, Airbus pilots should know. Otherwise better get your book out and study them. No FDs? No worries because the VSI is the main instrument/cue. Just fly to the green zone. Simple. Once clear of conflict, reverse the RA procedures. 1. FDs ON. 2. A/P ON. Please do not compare Boeing, Cessna or Bombadier procedures with Airbus. There are of different make and philosophy. Cheers!

[Blues&twos]

LVRTO,

I understand where you're coming from, but when you write software or design a control system which has a safety related function you have to design it assuming that the "human link" is going to fail, for whatever reason....

I'm a low hour student pilot...so no expert. But if the response to an RA is always "disconnect the autopilot, switch off the FDs and fly as advised" then surely an automatic AP/FD disconnect simply means that's something additional which the pilot doesn't have to waste their time doing? I know people don't like it when there's too much control given over to the computer (even I'm wary of this, and I'm a controls engineer), but simply doing something which a pilot always has to do anyway is just saving time and thought processes, allowing them to concentrate on pushing/pulling the yoke and maintaining control of the a/c.

Please let me know if I've missed something or have misunderstood!

[LowVizRTO]

Please let me know if I've missed something or have misunderstood!
Blues&twos - no you haven't missed my point at all, and I'm not missing your point either. Having thought about the whole process and the allergic reaction most of us long term pilots have to automation taking control of the aircraft, I am more open to some changes. For instance, there is no reason why the FDs don't get automatically turned off in the event of a TCAS RA. That could easily be accomplished with software and maybe some pin changes. That would be an improvement in my opinion. We can agree on that point for sure.

However, the jury is still well and truly out for me (and most other pilots) as far as the AP automatically flying an RA maneuver on its own. The aversion is not to automation making the job easier, but to being relegated to SLF (self loading freight) status by computers flying in loose formation.

If my understanding is correct about history, the early Airbuses killed some of their own test pilots because of the priority given to the computers over the pilot's control input. Airbus learned that lesson the hard way because the engineers had too much control over the final design. Pilots (humans) are indispensable to the safe operation of commercial aircraft. At the risk of repeating myself from an earlier post, computers are not afraid to die - pilots are; and so pilots will do whatever it takes to stay alive. Unfortunately, they do get tired, fatigued and tend toward laziness and occasionally overlook things. This is where the rub occurs. Just trying to be realistic here.

Having said that, there have been some tremendous advances in automation over the years to make the job of the pilot much easier and do the heavy lifting of monitoring and taking initial action steps in the event of an emergency. The latter improvements did away with the Flight Engineer requirement altogether. Where does computerization and automation stop though? Can we fly with only one pilot in a commercial aircraft in the near future, or without any pilots whatsoever?

Maybe it stops only after thousands of people are killed by the computers flying aircraft into each other, following the logic of engineers that can't possibly think of everything that could possibly ever happen. Interestingly enough, the accident, incident and death rates are the very same reasons for taking the pilot out of the loop now. Hopefully some common sense, which seems to be not so common anymore, will prevail before the above scenario comes to fruition. Please do not misunderstand me here, I am not taking a shot at you or anyone else - I'm being far more global than that.

I've even considered the possibility that I'm suffering from Alvin. Toffler's 'Future Shock', or just passed my 'use by date', but I'm certainly not senile or stupid. There are natural limits to the human being's capacity, as there there are natural limits to computerization. We need to go slowly down that slippery slope is all I'm saying.

This discussion brings to mind the old joke about this very issue on future aircraft designs. It goes like this: 'The totally computerized aircraft of the future will have one pilot and one large dog. Everything will be fully automated and the pilot only needs to monitor the progress - the computers do everything from takeoff, cruise, descend and land. "What you may ask, is the role of the dog?" The dog is there to bite the pilot if he touches anything.'

I thought it was funny when I first heard the joke, but it seems to be materializing as we speak. Now that's not so funny. Cheers LowVizRTO

[Blues&twos]

Thanks for that LVRTO. I entirely agree with you. I wouldn't be in a hurry to get on a fully automated aircraft....I still get nervous on the Docklands Light Railway.

[LowVizRTO]

The below quote is word for word copy-pasted out of Airbus A330 FCOM 1.

– If no AP/FD pitch mode is engaged, the A/THR mode reverts to SPD/MACH mode. In other words, the selection of a pitch mode determines the associated A/THR mode.

Because there seems to be some confusion or incredulity that the FD left on makes any difference in an Airbus TCAS RA maneuver, hopefully this will put it to rest. Unless you have been trained on Airbus, you wouldn't have any frame of reference to such interrelationships on the Bus. It is a different kettle of fish to other aircraft manufacturers, and it works just fine, albeit differently.

Effectively, what happens when you descend off a cruise altitude with A/P off and FD(s) on, the thrust will increase resulting in higher speeds than desirable. Conversely, climbing with the A/P off and FD(s) on, the thrust will decrease reducing speeds to lower than that desirable.

With both A/Ps and both FDs off, the AutoThrust reverts to Speed mode and maintains the correct thrust for the current Speed/Mach whether in a descent or climb throughout the TCAS maneuvers.

There is no question that the FDs have to be turned off in a TCAS RA maneuver, and not just because Airbus says so - there's very good reason for doing it. Whether that occurs automatically (which is currently not programmed in Airbus) or occurs manually (which is the current Airbus procedure for dealing with a TCAS RAs), it doesn't matter, the FDs have to be turned off to avoid a potentially dangerous secondary situation from developing during a TCAS RA maneuver.

The quote below from the A330 FCOM 1 demonstrates that automatic disengagement and removal of the FDs currently occur in various situations, and so would more than likely not be a major issue to include the FD removal during TCAS RAs.

AUTOMATIC DISENGAGEMENT DUE TO SPEED PROTECTION
When APs are not engaged and if you do not fly the FD bars, an automatic disengagement of FDs and corresponding FMA modes will occur if the aircraft speed reaches VMAX in climb with CLB or OP CLB mode engaged or if the aircraft speed decreases to VLS in descent with DES, OP DES mode engaged.
Refer to Automatic speed protection in this chapter.

AUTOMATIC FD REMOVAL
– The FD pitch bar is removed when no vertical mode is engaged or when ROLL OUT is engaged.
– The roll FD bar is removed when no lateral mode is engaged or when the RWY or ROLL OUT mode is engaged.
– Both FDs are removed when the aircraft pitch exceeds 25° up or 13° down, or bank angle exceeds 45°.

Although this is bit of a rabbit trail, I hope this is of some interest to those who have no Airbus technical exposure. Fellow 'Bus drivers' already know this in spades.

Cheers, LowVizRTO

[Capn Bloggs]

Don't want the AP to fly an RA ie gentle wings-level push down or pull up over a few hundred feet but are quite happy to let it take you to within metres of the ground in complete IMC. Something not logical there. Let's face it, following an RA would hardly be rocket science for a FD/AP.

An RA is an emergency manoeuvre that could spring up at any time with very little notice, perhaps preceded by "Traffic Traffic". I for one would be quite happy to let the AP do it. Had the Russian jet had this feature (I know that particular model couldn't have due to old technology, but had it been a modern jet...), all would have lived. This is one scenario where a crew's instant reaction was completely wrong and automation would have saved the day.

Man will still have his place in the cockpit way past the time when you and I retire. But I'm for APs to follow RAs.

[LowVizRTO]

Man will still have his place in the cockpit way past the time when you and I retire. But I'm for APs to follow RAs.

Capn Bloggs (http://www.pprune.org/members/48133-capn-bloggs) - W'ell have to respectfully disagree on that point. Over the last ten years, i have had a half dozen FALSE TCAS RAs. Don't know why they happened but in several cases a TA rapidly progressed into RAs and we took the required maneuver but there was no one there, not visually, not on TCAS after the event and not on radar with ATC.

Capn Bloggs, what do you do with such failures of the TCAS systems when you hook an AP up, to vault one way of the other at will, and possibly into oncoming traffic in RVSM airspace just 1000' above or below you?

Auto FD's off - YES. Auto AP maneuvers - NO thank you.

Cheers, LowVizRTO

[rodthesod]

There's a simple formula to calculate the pitch change required to achieve the climb or descent required for a RA: 1000/TAS (in knots) = Pitch Change. I used to train all my pilots to be aware of their approx TAS at all times in readiness. The maximum pitch change is obviously at low speeds and in an intermediate approach phase usually around 5-7 degrees. 5 seconds to adjust 7 degrees is not aerobatic and hardly enough to spill coffee. At high level cruising speeds the pitch change required is usually in the order of 1-2 degrees. When properly executed by both conflicting aircraft (initially in level flight) an RA should not cause either aircraft to diverge more than about 300 feet (i.e. they miss by 600 feet vertically).

The above info was true for TCAS II - maybe things have moved on since I retired 4 years ago - if so I'm happy to stand corrected.

This incident was a GROSS over-reaction by the crew - don't they practice RAs in the simulator? In my last (cargo) airline all pilots had thorough initial TCAS training and testing and regular refreshers during LOFT sim exercises. I always thought both TCAS and windshear training served a useful secondary purpose in refreshing crews in the effects of controls (not coupled to an AP for a change).

[Mr Pax]

Having been on an aircraft that had a TACAS RA ( http://www.pprune.org/rumours-news/247503-ba-2166-tampa.html ) can I just say that as SLF I would rather suffer a few bumps and bruises than collide with another aircraft, I am sure that any Captain on receving that instruction for real would be very keen to take the nessesary action and quickly.
I am still very gratefull to the Captain of that aircraft, heavy handed or not for he may well have saved my life!:D

MP

[Capn Bloggs]

LowViz,
Over the last ten years, i have had a half dozen FALSE TCAS RAs. Don't know why they happened but in several cases a TA rapidly progressed into RAs and we took the required maneuver but there was no one there, not visually, not on TCAS after the event and not on radar with ATC.

Fair enough, but I can't see what erroneous RAs have to do with the method of following them. I hope you are not suggesting that pilots assess whether the RA is valid and then follow it if they believe it is?

what do you do with such failures of the TCAS systems when you hook an AP up, to vault one way of the other at will, and possibly into oncoming traffic in RVSM airspace just 1000' above or below you?

Again, I hope you are not suggesting pilots make up their own minds about the validity of an RA before following it. A typical RA would not use more than 500ft of altitude (unless the aircraft is already climbing and descending) and therefore the "third party" aircraft would still be adequately separated. I would expect that if it's TCAS thought there was an issue, it would issue an avoiding RA too. It is easy to create a scenario where TCAS would be overwhelmed, but in reality, does it happen often, and would we have been better off without it? I suspect the answer is "no".

It would also be interesting to know how many false RAs are now occcurring with Change 7 software.

[LowVizRTO]

Fair enough, but I can't see what erroneous RAs have to do with the method of following them. I hope you are not suggesting that pilots assess whether the RA is valid and then follow it if they believe it is?

Capn Bloggs - No, nothing of the sort. I always carried out the RA commands, false or otherwise, as there is no way to know if you are just not seeing the other aircraft or it is simply a false echo. I am merely pointing out that TCAS is man made and therefore fallible.

Additionally, I have no idea on the TCAS 7 failure or false RA rates. I am not against automation per sea, just against hooking the A/P up to TCAS RAs. You're entitled to your own opinion which is opposite to mine. That's fine. I remain with others on this thread that the solution is more TCAS training in recurrent Simulator sessions and not A/Ps. From your comments, I don't expect you to agree, and that's fine as well. Best regards. LowVizRTO

[ATC Watcher]

It would also be interesting to know how many false RAs are now occcurring with Change 7 software.

Version 7 is mandatory in Europe since 2001 (RVSM requirement) and false RAs do occur in Europe . So the answer is yes. In fact the false RAs are not related to version 6.04 or 7.0 ( there might be a 7.1 or even version 8 to cure some current problems , notably in sense reversals, but that is another story )

The "false RAs" are due to many factors, one of them was identified to a perticular brand of transponder ( fixed by now I presume ), another with the antenna picking up responses by its own transponder : hence the sudden appearance of a red square with 00 and an RA Descend descend without prewarings (TAs) . These also should be eliminated by now. But we heard of some other cases, difficult to replicate .

Back to the AP coupling issue, as it was said earlier, if you have or suspect a false RA, you'd better follow it first and ask questions later. So the AP coupling idea is still valid , wether it is following a false RA or not.
The only problem I have with the AP coupling is let's say below 10.000 ft.
But, to be fair, in the upper airspace there is no valid reason , other than resistance due fears of automation, to have the AP move you 3 or 400 feet in any direction.

As someone else said very correctly, you have no problem on automation following an ILS signal down to the ground in near zero visibility, (and we know how easy it is to garble or bent an ILS signal !) so why not automatically following an RA ?

[LowVizRTO]

Good info TwoOneFour
Collision risk demands rapid roll-out of new TCAS: analysis (http://www.flightglobal.com/articles/2008/08/22/315083/collision-risk-demands-rapid-roll-out-of-new-tcas-analysis.html)

As someone else said very correctly, you have no problem on automation following an ILS signal down to the ground in near zero visibility, (and we know how easy it is to garble or bent an ILS signal !) so why not automatically following an RA ?

ATC Watcher & Capn Bloggs - I currently fly CAT III A & CAT III B approaches both in the sim and 'in anger' but we're using two (2) APs coupled together in FAIL OPERATIONAL mode. B744s use three (3) APs for the same approach. CAT III B No DH has a 75 Meter visibility requirement and No cloud base DH, so I understand what you are saying about APs close to the ground, but still disagree.

Two or three APs engaged produces a huge difference in the margin of safety as compared to operating with only one AP. There's very good reasons why low viz approaches employ multiple (all the available) APs. So still not convinced guys. Cheers.

[Capn Bloggs]

Interesting that both of the TCAS events that are to be fixed by revision 7.1 would not be occurring if the AP was allowed to respond to the RA.

The sooner the AP is allowed to manage RAs the better, IMO. It's now not if but when.

[Magnet77]

While the FD could be automatically switched off during an RA and be an improvement for response time (possibly), I do not agree that the AP automatically following RAs is the answer. More TCAS RA Sim training is the answer.

[ATC Watcher]

This is getting interesting ;

LowvisRTO, thanks for the info on the 3 AP on CATIII, I was not aware of that. Indeed a big difference. However they will all 3 still be following an ILS signal, which is my opinion still the weakest link of the whole procedure.

Capn Bloggs : I agree with your last comment (not a question of “if” but “when” ) I think that with the next collision, the AP will most probably be mandated.

Magnet 77 : more training is the solution : we tried that, but it does not work, we are now more than 10 years after implementation and the percentage of “TCAS pilot errors” is still abnormally high. And , with TCAS you need everyone in the loop ( i.e. including the guy in front of you that you do not know ) trained with the same standards all around the world. Very , very difficult to accomplish.

Finally on the version 7.1 ( if they decide to call it that way ) my last info was that the FAA was against making it mandatory. ( as they would have to pay for it and they do not have a budget for this, plus some lawyers said that mandating a new version would be an admission that the previous versions were not safe , etc.. ) The European might do it , but not the USA.(in fact a repeat of the version 7.0 debate )

Whether this retrofit will be free (paid by Eurocontrol ?) or for a fee ( paid by the aircraft operator) is still under debate. But , the next collision will probably settle that issue as well. Sad, very sad.

[Jofm5]

I am purely SLF.... with a keen interest in aviation however as a software developer I have some questions to raise with the current discussion point of the a/p taking over.

Do we know if it has ever been simulated with the current a/p software that it would have coped adequately in a manner whereby firstly the situation was saved, secondly the pilot was able to quickly adapt to the change and whether it was an adequate response.

To explain the above, the first is obvious and the most critical as in no collision occured, the second is if the plane makes manouvers outside pilot control then when the pilot is expected to resume control there must be some sort of handover etc so the pilot knows exactly what occured. And my final was whether it was adequate i.e. was the response too extreme for what was required.

My basic understanding is that whilst on a/p there are a number of systems counter checking each other to come up with what is a considered acceptable manouvre, put in an emergency situation like this and you have those same systems crosschecking each other to see what is acceptable, add another plane into the equation and even more serious more than one then you have numerous systems cross checking their own paramters and not each others. I dont know enough to say that to leave it manually would be the best solution but at least if a visual ident can be made then the pilot would know at least where not to go.

Sorry but my understanding of TCAS is limited, I would hope they negotiate the best paths for avoidence between corrseponding planes on collission. If this is the case then it would make the a/p softwares (you never rely on a single set of code) easier to come to a decision.

I guess what I am saying is what seems like a perfectly reasonable request to avoid collision when handing control to the a/p can be over complicated by the different software each negotiating with each other which is the best solution. Combine that on two different platforms (planes) and should the software conflict without communicating with the opposing platform (plane) then in all possibility a collision may occur.

I have never looked at the software, so I would hope the above has already been addressed etc - just wondering if you pro's know more.