TCAS RA in VMC. Follow the RA or Not?

[Capt Pit Bull]

Well, I am certainly going to get shouted down on this one, but....

Whilst the general move towards people seeing compliance with RAs as the sensible default course of action is a good thing, I'm a little concerned that the pendulum is swinging to far in the other direction.

There are a number of misconceptions in the replies below. Folks, TCAS is not a perfect system. That is why the final decision about how to fly the aircraft should remain with the crew.

- It is EASY to think of circumstances where following an RA would be a bad idea. If you can't think of an example, you don't understand the limitations of the system well enough!

(off the top of my head I can think of 3 generic kinds of scenarios where full compliance with an RA would be more dangerous than not manouevering. 2 of the 3 are not particularly likely, but then again mid airs are well down the end of the bell curve anyway.)

But we've had these discussions before, and its late. The search facilty for "TCAS" should reveal plenty of information.

Previous to the DHL/Sibair midair collision, ICAO defined the response to a RA as 'should' be followed. Grammatically the word 'should' is NOT a command and therefore many states and individual operators technically allowed crews to ignore an RA if they percieved a 'better' solution. However the DHL/Sibair accident showed the vulnerability of this procedure.

No, I disagree. The original guidance notes from the JAA did indeed say an RA should be followed. However, they also included very specific guidance that manoeuvering in the opposite sense to an RA is highly dangerous. The accident was about the later, rather than the former, in that sibair could have done nothing (i.e. not complied with the RA) and the result would still have been safe. It was the maneouvre in the opposite sense (well, 2 in fact) that prevented TCAS from averting the disaster. The necessary guidance - i.e. don't manoeuvre opposite - was plainly stated by the JAA (as well as every manufacturers handbook I have ever seen).

We don't need more regulations. TCAS' introduction to Europe was preceeded by several years of operation in the USA which gave us the opportunity to learn many of the operational issues relating to it. Hundreds of incident reports were available from the FAAs TCAS transition program.

The lessons were all there.

Specifically, several incidents involving manoeuvres opposite to RAs occured, including some which would have been similar disasters if not for the fortuitous presence of a little bit of horizontal distance.

IMHO given that the exisiting (completely adequate) guidance material was not being implemented universally, whats the point of making it more inflexible?

Finally, with reference to the original question, I would think that with that vintage, it would be entirely possible that the non compliance would have been within the bounds of the companies SOPs.

CPB

[XXTSGR]

Assume TCAS 2 or better:-

If you don't follow TCAS, you are second-guessing what the other guy is going to do.

If you don't follow TCAS, you are assuming that you can see and avoid all traffic in your area.

If you don't follow TCAS, you have no legal excuse to offer ATC when your non-TCAS avoidance causes other problems.

If you don't follow TCAS, you are probably going to die.

All the above, of course, assumes that you follow the recommended method of use. If you don't, then you are probably going to die soon anyway, so why bother?

I haven't yet heard of any sensible reasons for not following TCAS. And yes, I do understand how it works and its limitations. Capt Pit Bull may be able to think of 2/3 unlikely scenarios that may result in not following TCAS being proved a bad idea. I can think of at least 2 ACTUAL scenarios where NOT following TCAS killed people.

Beat that.

IMHO given that the exisiting (completely adequate) guidance material was not being implemented universally, whats the point of making it more inflexible?

Errrr - perhaps to ensure that it IS implemented universally????

The necessary guidance - i.e. don't manoeuvre opposite - was plainly stated by the JAA (as well as every manufacturers handbook I have ever seen).

OK - so neither aircraft manoeuvres oppostie to TCAS advice and what happens? errrrrrr - they crash. No sh1t, Sherlock. So why do we have it? You tell me.

Sorry CPB, but as you say you were expecting to get shot down, and you proved exactly why you expected that - because you are talking a load of dingo's kidneys, and dangerous stuff at that. If you are going to pick and choose which regulations you obey, then you are far too dangerous to operate an aircraft. "Oh, sorry, I don't agree with those rules - I'm not going to obey those. These I might. Oh - and not those ones".

Not acceptable. Not professional. Not legal. Not safe. Not intelligent.

[POL.777]

YES!
I need to write a bit more in order to get this message posted.

[Capt Pit Bull]

XXTSGR

Not acceptable. Not professional. Not legal. Not safe. Not intelligent.

I'm somewhat disappointed to be characterised as such. I had hoped folks might use the search facility as I suggested as this is just regurgitation of issues that have come up before in numerous threads.

Just for the record, I've been singing the 'follow the RA' hymn for years, ever since being the project pilot for TCAS implementation at our company. At the time, SOP authority to disregard an RA in the event of visual spotting was available, however our training course emphasised in the strongest possible way the dangers in so (not) doing.

If you think an RA is unnecessary, so what. Follow it anyway.

However, if the RA is, in itself, dangerous, then I believe that as professional pilots we have the right to hold final authority.

I haven't yet heard of any sensible reasons for not following TCAS.

Like I said, TCAS is not a perfect system. Here are a few reasons why.

Someone earlier said "Only TCAS sees the whole picture". Well, thats not true. TCAS doesn't see non transponding aircraft, and effectively can't do anything about non altitude reporting aircraft either.

It also doesn't properly see terrain either. OK, RAs have various radalt based inhibitions, but that won't help you if you are flying towards a ridge / cliff / escarpment or over tall buildings / aerials.

Someone else said that GPWS will protect you from ground contact. Thats true, but now check your MEL. Can you dispatch with EGPWS inop? My current aircraft type can. Do you have any AIS related issues or QRH procedures that might require you to disable your EGPWS or its terrain function? My current aircraft type does.

I therefore put it to you that TCAS is quite capable of flying you into terrain or obstructions.

So the point is, that in a potential mid air collision scenario, NO ONE can see the big picture. Not TCAS, not ATC and not the mark 1 eyeball. Therefore any action taken by the crew has to be a composite of the available information, and it is something that ultimately requires judgement, and training.

Now, as it happens, the output of that process should, in 99.9+% of scenarios, be that following the TCAS is the best plan of action.

OK - so neither aircraft manoeuvres oppostie to TCAS advice and what happens?

Now, whilst I can think of scenarios where one aircraft might feel a manouevre in one sense is dangerous, I can not think of any where both aircraft would feel that.

e.g. if at low level over sharply undulating terrain one aircraft might balk at a descent RA, there would be no reason for the other aircraft not to climb.

So, 2 out of 3 of the general kinds of scenarios involve a collision risk with something TCAS doesn't know about - i.e. aircraft and terrrain.

That leaves my 3rd group. Namely that TCAS will command an unachievable rate of climb. Whilst it is correct to say that TCAS has performance based RA inhibitions, it should be pointed out that these are approximate only. With the TCAS equipment we had fitted to the ATRs I was flying at the time, the manufacturers handbook specifically stated that the pilot was responsible for protecting the flight envelope of the aircraft, if necessary to the detriment of following the RA. Simple really, if you are told to climb, but you stall, you descend, (a manouevre in the opposite sense) and therefore more dangerous than doing nothing.

Having looked at the performance inhibitions, they were quite optimistic. Certainly at heavy weights, especially in icing conditions plus actual ice on the airframe, you could definitely recieve an unachievable RA.

But if you are being directed up, then the other guy is being directed down and should have no problem achieving it.

So, in summary, there are reasons why TCAS RAs can be dangerous in themselves. A disregarded RA is NOT the end of the world - the other aircraft will be perfectly capable of solving the encounter and is not going to feel constrained in the same sense.

Therefore the idea that failing to follow and RA means you are "probably going to die" is overly alarmist. Its not supported by data either - there are many documented cases of RAs not being followed where the scenario was resolved by the other aircraft or horizontal separation.

Neither of which means that RAs should be disregarded casually. Defintely not so. Follow the RA unless there is an immediate, obvious, in your face danger.

And in fact, thats pretty much what it says in the links provided by Mike Jenvey below. Here is a quote:

One of the main messages for pilots to note from the ACAS training material is an emphasis on the need for pilots to ‘follow the RA’. Indeed, ICAO Doc.8168, ‘Procedures for Air Navigation Services – Aircraft Operations (PANS-OPS)’, Volume 1, Part VIII, Chapter 3, Paragraph 3.2 c) 1) of Amendment 12, states that: "in the event of an RA [Resolution Advisory], pilots shall respond immediately by following the RA as indicated, unless doing so would jeopardise the safety of the aeroplane".

I therefore put it to you that my contention is precisely in line with ICAO, and is not in fact a 'load of dingos kidneys'.

If you operate good performing twin jets into flat airports with poorly equiped GA aircraft kept outside of the airspace, your chances of meeting a 'dangerous' RA are slim. If on the otherhand you operate a more poorly performing turboprop into regional airports swarming with 'puddle jumpers' and blessed with cliffs / ridges etc the chances start to grow a bit.

Finally, what I can not emphasise enough is that it is manoeuvres opposite to RAs that are the killers.

Rather than saying "Always Follow the RA" it is better to say "If you possibly can, follow the RA. Failing that, get as close as you can. Whatever you do, never manoeuvre opposite."


Mike Jenvey

A couple of points from this incident. Would you not agree that if the second aircraft had done nothing this would have been resolved relatively safely. It is the manoeuvre in the opposite sense that made it so dangerous.

Secondly I find it profoundly depressing that there are still crews making up their own avoidance manouvres based on the Traffic Display. This sort of behaviour should have been trained out of the pilot workforce during the initial implementation of TCAS.

What this tells me is that as an industry we have got the training wrong. We've invested in the kit, but we have paid lipservice to the need to train our crews to understand the system properly. They don't have the right ammount of faith in it. Either no faith at all (as per the incident you report) or too much faith, such that companies view it as sufficient to have proscriptive SOPs.

Lack of training killed the folks at Ueberlingen. Trying to dumb it down is not the answer.

CPB

[Captain Stable]

While Capt. PB's comment

Rather than saying "Always Follow the RA" it is better to say "If you possibly can, follow the RA. Failing that, get as close as you can. Whatever you do, never manoeuvre opposite."

is admirable common sense, my personal feeling is there is not enough time to evaluate from the time the RA is activated. That would require looking at the situation, making an assessment, taking a decision, and acting on it. Too slow.

Like GPWS, the reaction has to be very fast indeed. And that leaves little or no time to think how useful you think the RA is.

Better by far simply to obey. (IMHO)

[alf5071h]

I agree with the consensus that the RA must be followed.

Given the extent to which the industry has researched the operation of such safety systems I continue to be surprised by the number of pilots who wish to out think them – EGPWS, Windshear, ACAS. These warning systems represent the only real need for memory drills - immediate action (excepting RTO or GA); their relative priorities, resultant flight paths, and range of scenarios have been well thought out. A RA will not fly the aircraft into the ground, it provides the safest option for the situation, including other aircraft and their flight paths; similarly windshear recovery systems will not stall the aircraft.

During development and initial in-service operation, many systems have glitches, ACAS more than most, but overall the industry has benefited from their introduction, flying is safer and will remain so provided crews do not invent or visualise highly unrealistic scenarios which they may never encounter. No system is perfect, but they are considerably more reliable than the crews who use them.

With GPWS there were many accidents which could have been avoided if crews had followed the warning; there were many background circumstances to these behaviours, but with ACAS the industry cannot allow these human issues to be repeated. With the advent of EGPWS and many reported ‘saves’, the industry not only has an improved safety record, but also an indicator that the underlying human error causing close encounters with terrain still remain. Thus EGPWS has transitioned from a terrain warning system to human error recovery system; ACAS should be viewed similarly.
--------------------
Unless specifically authorized everything else is forbidden.

[Capt Pit Bull]

Captain Stable

no time to think how useful you think the RA is.

<nods> I see, in the original TCAS guidance / authorisation to disregard an RA if the intruder is visually spotted, a concept of "you don't have to do it if you don't think its necessary". I don't advocate that at all.

I also agree that we should not be suggesting that people spend time thinking about. I'm just saying that it ought to be allowable to take account of clear obvious dangers.


alf5071h

A RA will not fly the aircraft into the ground.

Yes it can. You just need sharp enough terrain, or a high obstacle like a tall mast on a ridge, coupled with a non functional terrain mode on EGPWS. Plenty of places I operate to have topography that TCAS radalt based inhibitions will not protect against, so its down to EGPWS.

Mandate that EGPWS must be serviceable (rather than MELable), and that serviceability is contingent on having an up to date database. E.G. right now I fly to a destination where, inspite of the fact that a new runway has been in use for several months (and obviously planned for for years previously), we are turning our EGPWS terrain mode off because the runway isn't in the database. (so that aspect at least is NOT a highly unrealistic scenario)

If the regulators mandate these points, then I'd be a lot happier.

it provides the safest option for the situation, including other aircraft and their flight paths

Apart from non transponding / non altitude reporting traffic. There has to be the possibility of TCAS directing you into the flight path of such an aircraft. If you operate to regional airports this is a real threat.

Mandate altitude encoding transponders, and mandate they be used, and I'd be a lot happier.

With GPWS there were many accidents which could have been avoided if crews had followed the warning;

Agreed. But there is a big difference between GPWS and ACAS. The ground won't get out of your way under any circumstances. If it comes right down to it, a possible air hit is a worthwhile price to pay to avoid a certain ground hit - hence why EGPWS and Windshear have priority.

With RA non compliance by one aircraft only (because it has a threat in its RA sense, the other aircraft won't) the chances of a successful resolution should still tend towards 100%.


CPB

[alf5071h]

Cap’n P B, I support your calls for altitude encoding transponders and revision of the EGPWS MMEL; however, I wonder if you are overstating the weakness of ACAS.

My previous simplified position on the safety of ACAS should have balanced the probabilities of encountering such rare situations that you describe. For the scenario that you envisage, consider the following geometry; forgive the generalizations and any inaccuracies in the equipment specs, I no longer have access to these details.
Assuming that RA commands are limited to excursions of plus/minus 1000ft (no adverse manoeuvre by the intruder), then the host aircraft would have to be flying in an area with less than 1000ft obstacle clearance. This normally occurs during take off or approach. During take off the climb vector may be assumed to bias the probability of an alert towards a ‘do not climb’ or ‘climb’, after the ACAS becomes active above 500ft. During the approach the probability of obstacle or ground collision may only be of consequence when below 500 ft. The significant obstacle or terrain defining the MOC may not be on track, and then the RA has to be ‘descend’, which at low altitude requires unique collision geometry, particularly where the host aircraft should be descending on an approach to land. I would suggest ‘do not climb / maintain VS’ are more likely.
Therefore I would conclude that the chance of a CFIT accident is most unlikely, even more so when considered in conjunction with a MEL’ed EGPWS.

The point is that safety systems may not be able to cope with every scenario that can be envisaged, but from the balance of probabilities of encountering such a unique scenario, the safest course of action is to follow the RA. The dangers appear where pilots invent scenarios (areas on concern) that are of little consequence (probability of occurrence) and allow their concerns to bias decisions following an alert in other, and very valid scenarios.

The solution to these dangers involves classic airmanship; the need for discipline, control of hazardous attitudes (I know better, macho, press on). This also requires knowledge of the safety system, the operating logic, and regulatory rationale behind the approved procedures; for ACAS this appears somewhat deficient. The airmanship skills are those required to act quickly to an alert, the reaction should be a ‘trained for’ response in a predetermined, an announced situation – the RA. This is defence in depth, which should also include altitude encoding transponders and revision of the EGPWS MMEL.

I fear that too often some pilots outthink safety systems or unnecessarily challenge regulatory assumptions, and in doing so, create new hazards that the certification process should have accommodated on the basis of probability; but I stand to be corrected on my limited knowledge of the detailed workings of, and the certification of ACAS.

P.S. I am sure that you realise, but for other Ppruners, turning the EGPWS Terrain mode off does not inhibit the basic GPWS functions. Also, are Honeywell addressing the lack of runway data? Their response to data omissions and the on-line update service is normally very good.

[None]

From a Company Newsletter

This paragraph was recently put out in a newsletter:

"TCAS:

As a reminder, Change 7.0 desensitizes the alerting algorithms to eliminate nuisance alerts and to accommodate RVSM. Therefore, it is imperative that crews react as commanded in a timely and effective manner.

Since the TCAS systems on the two target aircraft are continuously reacting to changing flight profiles throughout the encounter, the crews cannot rely on visual sight as a means to not execute the full extent of the commanded maneuver. Ineffective execution to commanded vertical speed will result in increasing vertical rate commands or a reversal of vertical direction to compensate."

[LEM]

I have been criticised sometimes for not having a religious respect for Company SOP's.

My answer has always been the same: SOP's are made by Captains X, Y, Z, who still carry with them personal habits, often pathetic ignorance, and often don't even agree with each other.

It would be great if SOP came from God himself.
The brass in my company doesn't even know of Pprune!

This is the c**p I have to live with:

" An RA may be disregarded only when the pilots visually identify the potentially conflicting traffic and decide that no deviation from the current flight path is necessary"

[alf5071h]

LEM try some subtle education for your managers; show them some of the Eurocontrol documents on safety subjects (or leave them around for them to find).

See:- Downloads

See the ACAS II Bulletins here: -
ACAS Safety Information web page ACAS Safety Information
‘Incorrect use of the TCAS traffic display'
‘Controller and Pilot ACAS regulation and training’.
‘TCAS II and VFR traffic’.
‘Wrong reaction to “Adjust Vertical Speed” RAs’
‘RAs and 1000 ft level-off manoeuvres’.
‘Follow the RA!’

Do request a copy of their Level Bust Tool Kit – and use it.

Eurocontrol ’Hindsight’ safety magazine

[LEM]

alf5071h Thanks for your valuable suggestions and links.

Sometimes it's hard to always fight alone.... like being the only one (!) in our base to routinely reduce the vertical speed to 1000 in the last 1000 ft (both climb and descent, of course).

Cheers, LEM

[Narada]

Interesting thread (and there have been a few before, and there will likely be more in the future). After reading several related materials (including ICAO's and FAA's), the gist of the current regulations as I understand them (and agree with) is:

Follow the RA - unless there is positive determination that following the RA will endanger the safety of aircraft (in that case, don't manuver opposite of RA).

The only exception to following the RA is when doing that will definitely put the safety of the aircraft in jeopardy (like flying into terrain or another aircraft visually confirmed as in conflict).

It is the understanding of the wording of these exceptions that appears to cause some confusion. These words are NOT to be interpreted as "don't follow the RA if you have visually identified the conflicting traffic *you think TCAS is telling you about*." They are to be interpreted as "don't follow the RA only if you have visually confirmed aircraft/terrain/obstruction you will run into while folllowing the RA, and even in that case don't maneuver *against* the RA."

==

The exceptions to RA noted are basic common sense pilot responsibilities that don't require any special "judgement." Once that is understood, "follow the RA" is the simple rule to remember.

[edited to remove typos]

[Firestorm]

Follow the RA. I nearly lost a friend (not to mention the crew and pasengers of his aeroplane, and the other aeroplane) because the other guy didn't.

[saywhat]

If you ain't gonna follow it, you might as well switch it off and count to 20. If you're still alive, you might have done the right thing. (or you might have another 10 seconds to count)

[rubik101]

RA at Max FL

Tooling along at max weight at your max certified FL, close to the buffet margin and you get RA.......CLIMB... Poor vis and no sign of traffic.
What do you do?

[LEM]

You climb! No problems at all.

[rubik101]

All agreed? Bust the certified ceiling by 500' not a problem for us?
I agree, that's what I would do, a 1000' if required.

[Right Way Up]

For all those that think they know better (and lets face it if we were that bright we probably be earning stacks of money in a proper job), could you please set your transponders to TA only. At least all of us that value our lives may have a chance of surviving a conflict.

[rubik101]

Right Way Up, would you care to elaborate? Why set TA only? I don't follow your logic, if there is any.