Mid-Air Collision over Southern Germany (merged)

[411A]

Absolutely spot on, Devils Advocate, IMHO.
Is it likely to happen?....don't hold your breath.

[Standard_Departure]

TCAS Control Question...

Can anyone tell me if there are any present or future plans that will allow autopilot to "execute" TCAS RA's, so that the Pilot must override the manoeuvre should this be required?

[garp]

Devil's advocate,
Perhaps you should wait to see what the exact tracks of the concerned a/c were just before impact. Perhaps you will see that one of the two a/c made a turn. I suppose it will not take very long before the charts with the plotted tracks will be released to the public. After that you might want to reconsider your proposal.
Garp

[Standard_Departure]

Ho hummm,

Methinks that garp has inside information ??

[garp]

Latte Tester,
I'm sure that if you don't get any traffic information it simply is because the controller has other priorities. Whenever possible I will give traffic information but when it gets too busy you have to throw that overboard. Establishing separation is the main goal, getting traffic information is like receiving the Cognac on the house after a good meal, nice gesture but you shouldn't expect it every time you go to the restaurant.
Garp

[Standard_Departure]

The Dubai CTA has recently changed, instead of one big airspace, it has been split into Arrival and Departure Sectors. These sectors are spread all over, some arrival sectors on top of departure sectors, some departure sectors on top of arrival sectors, and some sectors that are vertically exclusive to one or the other.

The point is, there are often times that aircraft are descending or climbing toward other aircraft that are not on the same frequency.

To avoid the pilots continually asking, or at least to allay concerns about the TA they are just about to receive, I will attempt to always give traffic information, including relative bearing, distance direction of cross and restricted above/below.

This seems to work best all around. (But as garp says, when the frequency gets loaded, that the first to go)

[ferris]

I have been thinking about this accident a lot, and the one thing that stands out to me is a system whereby at some indeterminable point, the two aircraft were operating under different control regimes at the most critical time. That is, it appears the TU was under ATC, and the DHL was under TCAS.
That, to me, is the fundamental flaw.
What currently happens is everyone accepts ATC instructions until some point where they take matters into their own hands, and as we have seen, that didn't produce the required result.
In the few brushes with TCAS that I have had, the situations were resolved by the aircraft reporting 'TA' to me, then me providing the info and what the plan was. In one case, the a/c BOTH ignored RA's and followed my instructions (and ended up with a much better result- in my opinion).
I don't think anyone doubts that if the RA had not been complied with, there would have been no crash.
Maybe this aspect should be focussed on; as someone else said "TCAS used as an AID, not as gospel"

[Devils Advocate]

garp - it will indeed be interesting to see the tracks ( uhm, so how come you're privy to them then, and will they pop-up in the ATC forum perhaps ? ). That said, would you please explain how it is possible for two aircraft which are apparently on a collision course (LOCB), at which if one or both turn right ( to remove the LOCB) that they can still hit each other ? (given that they're roughly doing similarly constant speeds, etc); and why, for years, if a collision risk exists (LOCB) have pilots relied on each other to each turn right (to take away the LOCB) in order to avoid hitting each other - or was this just BS ?

Maybe what you're alluding to is that if ( and I'm HYPOTHESISING here ) two aircraft (A & B) are on a collision course, each gets a TCAS RA, aircraft A descends, aircraft B should climb but - for some reason known only to its crew - also descends, aircraft A turns right, aircraft B turns left, both collide - this is known as the "You're having a very bad day - aka. Sh!t Happens !" scenario.

[BOAC]

D A - TCAS does NOT use bearing information. Cpt Pit Bull has posted some excellent info here and on tech log about TCAS which you should read. The bearing info (often not reliable) is to aid YOU in visual acquisition of target. It works on closing rates, range and altitude differentials, and has NO concept of collision courses except that on such, at conflicting altitudes, the closing rate and range become significant. You do NOT have to be on a collision course to trigger an RA - only the separation minima have to be threatened.

Regarding turning to the right - what if the target is slowly closing with you from the right hand side on a converging course and alongside at your altitude - say half a mile away in cloud?

[garp]

Imagine the scenario were you have a very fast jet (500kts) and a slow moving turboprop (200kts). Bearing in mind the inaccuracy of TCAS in the horizontal plane would it not be possible that the jet by turning right would go straight into the turboprop. Let's say they're 10 NM apart on a 90° angle. A turn of the jet would bring the two together since the turn to the right of the turboprop is barely doing anything due to the slow speed. If they would keep their headings the jet would be flying ahead of the turboprop. Please correct this theory which I have seen many times in practice.

Garp

[Standard_Departure]

VDL Mode 4

I cannot wait for the wide implimentation of VDL Mode 4, I suspect that apart from some other really neat features, the whole concept of ACAS's will improve in leaps and bounds.

[Ghostflyer]

Having spent half my life trying to create collision geometry prior to releasing a weapon or entering a visual merge, air-to-air geometry has been of more than a passing interest to me.

I am a great believer in offset tracking (lets do it) but it only helps in head to head or head to tail passes. With other collision geometry it may turn a miss into a collision or vice versa, it just moves the collision point randomly.

As to semi-circular levels, they don't always help either, 2 aircraft can both be at perfectly legitimate semi-circular levels and still be on a collision course.

Visual aquisition out of the window is very useful for lateral deconfliction but not necessarily for vertical deconfliction. Next time you fly down an airway watch the opposing traffic a level below you visually. It will always appear to be slightly high on you until the sightline rate breaks and then it will pass below.

Lateral deconfliction is a simpler thing visually. If an aircraft is stationary in the wind screen you will hit him; if he is at the same level. To break the collision all that is necessary is to turn towards his aircraft to generate a sightline rate. Turning right doesn't work all the time unless you both do it.

As an aside, you also have to be certain that you have acquired the right aircraft in our ever denser ATC environment. I know of one mid-air where the pilot concerned was looking at what he thought was the conflicting traffic as he hit his wingman. Those 2 aircraft were over 2 nms apart.

So what are the keys to collision avoidance. Well this accident seems to be the classic unbroken chain.

The 2 aircraft were at the same level.
The Swiss controllers collision warning system was turned off.
The collision geometry was perfect.
There were communication difficulties.
Neither aircraft saw the other except on TCAS.

If any of these links had been broken the accident wouldn't have happened.

Of one thing that I am certain, sitting in the cockpit with the lights down on long-haul flights is not the answer. After quite a short period of time attention drifts and then empty field myopia sets in because there is nothing to look at. Shortly there after everyone on the flightdeck is asleep.

During the day, even with an Air Defence radar showing where the target is it is very difficult to see it, if it is on a collision, until the Target Designator Box in the Head Up Display becomes active. (TD Box - A circle that surrounds the point where the target is in space to give a visual acquisition cue.) Guess what, fighters point at other aircraft to avoid visual detection.

On a clear night visual detection is easier but it is hard to establish where the aircraft is relative to you if it will pass close by. The other drawback is that if you stare at the sky long enough all the stars start to move and some of them appear to have anti-coll lights. (Whoo get the straight jacket out) Put a cloud in the way and you are stuffed anyway.

In an airliner you are worse off. The radar isn't telling you where the target is TCAS is. Sadly, TCAS does not have particularly good lateral resolution which could lead to you looking out the wrong side of the aircraft based on a TCAS cue. The vertical resolution is good and one of the keys must be to monitor the TCAS display and listen to ATC instructions that are given to the aircraft around you. But practice like the fighter mates do! Use the TCAS to acquire other aircraft visually. Find out where to look for an aircraft that is 60° left on the scope and 4000' low but climbing at 25nms. The more you do it, the better you get. Then you will have a better chance of spotting the intruder when you really need to.

The best way of building situational awareness is not to lose it in the first place. In a dense air traffic environment it is essential to listen to the other aircraft on freq around you and try and work out what they are doing. But even that won't work in some airspace, try blundering through Bombay's airspace, at night, during the monsoon, whilst trying to talk to the mighty Mumbai ATCers on HF. The chances of you having a mental plot of any aircraft within 1000nms is zero.

If you do spot that the TCAS shows an aircraft at your level and it looks like a collision start talking about it. Don't assume, check! If you have the room in your airway to take avoiding action laterally do so. If not ask for it. If you are still uncertain do something, being alive and having to write an ASR beats the alternative. Remember the rules only work if every one folows them, when someone has dropped the ball that is the time to earn all that Flightcrew dosh. If you are an ATCer feel uncomfortable whenever blips might have to merge on your scope and if you can avoid it do. Nobody really cares how slick the feed in is as long as it is safe.

The bottom line is that this is an unfortunate avoidable accident. Had anyone of the players involved spotted what was about to happen and taken timely avoidance action we wouldn't be discussing this now. How often do close airmisses generate this kind of response.

All that we can do is send our condolences to those that have been lost but then as professional aviators make certain that we try to avoid the same mistake in the future. I don't know whether I would have done anything differently were I in their shoes. What I do know is that quite often the information is there but it is either missed, ignored or in the worst cases not acted upon until it is too late.

As they used to say in Hill St Blues, 'Lets roll and lets be careful out there'

Ghost

[Standard_Departure]

Ghost

Good Posting, thanx

[Devils Advocate]

W.r.t. what if the target is slowly closing with you from the right hand side on a converging course and alongside at your altitude - say half a mile away in cloud? - surely in that instances the potential infringement of the TCAS bubble around each aircraft will cause a RA - where one should climb and one descend - so making a turn should have very little effect on that.

Also, and certainly when two aircraft operate at fundamentally different speeds (e.g. fast jet versus turboprop - as posed by 'garp' ) turning might be inappropriate, but surely only if the aircraft stay at the same altitude (which is what TCAS is supposed to prevent).

All I'm wondering is that where aircraft are subject to a RA, and as such they should be diverging w.r.t. their altitudes, would a turn maybe also be of some value - albeit not in all 'normal' RA instances, but instead just perhaps of benefit for those instances when somebody does not comply with their RA - after all, what we're looking at here is a scenario where ATC has failed to keep you apart, TCAS is telling you both to do something but one of you isn't, so what other choices do you have ?

Admittedly though, there's no two ways about it, but due to the 'uncertainty' of the position and movement of each aircrafts respective TCAS bubble - certainly from the perspective of the pilots - it's probably a tough choice as which direction you should turn, if at all.

Of course if the answer is to throw technology at the problem maybe what's really needed is TCAS that's hooked up to GPS (make it compulsory) that way you'll get highly accurate heading, speed, height information and as such the TCAS unit should be able to triangulate a position and so produce a more appropriate RA maybe with 'roll' guidance - uhm, would we then still need ATC ? …… only kidding


Ps. (hence the edit) Ghost - bravo, an excellent posting !!!

[Seriph]

Ah we are down to the sanctimonious and self rightous preaching for which these forums are noted. Experts suddenly everywhere, so you can't stay awake with the lights down eh Ghost, prefer to insulate yourself from your environment, well it's your choice but don't preach it as gospel. As for lateral separation, I crossed France yesterday on three directs, crossing traffic and opposites everywhere, how do you arange offsets? and if everyone sticks to the airways then you just congest further.

[Ghostflyer]

Seriph,

Nope I'm not snctimonious (see can't even spell it) or self righteous, I don't know if anything I have said helps but I hope it does. I do know of a lot of sanctimonious arrogant 'professional' pilots that go around with there head up their own a***s rather than try to learn from others. They spend their whole days bitching about problems without ever trying to make anything better.

I have learn't a lot from the contributions of other people on these forums and if just one snippet prevents anything like this from happening in the future, I for one will be greatful.

What would you have done put the blinds up and prayed to the big sky principle. Sounds like the sort of preaching that went on in the dark ages.

Holy s**t, I just bit..... oh well feel better now,

Ghost

Mr DA,

The closure issue is interesting but effectively all that happens is the exact collision point moves. I once tried intercepting a Sea King doing 80 kts IAS with a ground speed of about 20, it was a a horrible nightmare only to be discussed in the bar. Against slow speed targets, any avoiding action by you ensures a wider miss distance.

Putting the boot on the other foot, lets say you are in a turbo prop and configured whilst trying to avoid a fast jet that is not on your centreline but still on a collision. A turn towards the intruder that points your nose behind him uses his speed and geometry to create the miss. You have to turn more than he would to get the same miss distance because most of the closure rate is being generated by his platform. However, if you both turn it can get mighty confusing hence the right hand traffic rule. The key is to do it early and decisively so everyone knows what is happening and more importantly share your information with the controllers.

As to turning to avoid a collision to provide a 'belts and braces' miss once all else has failed; it depends on when you do it. If you have TCAS or visual SA of the approaching threat and can clearly see how to avoid him then taking that action early, even before a TA cannot be a bad thing. If he is stationary in the windscreen and the TCAS shows him at your altitude even a slight turn at range helps immeasurably.

Getting to worst case scenarios, even a 10° turn that generates a line of sight rate will produce an acceptable miss distance if actioned when you get the TA. On a head to head at the TA (40 secs to impact) assuming both aircraft at 0.8 mach you have 16 miles per minute of closure.

So if you actioned the turn at the TA and had it completed with 30 secs to go ( 8 nm) you would miss the other aircraft by about 4,000 feet laterally using the 1 in 10 rule.

If you wait to start a turn until you get an RA at about 25 secs and again took 10 secs to recognise and complete the turn you you would get a miss of around 2,000 feet laterally but at some cost of vertical performance.

Warning To do something this late, you have to be certain that you are carrying out the right action. For that to be the case you must have sound SA.

If you really do have the SA though, do something early! Don't wait for an RA to make your move. Avoid the TCAS conflict completely if you can by firstly talking to ATC and secondly maneuvring. By the time you have got to an RA you are in an emergency situation and want to have the entire kinematic performance of the aircraft available to you in the vertical plane.

My companies SOPs say try to acquire the aircraft visually if you get a TA but do not attempt to maneuvre based on a TA alone.
Well if I got the TA, spotted the traffic, assessed the geometry as a collision that would result in a subsequent RA and could turn to avoid it, I would. (But you don't have much time and again you have to be sure and decisive.)

However once I got to the RA, I would follow the TCAS commands religiously in case of coordinated avoidance. Some trainers will say take the time to disconnect the AP and turn off the FDs in the time between the TA and RA, boll**ks get your head out and try to sort the problem out.

Quality decisions are normally made in good time, the TCAS conflict scenario is something that can be avoided by timely avoiding action at range. Like most of us I listen to traffic climbing and descending and try to tie it into TCAS returns around my aircraft. If we do have the SA it is essential we give that to the men who can fix it in good time - ATC.

If used correctly with the display range and altitude coverage set correctly and if we regularly scan it there shouldn't be too many surprises. Almost always the intruder will have been in the TCAS display system for some time before generating a TA. In the future I know that I will be taking as much interest in the TCAS approaching waypoints as I do now when I get into terminal airspace.


Ghost

[HotDog]

I was severely admonished by my captain one day in cruise east bound for reading a newspaper between my regular panel scans. Reason given was degregation of lookout alertness. Next day after an overnight stop in Japan, we took off and climbed to cruise level west bound late afternoon, whereupon same captain plastered his windscreen with his Jeppeson enroute charts to keep the sun out of his eyes. I must admit, this was before the CRM concept. What to do?

[BOAC]

Going a bit off an important thread, here, HotDog, but to answer your point - keeping the sun out of one's eyes degrades SA a lot less than reading a newspaper.

[vanman]

I have to say that the argument for flying a couple of miles right of each track as an offset sounds so obvious and relatively easily accomplished I can't understand why it's not already here. TCAS is rocket science by comparison. It clearly would not have saved them in this case though, as the two a/cs tracks were basically 90deg off (Skyguide press release picture of a/c tracks )

Ferris recently said:-

What currently happens is everyone accepts ATC instructions until some point where they take matters into their own hands, and as we have seen, that didn't produce the required result.

No - they accept ATC instructions until ATC fail them, then they (must) put there faith into the TCAS only. It would seem the conflict between TCAS and controller instructions may have caused an incorrect response from one aircraft. The controller had been informed that TCAS RA was being actioned and at that point I would argue he should have said no more until the conflict was avoided.

In one case, the a/c BOTH ignored RA's and followed my instructions (and ended up with a much better result- in my opinion).

Hmm. Don't be too alarmed if I chose to ignore your 'advice'. This is exactly the type of situation which could generate another such situation. Don't forget that, if as Standard Departure (do your mates call you Sid?) said:-

An important thing to remember, is that as ATC's, we get a reduced re-fresh rate (sweep update) on the SSR, that is to say relative to the aircraft equipment, and it may well transpire that what you and I see on the Mode C readout is out of date by 5 seconds, and could imply that aircraft have already crossed levels and cleared each other.

...then I for one will take my chances with the TCAS and Mk1 eyeball thanks.

I for one found an urge to re-read about TCAS in the light of this sad event. (can't think why ) Found this Honeywell TCAS II Pilots guide (2Mb) very useful, others may too. It certainly confirms that an a/c will receive an RA (not just a TA) if the other a/c has unserviceable TCAS but functioning transponder (mode C).

Also worthy of note is that a pilot is expected to react within 5 secs to the RA and only 2.5 secs to the RA reversal (which would presumably have been received in the 757 during the original TCAS descent). Possibly particularly useful info for Orca Strait who said:-

Upon receiving a TA/RA, my first reactions are; check my altitude, eyeballs outside - then lights on, day, night, VMC or IMC. The gee whiz marvels of today’s technology are tools that still need old fashioned airmanship for backup.

...just make sure you do it quickly, it might be me coming the other way.

Lastly, there have been serveral enquiries (including my own) about what happens when only one controller is on duty (should he become incapacitated (as this poor individual probably was after watching this disaster unfold before his eyes). Not much of a response as yet, anyone do better?

Cheers

[purr]

regarding aircraft closing in at an angle the solution has been around for some time in the shipping corridors infact officers have an oral exam on this and the same procedures are in force the world over .