Lake Constance midair collision - why no hard turns ?

[Guest ThrottleUp]

Hello all :)I just got to see a Discovery Channel programme on the tragic July 2002 midair collision between a Tupolev-154 & a DHL 757 over Lake Constance at FL350 and came away with a burnng question in my mind that I wanted to put to all of you.The investigation focused on two main factors:(1.) The overworked ATC controller using standby equipment while a colleague was on a break.(2.) The TCAS system. The pilots were alerted to the collision danger and the Tupolev was instructed by ATC to descend but its TCAS was screaming at it to climb. The TCAS in the 757 also commanded a descent. In the event the Tupolev crew chose to listen to ATC instead of their TCAS and *descended*. The 757 crew listened to TCAS and descended -which began to scream for an increased rate-of-descent (because the TU-154 was also descending)Tragedy occured as both airplanes collided at 35,000 ft over Lake Constance.I came away from this programme thinking one thing:Why TCAS did not command TURNS in such dire circumstances.The B757 was on a heading of 004 & the TU-154 was on a heading of 274. If the B757 had turned hard left during its descent & the TU-154 hard right - this would not have happened.I guess Im asking wether TCAS only allows a straight-ahead descent or wether it can also command hard turns?What do you all think?

[michal 44]

I don't know whether TCAS can command turns.But it is not an issue here. TCAS worked very well in this case. The main problem was Rusian pilot's decision to follow directives from the controller. It was a big mistake, why? Because pilots in the western world (not in Russia where TCAS is still relatively unknown) are correctly taught to disregard ATC controller in cases where speration minimums have been breached and TCAS activated. If seperation minimums have been breached it means human controllers failed miserably and you are left with automation to get you out of the situation. When 2 airplanes are already so close (seconds from collision) there is usually no time for controllers to correct their errors - at least they should no longer be trusted as the reasoning goes. It is hard to argue with this logic.Michael J.WinXP-Home SP2,AMD64 3500+,Abit AV8,Radeon X800Pro,36GB Raptor,1GB PC3200,Audigy 2

[Guest BobLee]

Since TCAS is transponder based, i.e. it reads the altitude of the target, it can only issue climbs or descents as resolution advisories.

[michal 44]

> Since TCAS is transponder based, i.e. it reads the altitude>of the target, it can only issue climbs or descents as>resolution advisories.I have no reason to question the above however the actual TCAS display does show the correct spatial relation (bearing) of both airplanes. So one might think that the 'system' could be smart enough to figure turn commands. However it is possible that turns are more likely to result in secondary warnings whereas altitude change commands are a bit safer in this regard.Michael J.WinXP-Home SP2,AMD64 3500+,Abit AV8,Radeon X800Pro,36GB Raptor,1GB PC3200,Audigy 2

[Wolfko 0]

I guess one of the reasons, that TCAS only advises "climb" and "desenct" is, that these commands are simple and easy to execute.If a turn was involved, those turns have to be calculated, the current headings of the aircrafts, thier speeds have to be considered etc. And also for the pilots it is easier to just execute a climb or a decent, rather than to make a turn to a specific heading.Just pulling or pushing the stick, that's the best, most effective and simpliest solution. Everything else was too complicated in such a stressful situation.Wolfgang

[michal 44]

>I guess one of the reasons, that TCAS only advises "climb">and "desenct" is, that these commands are simple and easy to>execute.You are absolutely right. As a matter of fact with some converging angles (try for example 90 deg) it would be extremely hard to provide reliable turn command. Altitude change is much simpler and gets the job done reliably. I happen to work at NASA Ames at the Air Traffic Management division and recall attending an 1-hour presentation on aircraft traffic conflict resolution and it was crystal clear after listening to and watching various shown scenarios that heading change is quite difficult and full of nasty surprises.Michael J.WinXP-Home SP2,AMD64 3500+,Abit AV8,Radeon X800Pro,36GB Raptor,1GB PC3200,Audigy 2

[jagabom 21]

Michal, I'am looking at getting a puter like yours. How's it work?JimCYWG

[michal 44]

>Michal, I'am looking at getting a puter like yours. How's it>work?>Jim>CYWGI love it. I had some problems at the very beginning with excessive CPU temperatures but BIOS change fixed that.Michael J.WinXP-Home SP2,AMD64 3500+,Abit AV8,Radeon X800Pro,36GB Raptor,1GB PC3200,Audigy 2

[Guest w4p]

I agree with the complication issue, but with ac of that size a turn (even a hard turn) would take to long to execute in an emergency situation. An ac can dive and (at cruise speed) climb rather quickly when compared to turns.

[Guest bisinchi]

Hi,It looks like at the moment TCAS only gives advisories to the pilot in the vertical field.TCAS works with the transponders and at the moment, the number of information given by the transponder is pretty limited and doesn't include heading. With Mode S and enhanced Mode S transponders, more information will be transmitted, which will incude heading, rate of turn, etc and when these information will be available, TCAS surely will allow advices to pilots to make turns and to climb/descent to avoid a collision.Francois

[michal 44]

>of information given by the transponder is pretty limited and>doesn't include heading. Excellent point Francois.Michael J.WinXP-Home SP2,AMD64 3500+,Abit AV8,Radeon X800Pro,36GB Raptor,1GB PC3200,Audigy 2

[Guest ThrottleUp]

If Im not mistaken fighter planes like F-15, Rafale & Gripen have displays that can show an enemy planes altitude, heading & speed....can't they put something similar in civilian planes?Or can the above only be provided by a hi-power radar...which I guess would be too expensive...

[michal 44]

I think that in case of ground based ATC radars aircraft heading is simply derived from her position before/after. Maybe sophisticated Doppler radars can tell the heading (or rather track) in a more direct way. If you want to build a reliable last defense against midair collisions I think it is prudent to stay away from unnecessary complications.Michael J.WinXP-Home SP2,AMD64 3500+,Abit AV8,Radeon X800Pro,36GB Raptor,1GB PC3200,Audigy 2

[Guest bisinchi]

With Mode S transponders, more data will sent to the ground controlers, including heading, rate of turn, etc, which means controllers will have data straight from the plane and hence, will have a pretty accurate picture!!At the moment, the picture is only drawn from radar echo returnsFrancois

[Tom Allensworth 1,532]

"Or can the above only be provided by a hi-power radar..."Correct. Most are in the "pulse doppler" category and they are also fed (at least in NATO and US a/c) with input from AWACS or equivelant radar platforms which is used as additional targeting inputs. And they are VERY expensive to purchase and their through-life support costs aren't anything to sneeze at either.