Radio altimeter in Turkish crash
Thread Starter
Join Date: Nov 2005
Location: Kildare, Ireland
Posts: 41
Likes: 0
Received 0 Likes
on
0 Posts
Radio altimeters on approach
Just a question regarding the findings from the Turkish crash at Amsterdam. I thought I'd better post it here rather than get torn to shreds by Rainboe et al in the accident thread. 
Am I correct in thinking that there are two radio altimeters, the left of which is linked to the autopilot (or is it the autothrottle??) and was faulty in this case (out by nearly -2000ft). My question is why is a radio altimeter being used so far out on an approach? Should such altitude data not be taken from the conventional altimeter or GPS data instead, and only the radio altimeter take over when say over the threshold? What happens if your final approach takes you over high terrain, as in eg Toncontin Airport in Honduras YouTube - Crazy Landing It just seems strange to me why such a system exists and in this case failed miserably.
Am I correct in thinking that there are two radio altimeters, the left of which is linked to the autopilot (or is it the autothrottle??) and was faulty in this case (out by nearly -2000ft). My question is why is a radio altimeter being used so far out on an approach? Should such altitude data not be taken from the conventional altimeter or GPS data instead, and only the radio altimeter take over when say over the threshold? What happens if your final approach takes you over high terrain, as in eg Toncontin Airport in Honduras YouTube - Crazy Landing It just seems strange to me why such a system exists and in this case failed miserably.
Last edited by Otto Nove Due; 8th Mar 2009 at 21:03. Reason: Change title
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
892 - you may wish to look at http://www.pprune.org/tech-log/25442...altimeter.html.
The RA feeds this 'tuning' information to the autopilot on its side (if in use). It is good enough over rough terrain, and could not work easily on pressure altimeters since we use height above sea level as our datum and not above ground. GPS would be possible, but not always available, and would need to be programmed with airfield elevation as well. Believe me - it normally works fine. We have an 'oddball' at AMS.
The confusing complication is that RA #1 feeds the autothrottle, regardless of which A/Pilot is in use.
Keep your head down - he might find you here
The RA feeds this 'tuning' information to the autopilot on its side (if in use). It is good enough over rough terrain, and could not work easily on pressure altimeters since we use height above sea level as our datum and not above ground. GPS would be possible, but not always available, and would need to be programmed with airfield elevation as well. Believe me - it normally works fine. We have an 'oddball' at AMS.
The confusing complication is that RA #1 feeds the autothrottle, regardless of which A/Pilot is in use.
Keep your head down - he might find you here
Join Date: May 2000
Location: Seattle
Posts: 3,196
Likes: 0
Received 0 Likes
on
0 Posts
The radio altimeter is simply one of many sensors that provide input into the FMS (flight management system) and autoflight systems (autopilot, flight director, and autothrottles). It also provides input to the GPWS (ground proximity warning system.
In general, one sensor in a redundant array (e.g., 2 or 3 radio altimeters) is the "master" and provides the input into the autoflight systems under normal conditions. Depending on the airplane/installation (I'm familiar with the 747, not the 737), the system will detect a fault in a bad "master" RA and switch to a good one, or give a warning to the crew. In a Cat 3 "fail operational" system, no single failure of a critical component will cause a "bad thing" to happen.
In the case of the Turkish 737, it appears there were only 2 RAs installed, not 3, and one of them (the "master") was known to have problems. Still, if the crew was aware of this and was properly monitoring the approach, even a premature "Flare" signal to the autopilot or autothrottle could be easily dealt with -- either revert to manual control and continue the approach, or hit the "Go Around" switch and deal with it later.
In general, one sensor in a redundant array (e.g., 2 or 3 radio altimeters) is the "master" and provides the input into the autoflight systems under normal conditions. Depending on the airplane/installation (I'm familiar with the 747, not the 737), the system will detect a fault in a bad "master" RA and switch to a good one, or give a warning to the crew. In a Cat 3 "fail operational" system, no single failure of a critical component will cause a "bad thing" to happen.
In the case of the Turkish 737, it appears there were only 2 RAs installed, not 3, and one of them (the "master") was known to have problems. Still, if the crew was aware of this and was properly monitoring the approach, even a premature "Flare" signal to the autopilot or autothrottle could be easily dealt with -- either revert to manual control and continue the approach, or hit the "Go Around" switch and deal with it later.
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
Guys - there is a distinct danger that we will start rehashing all the misconceptions and theories from the AMS thread again here. This is a good place to ask questions about Radalts etc, but PLEASE take the time to read through the AMS thread which will answer a lot of your questions about what how and when there.
Thread Starter
Join Date: Nov 2005
Location: Kildare, Ireland
Posts: 41
Likes: 0
Received 0 Likes
on
0 Posts
No Rainboe, this is a question from someone who doesn't know much about these systems and was good enough not to "contaminate" your and others' thread with "stupid questions". I have read that thread but it is more geared towards pro pilots who already know these systems well so it's sometimes hard to break it downinto lowly PPL level. As you rudely told me in the AMS thread, go ask these questions in the Questions thread...which is what I've done but still your not happy. Why don't you just go back to the other thread and spout some more there and leave us idiots in peace here to try and broaden our knowledge to somewhere even close to your obvious encyclopedic databank. We're not worthy.....
BOAC and Intruder - thanks for giving me a civil answer, I understand it now. In hindsight I shouldn't have mentioned Turkish in the title as I meant it as a general question, brought about by the Turkish crash, but in no way meant to be a rehash of the thousands of posts in that thread. Mea culpa
BOAC and Intruder - thanks for giving me a civil answer, I understand it now. In hindsight I shouldn't have mentioned Turkish in the title as I meant it as a general question, brought about by the Turkish crash, but in no way meant to be a rehash of the thousands of posts in that thread. Mea culpa
Join Date: Apr 2008
Location: United Kingdom
Posts: 14
Likes: 0
Received 0 Likes
on
0 Posts
I thought I was safe for a moment - and then raging Rainboe came and found us!
I just hope I'm never a passenger on his plane - his blood pressure must be borderline volcanic!
I just hope I'm never a passenger on his plane - his blood pressure must be borderline volcanic!
Warning Toxic!
Disgusted of Tunbridge
Disgusted of Tunbridge
Join Date: Jan 2005
Location: Hampshire, UK
Posts: 4,011
Likes: 0
Received 0 Likes
on
0 Posts
Fair enough, and actually a reasonable question. What was not in the answers is one of the primary functions of the RA is in terrain avoidance, preventing CFIT, so it only comes alive below 2500' above terrain. In this time, it is hoped there is enough time to provide GPWS warnings over any terrain. The only other time it becomes useful is on final approach, usually below 1000' onwards for stabilisation checks, and is totally relied on for autolands and all the associated calls and automatic actions that take place like commanding throttles closed and flare to commence.
Guest
Posts: n/a
As a dinosaur the 737 has "only" 2 Radar Altimeters. As a law, a radar altimeter is not required for any approach, but CAT II / III.
The terrain warning at 2500 ft should be an SOP trigger to cross-check both RA indications and will help to identify any discrepancy.
As mentioned before, the #1 RA is used for further inputs to the system, and it starts with the integrity check of the systems when AP 1 and 2 are selected during the approach.
Rainbow, this starts of course before 1000 ft .... (I know you will tell us now that you knew that.)
The terrain warning at 2500 ft should be an SOP trigger to cross-check both RA indications and will help to identify any discrepancy.
As mentioned before, the #1 RA is used for further inputs to the system, and it starts with the integrity check of the systems when AP 1 and 2 are selected during the approach.
Rainbow, this starts of course before 1000 ft .... (I know you will tell us now that you knew that.)
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
What I think is coming out of the AMS accident is the need for a heightened awareness of the importance and function of RA1, included in the OM/FCOM. Personally I am content that there are only 2, that there is no comparator and that crews SHOULD be able to cope with malfunction. Perhaps this will change now.
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
This post is to try and entice 'Safety Concerns' over here!
Since the defect was apparently NOT reported, all your comments on MEL simply cloud and confuse the issue on R&N. Few dispute your points and your concerns are echoed in most pilots, although your methods are less than well received.
Some points from a pilot's perspective, SC, on your posts:
1) Do you guarantee that pilots would notice the (transient) occasional defect?
2) If they did, and reported it as such, is there a non-volatile, interrogatable bite on No 1 RA which would contain all faults of the last 25 hours?
3) If not, the chances are extremely high that a 'bite check c/o, NFF, please report further' would ensue (9/10 'results' in my experience). The way engineers have designed tech logs, that 'request' will probably only last as 'visible' for one sector.
I would hope that AEI will responsibly and properly address the apparent difference in MEL actions between operators for this defect as well, by affirmative action rather than internet and press 'leaks'. Presumably you have access to the Boeing master MEL?
Since the defect was apparently NOT reported, all your comments on MEL simply cloud and confuse the issue on R&N. Few dispute your points and your concerns are echoed in most pilots, although your methods are less than well received.
Some points from a pilot's perspective, SC, on your posts:
1) Do you guarantee that pilots would notice the (transient) occasional defect?
2) If they did, and reported it as such, is there a non-volatile, interrogatable bite on No 1 RA which would contain all faults of the last 25 hours?
3) If not, the chances are extremely high that a 'bite check c/o, NFF, please report further' would ensue (9/10 'results' in my experience). The way engineers have designed tech logs, that 'request' will probably only last as 'visible' for one sector.
I would hope that AEI will responsibly and properly address the apparent difference in MEL actions between operators for this defect as well, by affirmative action rather than internet and press 'leaks'. Presumably you have access to the Boeing master MEL?
