Air NZ Erroneous Glide-slope Incident
Thread Starter
Join Date: Sep 2002
Location: Auckland, New Zealand
Posts: 2
Likes: 0
Received 0 Likes
on
0 Posts
Air NZ Erroneous Glide-slope Incident
For those of you who have been following Air NZ's 767 incident into Apia Western Samoa in July 2000, which resulted in the term "erroneous" glide-slope being adopted. you will be pleased to know NZCAA have just released their report on the incident.
You can find it at http://www.caa.govt.nz/. Just follow the link from Incidents and Accidents
You can find it at http://www.caa.govt.nz/. Just follow the link from Incidents and Accidents
Moderator
Thanks for that C, .... essential reading for all ... I hesitate to expose ANZ to a million requests .. but would it be feasible for you to indicate how one might obtain a copy of the video ?
Because of the proactive manner in which ANZ investigated the incident and the general importance of the technical aspects, I have taken the liberty of putting a direct link to the report in the Tech Log url thread.
Having now had time to scan the report, I have also put an FYI thread similar to this one in Tech Log ... the whole thing still sends shivers up my spine ....
Because of the proactive manner in which ANZ investigated the incident and the general importance of the technical aspects, I have taken the liberty of putting a direct link to the report in the Tech Log url thread.
Having now had time to scan the report, I have also put an FYI thread similar to this one in Tech Log ... the whole thing still sends shivers up my spine ....
Last edited by john_tullamarine; 17th Sep 2002 at 11:15.
Thread Starter
Join Date: Sep 2002
Location: Auckland, New Zealand
Posts: 2
Likes: 0
Received 0 Likes
on
0 Posts
John,
Thanks for the note.
If anyone wishes to purchase a copy of the video Air New Zealand made after the incident, they are free to e-mail me at the address specified at the end of the link from the NZCAA site. The video is described on:
http://www.airnz.co.nz/flighttrainin...e.jsp?pid=4004
We have recovered our costs on the project and still have a few copies.
Fly safe.
Thanks for the note.
If anyone wishes to purchase a copy of the video Air New Zealand made after the incident, they are free to e-mail me at the address specified at the end of the link from the NZCAA site. The video is described on:
http://www.airnz.co.nz/flighttrainin...e.jsp?pid=4004
We have recovered our costs on the project and still have a few copies.
Fly safe.
Join Date: Mar 2000
Location: Arizona USA
Posts: 8,571
Likes: 0
Received 0 Likes
on
0 Posts
For those that are interested...
The original design and patent on the Instrument Landing System was filed in 1937 and held by Reed Pigman...yes the very same fellow who did the dirty dive from low station inbound on a VOR approach to Ardmore Oklahoma years ago in a Lockheed Electra, not wearing his shoulder harness (not too bright)...the results are well known.
THE problems with "false" glidepaths have been well known for over sixty years.
One can only wonder...what is new?
Does not "anyone" read history?
Good grief...
The original design and patent on the Instrument Landing System was filed in 1937 and held by Reed Pigman...yes the very same fellow who did the dirty dive from low station inbound on a VOR approach to Ardmore Oklahoma years ago in a Lockheed Electra, not wearing his shoulder harness (not too bright)...the results are well known.
THE problems with "false" glidepaths have been well known for over sixty years.
One can only wonder...what is new?
Does not "anyone" read history?
Good grief...
Last edited by 411A; 17th Sep 2002 at 22:41.
Moderator
.. the problem here was not a traditional false GP, with which we are all familiar .. there was a set up problem with the ground equipment which caused an on-GP reading when there ought not to have been one .. the investigation coined the term "erroneous" GP to differentiate from the false GP. Apart from the DME discrepancy, which the crew finally picked up, it all looked reasonably normal .. wherein lay the trap.
My interest in the report is that it gives quite a bit of technical background which many, if not most, pilots may not be quite on top of ... the underlying message is "don't trust gadgets without independent checks if such are available" ...
My interest in the report is that it gives quite a bit of technical background which many, if not most, pilots may not be quite on top of ... the underlying message is "don't trust gadgets without independent checks if such are available" ...
PPRuNeaholic
Join Date: Jun 2000
Location: Cairns FNQ
Posts: 3,255
Likes: 0
Received 0 Likes
on
0 Posts
Test Mode???
Well said Captain S!
Good grief. Either the GP radiates or it doesn't. If it's radiating "on test", there is - or should be - a Notam to that effect, so that the signal isn't used. This issue relates to the equipment having been incorrectly setup in the first place.
The setup is checked by a specially equipped calibration aircraft. I haven't yet read the report - 203 pages is a bit costly for me to download on the pay-per-minute program used in this country - but it occurs to me that the setup hadn't been flight checked at the time of the incident. Or it had been "reset" by the navaid technicians, in between visits by the calibration aircraft.
In any event, the actual cause doesn't matter as much to those of us who have to use the equipment. We're more concerned with seeing our drinking buddies in the bar at the end of the flight. Thus, we really need to concern ourselves with how we can recognise the fault.
411A, you're right to suggest that "street smart" pilots are able to detect a GP discrepancy. I suspect the point being made here is that we all need to be on our guard ALL the time. Accept nothing at face value and we'll all live longer.
Good grief. Either the GP radiates or it doesn't. If it's radiating "on test", there is - or should be - a Notam to that effect, so that the signal isn't used. This issue relates to the equipment having been incorrectly setup in the first place.
The setup is checked by a specially equipped calibration aircraft. I haven't yet read the report - 203 pages is a bit costly for me to download on the pay-per-minute program used in this country - but it occurs to me that the setup hadn't been flight checked at the time of the incident. Or it had been "reset" by the navaid technicians, in between visits by the calibration aircraft.
In any event, the actual cause doesn't matter as much to those of us who have to use the equipment. We're more concerned with seeing our drinking buddies in the bar at the end of the flight. Thus, we really need to concern ourselves with how we can recognise the fault.
411A, you're right to suggest that "street smart" pilots are able to detect a GP discrepancy. I suspect the point being made here is that we all need to be on our guard ALL the time. Accept nothing at face value and we'll all live longer.
Join Date: May 2002
Location: Australia
Posts: 2,242
Likes: 0
Received 0 Likes
on
0 Posts
If a Nav Aid is transmitting on 'test' then the ident is normally prefixed by the morse letter 'T' or some other difining ident, if this had been a simple case of pilot error then I doubt ANZ would have gone to such great length, the problem would have laid within and been dealt with. Clearly this was a possible flight safety issue and ANZ went public.
[When approaching 27R at FRA, from the South, it is possible to pick up a false LOC and the AP will turn you on, also worth remembering]
[When approaching 27R at FRA, from the South, it is possible to pick up a false LOC and the AP will turn you on, also worth remembering]
Moderator
According to the report, there was, indeed, some finger trouble on the ground. The lesson, though, relates more to the fact that the cockpit presentations were substantially, if not totally, normal and the crew fell into a trap which, as they say, "there, but for the grace of God ....".
To me the main value of the investigation was that the ANZ safety guys were pretty aggressive at trying to get to the guts of what happened as opposed to the (often seen elsewhere) traditional "shoot the pilot" fix. In the ensuing sniffing around they uncovered a few electronic gotchas which the average pilot is not up with and these are described in an internal ANZ report and the NZCAA document.
During discussions at the time with one of the safety guys whom I knew I found out that there was more than a small suspicion that these sorts of problems may have been implicated in previous approach accidents and incidents.
It should be noted that OzExpat has expertise in these matters of a somewhat more detailed and expert level than the average instrument pilot. For the rest of us it has to be said that technical knowledge levels vary to a greater or lesser degree.
The CAA report, unfortunately, is set up for download in chunks which aren't really useful for getting to the interesting bits and pieces easily. The following quotes pp 146-148 might whet appetites sufficient to do the download
" .... When the aircraft captured the glideslope the crew were presented with a situation that was outside their knowledge, experience or expectation. Any warning that the crew could reasonably expect to be displayed was not presented to them. The scenario was totally unexpected and untrained for.....
....... it is the view of the investigation that a high proportion of line crews would have made the same decision at glideslope capture. Human error caused the incident but it must also be recognised that human factors prevented a more serious outcome.
This investigation has determined that one glideslope check cannot be relied upon as an absolute error trap for a glideslope radiating erroneous information. It is still possible that the glideslope/altitude check could be carried out and an erroneous glideslope not be detected by the crew....
Alternative means should be instituted to ensure the crew verify the glideslope for 'reasonableness’ prior to acceptance / during approach. The glideslope check should be used for what it is designed; to validate altimetry and trap the potential error of mis-set or erroneous altimeters to ensure that, on a precision approach, the aircraft does not infringe Obstacle Clearance Altitude/Height at DA/H and through the missed approach path. This investigation has established that, if an anomaly with the altimeter check is noted, the crew may not be able to immediately identify where the problem lies....
Once the error chain is established total reliance may be placed on crew to detect an anomaly and take the appropriate action. Faced with information that they have been trained and conditioned to accept and trust, they may have to detect subtle clues and analyse the problem while confronted with compelling information that will probably over-ride the conflict.... "
I guess I was lucky in that my apprenticeship was under training captains who encouraged a "don't trust it too much" philosophy ... but there are many, particularly among our younger PacMan brethren who, perhaps, are more susceptible to GIGO problems with computers. I see this report as being a very useful wakeup call to all of us who necessarily have to put our trust in computers over which we may have little direct control.
OzEx, I'll save a copy of the CAA file for you and present it to you with due ceremony over an ale at the Barossa next year ......
To me the main value of the investigation was that the ANZ safety guys were pretty aggressive at trying to get to the guts of what happened as opposed to the (often seen elsewhere) traditional "shoot the pilot" fix. In the ensuing sniffing around they uncovered a few electronic gotchas which the average pilot is not up with and these are described in an internal ANZ report and the NZCAA document.
During discussions at the time with one of the safety guys whom I knew I found out that there was more than a small suspicion that these sorts of problems may have been implicated in previous approach accidents and incidents.
It should be noted that OzExpat has expertise in these matters of a somewhat more detailed and expert level than the average instrument pilot. For the rest of us it has to be said that technical knowledge levels vary to a greater or lesser degree.
The CAA report, unfortunately, is set up for download in chunks which aren't really useful for getting to the interesting bits and pieces easily. The following quotes pp 146-148 might whet appetites sufficient to do the download
" .... When the aircraft captured the glideslope the crew were presented with a situation that was outside their knowledge, experience or expectation. Any warning that the crew could reasonably expect to be displayed was not presented to them. The scenario was totally unexpected and untrained for.....
....... it is the view of the investigation that a high proportion of line crews would have made the same decision at glideslope capture. Human error caused the incident but it must also be recognised that human factors prevented a more serious outcome.
This investigation has determined that one glideslope check cannot be relied upon as an absolute error trap for a glideslope radiating erroneous information. It is still possible that the glideslope/altitude check could be carried out and an erroneous glideslope not be detected by the crew....
Alternative means should be instituted to ensure the crew verify the glideslope for 'reasonableness’ prior to acceptance / during approach. The glideslope check should be used for what it is designed; to validate altimetry and trap the potential error of mis-set or erroneous altimeters to ensure that, on a precision approach, the aircraft does not infringe Obstacle Clearance Altitude/Height at DA/H and through the missed approach path. This investigation has established that, if an anomaly with the altimeter check is noted, the crew may not be able to immediately identify where the problem lies....
Once the error chain is established total reliance may be placed on crew to detect an anomaly and take the appropriate action. Faced with information that they have been trained and conditioned to accept and trust, they may have to detect subtle clues and analyse the problem while confronted with compelling information that will probably over-ride the conflict.... "
I guess I was lucky in that my apprenticeship was under training captains who encouraged a "don't trust it too much" philosophy ... but there are many, particularly among our younger PacMan brethren who, perhaps, are more susceptible to GIGO problems with computers. I see this report as being a very useful wakeup call to all of us who necessarily have to put our trust in computers over which we may have little direct control.
OzEx, I'll save a copy of the CAA file for you and present it to you with due ceremony over an ale at the Barossa next year ......
Last edited by john_tullamarine; 18th Sep 2002 at 11:41.
PPRuNeaholic
Join Date: Jun 2000
Location: Cairns FNQ
Posts: 3,255
Likes: 0
Received 0 Likes
on
0 Posts
OzEx, I'll save a copy of the CAA file for you and present it to you with due ceremony over an ale at the Barossa next year ......
Meantime, I'll try to consult the BASI guys here - they might have hard copy on it. I'm now thinking that the setup error might have some lessons for us to take on board! 
PPRuNeaholic
Join Date: Jun 2000
Location: Cairns FNQ
Posts: 3,255
Likes: 0
Received 0 Likes
on
0 Posts
I've just seen the write-up in NZ CAA's mag "Vector". The relevant bit of this very highly condensed report is as follows :-
====================
Analysis
Subsequent analysis of the flight data recorder information established that the aircraft had descended on a glide path of approximately 3.5 degrees to a point approximately 5.5 miles short of the runway with "normal" localiser and glideslope indications displayed on the flight instrumentation.
It was established that the ILS glide path transmitter had inadvertantly been left in control (monitor) bypass mode, with the unserviceable transmitter selected. In the bypass mode, the glide path transmitter executive monitor was unable to shut down the faulty transmitter or to transfer to the serviceable transmitter. The result was the radiation of invalid glide path information.
A false glideslope is a normal byproduct of the glide path. If it is intercepted and if it can be followed, it will guide the aircraft to the source of the glide path.
An erroneous glideslope, on the other hand, is not well known. It is the result of a faulty or partial signal being transmitted. It will indicate to the aircraft that it is 'on slope' irrespective of where the aircraft is in space, and it will not lead the aircraft to the source of the glide path. Erroneous glideslope signals are occasionally transmitted for maintenance purposes.
====================
If, in this instance, it HAD been radiated for maintenance purposes, there should already have been a Notam about it, or the correct signal should have been restored before any operational use could be made of it.
Indeed, as it seems that the standby GS Tx was serviceable, it would seem to border on criminal negligence to have left the monitor in bypass mode. It was, at very least, reckless and could have had far more serious consequences if the ANZ crew hadn't been alert to the disparity between the GP indication and the DME Distance Vs Altitude scale - among other cues that may not have existed if the weather had been worse.
The report makes the understatement of the century when it states that the fault "is not well known". This is simply because it cannot happen in a proffesionally disciplined maintenance environment. This seems like the sort of thing that COULD happen in some other parts of the world.
====================
Analysis
Subsequent analysis of the flight data recorder information established that the aircraft had descended on a glide path of approximately 3.5 degrees to a point approximately 5.5 miles short of the runway with "normal" localiser and glideslope indications displayed on the flight instrumentation.
It was established that the ILS glide path transmitter had inadvertantly been left in control (monitor) bypass mode, with the unserviceable transmitter selected. In the bypass mode, the glide path transmitter executive monitor was unable to shut down the faulty transmitter or to transfer to the serviceable transmitter. The result was the radiation of invalid glide path information.
A false glideslope is a normal byproduct of the glide path. If it is intercepted and if it can be followed, it will guide the aircraft to the source of the glide path.
An erroneous glideslope, on the other hand, is not well known. It is the result of a faulty or partial signal being transmitted. It will indicate to the aircraft that it is 'on slope' irrespective of where the aircraft is in space, and it will not lead the aircraft to the source of the glide path. Erroneous glideslope signals are occasionally transmitted for maintenance purposes.
====================
If, in this instance, it HAD been radiated for maintenance purposes, there should already have been a Notam about it, or the correct signal should have been restored before any operational use could be made of it.
Indeed, as it seems that the standby GS Tx was serviceable, it would seem to border on criminal negligence to have left the monitor in bypass mode. It was, at very least, reckless and could have had far more serious consequences if the ANZ crew hadn't been alert to the disparity between the GP indication and the DME Distance Vs Altitude scale - among other cues that may not have existed if the weather had been worse.
The report makes the understatement of the century when it states that the fault "is not well known". This is simply because it cannot happen in a proffesionally disciplined maintenance environment. This seems like the sort of thing that COULD happen in some other parts of the world.
