An air carrier crew described a confusing VHHH Runway 25L ILS approach. The ILS was predictably unreliable at their location; ATC communications were misleading and the crew was fatigued thereby not understanding what the flight instruments were depicting.

2010-06 · NASA ASRS report 894153

Date: 2010-06 · Aircraft: Widebody; Low Wing; 4 Turbojet Eng

Anomalies: deviation-altitude-excursion-from-assigned-altitude|deviation-discrepancy-procedural-clearance

Synopsis

An air carrier crew described a confusing VHHH Runway 25L ILS approach. The ILS was predictably unreliable at their location; ATC communications were misleading and the crew was fatigued thereby not understanding what the flight instruments were depicting.

Narrative

On approach to Runway 25L at VHHH; we were cleared to 8;000 MSL and were proceeding direct to TD VOR; and cleared for the ILS Runway 25L approach. We were in a descent to cross the TD VOR at 8;000 MSL and had arrived at altitude slightly before the fix with VNAV engaged I began a speed reduction and flap selection to 180 KTS; I called for 4;500 to be placed in the MCP altitude window which was the altitude restriction at LOTUS the join point on the ILS. At approximately 7;500 MSL the ATC Controller advised us there was a possible altitude fluctuation; however VNAV VPI showed us on path; this transmission contributed to us feeling as though we were high; though I believe he merely saw us in a slowing segment. To be sure I would make the restriction I selected V/S and placed the altitude predictor approximately 1 mile prior to LOTUS. I then selected approach mode and planned to arrive at LOTUS at 4;500 MSL capture the altitude; and await localizer and glide slope capture. As we approached LOTUS we received an ALT CAP on my FMA and I checked the glide slope; to my surprise it showed us 1 dot high on the glide slope; which surprised me because at LOTUS at 4;500 MSL we should have been below the glide slope and have been awaiting capture from above. (The profile view indicated glide slope intercept at .9nm inside of LOTUS). At that point I tried to capture the glide slope by selecting a lower alt and engaging V/S. We captured the glide path and continued down; for about 30 seconds when I received a degraded glide slope signal FMA; followed sometime later by complete loss of the glide slope indication. Since we were in VMC conditions at that point and I thought it would only be a momentary loss of signal by a vehicle or something similar I selected V/S and continued the descent awaiting re-acquisition of the glide slope signal. At that point we received a transmission from ATC asking our current altitude; we informed him it was 3;700 FT and he informed us that that altitude was below glide slope for that distance. Almost simultaneously we reacquired the glide slope and it showed us 1.5 dot low on the glide slope; I immediately disconnected the autopilot and hand flew the aircraft back to glide slope and reengaged the autopilot. We landed without further incident. In reviewing the chain of events that lead up to and contributed to this incident I realized there were several factors. 1. First; I believe there was an element of mental fatigue; as it was the end of an 8 hour flight arriving in VHHH; this I believe hindered our recognition of what was happening. 2. My approach briefing did not emphasize the importance of 3 key elements: A. glide slope ball note-- glide slope signal unusable beyond 7 degrees left of LOC course. B. The profile view shows that there is actually a level segment at 4;500 FT for .9 NM before glide slope intercept; so any capture or indication supporting a descent prior to 14.1 nm should have been disregarded. C. Taking those factors into account I should have briefed that I would only select localizer until sure of a reliable glide slope signal 3. Another contributing factor was the ATC transmission we received notifying us of 'possible altitude fluctuation' as it served to contribute to the crews belief we were high; supported by a false high glide slope indication; when in fact we were not high. This combined with an over reliance in glide slope indication vs. approach plate altitude discipline influenced my decision to start a descent to capture the glide slope. D. Lastly; in retrospect with confusion on why we were indicating above glide slope; at a point in the approach where I knew we should have been below the glide slope; followed by loss of glide slope indication I realize that I should have executed an immediate go around to determine the inconsistencies regardless of the VMC weather conditions.

Second reporter narrative

WHAT HAVE WE LEARNED; WHAT CAN BE DONE? I feel that we could have from the start requested clarification from Hong Kong Approach as to what he meant by; 'Flight path fluctuation.' So many times operating internationally do we hear messages such as this and while usually figuring out what the Controller means through an understood language barrier; we don't always seek clarification. Usually we ask; 'Say Again;' but sometimes we are task saturated and 'assume' we know what they meant. Secondly I feel that more attention to the approach plate note for false or unreliable glide slope signal could have been emphasized in our brief. We were all to blame and established habits led us in to selecting 'APPR' mode armed to capture glide slope and localizer; INSTEAD of selecting just 'LOC' armed; flying the profile; and THEN selecting full 'APPR' mode armed after fully capturing the localizer. But; we were feeling rushed and didn't want to forget to arm the approach. While in our state of disorientation and feeling of being high the Captain selected 'APPR' armed so as not to forget. The intermittent and false glide slope signals were similar to what you see when another aircraft or vehicle interrupts the glide slope signal momentarily; and in fact this is what I believed may have been happening. The interpretation led us to believe we were high. Not the case; as we found out. All in all we should have just gone around after reaching 3;700 FT. We were not conflicting with terrain or other traffic yet; but this is always the safest thing to do; but seems the hardest to ACTUALLY do. It is pilot nature to try to fix what is presently wrong and we probably had some 'get home itis' occurring. This was a learning event and one that I'm taking away a lot of in the way of experience. It also shows how we must be prepared; interpret multiple sources of information and always take the safest course of action.

Source: NASA Aviation Safety Reporting System (public domain). Reports are voluntary submissions and are not verified by NASA.

Loading the flight search…

Frequently asked questions

How do I search flights by aircraft type on FlightFinder?

Pick an aircraft model — Boeing 737, Airbus A320, A380, Boeing 787 Dreamliner and more — enter your origin airport, and FlightFinder shows every route that plane flies from there with live fares.

Which aircraft types can I filter by?

We support Boeing 737/747/757/767/777/787, the full Airbus A220/A319/A320/A321/A330/A340/A350/A380 family, Embraer E170/E175/E190/E195, Bombardier CRJ and Dash 8, and the ATR 42/72 turboprops.

Is FlightFinder free to use?

Search and schedules are free. Pro ($4.99/month, $39/year, or $99 one-time lifetime) unlocks the enriched flight card — on-time stats, CO₂ per passenger, amenities, live gate & weather — plus My Trips with push alerts.

Where does the route data come from?

Live schedules come from Amadeus, AeroDataBox and Travelpayouts. Observed routes (which aircraft actually flew a given city pair) are crowdsourced from adsb.lol ADS-B data under the Open Database License.