Thai Air B777 Melbourne NDB approach

[HIGH n MIGHTY]

"On 4 November 2007, a Boeing Company 7772D7 aircraft, registered HS-TJW, was being operated on a scheduled passenger service from Bangkok, Thailand, to Melbourne, Vic, with 17 crew and 277 passengers on board. During a non-directional beacon (NDB) non-precision approach to runway 16 at Melbourne Airport, the aircraft descended below the segment minimum safe altitude at 6.8 distance measuring equipment (DME, a measure in nautical miles). Soon after, the crew received two enhanced ground proximity warning system (EGPWS) cautions. The crew then levelled the aircraft and conducted a visual approach and landing on runway 16.
The investigation is continuing."

Full report here...

http://www.atsb.gov.au/publications/investigation_reports/2007/AAIR/pdf/aair200706727_prelim.PDF

[Goldfish Jack]

I was led to believe that modern a/c like the 777 and some of the newer generation airbus's do not have ADFs on them....

[HIGH n MIGHTY]

all Boeings do I think but the new gen Airbusses (Airbi.??) dont. Caused a few dramas the few months the ILS was out Im told...

[ratarsedagain]

all the latest Airbus i flew, had a single ADF on them, so still able to fly ndb approaches.

[mustafagander]

Not to MEL 16!! It requires 2 ADF.

[faheel]

nope it does not .
in the old days it was a twin locator approach, not anymore.

[hawker750]

Why do airlines/manufacturere penny pinch on the relatively small extra cost of a second ADF. There are may instances of incidents/accidents caused by the lack of a second one. The USAF 737 fatal at Dubrovnik being just one. Trying a twin locator (ADF) approach with only one is iffy to say the least.
Or am I just behind the times?

[HIGH n MIGHTY]

The Jepp chart now states DME and both NDBs required even though its now an NDB rather than a Twin Locator.... Now, whether that means both need to be monitored simultaneously or not is another question..

[kotakota]

No go-around ?

[Wiley]

There wouldn't be any NDBs in Australia today (and for quite some years now), saving tens of millions of dollars a year in maintenance and upkeep, if some nameless person in Canberra hadn't screwed up royally when writing the GPS specs of the J model Herc.

But that's another story...

Non precision approaches in big aeroplanes are seen to be so incredibly dangerous by some airline managements that they forbid their aircrews from practising them on the line. (Not just NDBs, but VORs and LLZRs as well.) Total exposure to NPAs is once every six months in the sim - unless a line pilot has to do one for real on the line because a precision approach is not available.

I don't know if Thai follow this... shall we say rather remarkable practice. However, when aircrew who aren't allowed to practise have to do an NPA in anger, their management wonders why they sometimes get it wrong.

[BottyTotty]

Join Ryanair, and you will likely shoot a NPA every or every other day. All flown in VNAV now though so no massive rates of descent in V/S, which is where the danger lies.

They are not dangerous if you are well trained, diligent and practise regularly.

[focault]

Hello everyone, first time post!
NPA are normal ops on charter flights, I fly either A330 and A321 both equipped with two ADF extremely usefull for landing in places like Boavista, Capoverde or Banjul, Ghambia. If you are trained its fun tough! :}

[Capn Bloggs]

The crew reported that the approach briefing included an intention to conduct a constant angle approach path using the vertical navigation (VNAV) mode of the automatic flight control system

If they did indeed use VNAV and it flew them low there'd better be some flack-jackets issued for the database boys in Jepp...

On that point (and never trusting Bill Gates oops I mean computers in general) it is interesting that two profile points, 8DME and 7DME, are not printed on the approach chart. How on earth is a crew supposed to easily monitor the VNAV (or do it themselves) if the chart hasn't got the complete profile on it? The two that are not there cover the height the aircraft got to when the EGPWS went off. "Thanks for nothing, system." Really nice pic of the runway though...

[ACMS]

Firstly: most current 777's have 2 ADF's fitted. However the next bunch of 300ER's at CX will NOT come with ADF's

Second: I'm pretty sure the 16 Twin Loc app in MEL is not in the Honeywell FMC database. You can build the legs if you want to and use L Nav ( monitoring the raw data for tracking ) BUT you cannot use Vnav for the arrival. Only Data base approaches can use V Nav for guidance. ( ILS's VOR/DME GPS etc all ok, there are no NDB approaches at all in a Boeing FMC )

[blueloo]

...but in this instance you can select the ILS - and use it as an overlay. LNAV VNAV should then work, with appropriate raw data and profile height checks.

I dont see why the honeywell FMC cant be as flexible as the smiths FMC.

Smiths you can just build rwy extension and waypoint on waypoint, plug some altitudes in and watch LNAV VNAV fly it for you.

[PositiveRate876]

Smiths you can just build rwy extension and waypoint on waypoint, plug some altitudes in and watch LNAV VNAV fly it for you.

But would you really fly an approach that you've "built" down to minimums and be the first one to see if it works? Or doesn't. :ugh:

[Capn Bloggs]

in this instance you can select the ILS - and use it as an overlay. LNAV VNAV should then work, with appropriate raw data and profile height checks.
I dunno about the legality of that, Blueloo! VNAV below the MSA using an different approach to that being flown? Duck heads, over.

[AA717driver]

Does the 777 have the "Green Arc" showing descent point like the 757/767? I use that in VS a lot. (Yes, I know I'm a danger to myself and humanity in general for using vert speed. I fully expect the police to be at my door anytime now...:rolleyes: )

Descending below the min alt. for a given segment on a NPA can be done by anyone not paying attention. Certain individuals go below G/S without a peep from the other seat, too. TC

[woodpecker]

Even if, as they were, joining the V-Nav (FMC calculated glidepath) from above the initial MCP Alt window should have had 4000' until 12D when 3000' would have been entered until 9D and so on.

The FMC should be an aid, with the MCP Alt window still protecting the the crew from the "embarrassment" of FMC mistakes..

[blueloo]

"But would you really fly an approach that you've "built" down to minimums and be the first one to see if it works? Or doesn't. "

Most probably not would be my answer at this stage.

- my thoughts - if no non-precision approach - or overlay as appropriate is in the database why could you not build an LNAV track to replicate the approach. Your choices - a blank map and raw data - in a big boeing using HDG SEL and VS. Or build an LNAV track - let it fly LNAV monitoring raw data - you now have a display for situational awareness, and you shouldnt have to fiddle with the HDG knob every few seconds when the wind changes, or you havent got the HDG/drift quite right....



Capn Bloggs:
Remember you still must monitor raw data. The approach must be flown ie Alt constraints etc as per the NDB chart - (or company requirements) - all VNAV and LNAV are doing is reducing the workload, to allow you to monitor and manage the approach - an ILS overlay which has the same exact lateral tracking should be fine (the melbourne ILS overflys BOL and ROC) - and as ILSs tend to fly 3 degree approaches then VNAV should be appropriate.


Woodpecker:
"Passed the Cb, we were cleared present position direct to the VOR"

I must have missed something in the cleareance - or was it "cleared to the VOR to resume the STAR"?

I am not familiar with the area you are talking about - the terrain ATC or otherwise.

If you are cleared direct - surely you are no longer on the STAR. Which presumably means the Atl constraints are no longer valid. So you need minimum safe altitudes from somewhere.....the original STAR constraints may not protect you if you are off track - they may in fact be higher.

[woodpecker]

Bluetoo,

The basic point is, under V-Nav and L-Nav, should you rely on waypoint altitude protection or should the MCP Alt window reflect the minimum altitude that you can descend at you present position?

Forget the Daros 1A, have a look at the Melbourne plate (URL link), should the MCP Alt have been set to the locator height or to 4000 (until 12D)?

Which is the safe option, and which is SOP?

Airmanship suggests 4000' is safe, SOP's (in my airline) suggest the locator height.

[blueloo]

Must be some confusion here - my apologies - thought you were talking about the STAR and direct tracking on the STAR- not the Approach itself. Approach constraint would be different entirely, and I agree the in the Melbourne case 4000' is the minimum until BOL ( I am using the jepp chart).

[Capn Bloggs]

Our SOP, once "established", is to set the MDA. Then allow either the VNAV or use VS/FPA to follow the briefed or charted profile down. We don't wind in each limiting step because it increases workload to the point of distracting the crew, especially where there are many steps.

Re setting limiting steps on a STAR, if you did do this, then you couldn't set your actual level cleared-to by ATC. This would lead to the possibility of missing your assigned (as opposed to STAR-limit) level.

[411A]

Our SOP, once "established", is to set the MDA. Then allow either the VNAV or use VS/FPA to follow the briefed or charted profile down. We don't wind in each limiting step because it increases workload to the point of distracting the crew, especially where there are many steps.



Indeed so, Capt.
Many new crew use the automation so much they can't see the forest for the trees and, if crew can't keep a mental picture of what they are doing, they don't belong in the pointy end in the first place.
Are NDB approaches 'interesting', in some circumstances, especially with a large heavy jet?
Of course they are, but that is precisely why we are paid the big bucks.
Can't stand the heat, stay out of the kitchen.:rolleyes:

[xsbank]

Isn't this a non-event?

"The crew then levelled the aircraft and conducted a visual approach and landing on runway 16."

[aulglarse]

Xsbank, have you had a look at the profile on the full report?

One thing notorious for RWY16 is the undulating terrain associated with steep gullies around the 6-4nm final. This may set off an alert if not in the appropriate config. A few years ago I have had a terrain alert on a clear day with a slighly higher descent than normal with flaps 1 (airbus) selected.

[morbos]

Sure, the outcome was a non-event but history is rife with RT misheard followed by CFIT. Here is one example:

http://aviation-safety.net/database/record.php?id=19890219-0

In this ymml case, overhearing the cloud level elsewhere was 1700 when in fact it was 1000 @ymml. Presumably @1700 the crew must have wondered when they would break out.

I think the EGPWS was a tidge late wrt the initial deviation from the minimums but a timeline would help. Certainly the corrective action upon the second warning was welcome.

I would count this as an EGPWS 'save'. A valuable piece of equipment to say the least.

[Kapitanleutnant]

The 737 NG's (7,8 and 900's) do NOT have any NDB's on them, yet with the FMS's installed, we still shoot the NDB approaches (in the sim, never done one real time yet)

I think NDB approaches are in the emergency section, aren't they? :)

K

[permFO]

This incident has nothing to do with a vnav approach. The crew flew over the ndb, set the minima and then pushed LVL CHG or, for the Airbus minded, the equivalent of selecting open descent. A basic error but not a non-event. Certainly airline procedures will need to be addressed as the crew response to an EGPWS warning at night would probably not be considered ideal.

[Capn Bloggs]

This incident has nothing to do with a vnav approach.
Oh well, if they didn't do what they briefed they were going to do...

[Wing Root]

A jet decending below MDA is a non-event?

They didn't just get below profile they busted an MDA step. I also gather they were in cloud. It's an NDB approach, but it's a straight in runway aligned one with no reversal procedure. Coupled to LNAV it's basically a localiser approch from an altitude management point of view.
Also, why not use a FPA descent? at the top of the profile (11.5DME) set -3.0 degrees dialing in check heights along the way and let Mr. Boeing do the rest.

[RYR-738-JOCKEY]

They didn't start descending at the appropriate point, the FO hit Level Change to get down, and it certainly looks like they had briefed for a VNAV approach because of the 50' addition to MDA. Anyway, at 6,8 D they were a 1000' low on profile triggering GPWS. NDB or not....totally irrelevant, the point is the crew started descending in a non-standard mode and did nothing when passing through the VNAV path/CDA profile.
On a side-note, the crew responded correctly to the GPWS. Levelling and continuing visually is accceptable on GPWS caution.

[Lord Flashhart]

XSBANK - NOT a not event. The outcome was good (ie no CFIT), but they the aircraft was not going where it was supposed to go. (Lucky they were in VMC)

[permFO]

"On a side-note, the crew responded correctly to the GPWS. Levelling and continuing visually is accceptable on GPWS caution. "

I think this is acceptable in day VMC but this occurred at night time. Also there is a high probability that the crew's SA was not good. The only reason to ignore an EGPWS warning is if you are aware of the terrain and you can see it. I don't think that was the case in this incident.

[datkat]

my goodness still doin an ndb app. my question is why do we not get rid of the whole ndb app. i mean i can see if its used as an aid to other apps but to have an entire app solely based on ndbs?man its out of date and out of taste

[amos2]

How can anyone cock up a twin locator or ndb approach?

Especially into 16 at Mel which is as simple as they come!

Strewth!...didn't we all learn this in a link trainer when we were kids?

Gotta wonder about the standards here, guys!! :sad::sad:

[FlexibleResponse]

How utterly disgraceful!

How utterly third world!

How utterly primitive!

Why in the name of the heavens is any international airline aircraft required to carry out an NDB approach into any civilised International Airport?

For God's sake, aren't the passengers taxed enough on their tickets and the airlines on landing charges to expect 21st friggin' century technology approaches at destination?

Even your piece of crap shopping trolley car has a reliable GPS these days...

Shame Australia! Shame and again Shame!

I am embarrassed to be Australian. Does anybody remember that Australia was involved with the invention of the Microwave Landing System for example?

I've seen much better service as most of the poor third-world destinations in Asia.

I say shoot the bastards collecting the money at Melbourne Airport for their miserable failure and incompetence in maintaining a safe aviation facility at a minimum acceptable standard that might be expected by a reasonable man for the operation of International RPT!

[blueloo]

There is no way this particular twin locator will be removed. Qantas Longhaul is obsessed with twin locator approaches and no doubt will pay for this one to remain in service.

Oh wait a minute - I hear the quote - "its not qantas - its the casa matrix!"

(Which is in laymans terms "I am to lazy to change it - or spend money and install new technology and equipment into our ageing fleets")

[7times7]

BIG difference between a CAUTION and a WARNING. ;)

[Capn Bloggs]

Flex, I gotta agree with you there...

[Shot Nancy]

Hear, hear Flex.

[Wiley]

Why in the name of the heavens is any international airline aircraft required to carry out an NDB approach into any civilised International Airport?See post 10 on page 1.

[permFO]

I hope the ATSB report will address why the runway 09 VOR/DME or RNAV approach was not nominated by ATC. They seem to be very reluctant to use that runway even when a tailwind is present on 27. The approach to 09 is over paddocks and farms so surely noise abatement is not the issue. Also how difficult is it to design an RNAV approach onto runway 16?

[pattern_is_full]

"...my question is why do we not get rid of the whole ndb app. i mean i can see if its used as an aid to other apps but to have an entire app solely based on ndbs?"

"Why in the name of the heavens is any international airline aircraft required to carry out an NDB approach into any civilised International Airport?"

From various points in the report: "In Sept 2007 work commenced on the MEL RY 16 ILS....A NOTAM was issued 4 Oct...advised that the ILS would be unavailable from 8 October to 22 Nov....The crew had received this NOTAM prior to departure from Bangkok...the controller transmitted to the crew 'Descend to 4,000 cleared to runway 16 NDB approach."

The ILS was broke. The wind was 170 @ 25G35

Either fly the NDB or try and put a trip-7 down in 2300 meters on RWY 9-27 (ILS or VOR) with X-wind 25G35, or a VOR to 34 with tailwind 25G35.

[GXER]

Why in the name of the heavens is any international airline aircraft required to carry out an NDB approach into any civilised International Airport?

I'm not a pilot so I may have this completely wrong - but, it seems to me (having read the report), that the question "why was the crew required to (or chose to) execute a NDB approach" spectacularly misses the real point.

Surely the real issue is "why did the crew fail to execute the NDB approach safely/effectively)" (whether because of error or inadequate training or practice).

[Capn Bloggs]

The ILS was broke. The wind was 170 @ 25G35

I can only reiterate what Flex said.

Sure we all know NDBs "should" be capable of being flown. But the fact is they are very rarely flown by these types of crews. MEL is a MAJOR international airport. It is a scandal that if the ILS goes down an international crew has to do a raw-data NDB approach.

Why on earth is there no RNAV approach for 16?

Why on earth isn't the NDB 16 in the FMS database?

Either of these options would have prevented this near-disaster.

[Capt Fathom]

Why on earth is there no RNAV approach for 16?

How many pilots are qualified for RNAV Approaches? A lot of GA pilots are, but most of my airline acquaintances are not!

Figure that one out!

[Capn Bloggs]

How many pilots are qualified for RNAV Approaches? A lot of GA pilots are, but most of my airline acquaintances are not!
I have heard Virgin Blue, QF and Emirates pilots doing them, so I assume that there are quite a few.

But even discounting the international drivers who may not be qualified to do them, an GPS NPA RNAV approach in Aus costed $7000 a couple of years ago. Surely the safety benefit to pilots who are qualified to do them (in preference to the "twin" NDB) would outweigh that sort of outlay?

[Bullethead]

G'day Gents,

The RW16 ILS was out of service, and NOTAMed, because it was being replaced with a CatII/III facility, so it wasn't just broken.

The RW16 approach is a perfect overlay for the RW16 NDB approach both laterally and vertically. The DME/Altitude scales are the same.

The crew either forgot to engage VNAV or deliberately didn't for some reason and then failed to properly monitor the approach in any case.

I flew the RW16 NDB approach, in a B767, several times during the ILS outage and it worked just fine in LNAV/VNAV.

This sort of thing doesn't only happen in MEL, I was on approach to FRA years ago and had a RW25L VOR approach sprung on me at the last minute, in IMC, due to an ILS failure and had no problem. So, non-precision approaches to major airports do happen from time to time and sometimes due to the prevailing weather conditions are unavoidable.

So the relevant question is not WTF were the MEL airport authorities doing with the ILS, i.e. building a better one, but WTF were the crew doing flying a poorly executed approach in IMC?

Regards,
BH.

[Capn Bloggs]

For those of you (QF?) who seem to think flying, in LNAV/VNAV, the ILS from the database when in fact you are flying an NDB approach, I am sure you would find CASA would have a fit if it knew you were doing it. Is it in your company's SOP? Doing such a thing, without a significant amount of local knowledge and extensive analysis of both approaches, is fraught with danger.

[Bullethead]

G'day Capn Bloggs,

It is the way we do it and it is CASA approved and what is the problem if you are monitoring the raw data, which is of course a requirement? In fact it's a CASA requirement, as part of our sim programme, to carry out a non-precision approach in exactly this manner.

You could also fly the approach in Hdg Select and V/S, but that is a higher workload than using LNAV/VNAV and you always have the option of changing modes if the map is inaccurate or going around if it really gets pear shaped.

Regards,
BH.

[Pilot Pete]

You can build a suitable LNAV track or use an overlay approach that is the same. As pointed out earlier, you can't construct a VNAV profile, it MUST be supplied in the FMC database and you must be trained and authorised to do VNAV approaches.

We (can) use LNAV for non-precision approaches, with raw data monitoring, which means you can concentrate on getting the descent profile accurate without having to keep adjusting the heading using HDG SEL. We use VS for the descent path, having calculated and cross checked suitable step downs for each DME mile, interpolating between supplied hard altitudes on the chart, if required.

So I am not sure I agree with Flex and his rant about 'big' airlines having to fly non-precision approaches. They are a fact of life in many parts of the world and you can fly them safely in a big aeroplane if you are trained, practised and you fly a constant descent from the right point to the right minima. They are more challenging than an ILS and the risk factor is higher, but for once I agree with 411A, that's what you get paid to do, so learn how to do it.

RYR-738-Jockeyit certainly looks like they had briefed for a VNAV approach because of the 50' addition to MDA. Not sure why you would come to that conclusion. Many airlines are authorised by their authority to add 50' to published MDA on a non-precision approach in order to fly it as a CDA and not to bust MDA in event of a go around. Indeed this reason is mentioned in the report as to why they did this. Nothing to do with VNAV being used.


From the report
The database did not contain the runway 16 NDB approach. The copilot reported that he entered into the FMCs the final approach fix waypoint at 6 DME with an altitude constraint of 2,100 ft above mean sea level (AMSL). Other segment minimum safe altitude constraints associated with the NDB approach were not entered into the FMCs.

So unless they can enter VNAV approaches in manually, they were not following correct procedure.

The F/O put 2,100ft in at 6D, whereas the procedure crossing altitude at that point should have been 2,250ft according to the chart (LIDO, which I am not overly familiar with).

They started descent from 4,000ft at 9.9dme instead of the published 11.5dme, that is 1.6nm later than published, so probably somewhere in the region of 170ft high on profile at that point. They were not fully configured for landing at that point either.

They descended at an average of 1500fpm, and recorded 1808fpm at 8.5dme. When they descended out of cloud all three of them were 'looking out' and the PF was not concentrating on 'looking in'. They got to 513ft above terrain level with 2 EGPWS cautions. The F/O then disconnected and climbed, then flew level before following the PAPIs, stating they were fully stabilised by 500' ARTE.


So to speculate;

I would say

1. They were not adequately briefed for this non-precision approach.

2. I suspect they did not follow correct procedures and constructed an (incorrect) VNAV descent path for the final approach. Nobody cross checked the FMC entries adequately.

3. The increased workload lead to a rushed approach. The aircraft was not configured for landing before starting the final descent from 4000' and the descent point was missed.

4. Neither (or infact none of the three) pilots recognised the high workload and dealt with the overload that the F/O found himself in (when he missed the descent point and was unsure of their clearance to fly the approach) and when he conducted the final approach contrary to the briefing he had given (he used LVL CHG and not VNAV).

5. An high rate of descent after missing the descent point further increased the workload and lead to less than optimum aircraft handling and ultimately a descent well below procedure safe altitude when the EGPWS cautions occured.

6. Situational awareness had broken down and there seemed to have been little active monitoring and assertive communication from the captain, who should have had the spare capacity and could have just said 'Go around' and corrected the situation.

These are just my thoughts though.....

PP

[stilton]

We still have NDB'S on our 757's but none on the 76's.

As was mentioned there are no NDB approaches in the database, cleared for one, however we are certified to use an overlay of, for example an ILS approach (no glideslope) to the same runway and LNAV / VNAV down to, either a 'derived decision altitude' of MDA + 50 feet or, as noted on the chart, in some cases MDA itself as your DA or decision altitude.

There is no better method to fly a NPA, stabilized, almost as good as an ILS.

[737only]

I see it once in a while beeing an instructor in the sim, crews start the descent late, are using an idle descent mode (level change, as the thai guys did) to chase the path, and do not have any gate to stop the massive rate of descent. Furthermore they are not using the vertical profile displays to avoid coming low. Just a nice example of THREE pilots who lost their situational awareness completely.....
Non prec approaches are nowadays as easy as ILSes! It just needs a little bit of preparation and maximum use of the modern instrumentation we have nowadays...:ok:

[Dan Winterland]

As QANTAS parked one on the golf course at Don Muang a few years back, perhaps Thai thought they'd return the favour! :eek:



ADFs are usually an option. My company's Airbus all have twin ADF fits.

And it's no use these days saying that a failure to fly an accurate NPA is down to pilot proficiency and we all managed it safely in the old days. Things have progressed and safety is more of an issue, hence the tendancy for ILSs. My company is allowed to do fully managed NPAs providing we have GPS primary. It takes away all the risk, the approach is not much different to an ILS - and it is very safe.

[Capt Fathom]

Capn BloggsFor those of you (QF?) who seem to think flying, in LNAV/VNAV, the ILS from the database when in fact you are flying an NDB approach, I am sure you would find CASA would have a fit if it knew you were doing it

BulletheadIt is the way we do it and it is CASA approved

Oh dear Capn Bloggs := Do keep up!

[Capn Bloggs]

737Only,
Furthermore they are not using the vertical profile displays to avoid coming low
What vertical profile display?

Fathom,
Oh dear Capn Bloggs Do keep up!
Keep up with what? I don't know of any rule that says I can load an ILS into the FMS and then fly a NDB just because I looks like it's the same approach...

[Pilot Pete]

Captn Bloggs

you can construct a lateral profile to follow using LNAV IF you monitor raw data. So, if you drew an inbound course to the NDB as promulgated on your approach chart this would be permitted (as long as it complies with company SOPs). If that inbound course to the NDB was exactly the same as the ILS inbound course what would the difference be if you just selected the ILS in the FMC and followed the lateral portion using LNAV? That is what posters are saying. Remember, the lateral navigation is being done on the raw data display, you are merely using the LNAV capability to keep the raw data tracking correct. If it shows an error, then you MUST follow the raw data info and use another mode, HDG SEL or G/A.

The VNAV portion of that approach is what stands out to me as being inappropriate. As I stated earlier, in my company you are NOT allowed to construct the VNAV profile, it MUST be in the FMC database and thus the profile is generated when you select the approach in question. Perhaps the incident demonstrates why my company do not allow the vertical profile to be constructed and flown in VNAV (the input error that was not noticed by any of the pilots) and then they never used it anyway, selecting a high rate of descent in LVL CHG due to being above profile and unstable!

I think it all comes down to the approach not being planned well enough with a clear enough picture of how it was going to be flown.

As an aside, we've all heard an approach brief where a pilot rattles through the detail on the plate, but neglects to say HOW he is going to fly the approach, what modes will be used, when he will configure the a/c and what cross checks he will be requiring from the NHP to ensure compliance with the step downs/ profile. This is more critical when flying a non-precision approach and 'getting down and dirty' at the platform and being fully configured to start the final descent at the correct point is the safest way to conduct this sort of approach IMHO. I have seen many trying to fly an NDB approach like an ILS, keeping the speed inappropriately high and leaving the config until late and invariably the approach is untidy as the workload goes up exponentially and you don't have the luxury of LOC/GS to reduce the workload to free up grey cells.....the bit that usually goes wrong? The descent profile. With some 'pink string' or a beam bar display the lateral portion is visually easier to monitor. The vertical profile needs concentration and use of some brain power to maintain a stabilised path as you need to constantly keep checking crossing altitudes and speeds. This is where VNAV approaches can help to free up capacity, in effect making the approach more like an ILS, but you need to make sure that the VNAV profile is correct before you start as there is no raw data backup, just you and your colleague....

PP

[GlueBall]

hawker750 . . . if you re-read the report carefully you would have known that the airplane was equipped with dual ADFs, and according to the report the airplane was delivered new from the factory just the previous month. So it's highly likely that both ADFs were in working order.

permFO . . .the event occured in daylight hours, shortly after 1 o'clock PM.

[Phantom Driver]

Pilot Pete.

I must say, excellent analysis of what should be SOP on any non precision approach; i.e be fully configured with all checks completed latest 2 miles before start of the chart "glidepath", and most importantly, raw data monitoring of lateral and vertical profile by the PM. The profile is invariably printed on the chart. How difficult is it for PM to sing out the required ht against DME range and for the PF to adjust accordingly? The ND "profile guide" is a good crosscheck, if the coded approach has not been modified (again another SOP). Some glides are steepish, others (e.g SHJ) are shallow,but all close enough to a standard 3 degree glide; any such anomalies well covered in company airfield brief.

However, having said all this, it is often easier said than done, especially with regard to the "fully configured" bit, when ATC vectors you in tight, there is a tailwind, or finger trouble with MCP selections, all resulting in not being at the correct platform ht for VNAV PTH descent (talking Boeing 744 here); mode goes to VNAV ALT(thats if if you remembered to reselect VNAV from whatever you were using before- VS or FLCH) and then you're scrambling. I like to practice VNAV PTH descents whenever possible (nice clear days!), as we don't get to do too many non precision approaches in our operation, but I have to admit that in an IMC environment,(and sim checks!) I will take LNAV (raw data monitored) and VS (raw DME/chart ht back up) any day. Too many gremlins lurking in VNAV. As for FLCH on finals, don't think so somehow!

[Pilot Pete]

Phantom

Nice comments and some real practical advice about V/S over VNAV. I think V/S is better myself, as you are much more in control and 'in the loop', but you need to get configured and all checks out the way as you say, to concentrate on the profile.

When I fly an non-precision approach I write out check altitudes for every DME mile on a bit of paper (someone mentioned that a couple were omitted on this approach chart, which indeed they are), especially when all I am offered is a platform and a marker crossing altitude. I get the other pilot to check these in the cruise to look for errors. I then brief the PM that I want a call at every DME mile plus half (i.e, at 8.5d, 7.5d, 6.5d) with the NEXT crossing altitude, i.e. passing 8.5d - "8d - 2890ft", then as I pass 8d a call such as "50ft high".

The other benefit of getting fully configured before descent is you can cross check groundspeed and aim for an appropriate V/S (shown in the box to the right of the vertical profile on this LIDO chart.) If you aren't fully configured and are going to try to slow down and go down the final approach what V/S are you picking at the descent point? Invariably things are starting to get a little rushed, a V/S is selected (usually too much) and then with config and speed reduction the attention is lost from the V/S and the approach is already getting unstable with a high rate of descent (100fpm+) which is not checked until (usually at least) 2nm further on, when it is realised that you are going low.:rolleyes:

However, regarding it is often easier said than done, especially with regard to the "fully configured" bit, when ATC vectors you in tight, there is a tailwind, or finger trouble with MCP selections, all resulting in not being at the correct platform ht Well, I would argue that is the airmanship element. Brief it FULLY and PROPERLY, get ahead of the game and stay ahead. Think about the tailwind and slow earlier. Tell ATC what you require, don't necessarily just accept being turned in too early. NEVER forget about that fantastic speed reducer that is the landing gear. So many pilots seem reluctant to use gear really early when it will regain the upper hand easily. Forget slick, min fuel burn approaches when doing non-precision approaches; get down and get dirty!!! How many pilots consider taking the landing gear on a radar vectored base leg to a non-precision final approach when you can see they are only just going to make the platform altitude at the descent point? They only have a couple of stages of flap and need to configure fully and slow down still, but many try to fly it just like it is an ILS. Never ceases to amaze me. ATC invariably vector you tight as they think that is efficient and what you want (which it is for an ILS!!!) Brief this and then it is no surprise, drop the gear and stay ahead of the game!

I have flown 75/76 and now fly both genres of 73 in charter ops, so probably get much more exposure to non-precision approaches than the average 777/747 pilot, but that says to me that longhaul pilots need to give it even more thought and get slowed down and configured EARLY!!:ok:

PP

[woodpecker]

As for FLCH on finals, don't think so somehow!

Sums it all up for me as well! V/S, FPA but not FLCH.

[Phantom Driver]

PP
When I fly an non-precision approach I write out check altitudes for every DME mile on a bit of paper (someone mentioned that a couple were omitted on this approach chart, which indeed they are), especially when all I am offered is a platform and a marker crossing altitude. I get the other pilot to check these in the cruise to look for errors.

Agree entirely; into critical places like KTM, used to write the numbers on my crib in extra large font, especially at night, eyes/glasses not being what they used to be. I wonder how many people know/remember what happened to PIA A 310 landing there some years back; one step adrift on the profile calls and flew into the hill on finals.

Re: Tailwinds on finals; love to brief it in the cruise, and usually do as a "contingency", but to quote erstwhile defence secretary Rumsfeld, it is one of those "unknowables" that are thrown at you, usually as you are turned from base onto finals just outside the outer marker. AMS are just as guilty as the guys at MAA! But can we blame them? Surface winds reported as light and variable can be a moveable feast when compared to winds at platform height.

[Pilot Pete]

Some more interesting reading. Watch out with those NPAs! Brief, check, and double check!

http://www.aaib.dft.gov.uk/sites/aaib/publications/formal_reports/7_2007_f_ojhi.cfm

http://www.aaib.dft.gov.uk/sites/aaib/publications/formal_reports/5_2007_g_medg.cfm

PP

[Capn Bloggs]

Very interesting links, Pilot Pete, and should be required reading.

From Birmingham,
Safety Recommendation 2007-111: It is recommended that Mahan Air should expand its FMS database to include all approaches relevant to their route structure.
I hope the ATSB (or whoever investigates this) comes to the same conclusion with Thai.

From Khartoum,
The investigation was unable to identify any formal CAA policy, in place prior to the incident, regarding the implementation of MNPA operations. Instead, agreement was reached between individual operators and their respective CAA Flight Operations Inspector on how they might implement such operations.
That was the point I was making before. I'm as good as anyone at "build-your-own-approaches", but are they legal, even if you are only ceating and using a lateral path and monitoring it with raw data?

I also noted the old "MDA on the ALT SEL" chestnut; at Khartoum the MDA didn't appear to be set on the ALT SEL as the aircraft went straight thru it, with the autopilot still engaged in V/S?

[Pollution IV]

Some great advice seen here re sound flying procedures for NPAs.

The self built vs database appch consids are important, as Airbus at least does not endorse any changes to a database appch after the FAF, thus to fly a managed self- built rwy appch would be a definite no-no.

The idea of using an ILS overlay to fly an NPA is certainly creative, but I doubt if the acft or FMS manufacturer would support such an activity. Those who trust their airline SOPs to support them in the event of an incident/accident during such an appch may find themselves out in the cold. Is it actually written in balck and smudge somewhere? I would certainly doubt that CASA would go out on a limb and endorse such a procedure. Perhaps a creative Airline Trng manager is writing cheques that his position can't actually cash.

In any case a database appch will not guarantee a correctly flown managed appch as per the plate, as they are often riddled with mistakes, do not align with the raw data all that closely and are sometimes unable to fly steep vert profiles with changing descent gradients and thus undercut the steps on final. Additionally, you may suffer a 'map shift' or loss of GPS primary & thus sufficient nav accuracy, thus the green/magenta line is no longer accurate. It is human nature to follow the neat fltpln Nav Display line, so very few pilots will manually select lat navigation in such cases, even with the raw data out of tolerance (automation complacency). Those who don't believe me simply haven't seen enough examples in the sim. All the above instances are commonplace in the area of the world where I commit aviation. I am yet to see a managed database appch here that flies with reasonable accuracy according to the IAP plate and raw data upon which it is predicated. Infact, due to the above safety consids, many of the pilots in my company will only fly an NPA in selected modes vertically and laterally. I think more emphasis needs to be placed on these types of simple real world problems during airline sim training.

The bottom line is, some NPAs are too tricky (read poorly designed) to be coded correctly thus the automation won't handle it effectively, that's why we pilots still have jobs! When PF on an NPA, I would say knowing your vertical profile is paramount and don't expect the other guy to feed you all the info as his mind might be elsewhere, as seen in the Thai incident.

I suppose the real bottom line is for regulators to hurry up and approve GPS NPAs at all airports.:ok:

[Capt Fathom]

The idea of using an ILS overlay to fly an NPA is certainly creative
I would certainly doubt that CASA would go out on a limb and endorse such a procedure

It would appear that CASA have approved exactly that!

Using the aircraft's GPS capability to fly a NDB approach. How novel!

If only the USAF Boeing 737-200 at Dubrovnik (http://www.pprune.org/forums/showthread.php?t=302409) had a GPS/FMS or a 2nd ADF!

[alf5071h]

This incident has many similarities with those described in the link below.
For those who believe that this was a minor incident, then consider if your biased perception might lead to a more hazardous position – “I know better, it’s not a real warning, no need to pull up etc, etc”. Of course you could share with us the secret of not suffering error in this type of operation.

Use of VNAV etc would overcome many of the problems (and open new opportunities for error), but this is not always an available solution – timescale and in some countries cost or ability to produce the procedure.
The crew can use defences; the use of the basic crosschecks, an altitude-range table, the required vertical speed, timing. All of these are on the chart – were they briefed, were they used? An NPA requires a good plan, a good briefing, and then an error tolerant operation involving crosschecking / monitoring. … are Captains good monitors?

The NPA and chart have opportunities for error. The use of overlapping range scales – the hazard of the DME displaced from the threshold. If GPS is used which range datum should be used? Will the threshold be visible at MDA in minimum visibility? Thus in addition to error in operating modern technology, crews are exposed to further opportunities for error during the procedure.

One critical issue is that MDA is not a safe altitude; it’s only a minimum altitude during the latter part of a correctly executed procedure; this strengthens the necessity to pull up even with an amber alert.
Although an operator’s procedure for an EGPWS ‘amber’ alert might only require ‘an adjustment to the flight path’ (certification terms to differentiate from a red pull up warning) the adjustment can and should be a climb. EGPWS alerts a situation where there an error has occurred (crew, ATC, procedure, etc), which has led to a reduced safety margin; the crew are unlikely to know why this has occurred, so climb and start again. Being visual with the runway does not guarantee safety particularly if you don’t start for a safe position – note the hazards of poorly defined obstacles or an illusion.

Ref EPGWS events and analysis. ( www.icao.int/fsix/_Library%5CTAWS%20Saves%20plus%20add.pdf)

[BOAC]

alf - I have not been into this chart/approach in detail, but I have to query some of your comments:
One critical issue is that MDA is not a safe altitude; it’s only a minimum altitude during the latter part of a correctly executed procedure; - surely once established on the i/b track, and past FAF, you can (should you wish - not my way!) SAFELY 'dive and drive' down to MDA in the absence of stepdown limits? Would this not have been a technically 'safe' option if correctly flown here? Terrain does not look like a serious issue. Indeed, our friend 411A would probably have been steaming in at MDA for miles:)the adjustment can and should be a climb. - 'can', indeed, but need not be 'should', surely? Not if it is simply high terrain closure rate? I would hazard a guess that was the 'warning' here, and I guess they half expected it too!

NB I do not defend the crew who appear to have grossly mishandled this particular approach

[betterave]

The ATSB report was an interesting read...

I can't speak for the B777, but at our company the use of VNAV on a CDAP approach is not allowed on the B763/752. If the non-precision approach is not in the FMC, we can overlay a precision approach (such as ILS for NDB if the tracks coincide with each other) to fly the approach in LAV. One pilot must, however, monitor the appropriate raw data from no later than the FAF to the MAP , and step-down fixes in the final approach segment must be verified with raw data.

Our procedure is to be in landing configuration by 2 NM prior to FAF; DA is MDA + 50'; after ALT CAP at FAF altitude, set TDZE in the MCP (rounded up to the next even 100'), at .2 NM to the FAF, selected the pre-briefed descent rate with V/S (we are allowed to bracket the pre-briefed descent rate by +/- 300 fpm, but a rate greater than 1200 fpm is not allowed) and use all available info to monitor the descent rate. Our techniques for monitoring involves: 1) the green arc should be approximately on the runway threshold (once TDZE is set in the MCP); 2) VDI on the HSI showing on profile; and 3) monitor Vertical Track Error on PROG page 2/2.

Admittedly NDB approaches are being removed from most airports in the US so the chance of flying one domestically is almost nil, but we still practice it during PC/PT just in case we encounter one internationally. We do fly other non-precision approaches occasionally (SAN LOC 27 comes into mind) and we usually brief it to death during cruise just because we don't fly one very often. Let's hope that we can learn from this incident and avoid a potential CFIT event in the future.

BTW, if I did the math correctly, the relief pilot started his B777 course with 594 hrs total flight experience! During the last hiring spree by our company, there were probationary pilots getting B744 and B777 F/O bids but even then our new-hires would have thousands of hours of flight time with at least 1000 hrs PIC. Now that our company is hiring again, I understand the new-hires are being restricted to B737 and A320 F/O assignments (even when the B763/752 fleet is very short of pilots) because management does not want "inexperienced" pilots flying internationally. I'm not saying that experienced pilots do not make mistakes (and of course there have been lots of screw ups by experienced crews in our company), but perhaps more flight time as actual pilot flying (as opposed to a relief pilot not touching the controls during T/O and landing) before being assigned as a large jet F/O would reduce the number of incidents/accidents.

Cheers.

[BuzzBox]

Irrespective of whether or not it's legal/sensible to use an ILS overlay to fly a non-precision approach, it seems to me the problem here was an inappropriate choice of autopilot vertical mode to carry out the approach.

Having missed the descent point for the NDB approach, the FO selected FLCH to commence the descent. Having done so, the only terrain protection available was the 1,190 ft (MDA+50ft) set on the MCP. If they'd stayed in VNAV and simply pressed the Altitude Selector button on the MCP, the aircraft would have commenced the descent in VNAV SPD, captured VNAV PTH when the correct descent profile was eventually intercepted, and respected any intermediate altitude constraints entered in the FMC.

VNAV is a great tool, but there are plenty of traps.

[alf5071h]

BOAC, in principle, (as I understand PANS OPS) the terrain / obstacle clearances for NPAs are divided into three segments; IAF - IF, IF – FAF, and FAF – MAP. These segments have clearances of 1000ft, 500ft, and 250ft respectively. Thus, your assumption would be correct (but still bad practice). However, the chart shown in the ASTB report shows two ‘minimum’ altitudes after ‘F’ (6 DME), 1250ft and the MDA of 1140ft. If ‘F’ is the FAF as indicated by the chart, then there is some ambiguity between PANS OPS and the chart being used (another chart hazard?). It would appear more logical that the FAF was at ‘ROC’ (4.2 DME) beyond which MDA applies. Perhaps someone with access to the originating Australian procedure chart could comment.

Re ‘can / should’ be a climb. The history of the certification wording for EGPWS will show a valid, but exceptional, reason why a turn vs climb maneuver is allowed at the amber level. However, the industry’s experience and continued improvements to the EGPWS alerting algorithms might now show that a climb is always a suitable maneuver, hence IMHO, the same simple procedure for an alert and warning – pull up, can be taught and practiced. This also avoids any crew deliberation on what to do after and amber alert, perhaps minimizing the risk of debating the reason for the alert – it doesn’t matter, the aircraft should not be where it is. Furthermore, if a turn was to be flown, it implies the use of the terrain map, which may not be as accurate as the basic navigation aid (including ADF in remote areas) unless the EGPWS uses an internal GPS receiver – i.e. beware map shift (we get complacent when using high quality technology in areas well equipped with navigation aids).

I think that the industry is too harsh in attributing connotations of blame on the crew, i.e. ‘mishandled the situation’, etc. Lets see what the investigation discovers; several issues have already been identified that can lead to an error, and the use of VS might have been a simple mistake, or slip (cf Strasbourg). This is not to excuse the crew / operator; a professional operation should have detected one or more the errors (and briefed the hazards). Crews are accountable for a safe operation (before the event) and on the evidence so far this operation appears not to have achieved the required level of safety, but we only believe that due to ‘an error’ (identifiable after the event).

Edit; I note that there are procedure differences between PANS OPS and TERPS. TERPS uses 250 ft, whereas PANS OPS minimum are 295 ft without a FAF and 246 ft with a FAF. This might answer the ambiguity in minimum altitudes before / after ‘ROC’, but it is still not clear to me where the FAF is.

[blueloo]

This is the Australian Chart. I have no idea how to read it - its to hard compared to Jepps!

http://www.airservices.gov.au/publications/current/dap/MMLLO01-113.pdf

[Capn Bloggs]

Airservices Australia have forgotten more about approach charts than Jepp will ever know...

[Bullethead]

alf5071h

The final fix for that approach is at 6DME, on the Jeppe chart, near 6D on the plan is a note 'FF16' and on the profile is a maltese cross at 6D which I believe is standard depiction for an approach final fix.


blueloo

I agree that the AUS AIP charts are a little more difficult to read than the Jeppes, but that's only because I've been using Jeppes for over twenty years and rarely use the AUS AIP charts. All depends what you're used to I suppose.

Regards,
BH.

[alf5071h]

blueloo, thanks for the chart, it appears to originate from Airservices Australia, which I believe is the originating legal reference. Note that the currency date (22 Nov 07) is after the date of the incident.
At least this chart only uses one range scale and defines the GPS range reference, thus avoiding any ambiguity in the use of altitude-range checks. I would prefer that the check table be formatted as ‘altitude over range’ as the former is far more important in achieving safe terrain clearance. Also adding (ML) next to ‘DME’ would further help to avoid misinterpretation.
The FAF is depicted – highlighting the weaknesses of the incident chart (but it’s their symbology – so much for world standardisation). The minimum approach altitudes appear to support the notion that the higher PANS-OPS value is used before the final nav aid (ROC), and that MDA applies thereafter. Both BOL and ROC ADFs are required.
Thus BOAC’s belief in this instance, is incorrect – beware those who think that dive and drive is safer! MDA is not a safe altitude inside the FAF.

"… are Captains good monitors?" No takers … of course there were three crew members, were they all looking outside?
This incident is a good example where investigations conclude ‘CRM failure’ without actually identifying which aspects or what the underlying causes for the human failures were. I have more confidence in ASTB, who show a keen interest in the HF aspects of events than do most investigators. We have much to learn from what initially appeared to be a 'minor incident'.

[safetypee]

Its interesting to see how many people propose ‘high tech’ solutions to the problems identified by this incident, but the use of LNAV / VNAV, Vert profile, GPS, etc, has many ‘gotch yers’ which start with a simple mistake or poor system knowledge. These systems represent increasing complexity in the modern world, which is often linked with higher workload, the need for wider knowledge, and much more crosschecking to avoid error and complacency.

The NDB 16 approach is a simple procedure; a 3 deg slope starting at 11.5 nm; in most respects just like a visual straight in. In IMC the visual cues have to be replaced with an electronic track – clearly depicted on an EFIS, and altitude checkpoints combined with vertical speed to confirm the required glidepath. Altitude checks can be made approximately every 30 sec (every nm @ 120kt GS). Thus the crew task is to use a scan pattern to include these aspects, together with airspeed etc to maintain safe flight. Crosschecks can be provided with timing, beacon crossing, comparison with the briefing (the plan), and the approach chart.
What could be easier; the modern jet transport has the latest display aids, further safety crosschecks provided by RNAV / GPS distance / position, and EGPWS terrain display, yet many of these aircraft get dangerously close to the ground short of the runway – see the link at #69.

Procedures remain simple (charts may need improvement) and aircraft are better equipped (safer), so why are these incidents still occurring? (The prevalence of events might be due to the identification and reporting from EGPWS data, another positive safety function of the system).
Some people cite that modern aircraft are more complex to operate (still simple to fly), but this complexity might be by choice where operators (management, even regulators) elect to use the high tech equipment without thinking about likely situations and consequences of error. In some circumstances it might be safer not to use the 'high tech' solution as the first choice (easier to fly, less opportunity for error), but still use it as a crosscheck.
Another variable involving people is the crew, the standards of training, professionalism, experience all appear to be changing and not necessarily for the better. These are the people who decide on the tactical use of the ‘high tech’ systems. They rely on a technical solution whereas some simple thought and preparation would provide a less complex and safer solution.
Perhaps its time to put people back into the flying loop (thinking and doing), which might improve airmanship and experience.

[Centaurus]

How difficult is it for PM to sing out the required ht against DME range and for the PF to adjust accordingly

The PF should already have the chart in front of him and he should be referring to it. The support call by the PNF is a back-up to what the PF should already know. To rely on just one pilot calling the chart details is poor airmanship as PNF's have been known to call erroneous heights versus DME distances. The PIA A310 crash at Kathmandu, the Air Manila (?) 707 crash 20 miles from Manila and many other similar CFIT accidents attest to that.

[Tee Emm]

Perhaps its time to put people back into the flying loop (thinking and doing), which might improve airmanship and experience

It will never happen. With the major manufacturer's pushing for full use of automation and major airlines blindly accepting this advice to a often foolish degree, you will rarely see a return to basic instrument flying skills. The Twin NDB approach into Melbourne runway 16 is dead straight simple but it took a bunch of automatic monkeys to cock it up. No doubt it will happen again while pilots blindly twiddle the knobs and gaze with awe at the beautiful picture on MAP. Would you conduct a instrument approach in a light twin with a WAC series map on your lap to see where you are going? I doubt it - but in effect that is what MAP is - a sophisticated WAC chart.

[Capn Bloggs]

As I said at the start, the Lido chart has two missing profile heights, 7DME and 6DME. While the 6DME alt is below, in the heat of the moment...

How about the system gives people a fighting chance?!

[Pilot Pete]

Yeah, but Capn Bloggs, many, many non-precision approach charts don't have all (or any) check altitudes. This is not a major factor, just something that needs careful consideration and thorough briefing. Like I said before, write them on a bit of paper and get the other guy to cross-check them for accuracy, or better still, get them to work out the altitudes independently and then compare your results.

PP

[Capn Bloggs]

I've been doing that for over 25 years, PP. The point is, these "automatic monkeys" as TeeEmm calls them probably have never even heard of the concept. Whose fault is that? Certainly not there's. They cannot reasonably be expected to invent a system that they may well have not even heard about, let alone perfected and practiced often. The system sucks, but who cares, only a couple of hundred lives are at stake, just because an ILS is off...

[Pilot Pete]

But surely the point is that if the pilots aren't up to the concept of working out their descent path and then managing the flight profile then there is something wrong with the system of pilot selection/ company training and oversight? Surely we can't just say that because two dme check altitudes were omitted from a chart is a defence for what happened? I wouldn't consider having these two checks as being something that would have saved this approach or even something that would have 'given them a fighting chance'.

The point is, these "automatic monkeys" as TeeEmm calls them probably have never even heard of the concept. If they are not capable of flying an non-precision approach then they shouldn't have flown one. Fact is, I bet they have been trained how to fly them and they do during recurrent training and checking, just like we all do. If they don't fly them very often then all the more reason to think clearly about it and revise required actions and brief thoroughly. Perhaps their training department could teach it more thoroughly, but this is pure speculation that their training wasn't up to the required (regulatory authority) standard. They cannot reasonably be expected to invent a system that they may well have not even heard about Nobody is saying they have to invent a system, just fly a non-precision approach the way their books say to. Everything offered on here has been how some of us do it to ensure least risk of a vertical profile cock-up. They WILL have been trained in the use of VNAV final approach descent (or if not they were breaking the rules). Problem is they didn't do it properly. There are sure to be many factors, but in my opinion they didn't minimise the risks before starting the approach and then didn't execute the briefed approach that well.

There is no perfect system and ILSs have to be taken off for maintenance every once in a while. If all that is left is a non-precision approach, so be it. Remember, they had plenty of clues that their approach was unstable and a Go Around could have (should have) been flown well before the EGPWS saved them (easy to say with hindsight, I know).

What we should all try to learn from such an event is

1. Non-precision approaches are inherently less safe that precision approaches.

2. They therefore require careful consideration before being flown and a clear plan of how you are going to fly them.

3. All FMC entries, but espcially manually entered ones need thorough cross-checking before entrusting the safety of the aircraft to them.

4. Flight procedures MUST be learn and understood and reviewed (even more regularly for infrequently used ones).

5. Fly within the SOPs. Follow the charts precisely.

6. Monitor closely the flight path during all approaches. PM call deviations and support PF fully.

7. Never be afraid to admit defeat and GO AROUND.

PP

[Phantom Driver]

PP;

Once again, you have hit the nail precisely on the head. Trouble is, I wonder how many people are actually reading and digesting your posts, because we seem to be going around in circles here, reinventing the same old wheel.

For example,our colleague Centaurus states;

The PF should already have the chart in front of him and he should be referring to it. The support call by the PNF is a back-up to what the PF should already know. To rely on just one pilot calling the chart details is poor airmanship as PNF's have been known to call erroneous heights versus DME distances. The PIA A310 crash at Kathmandu, the Air Manila (?) 707 crash 20 miles from Manila and many other similar CFIT accidents attest to that.

Who on earth has suggested otherwise? Basic airmanship (or whatever they call it these days) has always called for this, which is why the wise PF must monitor whatever the PM is calling out. Both you and I have mentioned the good habit of writing down the profile checks (in large font in my case) as an independent crosscheck. My reference (post #64) to the PIA accident in KTM was intended to illustrate that very point of how easy it is to get out of step with callouts;fatal in hilly terrain. The PF had better be in the loop, or else!

One of my favourite quotations has always been-"Man has oft more need to be reminded than informed." (Samuel Johnson?). In our business, we know that complacency is the biggest killer,but it seems we sometimes forget easily.

To that end, I found it worthwhile reading the comments of betterave (post # 71);sound operational technique for flying NPA's in todays glass cockpits. I believe the vast majority of professional pilots do maintain the highest standards. One would hope that what happened in MEL was an isolated aberration.

p.s. ending on a slightly more controversial note, I know we all love to show what aces we are at manually poling the big jet around, and yes, we do need to keep our "rusty" handling skills in shape; (another favourite saying-"The older I get, the better I was"), but dare I suggest that, in todays automated, high density aviation environment, there is a time and place for such things (e.g the simulator?).Apart from anything else, it sure gives the PM a hard time as you decide to manually fly a typical SID! Our SOP says don't do it,but some guys have mental lapses. Like it or not, automation is the way of the future,and getting to proper grips with it is not always easy or pleasurable, yet it has to be done, and that can be a sobering experience. Which is why, in daily operations, I often encounter my third favourite quote-"Experience is something you think you have, until you get a little bit more!"

[betterave]

PP and Phantom,

Well said! Let's hope you are preaching to the converted and others will learn from this incident to avoid making similar mistakes in the future.

NPAs flown as CDAP are supposed to be safer than the old "dive and drive" method. Either way the pilots should be staying ahead of the aircraft to make the approach and landing a routine event. I guess the question is how do we teach "airmanship" to the new pilots who have never flown "steam gauges" and rely mostly on the glass technology for their situation awareness?

As for the topic of using AP for most of the flying. IMHO, it should be a matter of airmanship/common sense again. If the SID/STAR involve many altitude constraints and ATC is very busy, of course the AP should be used even in VFR conditions. Personally I like to hand fly the departure and arrival as often as I can to maintain the "handling skills" but in a busy environment it would take the PM away from his/her duty of monitoring when lots of MCP manipulation is required.

Let's hope I never resemble the debrief I once heard a military instructor said about another student: "he was so far behind the aircraft that had he crashed he wouldn't be hurt..."

Happy Landings!

[Capn Bloggs]

These days, Airmanship is not taught and "the good habit of writing down the profile checks" is not something that comes naturally to a lot of new pilots, especially if they have little time doing it themselves SP IFR.

Strong, detailed SOPs that cover the above points raised by you all are what is required, so that a totally inexperienced, junior crewmember can usefully and safely participate in keeping the pax alive. This I suspect will be a factor in this incident. Will it be fixed? Probably not. Will the Chief Pilot frown on his boys wanting to practice "out of the square" operations? Probably yes. So what hope have these crews got?

I contend that doing an practicing NPAs only in the sim is not frequent enough. They must be practiced regularly on the line as well (quite apart from the recency requirements) but then even the manufacturers frown on that!

[faheel]

Nobody is saying they have to invent a system, just fly a non-precision approach the way their books say to. Everything offered on here has been how some of us do it to ensure least risk of a vertical profile cock-up. They WILL have been trained in the use of VNAV final approach descent (or if not they were breaking the rules). Problem is they didn't do it properly. There are sure to be many factors, but in my opinion they didn't minimise the risks before starting the approach and then didn't execute the briefed approach that well.


I agree with the above statement, however if they had been trained in the use of vnav final approach descent then they cocked up because it cannot be used for an approach not encoded in the fms data base.As such they are prohibited from using vnav and should have been vs or fpa.

[Sub Orbital]

Can we please get back to basics.
Absolutely NO aircraft should be doing 1800'/min on approach - period.
Common sense will tell you that it should be around 800 - 900'/min.
Any so called captain that lets his aircraft do that rate of descent at that phase of flight should never be allowed near an aircraft.
An aircraft with that rate of descent and (as reported by the ATSB) with ALL flightdeck members looking outside simply reinforces to me (from past experience) that pilots from that region haven't the faintest idea how to fly non ILS approaches.
They had no idea what they were doing and were desperate to get visual.
Lets forget political correctness and tell it as it is.

[Pilot Pete]

Sub Orbital

pilots from that region haven't the faintest idea how to fly non ILS approaches. I think you are making far too general a statement there mate. We all know of the short-comings of certain cultural backgrounds with regards to flying aircraft, but you cannot tar everyone with the same brush. I bet there are pilots from that region who fly non-precision approaches perfectly acceptably, just as there are Western pilots who cock them up, and I speak from experience with regards that point. So let's not get into that kind of bigoted debate. They had no idea what they were doing and were desperate to get visual I think that is wild speculation. Indeed they got the approach all wrong and unstable, but how can you speculate that they did it in a desperate attempt to get visual? They lost situational awareness with a high rate of descent on, I can't see that it was a planned attempt to dive below cloud to get visual.....:confused:

PP

There but for the grace of......

[A. Le Rhone]

Why is an aircraft like a 777 even doing an NDB approach? A VOR is bad enough but an NDB with paying passengers?

Yes we're all heroes and could do these blindfolded when we were in our twentiesin Barons in Africa but in this age of triple IRS's and GPS's there is no place for pre-WW2 nonsense like NDB's.

The sooner such passenger-threatening rubbish is banned the better.

[Pilot Pete]

Yeah and just ban flying as it is so dangerous, take a boat instead.:rolleyes:

PP

[skiesfull]

It is my understanding that unless the NPA is already in the data-base, then VNAV is not recommended and V/S preferred. Since MEL rwy16 does not have an NDB approach in the data-base, then VNAV should not be used and FL-CH definitely not to be used. Regarding the F/o's low hours -welcome to Asian airline experience! At least they got away with it and hopefully learned from it.
By the way, what is CASA's aversion to ILS runways, especially CAT 2/3????
Stuck in the 20th century - or is it not deemed "a requirement"????!!!!

[FlexibleResponse]

I believe that,

"Non-precision approaches are inherently dangerous".

You may wish to disagree, but you have that privilege and also you are alive to exercise it.

I think that the spirited discussion from such obviously highly qualified and experienced aircrew over the last few pages tends to back me up.

I do not disagree that NPAs can be executed safely and with a high level of skill.

I also agree that any International ATPL licence holder should be able to execute any approach that he is qualified to carry out.

I suspect that given a sample of said International ATPL holders that there will be a much wider scatter in the results of conducting an NPA as opposed to the results of conducting a precision approach.

That is the long and short of it.

[Tee Emm]

One critical issue is that MDA is not a safe altitude; it’s only a minimum altitude during the latter part of a correctly executed procedure

Pardon? Do you mean to say for the past umpteen years I have trusted the chart designers with my life to give me a published MDA and now you say it isn't safe...

[Pilot Pete]

Tee Emm

I think what they mean is you can't 'dive and drive' down at MDA. If you fly the profile correctly and end up at MDA then it is safe to the missed approach point. What many operators do now is fly a non-precision approach as a CDA, adding 50' to the MDA to in effect make it the same as a DA; the point at which you initiate a missed approach if you don't have the visual reference, thus no level segment at MDA and then a dirty dive as you see the runway at the last minute.

Another point worth considering is what you will see at the published minima on a non-precision approach. For instance, many non-precision approaches have an MDH of say 700', with a required visibility of something like 2000m. If you are going to adopt the aforementioned procedure of not flying level, destabilising the approach and then diving for the runway, then you are not going to see the required visual references if the vis is on the minimum of 2k. At 700'AGL you will be somewhere like 2.1nm from touchdown on a nominal 3 degree descent. 2.1nm is approximately 3300-3400 metres. Another point to consider. Just because you have the legal minimum visibility, is it worth making an approach like that?

PP

[Spooky 2]

For what it is worth, Boeing/Alteon does not teach the drive and dive method unless an operator specifically asks for it based upon their respective ops specs. All "Non ILS" approaches are built using VNAV.

[alf5071h]

TM, MDA is only ‘safe’ after the FAF and up to MAP (plus a small distance for climb entry).
One of the points to be made was that crews should not consider MDA as a safe altitude as in same level of safety associated with MSA.
With an EGPWS warning, a climb to or maintaining MDA is not an acceptable solution, the climb should be made to the area / sector safe altitude (SSA). One of the reasons for a warning could be due to a navigation error, such that the aircraft is not within the area covered by MDA, thus the only safe manoeuvre is to climb to the higher SSA.
I guess that you understand this, but many pilots do not. The problem is aggravated by the differing uses of ‘safe’ without specifying the situation – my mistake in #77.

[Pilot Pete]

All "Non ILS" approaches are built using VNAV. I think that is a bit broad as a statement. Not all operators use/ authorise VNAV approaches. Many use traditional descent path methods such as V/S, with a rate worked out by the crew for the profile in question.

PP