Danger of letting down based on NDB false overhead. B707 crash 1974

[Centaurus]

https://web.archive.org/web/20000305044212/http://aviation-safety.net/reports/740422-0.htm

An accident report worth re-visiting. In 1974 a PANAM Boeing 707 was en-route Hong Kong to Bali at night. There was no DME or ATC radar at Bali in those days. The crew of the 707 planned on an NDB instrument approach on arrival. Forecast was fine but it was a black moonless night for arrival.
Approximately 30 miles north of Bali one of the two ADF's on the 707 swung and the crew misinterpreted this as an indication that the aircraft was overhead the real Bali NDB. They commenced the instrument approach as per the NDB chart. The aircraft crashed into a mountain.

Today blessed with highly accurate GPS position fixing this type of accident would be unlikely. But there are airports around SE Asia and in Australia still with an NDB as the let down aid. The PANAM 707 crash 43 years ago should remind pilots that ADF false overheads can still remain a trap to an unwary crew.

[White Knight]

Well, with the GPS usually comes the ability to fly RNAV, RNP, GPS approaches etc!

I can't even select or tune NDBs on my Maxibus; can only see them on the ND via the NDB button on the EFIS control panel.

Even GA operators seem to have all the bells and whistles these days! But yes. Nice to have a reminder that there are traps out there for the unwary in all aspects of flying!

[Bergerie1]

Centaurus,

Interesting. Do we know whether both ADFs were tuned to the same NDB. I do not know Bali and I do not know what other aids would have been available or needed. For example, did the procedure require another NDB to be tuned as well? If both ADFs were tuned to the same NDB, I would have been very wary of only one needle swinging round.

Also, do you know more about the weather at the time? Although the sky was reported to be clear with the stars bright and the sky moonless, presumably that was at the airfield. Were there any active Cbs close to where the needle swung and they turned onto the new heading? I have certainly seen ADF needles point to a Cb instead of the NDB.

[wishiwasupthere]

What's an NDB approach?:}

[Capn Bloggs]

NDB = Never Done Before! :ok:

[Moneymoneymoneymoney]

I was 12 years old and in an Ansett 767 cockpit when the captain explained to me the dangers of NDB's after I queried the needle oscillations. I never forget it since.

[ZAZ]

My AD still has an NDB and I get tested every year but I also recall on two occasions in past decade getting caught out
One failing to monitor NDB on last segment and have the old KR86 switches go noisy and giving false bearing. Saw the Grampians on RHS of the aircraft and had decided not to descend to 20 mile lsalt early like some people do was not on track and would have hit Mt William like a guy called Wills did decades back.
Second was at Temora flying by a thunderstorm needle pointing to cloud did not see station passage let down in valley approaching YOUNG between ranges.
Still alive and still doing NDB approaches and will stack them against NPA GNSS any day.
Tricks with NPAs are to do them correctly and how they send you 15 miles away from the airport whereas NDB is done from overhead and you often get cloud breaks and can get visual. FWIW 2 cents.

[Centaurus]

Friend of mine saw this thread and wrote thus:

"Back in the days before they had radar at Bali, I was ferrying an F 27 from Muscat to Christchurch. On board I had a trainee Captain and a low time First Officer. Most of the sectors were conducted with me supervising from the jump seat when the Captain was flying and from the left when giving the F/O a leg.

Inbound to Bali, at night, I was having a bit of a snooze down the back, with the intent to go forward for the approach, as none of us had been to Bali before. I heard the engines throttle back for the descent and thought to myself “hmmm...a bit early, maybe we had good tailwinds?” Went forward to note that they had set 4000 feet on the alt alert. “What’s happening?” Says I. “Cleared to the NDB, 4000” says they. So, of course I stopped further descent while looking up the MSA, which I knew to be high to the north. Lucky I did. Seems that back then, if there was no conflicting traffic ATC cleared aircraft down to the IAF or FAF without any regard to terrain. It was assumed that the crew would fly the necessary procedure – which in this case involved ‘shuttling’ down in the holding pattern.

Lucky I have a built-in time and motion meter, hey? "

[john_tullamarine]

I tucked a point away years ago at Hobart.

On the 727, flying with some chap I had never met before. CAVOK, night, and cleared for the usual .. but he decided to stay at LSA plus until overhead the aid and fly the full letdown using other aids to confirm position as available. Perhaps a bit overly cautious but, in post flight discussion, he observed that it was his way to get to retirement when faced by aerodromes into which he had never operated before.

I gave it some thought over an ale and figured that it might not be too bad a tactic for new places ?

[Bergerie1]

John,

I agree one hundred percent!

[Centaurus]

Lucky I have a built-in time and motion meter, hey? "

Post No.8 got me thinking about premonitions. In this case the resting F27 captain who on hearing the engines throttled back for the descent into Bali thought to himself "hmm..a bit early maybe we had a good tailwind." And thus saved a certain CFIT.

It reminded me of a similar feeling or premonition that happened to the captain (Centaurus) of an RAAF Avro Lincoln four engine bomber during a long distance flight from Darwin to Townsville via Cape York peninsula on 9 December 1959. The Lincoln departed Darwin at night and was scheduled to at Townsville about 0600 next day. Flight time was nine hours of which 30 minutes was day. There was a crew of nine aboard including two pilots and a navigator.

Some time after the aircraft turned the corner at Cape York for the southbound track off the coast towards Townsville, the captain gave control to the co-pilot and put him in the captain's seat and slipped down the back for a quick kip behind the main spar. He asked the navigator whose position was in a compartment immediately behind the pilot station, to keep an eye on the new co-pilot and to wake the captain when 30 minutes out of Townsville. Cruise altitude was 1500 feet over water. After all the crew were supposed to log all shipping they may see on the way

Some indeterminate time later the captain woke up with a start and quickly made his way back to the cockpit only to discover the navigator slumped asleep over his charts. Worse still, the co-pilot was also sound asleep and the aircraft was on autopilot.

In the Lincoln the co-pilot seat was a fold down type and when not in use was secured against the fuselage wall. This allowed crew members to walk or crawl into the nose compartment. On arrival at the cockpit the captain saw the altimeter showing 1500 ft and that the aircraft was in low cloud. There was no time to wake the co-pilot and get him out of the captain's seat. Although standing up in the aisle-way he disengaged the autopilot and applying full throttle to all engines pulled back the control column.

Within seconds the aircraft came out of cloud into clear air at 2000ft with dawn coming up on the left horizon. Dead ahead at 12 miles with only the summit poking clear of cloud, was Mount Bentley at 1798 feet elevation the highest peak of Palm island. If the captain had not woken sensing something was not quite right the aircraft would have flown into the hill within a couple of minutes.

[Centaurus]

Interesting. Do we know whether both ADFs were tuned to the same NDB. I do not know Bali and I do not know what other aids would have been available or needed. For example, did the procedure require another NDB to be tuned as well? If both ADFs were tuned to the same NDB, I would have been very wary of only one needle swinging round.


Bergerie1. Sorry but can't help you on that lot. Apart from the published official Indonesian investigation, unable to locate more pertinent info. Presumably there would have been a US NTSB investigation but could not find it on the internet.

[Bergerie1]

Centaurus,

Thanks for looking. The premonition you describe sounds just like those described by John Farley!

[clark y]

Have seen TS cells give station passage before.

With respect to the PanAm accident could terrain shielding be a possibility for a false station passage if the aircraft was at the LSALT while still well to the north?

[ampan]

Can I add another question re PA812: What about the INS? The accident report makes no mention of it.

[Tinstaafl]

Speaking (or writing) of getting to retirement, when I worked at Bankstown, I was asked to accompany a young, inexperienced bloke on a bank run. He'd been flying with another bloke on this run, but not yet experienced enough to do it on his own. The usual pilot was unavailable (ill, or something. I forget why). I didn't work for this company so wasn't familiar with the route or routine. This was pre-GPS.

We did the trip. There were some issues but the worst was the last sector. Young bloke in the LH seat, me supervising. Returning to YSBK at night in IMC over the Blue Mountains to the west. En-route, and without any fix whatsoever - not even a groundspeed check - he decided it was time to start descent into YSBK, using the LSALT step downs.

I had to stop him. He said that's what the other pilot did. The problem, in my mind, was that hadn't confirmed - or determined - he had passed the mountains! Not even a groundspeed calculation. Everything based on flight plan times.

Debriefing afterwards, I had to emphasize that were there headwinds stronger than forecast, or a TAS less than planned, we would have descended into the mountains - but he had no way of knowing that. I said that if he wanted to descend early he needed to pre-determine x-bearings (+ tolerances) from navaids to give fixes to allow that.

In my mind was a historical crash written about in the old ASD ('crash comics'. Nothing so useful in its abysmal replacement!) about a plane that used ded. reckoning for its descent and the winds weren't as expected. End result was a premature interaction with the Great Dividing Range.

[Bergerie1]

ampan, No INS on 707s at that time.

[megan]

Centaurus, this is all that's available on the NTSB site. Seems access to the file would have been removed in January 2001.

https://ntsb.gov/_layouts/ntsb.aviation/brief.aspx?ev_id=83155&key=0

Retrieval would have been tough work, caution, brief glimpse of body.

Zvdrz0UytGA

Although the report doesn't directly say so, it seems to hint that both ADF were tuned.

[waren9]

But there are airports around SE Asia and in Australia still with an NDB as the let down aid.

luckily the industry is slowly grinding forward. if only one step at a time.

whilst there may still be ndb approaches, its been a few years (and several jet types) since i was in an aircraft with an adf installed.

the premonition tales ring true. thanks for posting.

[Pinky the pilot]

I remember once doing a practice NDB approach in a Chieftain with dual ADF, both tuned correctly to the aid.

One showed station passage at least 10 seconds before the other!:hmm:

`Why is it so?`(In my best JSM accent:D) I queried the C&T Pilot in t`other seat.:confused:

`You`ve just learned something important!` was the reply.

ie; Cross check with DME (then) or GPS. (now):=

[ampan]

or stars ie; take a navigator instead of a second co-pilot

[Bergerie1]

In my view, navigation is 'the management of errors'. No single navigation aid (even GPS) is free from errors of some sort. The cautious navigator or pilot needs to use all available aids - whatever they may be at the time. And the greater the risk of error, the greater should be the safety margin.

See also this report:- https://aviation-safety.net/database/record.php?id=19660124-0

[megan]

navigation is 'the management of errors'. No single navigation aid (even GPS) is free from errors of some sort. The cautious navigator or pilot needs to use all available aidsOne of the passengers reported he noticed something was not right, being a frequent traveller on the route, he recognised the sun was not in the correct position.

https://en.wikipedia.org/wiki/Korean_Air_Lines_Flight_902

[Bergerie1]

megan,

Indeed, yes. The Mark 1 Eyeball (and the seat of his pants) is part of a navigator's equipment. Is the sun in the right place? Is Polaris (or Southern Cross) where it should be? Or, when sailing, why has the wind changed? Why are the waves coming from a different direction. Why is a cloud that shape, is it the top of mountain? Do those cumulus clouds indicate an island beneath?

Does the change of movement of my boat indicate land nearby - waves reflected off rocks or refracted down wind of an island?

There are similar signs when flying (perhaps the source of Centaurus' sixth sense). Every little indication should be used, however small.

[Wunwing]

There was INS fitted to 707s in 1974. Qantas fitted their first in 1971 and by 74 all were fitted. There was Doppler and navigators prior to that.
I'm not sure what Pan Am had by then but there is no doubt Bali was known as a challenging port.

Out of interest up to that time Qantas didn't operate to Bali but Pan Am did the SYD / DPS trip. From then Qantas did it 2x per week.

When QF got down to 2 x 707 we had about 2 years of the 3 or 4 day Bali trip and our boss couldn't figure out why we wouldn't bid for the 747?

[john_tullamarine]

and our boss couldn't figure out why we wouldn't bid for the 747?

A bit like the popularity of the AN PER and DRW trips with crew changes where crewing periodically forgot about folks and it took days to sort it out ..... of course, we always did the right thing and rang up ... after a while ... when it became absolutely obvious that they had forgotten all about us ...

[LeadSled]

Folks,
I see a number of comments containing the remark: "----- before BALI had radar".

So, now BALI has radar??

Would you actually rely on the radar for terrain clearance ---- and not just at BALI?? Do you believe ATC will not give you a clearance below LSA (by whatever name) or actively monitor you on a STAR that you might screw up??

I speak as somebody who knows the area ( and a much wider area) very well.

Tootle pip!!

PS: That was a bad year for "The World's Most Experienced Airline", they had several "experiences" that year +/-, including another B707 loss at Tahiti. Around about that era they put a B707 off the end of YSSY 34 (R now) in an abort.
PS2: PanAm B707-321 has two Bendix Doppler, with a common aerial, QF had two Marconi Doppler with a common aerial ---- they cold not be used as a position fix for a letdown, they were a DR position only. I know, I have flown them both.
PS3: To this day, there are far more "challenging" airfields than BALI. None of them are "challenging" as long as you stick to the SOPs., just some you have to work a bit harder.

[mustafagander]

Wunwing,

I don't think all QF B707 had INS by 1974. As I recall it 1 A/C was fitted around 1971 or 72 to get everyone used to it and gain experience for the B747 classic. I remember the one INS CDU being fitted in the OH panel to the left, a rather useless place to use it for real nav.

[witwiw]

Would you actually rely on the radar for terrain clearance Not rely on it exclusively however, yes, it would be another aid, providing I had a radar terrain clearance chart.(e.g AYPY 10-1R). Having said that, it relies on proper navigation and is a supplement not sole means.

[Wunwing]

Interesting. I had a debate last year about the single QF INS. Its good to see that someone else remembers it. It was 1971 as I worked on the project as a LAME and went directly to FE training from that.

IF the 2 INS were not fitted by 1974 it was close. In 74 I took my wife on her first staff travel trip and 2 of the most senior Navs were on it as a retirement trip so I guess they would be the ones to turn out the lights.

[ampan]

Seems strange to have had no navigator and no INS.


Was a northern approach to Bali standard? Just looking at Google Earth, wouldn't it be far less difficult to burn a bit more fuel and time by going past the airport, descending over the ocean and then approaching from the south?


Maybe they had a plane to catch.

[CurtainTwitcher]

Here's a link to a 1971 article in popular mechanics: The Brain that Tells a 747 Where to Go! (https://books.google.com.au/books?id=UtgDAAAAMBAJ&pg=PA108&lpg=PA108&dq=b747+carousel+ins&source=bl&ots=6U6hmJ3DUQ&sig=jioZn9F7Jiy6vLTKgf6IHdtJ3ks&hl=en&sa=X&ved=0ahUKEwiM-vK-3ezUAhXCKZQKHeodAXgQ6AEIMDAB#v=onepage&q&f=false) hailing the new miracle technology, the Carousel IV INS for the 747.

[Bergerie1]

ampan,

I can only speak for the UK legislation which specified that in certain areas of the world, where there were inadequate navigation aids, a qualified specialist navigator needed to be carried. Elsewhere, were there were adequate navigation aids, the pilots navigated themselves.

That has now been superseded by, first INS, and then GPS, and a combination of the two.

[ampan]

The International Society of Air Safety Investigators
VOLUME 10, NO.2 WINTER, 1977
PROCEEDINGS
of the
EIGHTH INTERNATIONAL SEMINAR
3-6 OCTOBER 1977
…

THE ACCREDITED REPRESENTATIVE
William L. Lamb
CHIEF, FIELD INVESTIGATIONDIVISION
BUREAU OF ACCIDENT INVESTIGATION
NATIONAL TRANSPORTATION SAFETY BOARD


…

The investigation of the Pan American World Airways, Boeing 707-321 accident on the island of Bali, Indonesia, illustrates the role of the accredited representative, and how I tdiffers from the role I normally held-Investigator-in-Charge. I was the United States accredited representative, and I hope that this .saga of my adventures-and mis-adventures may throw some light on the differences between the roles of the IIC and the accredited representative.
…
Pan American Flight 812 was a scheduled international flight from Hong Kong to Sydney, with an en route stop at Denpasar, Bali. The flight departed Hong Kong at 1108Z, with an estimated en route time of 4 hours 23 minutes. The flight was routine except for difficulty in establishing radio contact with Bali approach control. Communications were established and the flight was cleared to descend from its cruising flight level of 350 to flight level 100. The flight contacted Bali Tower at flight level 110 and requested a lower altitude; seconds later, it reported over the station. This report was made at 1519; although a corrected ETA of 1527Z had been reported a few minutes before the aircraft began to descend. The 8-minute early arrival should have been a cue to the crew to verify their position. The flight was cleared to descend to 2,500 feet and landing instructions for runway 09 were issued. Shortly before 1527Z , the crew reported level at 2,500 feet and were instructed to report the runway in sight. At 1527, the crew asked,"Hey Tower, what's your visibility out there now." This was the last radio transmission from the aircraft. The wreckage was found on the north slopeof the mountains about 37 nm northwest of the airport at the 3,000-foot level.The flight path through the trees indicated a direction of flight of 155° to160° and a 15° to 25° noseup attitude.
…
The CVR revealed that, following the problem connected with establishing radio contact with the Bali Approach Control, the crew identified being overhead the airport by the swing of one of the two ADF's that were tuned to the NDB on the airport. The comment that "one's swinging, the other decided to wait," followed by the overhead report and the commencement of the approach descent, indicates the crew's actions.

The descent flight path derived by a computer generated plot was in error, because in making up the cards for the computer, someone, the keypunch operator or someone, entered an extra digit in one card and this cancelled several heading changes. This error was not noticed until much later; however, this mistake would only move the start of the descent a mile or so to the west and the descent would still have been made behind the mountains.

It is interesting to note that the VOR referred to by the crew is about 6 nm south, or beyond the airport from the direction PA812 was approaching. The VOR was operating but it was not being used by the crew of PA812.


...

[Arsey Duck]

Twenty five years ago it was the same - on descent into DPS in an A330 from HKG and cleared to an altitude BELOW the MSA.
And it happened on several occasions!
On a slightly different note, 43 years ago I was ferrying an F27 from MEL to JNB and on the sector from SIN to Medan, approaching from the SE we were cleared to overfly and turn ‘right’ for a left downwind for RWY 05. Fortunately we had been warned that the controllers periodically had difficulty in translating their language into English and often got confused between their ‘left’ and ‘right’!
Needless to say, we ignored the instruction and made a safe landing.
Several months later a Garuda aircraft crashed into a mountain after following the ATC instruction to turn the wrong way!

[Dora-9]

Twenty five years ago it was the same -

Me too! Arriving in a B777 from the north, with a late change in landing direction, ATC vectoring straight towards the mountain...

[compressor stall]

About a year ago another Asian country I was cleared well below the 10k+ MSA. Took two queries before they realised.

[MK 4A Tank]

Meaconing and coastal refraction always almost the cause of NDB misdemeanors and accidents!

[krismiler]

I remember something about “night effect” from my CPL exam days and being told that at dusk or dawn you could kiss the ADF goodbye. My ATP examiner told me that he lost his crew chief in an NDB related accident in a DC3.

The NDB letdown into Bankstown under radar control from YSSY approach was probably the best approach but my main use for the ADF was to tune into a commercial radio station and use it to find the city I was going to.

[Ascend Charlie]

Our Hueys only had one ADF for navigation - our brilliant government paid to have the Tacan removed and a lead weight installed to keep the balance.

Doing IF training at Canberra, we would fly the Twin-NDB approach. Dial up the first one with the rotating dial (not numbered clicks, just roll till you hear the station) and fly the inbound leg. Overhead, mad scramble to dial the next one, identify, get back on the correct bearing, and hope for a high cloudbase. Luckily we were slow enough to have time for the nimble-fingered switch flicking between stations.

[340drvr]

There was this 1996 Croatia crash, definitely got my attention as I was just working on my instrument rating that year, with NDB as part of the training.
1996 Croatia USAF CT-43 crash - Wikipedia (https://en.wikipedia.org/wiki/1996_Croatia_USAF_CT-43_crash)

[Capt Fathom]

The twin NDB approach using just one ADF. What could go wrong?
Now you can do an NDB approach… without an ADF. 😱

[SOPS]

In my previous, previous life, I was flying a non glass, 737 300, to somewhere in Spain. I can’t remember where. However we were cleared for a NDB approach. Light cloud at 2000, about 4 Oktas. When we got station passage, the FO, who was flying, started the correct let down procedure. Looking out the window, and also the the DME, something suddenly did not add up. I told him to stop the decent. We actually got “station passage” about ( from what I could work out), about 5 nm South of the NDB. I have never figured out why. But it gave me a BIG SCARE.

Having read above, maybe it was night effect. I know it was scary. It was only the fact it was almost CAVOK, and I had been there many times at night, and it did not look right, triggered me.

[Big Pistons Forever]

Thank God the horrible NDB is now effectively a thing of the past. I spent the early years of my flying career driving a Piper Navajo around the rocks of the Pacific Northwest of North America. I had a near identical experience of a false passage indication with a probable CFIT accident if I had not listened to my spidy senses.

NDB’s did result in some innovative procedures though. One airport (CYCD) used to have a NDB approach with a big bend at beacon (FAF). Waiting for station passage put you well West of the remaining portion of the final track towards high ground and required you very quickly intercept the final approach track outbound. However it was often calm in the morning and before you reached the NDB you passed over a pre pollution control pulp and paper plant. The trick was a soon as you smelled the plant you put in 3 degrees of left bank which would mean you would be almost perfectly aligned with the runway at beacon passage