Revisiting the HS Trident Deep Stall accident in 1966

[Tee Emm]

Found this old report on the fatal accident to a HS Trident tri jet during an early test flight to investigate stall characteristics and recovery action. Readers may remember the Trident was one of the first of the T-Tail jet transports. The date was 3 June 1966. Many of todays pilots were not born then so it may be of interest to them; particularly reading of the phenomena of the Deep Stall. Shades of the Air France A330 crash in the South Atlantic?Quote:
The aircraft took-off from Hatfield at 1652 hours to carry out the first of a series of production test flights for the purpose of qualifying for a Series Certificate of Airworthiness. The schedule for the flight called for stalling tests should the aircraft and the flight conditions be suitable. After take-off the aircraft climbed towards the north-east and at about 1830 hours, after completing the greater part of the flight test schedule, the stalling tests were begun. Three approaches to the stall were made in order to check the aircraft's stall warning and stall recovery systems and the flight engineer's log shows that with the aircraft in the landing configuration the stick shaker operated at 102 kt and the stall recovery system at 93 kt. The fourth stalling run was made at a height of 11 600 ft with the aircraft still in the landing configuration but, in accordance with the requirements of the test schedule, the stall warning and stall recovery systems had been made inoperative.

Radio telephony communication with the aircraft consisted only of routine 11 messages until at 1834 hours when the pilot-in-command reported We are in a superstall at the moment". This was the last radiocommunication received. At about this time the aircraft was seen over Felthorpe flying very slowly heading south-west at about 10 000 ft. The nose was seen to go up 30 to 40 degrees and the aircraft began to turn to port; the starboard wing then dropped sharply and, following a short burst of engine power, the aircraft went into a flat spin to starboard. The spin continued, the aircraft turning once every 6 to 8 seconds until it reached the ground about a minute and a half later. All four crew members were killed
.Causes:During a stalling test decisive recovery action was delayed too long to prevent the aircraft from entering a superstall from which recovery was not possible. The flight was being conducted in accordance with an agreed test schedule. No evidence of pre-crash failure of the aircraft has come to light. During the final stalling run speed was reduced at a rate greater than 1 kt per second and recovery action was not initiated until the speed had fallen beyond the limit set by the test schedule. Unquote.

[DaveReidUK]

Shades of the Air France A330 crash in the South Atlantic?

Not really. The Trident ended up in a situation that was unrecoverable. AF447 was not in a deep stall.

[safetypee]

TM, This Trident accident was a harsh lesson for the design / regulatory process, and was a basis for modern standards for stall warning and avoidance.
This should not be confused with the ‘Staines’ Trident accident which was associated with Human interaction.

It would be misleading to associate the Trident stall with AF447. The differences are that the Trident involved a ‘locked in’ deep stall, where there were no further control resources available to recover, whereas AF447 may have involved a deep stall (depending on definition), it was not ‘locked in’ because some form of control remained for recovery. i.e. the difference between physically not being able to recover, vs having capability but not using it, or recognising the need to use it.

[dixi188]

The prototype BAC 1-11 crashed due to a deep stall in 1963. I am surprised that the Trident did not have a spin recovery parachute fitted to aid the crew.

[safetypee]

dixi, AFAIR the Trident accident was during a pre-delivery production test flight, thus additional experimental test-flight devices were not fitted.

Nowadays where pre-delivery testing requires a stall check either with or without stick push (stall ident) enabled, an additional glare-shield stall panel is fitted showing AoA calibrated for stick shake / push. Preflight the AoA vanes and indicator are crosschecked, so that if when inflight the shake / push do not occur at the expected AOA then the test is stopped.
The BAe 146 used such a panel, which may also have been approved for training stalls, providing the shake / push was enabled and checked / calibrated preflight.

[DaveReidUK]

dixi, AFAIR the Trident accident was during a pre-delivery production test flight

Yes, as per the first post:

The aircraft took-off from Hatfield at 1652 hours to carry out the first of a series of production test flights for the purpose of qualifying for a Series Certificate of Airworthiness. The schedule for the flight called for stalling tests should the aircraft and the flight conditions be suitable.

That's verbatim from the investigation report.

[bcgallacher]

I was informed that stall trials on the Trident 3b were protected by a rocket motor attached to the vertical bulkhead below engine number 2. Presumably it would have been fired if the aircraft settled into a deep stall.

[pulse1]

I can't find a link but there was a fascinating story in Pilot Magazine some years ago which was written by a Canadian test pilot. He was part of a team sent to Russia to evaluate a regional jet for the Canadian market. Before he went to Russia the Canadians and Russians had agreed a test programme which included stalling. The test pilot knew that this aircraft had no stall protection and was likely to go into a deep stall with no recovery possible.
He approached the aircraft for the first flight with some trepidation. His Russian copilot seemed prepared for the flight but, at the last minute before start up, the copllot came up with a problem which meant the flight had to be delayed. This went on for each flight, the last one of which even taxied out to the runway before some reason was discovered by the Russians for cancelling the flight. In the end they agreed that the aircraft could not be stalled and the programme was cancelled with relief all round.

[G0ULI]

There is only one possible method of recovery from a deep stall, which involves banking the aircraft 90° and pointing the nose down. Given enough altitude, sufficient airflow will develop over the elevators to enable the aircraft to be recovered. Potential pitfalls are overspeed, entering a flat spin, disruption of airflow into the jet engine intakes and entering another high speed stall while attempting a recovery pullout. This really is a last throw of the dice!

[Cunliffe]

At the time of the Staines accident my uncle was working in the Hatfield factory. He told me that the word in the factory was that Cunningham said he could get out of a deep stall and wanted to take a Trident up to demonstrate it. Naturally management refused the offer. Just saying.

[infrequentflyer789]

dixi, AFAIR the Trident accident was during a pre-delivery production test flight, thus additional experimental test-flight devices were not fitted.

Nowadays where pre-delivery testing requires a stall check either with or without stick push (stall ident) enabled, an additional glare-shield stall panel is fitted showing AoA calibrated for stick shake / push. Preflight the AoA vanes and indicator are crosschecked, so that if when inflight the shake / push do not occur at the expected AOA then the test is stopped.
I take it "nowadays" means after GXL888T?

[KenV]

There is only one possible method of recovery from a deep stall, which involves banking the aircraft 90° and pointing the nose down.By definition a deep stall means the wings are stalled, so no roll control, and the tail is blanked, so no pitch control. The rudder is also blanked so you can't use rudder to induce a roll. So how does one propose to "bank 90°" and "point the nose down" while in a deep stall?

[PDR1]

If the aeroplane has effective outboard slats then it's not at all unknown to still have some aileron effectiveness even when "fully stalled". It's actually quite rare for a wing to be fully stalled across the whole of the span anyway - it's just that usually the diffe3rence is academic. In this case it isn't.

PDR

[safetypee]

Ken, the situation may not be quite so absolute as you suggest.

A stalled wing does not mean no lift, just insufficient lift for the conditions; see Cl vs AoA charts.
So for roll ‘control’ all that might be required is to generate a difference in the remaining lift distribution between the wings; using aileron, spoilers, etc.
Similarly for the fin/rudder, although I would agree it would be less effective depending on aircraft type and AoA.
The ability to recover depends on what effect roll has in unmasking the tailplane/elevator, and/or changing the pitching moment, or AoA. Some have suggested asymmetric thrust.

AFAIR few if any modern commercial aircraft have a locked in deep stall characteristics.
Any DC9, MD 80, B717, B727, C5, C17, Dash 8, Embraer, ATR, Beech, TU, expertise out there; facts not fiction ?
BAe146, Avro RJ, although predicted, no deep stall situations were identified during flight tests.

[Mad (Flt) Scientist]

AFAIR few if any modern commercial aircraft have a locked in deep stall characteristics.

I would assume the same also assuming an "otherwise healthy aircraft"

I would put money that there ARE some types which have such characteristics if protection systems (such as pushers) were inoperative and the stall progressed far enough. The fundamental aerodynamics of a T-tail make this a not-unlikely hazard in such types, without something to counter the aerodynamics.

[slast]

I was informed that stall trials on the Trident 3b were protected by a rocket motor attached to the vertical bulkhead below engine number 2. Presumably it would have been fired if the aircraft settled into a deep stall.

I suspect that whoever "informed" you was confused by the RB162 boost engine on the T3B. I've never heard of a "rocket" being installed and the boost engine was really there because of the lack of main engine power in some conditions (you'll have to read the full history of the Trident to see how such a situation could have occurred - all due to commercial politics and not design skills at De Haviland ! )

I was on the crew for a number of Trident stall tests. Each was about 3 hours when we did full stalls just as safetypee describes. The first was a Trident 2 (GVFM, 11 November 1972) . This was during a "continuing airworthiness" C of A renewal. At that time it was still a requirement that aircraft in line service were tested to confirm that initial certification performance was still being met. The test was conducted by Gordon Corps who was I think CAA deputy chief test pilot in the left seat, the right seat pilot and commander was the Flight Manager Technical Mike Channing, I was the P3 (in the SPO/FE seat).

A few months later (4/5 April 73) we did the same stall tests in 5B-DAE which was a Cyprus Airways aircraft, out of Hatfield - I think these were pre-delivery tests, one with Gordon Corps for the CAA, the other with Pat Fillingham who was DeHaviland's (subsequently Hawker Siddely) Deputy chief test pilot. (That aircraft was subsequently written off after damage on the ground in Nicosia, along with another Trident, during the Turkish invasion of Cyprus. The sad remains of the other one 5B-DAB were still there in April this year according to Google Maps - apparently the only airframe on the abandoned airport.

And as other have pointed out, absolutely no connection with AF447. I suspect cunliffe's uncle's story can safely be discounted!

[blind pew]

Are you sure Channing was technical flight manager in 72 Steve?
I did one of my first line trips with the idiot in may? And he wouldn't let me touch the controls..three legs due to engineer's strikes to Keflavik with a massive concentration of birds on the runway which I wasn't aware of until we hit them..he said nothing as I was glued to the engine instruments.
months later he was all slimy wanted to be elected onto the balpa plc ..which happened and after selling us down the road the next week he was deputy to Batman.. another bloke whose ability didn't match his ego.
Modelled himself on Douglas Bader posing with a pipe...

[JammedStab]

I have to admit, I was surprised to discover that the 727 does not have a stick pusher to ensure that there is never a deep stall. ATR has one though.

[dixi188]

IIRC the first 2 B727s on the UK register (Dan Air) had stick pushers fitted as a CAA requirement. Later aircraft did not have them fitted.

[slast]

IIRC the first 2 B727s on the UK register (Dan Air) had stick pushers fitted as a CAA requirement. Later aircraft did not have them fitted.
Didn't DanAir later acquire at least one 727 ex Mexicana which had originally had the JATO bottle option? Identifiable by a fairing in front of #2 engine intake for stuff that had to be rerouted to accommodate the JATO bottles? Maybe bcgallacher's informant was also getting mixed up with that. Nothing to do with stall test/recovery though.

[Pugilistic Animus]

I have to admit, I was surprised to discover that the 727 does not have a stick pusher to ensure that there is never a deep stall. ATR has one though.
JammedStab the ATR is most likely fitted with a stick pusher in order to prevent spin entry rather than for deep stalls.

[megan]

dixi, don't know about later, but initial introduction saw the DAN AIR 727 fitted with stick pusher, stick nudger and stick shaker to offer full stall protection.

ATR 42 Model: 400/500 manualA stick pusher and a stick shaker are provided, preventing the aircraft from reaching a critical angle of attack. When the detected incidence becomes too high, the MFC sends a signal to an electric actuator which shakes the control column at stall alert thresholds.If angle of attack keeps increasing, a further threshold is reached and the MFC activates the stick pusher ; the complete pitch control linkage assembly is pushed forward. Note : There are two stick shakers, one for each control column but only one stick pusher actuator located on the captain pitch channel. In case of pitch uncoupling when the pusher triggering angle of attack is reached, only the captain control column is pushed forward.

[zzuf]

Typically, there is only one reason any aircraft is fitted with a stick pusher - the type cannot be shown to comply with the stall handling characteristics requirements of the applicable certification standard without it.
The stall handling deficiencies may included, deep stalling, unacceptable longitudinal handling or stability on the stall approach, unacceptable levels of "roll off" at the stall.
While, often, unacceptable characteristics may be corrected with aerodynamic fixes, that may be the more expensive option. It may be cheaper just to fit a known acceptable stick pusher system than to embark on a stall development programme with an uncertain outcome. Also, the aerodynamic fix may result in an unacceptable increase in stall speeds from a performance viewpoint.

[Pugilistic Animus]

There is only one possible method of recovery from a deep stall, which involves banking the aircraft 90° and pointing the nose down. Given enough altitude, sufficient airflow will develop over the elevators to enable the aircraft to be recovered. Potential pitfalls are overspeed, entering a flat spin, disruption of airflow into the jet engine intakes and entering another high speed stall while attempting a recovery pullout. This really is a last throw of the dice!
for an rapid airspeed loss and a pitch up the procedure is to turn near knife edge and allow the nose to drop then recover from the subsequent spiral dive...Nothing as said earlier will work you out of a deepstall

[ZFT]

for an rapid airspeed loss and a pitch up the procedure is to turn near knife edge and allow the nose to drop then recover from the subsequent spiral dive...Nothing as said earlier will work you out of a deepstall

The ATR can be recovered from a deep tail stall by lowering flaps (to 15) should you manage to get into one.

[fantom]

IIRC the first 2 B727s on the UK register (Dan Air) had stick pushers fitted as a CAA requirement. Later aircraft did not have them fitted.

They had little square holes in the centre panel which, I recall, was for the accumulator warning lights for the pushers. The CAA boss at the time (Davies - he of the famous book) had a thing about pushers. As soon as he went, they were ripped out.

The JATO aircraft was easily identified by the circuit-breaker panel - the JATO pair of cbs was still placarded. More obvious was the huge strengthening panel on top of the fwd fuselage - the JATO had set fire to the aircraft and it was extensively re-built.

[KenV]

Ken, the situation may not be quite so absolute as you suggest. A stalled wing does not mean no lift, just insufficient lift for the conditions; see Cl vs AoA charts.A deep stall requires a high AOA. So high that the tail is blanked by the turbulent wing flow. At that high of an AOA the wing is fully stalled and there is effectively no roll control. It is why during flight testing of deep stall conditions that the test aircraft is fitted with a drogue parachute to break the stall condition. You can't fly out of it. The roll and pitch maneuver is only effective to PREVENT a normal stall from progressing to a deep stall condition. If you've stalled, then do not quickly apply the correct procedures to prevent it from progressing to a deep stall, and find yourself in a deep stall, kiss your butt good bye.

During stall testing of the C-17 it was equipped with a recovery parachute and an emergency escape trunk. Even the recovery parachute might fail to break the stall in which case the pilots would use the emergency escape trunk to exit the aircraft and parachute to safety.

[safetypee]

KenV, I maintain my alternative view, #14. I would accept that the exact nature of the stall will, as ever, depend on the aircraft type, wing characteristics, control system, trim and the range of cg; and an ‘otherwise healthy aircraft’ (MFS #15).

The residual lift at ‘high’ AoA will be proportional to the wing characteristics, depending on how fast Cl falls away from Clmax with increasing AoA. The lack of either roll or pitch control generally discounts any residual ability of the control surface to ‘affect the airflow’, even although it is disturbed airflow, there still is airflow.

Other discounted factors might include any tendency for natural roll-off during the stall entry, span-wise airflow or disturbance which might affect roll, e.g. engine pod or pylon vortex shedding vs AoA or small values of sideslip.
Differing views might originate from specific type experience, theory or practice. Although not familiar with the C17, I suspect that my civil type experience is remarkably similar.

[zzuf]

Deep stall doesn't require extreme angles of incidence. It just requires sufficient incidence that the nose down pitching moments available to the pilot are insufficient for recovery.
I would be surprised if AF447 didn't achieve incidence angles far beyond those at which some T tailed types would be considered deep stalled, but it demonstrated roll oscillations showing that there were cyclic variations in lift across the wing.
During the early days of deep stall research it was found, for at least one transport aircraft type, that apparently benign sink in a level attitude could result in deep stall alpha with the pilot unaware of what was occurring.
Of course, many cases of deep stall do require extreme alpha to "lock in".
With regard to aerodynamic controls on a stalled wing, they are often shown to be effective, albiet with, probably, hugely reduced response.
I have no idea how "deeply" stalled the Hunter wing was during spinning, but on a daily basis demonstrations were made of the power of the ailerons to either recover from a spin, despite full pro-spin controls being held with the rudder and elevators, or prevent spin recovery despite full anti spin controls being held with the rudder and elevator.
This was a demonstration of the rolling moments the ailerons could produce which cross coupled to produce significant pro or anti spin yawing moments.

[zzuf]

dixi, AFAIR the Trident accident was during a pre-delivery production test flight, thus additional experimental test-flight devices were not fitted.

Nowadays where pre-delivery testing requires a stall check either with or without stick push (stall ident) enabled, an additional glare-shield stall panel is fitted showing AoA calibrated for stick shake / push. Preflight the AoA vanes and indicator are crosschecked, so that if when inflight the shake / push do not occur at the expected AOA then the test is stopped.
The BAe 146 used such a panel, which may also have been approved for training stalls, providing the shake / push was enabled and checked / calibrated pre-flight.

Mmmm... Twice, in another life I have terminated stall tests during a certification validation when the stall ident/stick pusher didn't operate at the scheduled speed. As a flight test team, our flight test engineer concentrated almost exclusively on the ASI during the deceleration. We allowed 1kt below the scheduled speed - if it were possible to read the ASI to that accuracy, before terminating the flight. Very stressful knowing that the pusher is required because of deep stall concerns, and no stall test instrumentation fitted.
These tests can be very demanding to fly. For the stall devices to operate on schedule, the correct approach must be flown. Any errors in approach rate or small G errors will negate the approach.
The TP often has little familiarity with the type and the manufacturer takes a dim view of the termination of flights - no pressure.